Riem

REVISTA IBRACON DE ESTRUTURAS E MATERIAIS
IBRACON STRUCTURES AND MATERIALS JOURNAL

Volume 6, Number 3 June, 2013 ISSN 1983-4195
Contents
Flexibility modeling of reinforced concrete concentric frame joints S. J. P. J. MARQUES FILHO and B. HOROWITZ Numerical and experimental study of a waffle slab designed to serve as a tennis court floor P. F. SCHWETZ, F. P. S. L. GASTAL and L. C. P. SILVA F Nonconventional concrete hollow blocks evaluation by destructive and non-destructive testing M.S. RODRIGUES, G.C.S. FERREIRA, L. SHIROMA and A.L. BERALDO Study of the influential factors on the rheological behavior of adhesive mortar available in the market M. R. M. M. COSTA, E. PEREIRA, R. G. PILEGGI and M. A. CINCOTTO Monitoring of crack opening in masonry and concrete using mobile platform A. P. MARTINS, J. C. PIZOLATO JUNIOR and V. L. BELINI Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load R. G. DELALIBERA and J. S. GIONGO Plastic rotation and tension stiffening effect analysis in beams using photogrammetry R. N. F. DO CARMO, J. VALENA and D. DIAS-DA-COSTA Numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models F.M. ALMEIDA FILHO, M. K. EL DEBS and A.L.H.C. EL DEBS
Editorial Board
Amrico Campos Filho (Editor, UFRGS, Porto Alegre, RS, Brazil) Jos Luiz Antunes de Oliveira e Sousa (Editor, UNICAMP, Campinas, SP, Brazil) Romildo Dias Toledo Filho (Editor, UFRJ, Rio de Janeiro, RJ, Brazil) Antonio Carlos R. Laranjeiras (ACR Laranjeiras, Salvador, BA, Brazil) Bernardo Horowitz (UFPE, Recife, PE, Brazil) Denise C. C. Dal Molin (Former Editor, UFRGS, Porto Alegre, RS, Brazil) Emil de Souza Snchez Filho (UFF, Rio de Janeiro, RJ, Brazil) Geraldo Cechella Isaia (UFSM, Santa Maria, RS, Brazil) Gonzalo Ruiz (UCLM, Ciudade Real, Spain) Guilherme Sales Melo (Former Editor, UnB, Brasilia, DF, Brazil) Ivo Jos Padaratz (UFSC, Florianpolis, SC, Brazil) Joaquim Figueiras (FEUP, Porto, Portugal) Jos Marcio Fonseca Calixto (UFMG, Belo Horizonte, MG, Brazil) Luiz Carlos Pinto da Silva Filho (Former Editor, UFRGS, Porto Alegre, RS, Brazil) Mounir Khalil El Debs (USP, So Carlos, SP, Brazil) Nicole Pagan Hasparyk (Former Editor, FURNAS, Aparecida de Goinia, GO, Brazil) Osvaldo Lus Manzoli (UNESP, Bauru, SP, Brazil) Paulo Helene (Former Editor, USP, So Paulo, SP, Brazil) Paulo Monteiro (Berkeley, University of California, Berkeley, CA, USA ) P.K. Mehta (Berkeley, University of California, Berkeley, CA, USA ) Pedro Castro Borges (CINVESTAV, Mxico, D.F , Mxico) Ronaldo Barros Gomes (UFG, Goinia, GO, Brazil) Rubens Machado Bittencourt (Former Editor, FURNAS, Aparecida de Goinia, GO, Brazil) Tlio Nogueira Bittencourt (Former Editor, USP, So Paulo, SP, Brazil) Vladimir Antonio Paulon (UNICAMP, Campinas, SP, Brazil)
We are now releasing the June 2013 issue of IBRACON Structures and Materials Journal (Volume 6, Number 3). This issue contains eight articles on remarkable topics of concrete materials and structures. The flexibility modeling of reinforced concrete frame joints is the topic for the first article. The second article discusses numerical and experimental study of a waffle slab designed to serve as a tennis court floor. The next article addresses the evaluation of nonconventional concrete hollow blocks by destructive and non-destructive testing. The fourth article draws on the influential factors on the rheological behavior of adhesive mortar available in the market. An Image-based method for monitoring of crack opening on masonry and concrete using mobile platform is presented in the fifth article. The sixth article presents a numerical analysis of two-pile caps with embedded sockets, subject to eccentric compression. The seventh article describes an application of photogrammetry in the analysis of plastic rotation and tension stiffening effect in beams. The issue closes with an article on a numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models We acknowledge the authors and reviewers contribution to this issue. Amrico Campos Filho, Jos Luiz Antunes de Oliveira e Sousa, and Romildo Dias Toledo Filho Editors Estamos publicando o nmero de junho de 2013 da Revista IBRACON de Estruturas e Materiais (Volume 6, Nmero 3). Este nmero tem oito artigos sobre temas importantes de materiais e estruturas de concreto. A modelagem da flexibilidade de ns de prticos de concreto o tema do primeiro artigo. O segundo artigo apresenta um estudo numrico e experimental de uma laje nervurada projetada como piso de uma quadra de tnis. O artigo seguinte aborda a avaliao de blocos vazados no convencionais de concreto por meio de ensaios destrutivos e no-destrutivos. O quarto artigo trata dos fatores que influenciam o comportamento reolgico de argamassas colantes disponveis no mercado. Um mtodo com base em imagem para o monitoramento da abertura de fissuras em alvenaria e concreto usando plataforma mvel apresentado no quinto artigo. O sexto artigo apresenta modelagem numrica de blocos sobre duas estacas com clices embutidos, submetidos a compresso excntrica. No stimo artigo uma aplicao de fotogrametria apresentada para a anlise de rotao plstica e da contribuio do concreto entre fissuras para o aumento da rigidez flexo (tension stiffening effect). O nmero se completa com um artigo sobre modelagem numrica, utilizando modelos de vigas, para estudar o comportamento da tenso de aderncia entre barras de ao e concretos auto-adensvel e convencional. Agradecemos aos autores e revisores por sua contribuio a este nmero. Amrico Campos Filho, Jos Luiz Antunes de Oliveira e Sousa e Romildo Dias Toledo Filho Editores
Reviewers
Reviewers are selected by the Editors among the IBRACON members with recognized competence in the specific field of each contribution. They are acknowledged at the end of each volume.
Cover: Courtesy:
Ville San Marino Victor Schwaner
Ibracon Structures and Materials Journal is published bimonthly (February, April, June, August, October and December) by IBRACON. IBRACON Instituto Brasileiro do Concreto Founded in 1972 R. Julieta do Espirito Santo Pinheiro, 68 Jardim Olmpia, So Paulo SP Brasil 05542-120 Phone: +55 11 3735-0202 Fax: +55 11 3733-2190 E-mail: arlene@ibracon.org.br Website: http://www.ibracon.org.br Editors Amrico Campos Filho (Brazil) Jos Luiz Antunes de O. e Sousa (Brazil) Luiz Carlos Pinto da Silva Filho (Brazil) Romildo Dias Toledo Filho (Brazil)
Volume 6, Number 3 June 2013 ISSN: 1983-4195
Flexibility modeling of reinforced concrete concentric frame joints S. J. P. J. MARQUES FILHO and B. HOROWITZ Numerical and experimental study of a waffle slab designed to serve as a tennis court floor P. F. SCHWETZ, F. P. S. L. GASTAL and L. C. P. SILVA F Nonconventional concrete hollow blocks evaluation by destructive and non-destructive testing M.S. RODRIGUES, G.C.S. FERREIRA, L. SHIROMA and A.L. BERALDO Study of the influential factors on the rheological behavior of adhesive mortar available in the market M. R. M. M. COSTA, E. PEREIRA, R. G. PILEGGI and M. A. CINCOTTO Monitoring of crack opening in masonry and concrete using mobile platform A. P. Martins, J. C. Pizolato Junior and V. L. Belini
360
375
392
399
414
Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load R. G. DELALIBERA and J. S. GIONGO Plastic rotation and tension stiffening effect analysis in beams using photogrammetry R. N. F. do Carmo, J. Valena and D. Dias-da-Costa Numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models F.M. ALMEIDA FILHO, M. K. EL DEBS and A.L.H.C. EL DEBS
Cover design & Layout: Ellementto-Arte www.ellementto-arte.com
436
475
499
Aims and Scope

The IBRACON Structures and Materials Journal is a technical and scientifical divulgation vehicle of IBRACON (Brazilian Concrete Institute). Each issue of the periodical has 5 to 8 papers and, possibly, a technical note and/or a technical discussion regarding a previously published paper. All contributions are reviewed and approved by reviewers with recognized scientific competence in the area.
works and products under development. The instructions to submit a technical note are detailed on IBRACONs WebSite. A discussion is received no later than 3 months after the publication of the paper or technical note. The instructions to submit a discussion are detailed on IBRACONs WebSite. The discussion must be limited to the topic addressed in the published paper and must not be offensive. The right of reply is guaranteed to the Authors. The discussions and the replies are published in the subsequent issues of the periodical.
Objectives
The IBRACON Structures and Materials Journals main objectives are: Present current developments and advances in the area of concrete structures and materials; Make possible the better understanding of structural concrete behavior, supplying subsidies for a continuous interaction among researchers, producers and users; Stimulate the development of scientific and technological research in the areas of concrete structures and materials, through papers peer-reviewed by a qualified Editorial Board; Promote the interaction among researchers, constructors and users of concrete structures and materials and the development of Civil Construction; Provide a vehicle of communication of high technical level for researchers and designers in the areas of concrete structures and materials.
Internet Access
IBRACON Structural Journal Page in http://www.ibracon.org.br
Subscription rate
All IBRACON members have free access to the periodical contents through the Internet. Non-members have limited access to the published material, but are able to purchase isolated issues through the Internet. The financial resources for the periodicals support are provided by IBRACON and by research funding agencies. The periodical will not receive any type of private advertisement that can jeopardize the credibility of the publication.
Photocopying
Photocopying in Brazil. Brazilian Copyright Law is applicable to users in Brazil. IBRACON holds the copyright of contributions in the journal unless stated otherwise at the bottom of the first page of any contribution. Where IBRACON holds the copyright, authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted for libraries and other users registered at IBRACON.
Submission Procedure
The procedure to submit and revise the contributions, as well as the formats, are detailed on IBRACONs WebSite (www.ibracon.org.br). The papers and the technical notes are revised by at least three reviewers indicated by the editors. The discussions and replies are accepted for publication after a revision by the editors and at least one member of the Editorial Board. In case of disagreement between the reviewer and the authors, the contribution will be sent to a specialist in the area, not necessarily linked to the Editorial Board.
Copyright
All rights, including translation, reserved. Under the Brazilian Copyright Law No. 9610 of 19th February, 1998, apart from any fair dealing for the purpose of research or private study, or criticism or review, no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of IBRACON. Requests should be directed to IBRACON:
Contribution Types
The periodical will publish original papers, short technical notes and paper discussions. Announcements of conferences and meetings, information about book reviews, events and contributions related to the area will also be available in the periodicals WebSite. All contributions will be revised and only published after the Editorial and Reviewers Boards approve the paper. Restrictions of content and space (size) are imposed to the papers. The contributions will be accepted for review in Portuguese, Spanish or English. The abstracts are presented in Portuguese or Spanish, and in English, independently of the language in which the paper is written. After the review process, papers originally written in Portuguese or Spanish should be translated into English, which is the official language of the IBRACON Structures and Materials Journal. Optionally, papers are also published in Portuguese or Spanish. Original papers will be accepted as long as they are in accordance with the objectives of the periodical and present quality of information and presentation. The instructions to submit a paper are detailed in the template (available on IBRACONs WebSite). The length of the papers must not exceed 20 pages. A technical note is a brief manuscript. It may present a new feature of research, development or technological application in the areas of Concrete Structures and Materials, and Civil Construction. This is an opportunity to be used by industries, companies, universities, institutions of research, researchers and professionals willing to promote their
IBRACON
Rua Julieta do Esprito Santo Pinheiro, n 68 ,Jardim Olmpia, So Paulo, SP Brasil CEP: 05542-120 Phone: +55 11 3735-0202 Fax: +55 11 3733-2190 E-mail: arlene@ibracon.org.br.
Disclaimer
Papers and other contributions and the statements made or opinions expressed therein are published on the understanding that the authors of the contribution are the only responsible for the opinions expressed in them and that their publication does not necessarily reflect the support of IBRACON or the journal.
Objetivos e Escopo
A Revista IBRACON de Estruturas e Materiais um veculo de divulgao tcnica e cientfica do IBRACON (Instituto Brasileiro do Concreto). Cada nmero do peridico tem 5 a 8 artigos e, possivelmente, uma nota tcnica e/ou uma discusso tcnica sobre um artigo publicado anteriormente. Todas as contribuies so revistas e aprovadas por revisores com competncia cientfica reconhecida na rea.
instituies de pesquisa, pesquisadores e profissionais que desejem promover seus trabalhos e produtos em desenvolvimento. As instrues para submisso esto detalhadas na pgina de Internet do IBRACON. Uma discusso recebida no mais de 3 meses aps a publicao do artigo ou nota tcnica. As instrues para submeter uma discusso esto detalhadas na pgina de Internet do IBRACON. A discusso deve se limitar ao tpico abordado no artigo publicado e no pode ser ofensivo. O direito de resposta garantido aos autores. As discusses e rplicas so publicadas nos nmeros subseqentes da revista.
Objetivos
Os objetivos principais da Revista IBRACON de Estruturas e Materiais so: Apresentar desenvolvimentos e avanos atuais na rea de estruturas e materiais de concreto; Possibilitar o melhor entendimento do comportamento do concreto estrutural, fornecendo subsdios para uma interao contribua entre pesquisadores, produtores e usurios; Estimular o desenvolvimento de pesquisa cientfica e tecnolgica nas reas de estruturas de concreto e materiais, atravs de artigos revisados por um corpo de revisores qualificado; Promover a interao entre pesquisadores, construtores e usurios de estruturas e materiais de concreto, e o desenvolvimento da Construo Civil; Prover um veculo de comunicao de alto nvel tcnico para pesquisadores e projetistas nas reas de estruturas de concreto e materiais.
Acesso via Internet

Pgina da Revista IBRACON de Estruturas e Materiais em http://www.ibracon.org.br
Assinatura
Todos os associados do IBRACON tm livre acesso ao contedo do peridico atravs da Internet. No associados tm acesso limitado ao material publicado, mas podem adquirir nmeros isolados pela Internet. O financiamento para suporte revista provido pelo IBRACON e por agncias de financiamento pesquisa. A revista no receber qualquer tipo de anncio privado que possa prejudicar a credibilidade da publicao.
Fotocpias
Fotocpias no Brasil, A Lei Brasileira de Direitos Autorais aplicada a usurios no Brasil. O IBRACON detm os direitos autorais das contribuies na revista a menos que haja informao em contrrio no rodap da primeira pgina da contribuio. Onde o IBRACON detm os direitos autorais, autorizao para fotocopiar itens para uso interno ou pessoal, ou uso interno ou pessoal de clientes especficos, concedida para bibliotecas e outros usurios registrados no IBRACON.
Submisso de Contribuies
O procedimento para submeter e revisar as contribuies, assim como os formatos, esto detalhados na pgina Internet do IBRACON (www.ibracon.org.br). Os artigos e as notas tcnicas so revisadas por, no mnimo, trs revisores indicados pelos editores. As discusses e rplicas so aceitas para publicao aps uma reviso pelo editores e no mnimo um membro do Corpo Editorial. No caso de desacordo entre revisor e autores, a contribuio ser enviada a um especialista na rea, no necessariamente do Corpo Editorial.
Direitos autorais
Todos os direitos, inclusive traduo so reservados. Sob a Lei de Direitos Autorais No. 9610 de 19 de fevereiro de 1998, exceto qualquer acordo para fins de pesquisa ou estudo privado, crtica ou reviso, nenhuma parte desta publicao pode ser reproduzida, arquivada em sistema de busca, ou transmitida em qualquer forma ou por qualquer meio eletrnico, mecnico, fotocpia, gravao ou outros, sem a autorizao prvia por escrito do IBRACON. Solicitaes devem ser encaminhadas ao IBRACON:
Tipos de Contribuio
O peridico publicar artigos originais, notas tcnicas curtas e discusses sobre artigos. Anncios de congressos e reunies, informao sobre reviso de livros e contribuies relacionadas rea sero tambm disponibilizadas na pgina Internet da revista. Todas as contribuies sero revisadas e publicadas apenas aps a aprovao dos revisores e do Corpo Editorial. Restries de contedo e espao (tamanho) so impostas aos artigos. As contribuies sero aceitas para reviso em portugus, espanhol ou ingls. Os resumos sero apresentados em portugus ou espanhol, e em ingls, independentemente do idioma em que o artigo for escrito. Aps o processo de reviso, artigos originalmente escritos em portugus ou espanhol devero ser traduzidos para ingls, que o idioma oficial da Revista IBRACON de Estruturas e Materiais. Opcionalmente, os artigos so tambm publicados em portugus ou espanhol. Artigos originais sero aceitos desde que estejam de acordo com os objetivos da revista e apresentam qualidade de informao e apresentao. As instrues para submeter um artigo esto detalhadas em um gabarito (disponvel no stio do IBRACON). A extenso dos artigos no deve exceder 20 pginas. Um nota tcnica um manuscrito curto. Deve apresentar uma nova linha de pesquisa, desenvolvimento ou aplicao tecnolgica nas reas de Estruturas de Concreto e Materiais, e Construo Civil. Esta uma oportunidade a ser utilizada por indstrias, empresas, universidades,
IBRACON
Rua Julieta do Esprito Santo Pinheiro, n 68 , Jardim Olmpia, So Paulo, SP Brasil CEP: 05542-120 Fone: +55 11 3735-0202 Fax: +55 11 3733-2190 E-mail: arlene@ibracon.org.br.
Aviso Legal
Artigos e outras contribuies e declaraes feitas ou opinies expressas aqui so publicadas com o entendimento que os autores da contribuio so os nicos responsveis pelas opinies expressas neles e que sua publicao no necessariamente reflete o apoio do IBRACON ou da revista.
Diretoria Binio 2011/2013

Diretor Presidente Tlio Nogueira Bittencourt
Diretor 1 Vice-Presidente Jos Marques Filho Diretor 2 Vice-Presidente Julio Timerman Diretor 1 Secretrio Antonio Domingues de Figueiredo Diretor 2 Secretrio Jos Tadeu Balbo Diretor 1 Tesoureiro Claudio Sbrighi Neto Diretor 2 Tesoureiro Carlos Jos Massucato Diretor Tcnico Ins Laranjeiras da Silva Battagin Diretor de Eventos Luiz Prado Vieira Jnior Diretor de Pesquisa e Desenvolvimento Ana Elisabete Paganelli Guimares A. Jacintho Diretor de Publicaes e Divulgao Tcnica Nelson Covas Diretor de Marketing Cludia Henrique de Castro Diretor de Relaes Institucionais Arcindo Vaquero Y Mayor Diretor de Cursos Iria Lcia Oliva Doniak Diretor de Certificao de Mo-de-obra Roseni Cezimbra
Conselho Diretor Binio 2011/2013

Conselheiros Individuais Tulio Nogueira Bittencourt Claudio Sbrighi Neto Augusto Carlos de Vasconcelos Antonio Domingues de Figueiredo Waldimir Antonio Paulon Geraldo Cechella Isaia Nelson Covas Luiz Prado Veira Jnior Ana Elisabete Jacintho Antonio Carlos Laranjeiras Conselheiros Mantenedores e Coletivos ABCIC - Associao Brasileira da Construo Industrializada de Concreto ABCP - Associao Brasileira de Cimento Portland ABESC - Associao Brasileira das Empresas de Servios de Concretagem CESP - Companhia Energtica de So Paulo CNO Norberto Odebrecht Escola Politcnica da Universidade de So Paulo - POLI Eletrobrs - FURNAS - Centrais Eltricas S/A L.A. FALCO BAUER HOLCIM Brasil S/A IPT - Instituto de Pesquisas Tecnolgicas do Estado de So Paulo Conselheiros Permanentes Eduardo Antonio Serrano Jos Zamarion Ferreira Diniz Paulo Roberto do Lago Helene Ronaldo Tartuce Rubens Machado Bittencourt Selmo Chapira Kuperman Simo Priszkulnik
Volume 6, Number 3 June 2013 ISSN: 1983-4195
Flexibility modeling of reinforced concrete concentric frame joints S. J. P. J. MARQUES FILHO and B. HOROWITZ Numerical and experimental study of a waffle slab designed to serve as a tennis court floor P. F. SCHWETZ, F. P. S. L. GASTAL and L. C. P. SILVA F Nonconventional concrete hollow blocks evaluation by destructive and non-destructive testing M.S. RODRIGUES, G.C.S. FERREIRA, L. SHIROMA and A.L. BERALDO Study of the influential factors on the rheological behavior of adhesive mortar available in the market M. R. M. M. COSTA, E. PEREIRA, R. G. PILEGGI and M. A. CINCOTTO Monitoring of crack opening in masonry and concrete using mobile platform A. P. Martins, J. C. Pizolato Junior and V. L. Belini
360
375
392
399
414
Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load R. G. DELALIBERA and J. S. GIONGO Plastic rotation and tension stiffening effect analysis in beams using photogrammetry R. N. F. do Carmo, J. Valena and D. Dias-da-Costa Numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models F.M. ALMEIDA FILHO, M. K. EL DEBS and A.L.H.C. EL DEBS
436
475
499
Volume 6, Number 3 (June 2013) p. 360-374 ISSN 1983-4195
Flexibility modeling of reinforced concrete concentric frame joints Modelao da flexibilidade de ns concntricos de prticos em concreto armado
S. J. P. J. MARQUES FILHO a sergiopriorifilho@hotmail.com B. HOROWITZ b horowitz@ufpe.br
Abstract
The intersections between beams and columns in a reinforced concrete building structure are called frame joints. It is a region with significant bending stiffness but subjected to large shear stresses. Appropriate modeling of the flexibility of reinforced concrete frames is essential to its design, in service limit states as well as in ultimate limit states. It has been shown, theoretically as well as experimentally, that the influence of joint flexibility may account for 20% of total structural lateral displacement. Models using only bar elements and rotational springs are proposed to consider the joint flexibility in linear analyses of building structures. In order to validate the proposed model, comparisons with experimental results found in the literature are made. Finally, the results of second order analyses using the proposed model are compared with those obtained by finite elements. Keywords: reinforced concrete, frame joint, building structural analysis.
Resumo
A zona de interseo de vigas e pilares em estruturas de edifcios chamada n de prtico. Trata-se de regio com rigidez significativa flexo, porm sujeita a grandes solicitaes de cisalhamento. A apropriada modelao da flexibilidade de prticos essencial ao seu dimensionamento, tanto nos estados limites de servio quanto nos estados limites ltimos. Foi verificado tanto teoricamente quanto experimentalmente que a influncia da flexibilidade dos ns pode alcanar 20% do deslocamento total da estrutura. Modelo utilizando apenas elementos de barras e molas proposto para levar em considerao a flexibilidade dos ns na anlise de estruturas de edifcios correntes. De forma a validar o modelo proposto, so realizadas comparaes com resultados experimentais encontrados na literatura. Por fim, os resultados da anlise considerando os efeitos de segunda ordem utilizando o modelo proposto so comparados com aqueles obtidos por elementos finitos. Palavras-chave: concreto armado, ns de prtico, anlise estrutural de edifcios.
Departamento de Engenharia Civil, Universidade Federal de Pernambuco, sergiopriorifilho@hotmail.com, Av. Acadmico Hlio Ramos s/n, 50740-530, Recife-PE-Brasil. b Departamento de Engenharia Civil, Universidade Federal de Pernambuco, horowitz@ufpe.br, Av. Acadmico Hlio Ramos s/n, 50740-530, Recife-PE-Brasil.
a
Received: 2a Sep 2012 Accepted: 14 Dec 2012 Available Online: 12 Jun 2013
2013 IBRACON
S. J. P. J. MARQUES FILHO | B. HOROWITZ
1. Introduction
The zone of overlapping of beams and columns in reinforced concrete buildings is called frame joint. This region has significant bending stiffness but is subjected to large shear stresses. In building framed structures subjected to lateral loading, zero moment can be considered at mid-height of columns and at mid-span of beams. Thus, the subassemblage highlighted in Figure 1 can be used to represent the interaction between beams and columns of a frame. This work studies the cases where the axes of beams intersect the axes of framed columns. Two types of joints are studied: complete joints, whose beams and columns have the same width, and concentric joints, where the column width is larger than the beam
width, as illustrated in Figure 3. As shown in Figure 2 the joints are grouped according to their geometry as interior cross-type joint, TLateral exterior joint, T-Top exterior joint, and L-type joint. The main objective of this study is to propose a simplified model of ready numerical implementation that uses only bars and spring elements to take into account the flexibility of joint regions which may account for 20% of total structural lateral displacement, as demonstrated by the experimental results of Shin and LaFave [1]. To simplify the development of the proposed approximate model the restraining effects of slabs and transverse beams are neglected. This is a conservative assumption since it tends to decrease joint stiffness. These additional effects will be the focus of the next phases of the present research. Unfortunately very few experimental results are reported in the literature of joints including those restraining members.
Figure 1 Bending moment diagram of a laterally loaded frame
Figure 2 Building subjected to lateral loads with all different joint types: (a) L-type joint; (b) T-Top exterior joint; (c) Cross-Type interior joint; (d) T-Lateral exterior joint
IBRACON Structures and Materials Journal 2013 vol. 6 n 3
361
Flexibility modeling of reinforced concrete concentric frame joints
Figure 3 Elevation and plan view of the example substructure (in cm)
second order analyses incorporating the model were performed on several multistory building frames. Obtained results are compared with those of second order, three-dimensional, finite element analyses.
2. Impact of joint flexibility

In order to gain an initial insight on the impact of joint flexibility on structural behavior consider the illustrative example of Figure 3 where two interior joint subassemblages are shown. The first is a complete type joint, consisting of beams with 5m span and a 3m high column. Both beams and columns have a 20cm by 50cm rectangular cross-section. The second joint is concentric and differs from the first only in the cross-section of the column which is 160cm by 25cm. Note that both columns have the same moment of inertia in the plane of bending. In both subassemblages horizontal displacement of the bottom section of the column as well as the vertical displacements of the end sections of the beam are prevented. The goal is to evaluate the required horizontal force applied to top of the column which produces a 1cm horizontal displacement. The modulus of elasticity of the concrete is 20,6GPa and Poissons ratio is 0,2. Initially consider as a basis the finite elements model of the first example, shown in Figure 4(a). This model uses quadratic hexahedron elements having twenty nodes with reduced integration implemented in the commercial code ABAQUS [3]. Concrete is modeled as a linear elastic homogeneous material. The discretization mesh follows the guidance given in Reference [4] for bending problems which suggests at least two elements through the width and shows excellent agreement with beam theory using four elements through the height, as well as experience gained throughout the present research. These techniques are used for the finite element models in the remainder of this study. For the finite element model of Figure 4(a) the value of the reaction force at the top section of the column corresponding to the imposed unit displacement is 81,5kN. Consider now the model shown in Figure 4(b) consisting only of bar elements where rigid links are specified for both beams and columns in the interior of the joint. In this case, the necessary force is 100kN. If no rigid links
Three-dimensional finite element models of beam/column subassemblages are submitted to unit displacements at one of the column end sections, and comparisons are made between the resulting reactions against those obtained using the proposed model. As a result of this study required parameters of the model are adjusted. Only elastic analysis is considered since this is the usual approach for nonseismic design of building structures, where second order effects are directly or approximately taken into account. In order to validate the adopted parameter values, comparisons are made between the predicted values of displacements against experimental subassemblage results for drift values compatible with the NBR-6118 code [2] limits for service lateral displacement. In order to assess the accuracy of the proposed simplified model,
Figure 4 (a) Finite elements model; (b) Bars model with rigid links in the joint region
362
Figure 5 (a) Joint deformed configuration; (b) Shear and normal stresses
are specified, with all bars having constant properties throughout their lengths, completely neglecting the stiffness of the joint region, the corresponding force would be 68,3kN. Computing the ratios of those different values of required forces to the finite element result one concludes that the stiffness of the subassemblage is either overestimated by 23% or underestimated by 16%. Figure 4(a) shows color-filled contours of shear stresses. It can be seen that there exists a high concentration of shear stress in the joint zone. These stresses result from the combined action of normal stresses due to bending moments of same sense applied by beams and columns to the joint. Figure 5(a) shows a representation of the moments applied by the beams to the joint. Figure 5(b) shows the normal stresses resulting from the applied bending moments and the required shear stresses inside the joint to maintain equilibrium. Joint shown in Figure 5(a) also suggests that the largest contribution to joint deformation is due to shear distortion, and not bending in the interior of the joint. For the second case, shown in Figure 3(b), of the illustrative ex-
ample observe from Figure 6(a) that the reaction obtained from the finite element model is 74kN, corresponding to an imposed displacement of 1cm at the top of the column. For a bar model with rigid links inside the joint the reaction force is 92,1kN. Figure 6(b), which shows a section through the symmetry plane of the joint, depicts again the concentration of shear stresses inside the joint. If no rigid links are specified the required force is 69,7kN. Despite beams and columns having the same moment of inertia, the joint of the second example has less stiffness than that of the first example. Among various factors, one of the largest contributors to the increase of the flexibility of the connection is the additional torsional deformation that exists in the column. This subject will be discussed in detail in section 4.1. From the above discussion it can be concluded that in usual situations, errors in joint modeling can result in lateral displacement errors in the order of 20%. Experimental results from Shin and LaFave [3] show that joint flexibility contribute with 24% of total lateral displacement for drifts of 1% of story height and 53% for drifts of 6% of story height.
Figure 6 (a) Interior frame joint substructure; (b) Concentration of shear stress inside the frame joint
363
Figure 7 Frame joint subjected to a lateral load, VC
due to contributions of beams, columns, and joint deformations. The deformation in the joint zone has normal, bending, and shear components. The most significant part is due to shear distortion that will be discussed in more detail in order to develop of an approximate expression of its stiffness where the important factors are included, with the correct power. In a second step the approximated expression is adjusted as a result of the conducted parametric study.
3.1 Flexibility of complete frame joints

Consider the subassemblage in Figure 7 subjected to a horizontal loading, VC , applied to the top of the column. Different modeling alternatives for joint flexibility of steel frames made of wide-flange sections are discussed by Charney and Downs [5]. They show that if on equates the lever arm between tension and compression stress resultants to the beam depth, the horizontal shear at the center of the joint, VN (see Figure 5(b)), is given by:
(1)
It can then be concluded that the contribution of joint flexibility to displacements is significant and must be taken into account in checking the limit state of excessive lateral displacements of building frames. In ultimate limit states second order global effects are also directly affected by joint flexibility, whose capacity to absorb the shear force must be checked. Where is the ratio of column width to beam span, L. The ratio of beam depth to story height, H, is equal to , as depicted in Figure 7. The average shear stress is given by:
3. Complete joints
Consider again the model shown in Figure 4(a), which corresponds to the complete joint of Figure 3(a). The top section displacement is Where: t=joint thickness and sN=volume of joint region.
(2)
Figure 8 (a) Adjusted rigid links model; (b) scissors model
364
Substituting VN in Equation (2) for its value in Equation (1), on gets:
For the scissors model the moment in the spring is VC H. The displacement at the top of the column only due to the flexibility of the joint region is given by:
(3) (6)
Using the unit dummy load method for computing the horizontal displacement of the upper section of the column one has that the shear stress corresponding to VC =1 is: Where: KNT = stiffness of the spring of the scissors model. Equating the displacements given by Equations (5) and (6):
(4)
Thus, the contribution, N , to the horizontal displacement at the top of the column due to shear stress at the joint is given by:
(7)
Comparative studies of steel interior joint subassemblages with experimental results and full planar frames, including P-Delta effects, reported in Charney and Downs [5] shows that the scissors model is effective for steel building structures in general.
(5)
Where G is the transverse modulus of elasticity of concrete. The two different alternatives for the modeling of joint flexibility discussed in the following sections are: Adjusted rigid links model. Scissors model.
3.4 Parametric study of complete frame joints

Since simplifying assumptions were adopted regarding the distribution of shear stresses and lever arm of the normal stress resultants, a correction factor, g, is needed in Equation (7). The resulting expression for the torsional spring is then given by:
3.2 Adjusted rigid links model

As previously discussed, the simple adoption of rigid links in the interior of the beam/column joint results in stiffness overestimation. A frequently adopted alternative is to adjust the length of the rigid links by multiplying by a factor 1, so as to shorten them, as can be seen in Figure 8(a). This model implicitly assumes that flexibility is due to bending deformation inside the joint. Parametric study conducted by Horowitz and Marques[6] shows that values varies between 0.44 and 0.75, and is highly dependent on the ratio of beam to column section heights, thus not being convenient for general use. This finding is corroborated by the recent study of Birely et al [7] where the adjusted rigid links model is used for seismic design and evaluation, based on a data base of experimental results including 45 cross type interior joints. Depending on frame effective stiffness approach and joint reinforcement design criterion values vary all the way from 0 to 1. Therefore this model is not pursued any further in this study
(8)
A parametric study was conducted keeping constant the width of all members at 20 cm and beam spans of 5m between col-
Figure 9 Distribution of shear stresses in the interior of the joint
3.3 Scissors model

Consider the model shown in Figure 8(b) consisting of rigid links at the ends of the members inside the joint, a hinge, and a torsional spring. The stiffness of the torsional spring represents the shear stiffness of the joint. This simple model was proposed by Krawinkler and Mohasseb [8] to consider the effect of panel zones of steel moment-resisting frames.
365
Figure 10 Variation of the parameter for 3 meters story height
Horowitz and Marques [9] suggests the following approach for the computation of the effective joint volume:
(9)
Where a and b are the joint dimensions, with a b, and t is the joint thickness. Figure 10 shows the surface that represents the variation of the computed correction parameter, g , with the depths of beam and column sections, for a story height of 3m. From the analysis of the obtained results the value of the parameter g for cross-type joints, g C , was taken as 0,45, which represents a value slightly below the average with a bias to larger flexibility. This is also representative for story height of 4m as shown in Horowitz and Marques [9], where the authors detail extensive parametric study of the scissors model applied to complete interior joints.
Figure 11 Anti-symmetric model in finite elements for T-Lateral and T-Top exterior joints
Figure 12 Finite element anti-symmetric model for L-type joints
umns. Story heights of 3 and 4meters were considered. For each height, values of 40, 60, 80, and 100cm were adopted for both beams and columns depths. Adopted values for the correction factor g for cross type interior joints, T-lateral and T-top joints, and L-type joints will be denoted by g C , gT and g L, respectively. 3.4.1 Cross type interior joints Initially a parametric study for the interior Cross-type joint is conducted. The boundary conditions applied to the three-di mensional model are the same as those used in the examples of Section 2. The adopted value of parameter g is obtained comparing results from the scissors and the finite elements models. The first issue to be discussed is the appropriate value of the joint volume, sN , that should be adopted for joints with general width/depth ratios. Figure 9 shows the shear stress contours for two different joint geometries: square, and the one in which the height is twice its width. It can be readily noticed that the behavior changes substantially. To approximate the geometric effect,
Figure 13 Deformed configuration of the complete cross-type joint
366
3.4.2 Complete T-Lateral and T-Top exterior joints Consider now the T-Lateral and T-Top exterior joints. Typical three-dimensional finite elements models are shown in Figure 11 with the imposed boundary conditions, where u, v and w are translations along global axes x, y and z, respectively. An anti-symmetric displacement field was applied in order to force a zero moment section at mid-height of the column and mid-span of the beams without resorting to multi-node artificial constraints. Similarly to cross-type joints the correction factor, gT , for T-type joints, is evaluated as the result of a parametric study. The final adopted value is, gT = 0,3, as reported by Horowitz and Marques [10], where the authors give a detailed analysis of the scissors model applied to complete T and L-type joints. 3.4.3 Complete L-type exterior joints The previously used modeling techniques are also employed for L-type joints as shown in Figure 12. As a result of the parametric study it was adopted, gL = 0,1, as reported in Horowitz and Marques [10].
14(a), subjected to a uniform displacement at its top section. The elastic restraint from strip B-B is less effective than that offered by A-A due to the column horizontal torsional deformations in the region of framing of the beams. In order to take into account this phenomenon, we consider a torsional member embedded in the column as show in Figure 14(b). The additional flexibility of the joint is that resulting from the flexibility of the torsional member, similarly to the proposed ACI code [11] equivalent frame formulation provisions for the computation of column stiffness in flat slabs. The differential rotation between the beam and the column, qconc, is given by:
(10)
The value of qconc can be taken as the sum of the differential rotation of the column at the point of incidence of the beam, qA, and the average rotation of the torsional member inside the joint, qt,average. As flexibility is the inverse of stiffness, and considering qA as being that corresponding to a complete joint, one has that:
4. Concentric joints
4.1 Modeling the flexibility of concentric joints
Consider the beam/column joint shown in Figure 13 subjected to an arbitrary displacement at its top section. One has that the difference between the rotation of the beam, qbeam, and the rotation of the column, qcol, is due to the shear distortion of the joint region, as discussed in Section 2. Consider now the concentric cross-type joint, shown in Figure
(11)
The value of Kcomplete was derived in Section 3.4 for the various types of joints. The term corresponding to the torsional stiffness is detailed below.
Figure 14 (a) Complete cross-type joint; (b) Cross-type with torsional member
367
4.1.1 Torsional stiffness Consider a column connected to a torsional member and subjected to a unit torsional moment, as shown in Figure 15(a). Assuming a linear distribution of moment per unit length and considering that the maximum value of the torsional moment at the center of joint is such as to provide a unit area under the diagram (see Figure 15(b)), it follows that the function that expresses the change of the applied moment to only one of the two arms of the torsion member, as shown in the discussion of the equivalent frame method in MacGregors textbook [12], is:
Function T(x) expresses the variation of the torsional moment along the member. From Figure 15(c) one has that the value of T at the point of framing of the beam on the column, point A of strip A-A in Figure 14(a), is given by:
(14)
Where bB is the width of the beams. The value of the derivative of the rotation in each section along the torsional member with respect to x is obtained by dividing the torsional moment at each section by CG, as shown in Figure 15(d), is given by:
(12)
where bC is the width of the column (see Figure 14(a)). From the loading distribution one can obtain the torsional moment diagram T(x) by integrating the function t(x), as shown in Figure 15(c):
(15)
Where G is the transverse modulus of elasticity of the material and C is the torsional constant used in the ACI code [11], given by:
(13)
Figure 15 Side view of the joint showing the torsional member; (b) torsional moment applied to the member; (c) torsional moment diagram in the torsional member; (d) derivative of rotation along the member
(16)
Where x is the least value between bC and bB, and y being the largest. The rotation at the end of the member is the integral of the derivative of the rotation along the member, and is given by:
(17)
A
Poissons ratio for concrete is taken as 0,2, thus the transverse modulus of elasticity is 0,42E. Substituting G for E/2, as suggested in the ACI code [11], and assuming that the average rotation of the member is one-third of the rotation at its end, one concludes that:
(18)
Torsional stiffness Ktor is the inverse of flexibility and therefore is given by:
(19)
368
Figure 16 Cross-type interior joint with its boundary conditions and load
parametric study, similar to the ones conducted for complete type joints. 4.2.1 Concentric cross-type interior joints In this new parametric study subassemblages are analyzed with story heights of 3 and 4 meters, beams with cross-section of 20x40, 20x60, 30x40, and 30x60cm and columns with crosssection heights of 20, 30 and 40cm and widths of 40, 60, 80, 100, 120, 140, 160, 180 and 200cm. As in the modeling conducted for complete joints, the horizontal translation in the bottom of the column and the vertical translation at the ends of the beams were prevented. The reaction force resulting from an imposed displacement of 1 cm applied at the top of the column of the finite element model is computed for each subassemblage. Figure 16 shows a crosstype joint with applied boundary conditions and external loading. As seen in Section 3.4.1, the value of correction parameter for complete cross-type joints is gC = 0,45. Using k=1 in Equation (20) to compute the spring stiffness of the scissors model results in differences of less than 5% when comparing with the finite element analyses. Therefore the correction factor of the torsional term for concentric cross-type interior joints is taken as, kC=1. 4.2.2 Concentric T-Lateral exterior joints Consider now T-lateral type joints. The three-dimensional finite element models are shown in Figure 17(a). Using gT = 0,3 and adopting kT-Lat = 1, the maximum difference obtained with the proposed model when compared with the finite elements results is in the order of 6%. 4.2.3 Concentric T-Top exterior joints The modeling of the T-Top joint is conducted in the same way as for the T-lateral exterior joints, as can be seen in Figure 17(b). If one uses gT = 0,3 and k=1 in Equation (20) for the computation of the torsional stiffness, the resulting frame model is 20% stiffer than the finite element model. Therefore based on the parametric study
4.2 Parametric study of concentric joints

Substituting the values of Kcomplete and Ktor in Equation (11) one arrives at the following expression:
(20)
Where g and k are correction factors. The values of g were obtained for each type of joint in Sections 3.4.1 through 3.4.3. In the following sections the values for k are obtained from a
Figure 17 (a) Model of the T-Lateral exterior joints; (b) model of the T-Top exterior joints; (c) model of the L-type joints
369
5.1.1 Cross-type interior joints
Figure 18 Experimental setup for the cross-type joint [13]
it was adopted kT-Top = 0,5 which results in a maximum difference of 10% with respect to finite element analyses. 4.2.4 Concentric L-type joints Figure 17 (c) shows the three-dimensional model used for L-type joints with its applied boundary conditions. The modeling of L-type joints using gL = 0,1 and k=1, as in the case of T-Top exterior joints, presents significant differences with respect to finite element results, in the order of 40%. As a result of the conducted parametric study it is adopted kL= 0,25. The use of this correction parameter decreased the maximum misfit in stiffness of the results from the two models to 20%. Since this type of joint is less frequent in the structures of buildings no significant loss of precision in global flexibility is expected.
5. Results and discussion

5.1 Comparison with experimental results
In order to calibrate the modeling of joint flexibility with real cases of structural buildings we initially conduct a comparison of the proposed scissors model prediction with experimental results found in the literature.
Consider the experimental arrangement shown in Figure 18 developed by Shiohara et al [13]. The columns have cross-section of 30x30cm, the beams cross section is 30x20cm, the height of the column is 1,47m and the length of the beam is 2,7m. NBR-6118 code [2] recommends maximum lateral displacement of H/850 for the wind action, representing a displacement of 0,12% of the story height. Using the experimental results of shear versus displacement, one reaches the conclusion that for a 0,12% drift, the corresponding horizontal force is 21kN. The concrete strength is equal to 28 MPa. The concentric scissors model is constructed with the following parameters: E=0,855600(28)0,5=2,52104MPa; G=1,05104MPa; a=0,111; =0,204; sN=2,7102 m3; Kcomp=272 MN-m/rad; Ktor=40783 MN-m/rad; Kconc=270 MN-m/rad. The following four structural models are considered for comparison purposes: n Finite Elements. n Bars with unadjusted rigid links. n Scissors model with uncracked bars. n Scissors model with effective moments of inertia according to the NBR6118 code: Icol,e=0,8Icol; Ibeam,e=0,5Ibeam; KNT,e=0,8KNT. Where Icol,e and Ibeam,e are the effective moments of inertia of columns and beams to be used in the structural analysis to approximately take cracking into account. The effective value of the torsional spring stiffness, KNT,e, is computed using the column reduction factor. The obtained results from the four models are shown in the Table 1. Results in Table 1 indicate that the scissors model with uncracked bars, reproduced quite accurately the results of the finite element model, demonstrating the suitability of the parametric adjustment. In order to reproduce the experimental results the gross moment of inertia of the members must be reduced to take cracking into account. The unadjusted rigid link model is twice as stiff as the scissors model with effective moments of inertia. 5.1.2 T-Lateral exterior joint From the experimental specimen of the T-Lateral exterior joint shown in Figure 19 [14], comparisons were conducted with the proposed model. The specimen consists of columns 2,70m high and a beam with span of 2,15m. The column has a cross-section
Table 1 Comparison of actuator force with numerical results Cross-Type interior joint
Model Finite Elements Unadjusted rigid link Scissor's Model with uncracked bars Scissor's Model with cracked bars Shear Force, Vc (kN) Theoretical Experimental 35,5 43,1 36,1 21,2 21 21 21 21
Figure 19 Experimental setup for the T-Lateral exterior joint [14]
370
Table 2 Comparison of actuator force with numerical results T-Lateral exterior joint
Model Finite Elements Unadjusted rigid link Scissor's Model with uncracked bars Scissor's Model with cracked bars Shear Force, Vc (kN) Theoretical Experimental 96,9 102,3 88,7 50,2 48,3 48,3 48,3 48,3
Figure 20 Experimental setup for L-type joint [15]
of 40x60cm and the beams 45x30cm, the concrete strength is 29,5MPa. In the experimental arrangement the columns ends are hinged in order to simulate the inflection points in a real building frame and cyclic loading is applied at the end of the beam. Using the experimental data provided by the senior author of Reference [14], it was found that for a displacement of 5,28mm, corresponding to a 0,25% drift, the applied force is 48,3 kN. The concentric scissors model is constructed with the following parameters: E=0,855600(29,5)0,5=2,58104MPa; G=1,08104MPa; a=0,093; =0,167; sN=0,102 m3; Kcomp=787 MN-m/rad; Ktor=26209 MN-m/rad; Kconc=587 MN-m/rad. The same four kind of models used in the previous section are compared in Table 2. The scissors model with uncracked bars presents a 8,5% error when compared to the finite element model. Considering the cracking of the model based on the reduction factors of the gross moment of inertia of the sections recommended by the Brazilian code, NBR 6118 [2], the prediction error for the proposed model with respect to the experimental results is 3,9%. Once again the unadjusted rigid links proved to be twice as stiff as the cracked scissors model. 5.1.3 L-type joint In the case of concentric L-type exterior joints, the proposed model was compared with experimental results obtained by Angelakos[15]. Figure 20 shows a schematic representation of the experimental arrangement. The specimen has a span of 1,325m and
height of 0,914m. The cross-section of the beam is 28cm by 40cm and that of the column is 40cm by 40cm. It was used a 1,83mm displacement representing the magnitude of the serviceability limit imposed by the NBR 6118 code [2]. The concrete strength is 31,7MPa. The concentric scissors model uses the following parameters: E=0,855600(31,7)0,5=2,68104MPa; G=1,12104MPa; a=0,152; =0,219; sN=0,064 m3; Kcomp=180 MN-m/rad; Ktor=35275 MN-m/rad; Kconc=179 MN-m/rad. The same four models previously used are compared in Table 3. Another feature is to be highlighted. Different experimental results were obtained for displacements that open or close the joint. This demonstrates the importance of the contribution of the reinforcement slip. Once more it can be seen the small difference, around 4%, between the proposed scissors model and the result of the finite element analysis. When the cracked scissors model is compared with the average of the experimental results, it is observed that the difference is only 2,9%.
5.2 Multi-story frames

In order to assess the accuracy of the proposed simplified joint model, six multi-story frames without slabs were analyzed and comparisons made between unadjusted rigid links, proposed scissors model and the finite elements model. In Figure 21 it can be observed typical geometry of the frames where L is the bay length, H is the story height, nbay is the number of bays and nsto the number of stories.
Table 3 Comparison of actuator force with numerical results L-type joint

Shear Force, Vc (kN) Experimental Opening Finite Elements Unadjusted rigid link Scissor's Model with uncracked bars Scissor's Model with cracked bars 12,65 19,74 13,1 8,59 10 10 10 10 Closing 6,7 6,7 6,7 6,7
Model
Theoretical
371
Figure 21 Schematic detail of the geometry and the loading applied to the frames
nonlinearity, a vertical load of 15 kN is applied to the outer columns and 30 kN to the inner columns at all stories. The load applied to the three-dimensional model of frame 1 can be seen in Figure 22. In the first two examples the frames contain complete joints only, while the remaining frames contain concentric type joints only. Table 4 summarizes the geometry of the frames adopted in the analyses. In Table 5 second-order analysis results are presented, where the percentage difference for the displacement response of each model is computed with respect to the finite element model. It can be readily seen that the scissors model presents the best results when compared to the finite element model. The maximum percentage difference is 5%, while for the model using unadjusted rigid links the difference increases to 16%. The last two columns contain lateral displacements of the scissors model with cracked members and the percentage increase due to cracking. It can be seen that average amplification of displacements is 46,5% which has a significant impact on bending moments of beams and columns.
6. Conclusions
Beam/column joint flexibility in usual reinforced concrete building frames, disregarding the presence of slabs and transverse beams, contributes around 20% of the total lateral displacement. Therefore joint flexibility modeling is needed to check the excessive lateral displacement serviceability limit state as well as the global second order effects in ultimate limit states. The NBR-6118 code [2] suggests the use of a rigid link model whose adjusted lengths are functions of the framing beam depths. Since joint flexibility stems from shear distortion, not bending inside the joint, this model may not always yield accurate results. An accurate yet simple scissors model, composed of bars and springs only, is proposed to take into account the flexibility of
In all cases a 10kN lateral load was applied in the first three and the last floors while a 15 kN load was applied to the intermediate remaining floors. The horizontal load was applied in the joints of the outer columns. In order to verify the effect of the geometric
Figure 22 Applied loads for frame model 1
372
Table 4 Geometry of analyzed frames

Complete joints frames nsto H (m) Beam section (cm) 14 14 3 3 20x60 20x80
Frames 1 2 Frames 3 4 5 6
nbays 3 1
L (m) 5 6
Column section (cm) 20x40 20x60
nbays 3 3 2 1
L (m) 5,3 5 4 5
Concentric joints frames H (m) Beam section (cm) nsto 16 16 16 20 3 3 3 3 20x60 15x70 20x60 20x60
Column section (cm) 100x30 130x30 100x20 100x30
beam/column joints including the usual case of column sections wider than framing beam sections. Simple expressions for spring stiffnesses are obtained from mechanical models affected by correction factors whose values are determined from extensive parametric studies. Comparative studies with experimental results demonstrate that the proposed model has adequate accuracy for design purposes if member stiffnesses are reduced according to the NBR- 6118 code [2] factors in order to take cracking into account. Maximum difference of 5% in displacements was observed between the proposed approximate scissors model and three-dimensional second order finite element analyses of six building plane frames.
7. Acknowledgements
The authors wish to recognize CAPES for financial support and Professor Hunk-Jen Lee for providing his experimental data.
8. References
[01] SHIN, M.; LAFAVE, J. M. Modeling of joint shear deformation contributions in RC beam-column connections to overall frame behavior. Structural Engineering and Mechanics, v.18, n.5, 2004; pp.645-669.
[02] ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Projeto de estruturas de concreto - Procedimento. - NBR 6118, Rio de Janeiro, 2007. [03] ABAQUS, Users Manual, version 6.3, Hibbitt, Karlsson& Sorensen, Inc, Pawtucket, RI, 2002. [04] ABAQUS, Getting Started with ABAQUS/Standard, version 6.3, Hibbitt, Karlsson& Sorensen, Inc, Pawtucket, RI, 2002. [05] CHARNEY, F. A.; DOWNS, W. M. Modeling procedures for panel zone deformations in moment resisting frames, EECS/AISC Workshop on Connections in Steel structures, Amsterdam, 2004.. [06] HOROWITZ, B.; MARQUES, S. J. P. J. Modelao da flexibilidade de ns internos de prtico em concreto armado. In: CMNE/CILAMCE, 2007, Porto-Portugal. [07] BIRELY, A. C.; LOWES, L. N.; LEHMAN, D. E. Linear analysis of concrete frames considering joint flexibility, ACI Structural Journal, V. 109, No. 3, May-June 2012, pp. 381-391. [08] KRAWINKLER, H.; MOHASSEB, S. Effects of panel zone deformations on seismic response, J. Constr. Steel Res., V. 8, 1987, pp. 233-250. [09] HOROWITZ, B.; MARQUES, S. J. P. J. Efeito da seo do pilar na flexibilidade de ns internos de prtico em concreto armado. In: 49 Congresso
Table 5 Results os second order analyses

Finite element Drift (cm) 1 2 3 4 5 6 6,23 9,12 5,36 4,23 13,84 29,12 Unadjusted rigid links Drift (cm) 5,26 8,23 4,64 3,56 12,11 27,07 Error (%) 15,5 9,8 13,4 15,8 12,5 7,0 Scissor's model uncracked bars Drift (cm) Error (%) 6,52 9,34 5,32 4,11 14,03 29,50 4,6 2,4 0,7 2,8 1,4 1,3 Scissor's model cracked bars Drift (cm) Increase (%) 9,24 12,34 8,52 6,60 19,38 43,08 41,7 32,1 60,2 60,6 38,1 46,0
Frames
373
Brasileiro do Concreto, IBRACON-CBC, 2007, Bento Gonalves, v.1, 2007. [10] HOROWITZ, B.; MARQUES, S. J. P. J. Modelao da flexibilidade de ns de prtico em concreto armado com peas de mesma largura. In: 50 Congresso Brasileiro do Concreto, IBRACON-CBC, 2008, Salvador - Brasil, v.1, 2008. [11] ACI Committee 318, Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary, American Concrete Institute, Farmington Hills, MI, 2008. [12] MACGREGOR, J., Reinforced Concrete: Mechanics and Design, Prentice Hall, 3ed, 1997. [13] Shiohara, H., Zaid, S. e Otani, S., Test of an Innovative Reinforcing Detail for High Performance R/C Interior Beam-Column Connection Subjected to Seismic Action, Proceedings of the Third International Conference on Concrete under Sever Conditions, Vancouver, 2001, pp. 739-746. [14] LEE, H. J., YU, S. Y., Cyclic Response of Exterior Beam-Column Joints with Different Anchorage Methods, ACI Structural Journal, v.106, n.3, 2009; p.329-339. [15] ANGELAKOS, B., The Behavior of Reinforced Concrete Knee Joints Under Earthquake Loads. Toronto, 1999, Thesis (Doctor of Philosophy) Department of Civil Engineering, University of Toronto.
374
Numerical and experimental study of a waffle slab designed to serve as a tennis court floor Anlise numrico-experimental de uma laje nervurada projetada como piso de uma quadra de tnis
P. F. SCHWETZ a paulete.schwetz@ufrgs.br F. P. S. L. GASTAL b francisco.gastal@gmail.com L. C. P. SILVA F c lcarlos66@gmail.com
Abstract
Reinforced concrete waffle slabs have become a common option for designers due to a need of rationalization in construction with reduction in costs and deadlines. To better understand the behavior of this structural system, and more realistically quantify stresses and displacements, a full scale waffle slab was tested. The structure, designed to serve as a tennis court floor, was submitted to a load of 12 kN/m2 and instrumented to measure strains and deflections at different locations. The loading process used the floors constructive base filling material and readings were taken at different loading stages and arrangements during the floors construction. Test data was compared to results obtained from the matrix analysis program Sistema Computacional TQS v11.0 and from the finite element model program SAP2000 v14.2.2. Slab behavior was as expected, with deflection and bending moments close to those determined by the numerical analysis. Keywords: waffle slab, reinforced concrete, matrix analysis of grids, finite element method.
Resumo
A necessidade de racionalizao na construo civil com a minimizao de custos e prazos vem fazendo das lajes nervuradas de concreto armado uma opo cada vez mais adotada pelos projetistas. Com o objetivo de compreender melhor o comportamento desse sistema estrutural e quantificar as tenses e deslocamentos da estrutura de uma forma mais realista, foi realizado um programa experimental em uma laje nervurada em escala natural. O estudo mediu deformaes no concreto e deslocamentos verticais em diversos pontos da estrutura, submetida a um carregamento de 12 kN/m2, correspondente a vrias camadas de enchimento necessrias para servir de base para quadras de tnis. Devido impossibilidade de aplicao do carregamento experimental de forma regular e em etapas, optou-se pela realizao de leituras em diferentes situaes de carga. Os dados coletados experimentalmente foram comparados com os resultados de duas anlises computacionais, empregando-se um modelo de anlise matricial de grelhas do programa Sistema Computacional TQS v11.0 e um modelo em elementos finitos do programa SAP2000 v14.2.2. O comportamento da laje esteve dentro do previsto, com deslocamentos verticais e momentos fletores prximos aos determinados pelas anlises numricas. Palavras-chave: laje nervurada, concreto armado, anlise matricial de grelhas, elementos finitos.
Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, paulete.schwetz@ufrgs.br, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil; b Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, francisco.gastal@gmail.com, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil. c Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, lcarlos66@gmail.com, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil.
a
Received: 10 Jul 2012 Accepted: 22 Feb 2013 Available Online: 12 Jun 2013
2013 IBRACON
Numerical and experimental study of a waffle slab designed to serve as a tennis court floor
1. Introduction
Sophisticated and rational structural solutions are on increasing demand in the daily work of structure designers due to the evolution of architectural designs and new construction management concepts. Therefore, it is necessary to evaluate a variety of structural systems, seeking to find the solution that can best deliver economic viability, speed and versatility of application. According to Dias R.[1], in multi-storied reinforced concrete structures, the slabs are responsible for a significant share of the concrete consumption. In the case of solid slabs, this share reaches almost two thirds of the total volume of the structure. Therefore, it is important to study different types of slab to be used in multi-story buildings to find solutions that are both technically viable and less expensive. From this perspective, the use of waffle slabs is proving to be an interesting option. This structural system can be defined as a set of ribs distributed in two directions, spaced at regular intervals and connected by a top concrete slab. The waffle slab structure was conceived with the reduction of concrete below the neutral axis, creating voids in a rhythmic pattern. This leads to a reduction in the weight of the structure and to a more efficient use of materials. However, the behavior of reinforced concrete waffle slabs still raises some questions due to their complexity, which has led some researchers to undertake numerical and experimental studies (Ajdukiewicz & Kliszczewicz, 1986 [2]; Selistre, 2000 [3]; Abdul-Wahab & Khalil, 2000 [4]; Dias, 2003 [1]; Soares, 2003 [5]; Schwetz, 2005 [6]; Schwetz, 2011 [7]). In Brazil, there has been a gradual increase in the use of waffle slabs, but conventional structural arrangements made of solid slabs and concrete frames remain the preferred model for reinforced concrete structures. However, it is necessary to develop procedures and standards to guide the use of waffle slabs constructive system. In fact, as the development of numerical algorithms that allow the design of this type of structure is quite recent,
Figure 2 Position of top and bottom reinforcement
many questions still remain with regards to the quantification of the slab stresses and their adequate modeling by the design methods and the mathematical models currently used. This study, which is part of an ongoing research project [7], seeks to improve the knowledge about reinforced concrete waffle slabs
Figure 3 Geometry of the waffle slab
Figure 1 Formwork developed for waffle slabs
376
P. F. SCHWETZ | F. P. S. L. GASTAL | L. C. P. SILVA F
and to verify whether their behavior is adequately simulated by the design methods and mathematical models widely used today. In order to achieve this aim, an experimental program was implemented, measuring strains and deflections in a full-scale waffle slab, designed to receive several layers of filling material necessary for use as the floor of a tennis court. The experimental results were compared with results obtained from numerical analyses performed using two different approaches to represent the structure (a grid matrix analysis and FEM analysis).
The loading adopted in the original design was the structures own weight of 4.8 kN/m2 with an additional dead load of 12 kN/m2, a live load of 3.0 kN/m2 and a masonry load applied directly on the edge beams. The high value of the additional dead load is due to the several layers of the base filling material of the tennis court. Only the additional dead load was used in the experimental program.
2.2 Instrumentation
The experimental program predicted measurement of strains and vertical displacements. In the concrete, strain gauges were placed in four slab locations, at the top and bottom surfaces (Fig. 4a). In the reinforcement, two strain gauges were placed at each instrumented location, protected with an epoxy based resin (Fig. 4b). A precision optical level was used to measure the vertical displacements (Fig. 5) at locations as shown in Fig. 6.
2. Experimental program
The floor analyzed was designed commercially by a structural engineering design office located in the city of Porto Alegre, Brazil. It is to be used as a tennis court floor. The formwork was built with plastic formwork developed for waffle slabs by Ulma Frmas e Escoramentos Ltda. This system is composed by a two-way ribbed structure that serves as support for the plastic molds. The ribbed structure is supported by a set of tubular metallic braces that are easy to disassemble (Fig. 1). The slab reinforcement was made with type CA-50. The weight of positive reinforcement used was about 16,000 kg and the amount of negative one was around 23,000 kg. Fig. 2 shows the generic position of the steel in the ribs. Besides the top and bottom reinforcements predicted in the numerical analysis, the structural designer considered the possibility of tensile stress on the bottom of the slab topping located between the ribs and placed a welded CA-60 steel mesh on the bottom part of the top slab.
2.3 Testing
The test began 63 days after casting. The loading was to be carried out in three stages and strain and vertical displacements were to be measured at the end of each loading stage. However, the load was distributed unevenly throughout the testing period. Some areas of the slab were used to store other types of material generating a non-uniform load (Fig. 7a and b). Hence, readings were made at different points in time and care was taken to register the actual load applied at each reading. The process of loading the structure lasted 87 days and readings were taken at 5 different times. Cylinder tests were also carried out to determine the modulus of longitudinal elasticity of the concrete (E) and the characteristic concrete strength (fck). The average value of the modulus of elasticity of the concrete E28 measured experimentally was equal to 28,45 GPa, and the average concrete strength obtained at 28 days was equal to 33,27 MPa, corresponding to a concrete characteristic strength fck, estimated equal to 30 MPa.
2.1 Geometry and loading

The geometry of the waffle slab floor is shown in Fig. 3. The designer used semi-inverted edge beams to increase stiffness in order to prevent excessive deformation. Around the internal columns, a 37.5 cm solid slab was used due to punching shear and high bending moments.
Figure 4 Position of strain gauges: (a) concrete and (b) reinforcement
CS
CI
A B
377
Figure 5 Vertical displacements measurement
3. Numerical analysis
Figure 6 Location of instrumentation
The numerical analysis was done with two computer programs traditionally used in the design and analysis of reinforced concrete waffle slabs. The results were compared to the experimental data obtained by Schwetz [7] in the actual structure, with the aim of verifying how close the mathematical models are to the actual behavior of the slab. The tested slab was numerically analyzed considering the values obtained experimentally from the test specimens for the parameters compressive strength (fck, ) and modulus of elasticity (E28). estimated
3.1 SAP2000
The first numerical analysis was done using SAP2000 software with finite elements. The SAP2000 is a structural engineering software for linear and nonlinear static and dynamic analyses of several types of structures, simulating their behavior when submitted to a wide range of demands [8]. The software has a library of finite elements. Among the existing options, the elements that were used in the numerical models proposed in this study are described below. 3.1.1 The Shell Element The Shell element is an area element used to model shells, membranes and plates in structures in two and three dimensions. The SAP2000 software defines two types of Shell element. They are denominated as Homogeneous Shell, used for homogenous materials, and as Layered Shell, used when the element is formed by heterogeneous materials or by more than one material. However, the software only allows nonlinear analyses when the element is a Layered Shell. Each element has its local coordinate system for the definition of material properties, loads and output of results. Tensions, internal forces and moments are determined by the Gauss quadrature method and extrapolated to the nodes.
378
Figure 7 (a) Unevenly distribution of the load and (b) Slab totally loaded
A
3.1.2 The Frame Element The Frame Element is a tridimensional element with six de grees of freedom per node: three degrees of freedom for translation and three degrees of freedom for rotation. It is used for two or three dimensional modeling of frames, trusses and grids. The formulation of the Frame element includes the determination of biaxial bending, torsion, axial and shear stress, through the integration of the tensions along the section. Those stresses
B
are determined at the ends of each element and at points along the element chosen by the user. In the FE model of the waffle slab under study, which is shown in Fig. 8, the top slab, the ribs, the edge beams and the solid slab regions were modeled using shell elements, totaling approximately 23,000 elements, and the columns were modeled with frame elements. A nonlinear analysis of the structure was carried out, applying three types of loads in the following sequence: self-weight, a masonry load applied directly on the edge beams and the additional dead load. The numerical results corresponding to the experimental load were considered as being the difference between the values obtained from the numerical analysis at the first and the last stages of the additional dead load application (Fig. 9).
Figure 8 Finite element model
Figure 9 Vertical displacements of instrumentation point PC concerning the loads applied on SAP2000
379
Figure 10 Grid model
3.2 TQS software

TQS is a software developed by TQS Informtica Ltda. for reinforced concrete structural design using the matrix analysis method for grids. This system, formed by a set of modules, creates a da-
tabase on the entire construction, grouping geometries and loads, managing the structural analysis, the transference and calculation of stresses and the detailing of all the elements that constitute the structure [9]. In the case of waffle slabs, the TQS software allows both linear and non-linear grid analysis. In linear analysis, the reinforcement is not considered in the determination of the cross sections stiffness. Nonlinear analysis calculates the vertical displacements considering the reinforcement used and the physical non-linearity of concrete due to cracking, based on the recommendations of the Brazilian Concrete Code (ABNT, 2004[10]). This software allows the structural designer to enter user defined values when determining stresses and detailing the reinforcement. In the specific case of waffle slabs, the most relevant choices are related to the structures flexibility. The user may change the torsional and/or flexural inertia of the bars, as well as adjust the stiffness of the connections between beams and columns. In this investigation, however, no reduction on torsional or flexural stiffness was used for the ribs or edge-beams. Additionally, the connections between ribs and beams and/or beams and columns were fully considered. This decision was made based on a parametric analysis, performed in a rectangular waffle slab, where the stiffness of the elements were varied and the results were compared to the ones obtained using the finite element program[7]. Given the necessary input data, the software automatically generates a numerical model of the slab, such as the one shown in Fig. 10. A nonlinear analysis of the structure was also carried out. However, in the case of nonlinear analysis, this computational pro-
Figure 11 Total load applied, excluded self-weight
380
Figure 12 TQS, SAP and experimental: deflections along line a shown in Fig. 6
Figure 13 TQS, SAP and experimental: deflections along line B shown in Fig. 6
gram considers only the sum of all loads applied on the structure, including its self-weight. This load can, however, be divided into a number of stages to be defined by the user. Hence, in this study, the results originated only from the consideration of the additional dead load were obtained through the difference between two different analyses: one applying the sum of self-weight load, additional dead load and masonry applied straight on beams and the other applying only the self-weight and the masonry load on the beams.
models and the experimentally measured data, showing that both numerical design strategies are fairly able to represent actual waffle slab deformed behavior.
4.2 Bending moments

In order to compare the numerical results with the ones obtained experimentally, the measured strains had to be transformed into bending moment estimates, considering the equilibrium of internal forces in the instrumented cross-sections [11]. Three values of bending moments were determined for each section, as shown in Fig. 14: n Experimental uncracked section with calculated s: Bending moment is calculated considering the uncracked cross-section, with the reinforcement strain inferred from a strain gradient defined from the experimental strain values measured; n Experimental uncracked section with measured s: Bending moment is calculated considering the uncracked cross-section, but takes into account the measured reinforcement strain. This estimate of the bending moment cracking onset concerns concrete contribution between cracks (tension stiffening effect); n Experimental cracked section: Bending moment is cal culated considering a cracked section with measured reinforcement strain.
4. Analysis of the results

This study presents the results obtained in the numerical analyses and those measured experimentally regarding the additional dead load applied as indicated in Fig. 11. Slab deflections and distribution of bending moments are compared using numerical and experimental results obtained along slab lines indicated as Line A and Line B as shown in Fig. 6.
4.1 Vertical displacements

Fig. 12 and 13 present deflection along slab lines A and B obtained with the nonlinear analyses of the structure as well as the values obtained experimentally. A similar behavior is verified between numerical results of the two
Figure 14 Strains used to determine experimental bending moments for (a) Uncracked section with s deduced, (b) Uncracked section with s measured and (c) Cracked section
381
Fig. 15 shows the distribution of bending moments obtained using both numerical models, compared to the estimates of bending moments derived from the experimental data obtained at location E1. Both numerical models present a similar shape, indicating maximum positive bending in the loading area and maximum negative moment in the solid slab region. At location E1, the bending moment obtained experimentally using uncracked section with measured s close to the values predicted by the numerical models, indicating a possible beginning of cracking around this section. Both models indicate a negative moment between the ribs and the edge beams. The value indicated by the model in FE is, higher. This difference is probably due to the fact that this model considers the top slab as a continuous plate, generating greater clamping effect on the edge beam, when compared with the grid model. Moreover, the FE model presents peaks of negative moments that are much more pronounced in the solid slab regions. This divergence may be associated to the different modeling of the solid slab regions in the softwares. At these locations, unfortunately no experimental data was collected, which could allow for a better assessment of the most adequate modeling for those connections and the influence of the solid slabs.
laborious, and difficult to use in the daily routine of design firms. 3. The grid analogy results in negative moments both in the slabedge beam connection and in the solid slab regions lower than those indicated by the finite element model. This difference in results may be related to the different modeling adopted by each computer program. Therefore, it is important the collection of experimental data that help define which of the models best represent this particular aspect of the behavior of a waffle slab.
6. Aknowledgements
The authors would like to thank CNPq (the Brazilian Research Council) and the companies Projetak - Tavares Eng. Associados S/C Ltda and Construtora Tedesco for the support they have given to this project.
7. References
[01] DIAS, R. H. Anlise Numrica de Pavimentos de Edifcios em Lajes Nervuradas. Dissertao (Mestrado em Engenharia) Escola de Engenharia de So Carlos, Universidade de So Paulo (USP), So Carlos, 2003. [02] AJDUKIEWICZ, A.; STAROLSKI, W. Reinforced- concrete slab-column stuctures. New York: Elsevier Science Publishers, 1986. [03] SELISTRE, S. L.C. Anlise Terico-Experimental de uma Laje Nervurada de Microconcreto Submetida a um Carregamento Uniformemente Distribudo. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, Porto Alegre, 2000. [04] ABDUL-WAHAB, H. M. S.; KHALIL, M. H. Rigidity and Strength of Orthotropic Reinforced Concrete Waffle Slabs. Journal of Structural Engineering, v. 126, n. 2, Feb., p. 219-227, 2000. [05] SOARES, Y.V. Anlise Experimental de Lajes Cogumelo Nervuradas de Concreto Armado com Pilares Metlicos. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, UFSC, Florianpolis, 2003. [06] SCHWETZ, P. F. Anlise terico-experimental de uma laje nervurada modelo reduzido sujeita a um carregamento linear. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, UFRGS, Porto Alegre, 2005. [07] SCHWETZ, P.F. Anlise Numrico-Experimental de Lajes Nervuradas Sujeitas a Cargas Estticas de Servio. Tese (Doutorado em Engenharia) Escola de Engenharia, Universidade Federal do Rio Grande do Sul, Porto Alegre, 2011. [08] COMPUTERS AND STRUCTURES, INC. CSI Analysis Reference Manual. Berkeley: University Avenue, 2010. [09] TQS Informtica Ltda. (a) . Sistemas Computacionais Engenharia Estrutural: CAD/Formas. So Paulo, s.d. vrias paginaes. [10] ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. NBR 6118:2003 : projeto e execuo de obras de concreto armado. Rio de Janeiro, 2004. [11] PARK R.; PAULAY T. Reinforced Concrete Structures. University of Canterbury, Christchurch, New Zeland, 1975.
5. Conclusions
This study presented a comparative analysis between numerical and experimental results obtained for a reinforced concrete waffle slab designed to serve as a floor for tennis courts. The results indicate that, with the two numerical methods used, grid matrix analysis and finite elements, there are still issues that require further studies. The results also show the importance of undertaking experimental studies with the aim of verifying and adjusting the numerical models, particularly when new elements are adopted. The main conclusions are: 1. The results of the numerical analysis of the waffle slab system, obtained through the grid analogy, represented here by the TQS software, show that modeling by this method is very effective, as long as the actual values for the moment of inertia are used with no reduction in the ribs, slab and beams, and rigid connections between the structural elements. 2. The finite element model, conceived in the SAP2000 software, proved to be equally adequate. However, when compared with the simplicity of an analogy to grids, the FE model proves to be very
Figure 15 Moments: SAP, TQS and experimental: bending moment along line A shown in Fig. 6
382
Numerical and experimental study of a waffle slab designed to serve as a tennis court floor Anlise numrico-experimental de uma laje nervurada projetada como piso de uma quadra de tnis
P. F. SCHWETZ a paulete.schwetz@ufrgs.br F. P. S. L. GASTAL b francisco.gastal@gmail.com L. C. P. SILVA F c lcarlos66@gmail.com
Abstract
Reinforced concrete waffle slabs have become a common option for designers due to a need of rationalization in construction with reduction in costs and deadlines. To better understand the behavior of this structural system, and more realistically quantify stresses and displacements, a full scale waffle slab was tested. The structure, designed to serve as a tennis court floor, was submitted to a load of 12 kN/m2 and instrumented to measure strains and deflections at different locations. The loading process used the floors constructive base filling material and readings were taken at different loading stages and arrangements during the floors construction. Test data was compared to results obtained from the matrix analysis program Sistema Computacional TQS v11.0 and from the finite element model program SAP2000 v14.2.2. Slab behavior was as expected, with deflection and bending moments close to those determined by the numerical analysis. Keywords: waffle slab, reinforced concrete, matrix analysis of grids, finite element method.
Resumo
A necessidade de racionalizao na construo civil com a minimizao de custos e prazos vem fazendo das lajes nervuradas de concreto armado uma opo cada vez mais adotada pelos projetistas. Com o objetivo de compreender melhor o comportamento desse sistema estrutural e quantificar as tenses e deslocamentos da estrutura de uma forma mais realista, foi realizado um programa experimental em uma laje nervurada em escala natural. O estudo mediu deformaes no concreto e deslocamentos verticais em diversos pontos da estrutura, submetida a um carregamento de 12 kN/m2, correspondente a vrias camadas de enchimento necessrias para servir de base para quadras de tnis. Devido impossibilidade de aplicao do carregamento experimental de forma regular e em etapas, optou-se pela realizao de leituras em diferentes situaes de carga. Os dados coletados experimentalmente foram comparados com os resultados de duas anlises computacionais, empregando-se um modelo de anlise matricial de grelhas do programa Sistema Computacional TQS v11.0 e um modelo em elementos finitos do programa SAP2000 v14.2.2. O comportamento da laje esteve dentro do previsto, com deslocamentos verticais e momentos fletores prximos aos determinados pelas anlises numricas. Palavras-chave: laje nervurada, concreto armado, anlise matricial de grelhas, elementos finitos.
Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, paulete.schwetz@ufrgs.br, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil; b Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, francisco.gastal@gmail.com, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil. c Programa de Ps Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, lcarlos66@gmail.com, Av. Osvaldo Aranha, 99 Trreo, CEP 90035-190, Porto Alegre, Brasil.
a
Received: 10 Jul 2012 Accepted: 22 Feb 2013 Available Online: 12 Jun 2013
2013 IBRACON
1. Introduo
Solues estruturais sofisticadas e racionais so exigncias crescentes no cotidiano de projetistas de estruturas, como conseqncia da evoluo dos projetos arquitetnicos e dos novos conceitos de gerenciamento das construes. Surge, dessa forma, a necessidade de avaliao de diversos sistemas estruturais, buscando a soluo que melhor represente viabilidade econmica, rapidez e versatilidade de aplicao. Segundo Dias R.[1], em estruturas de concreto armado de vrios pisos, as lajes so responsveis por elevada parcela do consumo de concreto. No caso da opo por lajes macias, esta parcela chega a quase dois teros do volume total da estrutura. Desta forma, torna-se importante um estudo para definir o tipo de laje a ser empregado em edifcios de vrios pisos, buscando-se obter solues tecnicamente viveis e mais econmicas. Dentro desta tica, a utilizao de lajes nervuradas tem se firmado como uma opo bastante interessante. Este sistema estrutural pode ser definido como um conjunto de nervuras, distribudas em uma ou mais direes, regularmente espaadas e integralizadas por uma capa de concreto. A estrutura em laje nervurada foi concebida a partir da eliminao da maior parte do concreto da laje macia, abaixo da linha neutra, criando vazios em um padro rtmico de arranjo. Desta forma, ocorre uma reduo do peso prprio da estrutura e um aproveitamento mais eficiente dos materiais. Porm, o comportamento de lajes nervuradas de concreto armado ainda suscita algumas dvidas devido a sua complexidade, o que tem levado alguns pesquisadores a realizar estudos numricos e experimentais (Ajdukiewicz & Kliszczewicz, 1986 [2]; Selistre, 2000 [3]; Abdul-Wahab & Khalil, 2000 [4]; Dias, 2003 [1]; Soares, 2003 [5]; Schwetz, 2005 [6]; Schwetz, 2011 [7]). No Brasil, tem se verificado um aumento gradativo na utilizao de lajes nervuradas, embora a preferncia ainda seja pelo modelo de lajes planas para estruturas de concreto armado. Entretanto, necessrio desenvolver procedimentos e normas que orientem a utilizao deste
Figura 2 Posio genrica das armaduras na nervura
sistema construtivo. De fato, dado o carter recente do desenvolvimento de algoritmos numricos que possibilitam o clculo deste tipo de estrutura, persistem vrias dvidas relativas quantificao dos esforos atuantes nas lajes e sua adequada reproduo pelos mtodos de clculo e modelos matemticos utilizados atualmente.
Figura 3 Geometria da laje nervurada
Figura 1 Sistema de formas e escoramento
384
O presente trabalho, que parte de uma pesquisa em desenvolvimento [7], tem como objetivo aprofundar o conhecimento sobre lajes nervuradas de concreto armado e verificar se seu comportamento adequadamente simulado pelos mtodos de clculo e modelos matemticos amplamente utilizados atualmente. Com o intuito de atingir este objetivo, foi realizado um programa experimental, onde foram medidas deformaes especficas e deslocamentos verticais em uma laje nervurada em escala natural, projetada para receber vrias camadas de enchimento necessrias para servir de base para quadras de tnis. Os dados obtidos experimentalmente foram comparados com os resultados de anlises numricas desta estrutura, utilizando dois mtodos computacionais distintos para representar a estrutura: a anlise matricial de grelhas e o mtodo dos elementos finitos.
2.1 Geometria e carregamento

A forma final do pavimento, com suas respectivas propriedades geomtricas, pode ser vista na Figura 3. Nesta laje, o projetista adotou um vigamento de borda com o objetivo de causar enrijecimento para evitar deformaes excessivas, sendo as vigas semi-invertidas em relao capa. Alm disso, na regio circundante aos pilares internos, foi considerada uma regio em laje macia com altura de 37,5 cm, devido a puno e momentos fletores elevados. O carregamento adotado no projeto original previa, alm do peso prprio da estrutura de 4,8 kN/m2, uma carga permanente de 12 kN/m2, uma carga varivel de 3,0 kN/m2 e uma carga de alvernaria aplicada diretamente sobre o vigamento de borda. O valor elevado da carga permanente deve-se a vrias camadas de enchimento necessrias para servir de base para as quadras de tnis. O ensaio foi realizado apenas com a aplicao da carga permanente.
2. Programa experimental
O pavimento analisado neste caso de estudo foi projetado comercialmente por uma empresa de projeto e clculo de estruturas na cidade de Porto Alegre, Brasil, para para servir de base para quadras de tnis. A frma para moldagem da laje foi montada com o sistema de cubetas recuperveis para lajes nervuradas desenvolvido pela empresa Ulma Brasil Frmas e Escoramentos Ltda. Este sistema composto por uma estrutura modular que serve de apoio para as frmas plsticas. A estrutura modular apoiada por um conjunto de escoras metlicas tubulares de fcil desmontagem (Figura 1). A armadura das lajes foi elaborada com aos tipo CA-50. O peso da armadura positiva utilizada ficou em torno de 16.000 kg e a quantidade de armadura negativa foi de aproximadamente 23.000 kg. A Figura 2 mostra a posio genrica da ferragem nas nervuras. Pode se observar que, alm das armaduras positivas e negativas previstas pela anlise numrica, o projetista da estrutura colocou na parte inferior da capa uma tela soldada de ao CA-60, considerando a possibilidade de esforos de trao nas fibras inferiores da capa nos vazios entre nervuras.
2.2 Instrumentao
O programa experimental mediu valores de deformaes especficas e deslocamentos verticais, apenas devido aos carregamentos de ensaio e, portanto, sem considerao do peso prprio da laje nervurada. A instrumentao para determinao das deformaes especficas no concreto e no ao foi realizada empregando-se extensmetros eltricos unidirecionais, colados em 4 pontos da estrutura. O concreto foi instrumentado nas duas faces, superior e inferior (Figura 4a). J no ao, para reduzir a possibilidade de perda de informaes, por falha nos extensmetros, optou-se por instrumentar cada seo com dois extensmetros. Para a proteo destes, durante o processo de concretagem, utilizou-se resina base de epoxi (Figura 4b). Com o objetivo de medir os deslocamentos verticais, foi utilizado um nvel ptico de preciso (Figura 5). A locao dos pontos de controle de deslocamento pode ser vista na Figura 6.
Figura 4 Posio dos extensmetros (a) no concreto e (b) na armadura
CS
CI
A B
385
Figura 5 (a) Leitura do deslocamento com nvel ptico e (b) Rgua posicionada durante a leitura
A 2.3 Ensaio da estrutura Figura 6 Planta de locao dos pontos de instrumentao
O ensaio comeou a ser realizado 63 dias aps a concretagem da estrutura. O carregamento deveria ser realizado em 3 etapas e, a cada etapa de carga finalizada, seriam feitas as leituras de deformao especfica e medidos os deslocamentos verticais. Porm, a carga foi sendo distribuda de maneira desigual na laje ao longo do tempo, por questes relativas ao funcionamento da obra. Alm disso, algumas reas da laje foram utilizadas para armazenar outros tipos de materiais, gerando um carregamento no uniforme (Figuras 7a e b). Desta forma, optou-se por fazer leituras ao longo do tempo, tendo-se tomado o cuidado de fazer o registro da carga real aplicada no instante das leituras. O processo de carregamento da estrutura durou 87 dias, com um total de 5 leituras. Foram realizados, tambm, ensaios complementares para a determinao do mdulo de elasticidade longitudinal do concreto (E ) e da resistncia caracterstica (f ck). O valor mdio do mdulo de deformao longitudinal do concreto medido experimentalmente foi E 28 = 28,45 GPa e a resistncia mdia obtida aos 28 dias foi f cj = 33,27 MPa, correspondendo uma resistncia caracterstica estimada de f ck, estimado = 30 MPa.
3. Anlise numrica
Neste estudo, a anlise numrica foi realizada empregando -se dois programas computacionais tradicionalmente utili zados no projeto e anlise de lajes nervuradas de concreto armado, adotando-se os valores medidos experimentalmen te para o mdulo de elasticidade longitudinal e resistncia compresso do concreto. Os resultados foram comparados com os dados experimentais obtidos por Schwetz (2011) em uma estrutura real com o objetivo de verificar o quanto os modelos matemticos encontram-se prximos do comporta mento real da laje.
386
Figura 7 (a) Laje durante o processo de carregamento e (b) Laje totalmente carregada
A 3.1 SAP2000
A primeira anlise numrica foi realizada adotando-se o programa computacional SAP2000, que utiliza elementos finitos. O SAP2000 um programa para engenharia estrutural que efetua anlises estticas lineares e no-lineares, bem como anlises dinmicas de diversos tipos de estruturas, simulando seu comportamento quando sujeitas s mais variadas solicitaes [8]. O programa possui uma biblioteca de elementos finitos. Dentre as opes existentes, os elementos que foram utilizados nos mode-
B
los numricos propostos neste trabalho esto descritos a seguir. 3.1.1 Elemento de Casca (Shell) O elemento de Casca (Shell) um elemento de rea, utilizado para modelar cascas, membranas e placas em estruturas no plano e em trs dimenses. O programa SAP2000 define dois tipos de elemento Shell. Eles so denominados Homogeneous Shell, utilizado para materiais homogneos e Layered Shell, adotado quando o elemento formado por material hetergeneo ou por mais de um material. Porm, o programa s permite a realizao de anlise no linear quanto o elemento Layered Shell. Cada elemento possui seu sistema de coordenada local para a definio de propriedades de material, cargas e interpretao de resultados.
Figura 8 Modelo em elementos finitos
Figura 9 Deslocamentos verticais do ponto de instrumentao PC relativas s cargas aplicadas no programa SAP2000
387
As tenses, foras internas e momentos so determinados pelo mtodo da quadratura gaussiana e extrapolados para os ns. 3.1.2 Elemento de Barra (Frame) O elemento Frame um elemento de barra tridimensional com seis graus de liberdade por n: trs graus de liberdade translao e trs graus de liberdade rotao. utilizado para modelar vigas, pilares, trelias e grelhas em estruturas no plano e no espao. O elemento modelado como uma linha reta conectada por dois pontos. Cada elemento possui seu prprio sistema de coordenadas local para a definio das propriedades dos materiais e da seo, carregamento e interpretao dos resultados da anlise. A formulao do elemento Frame inclui a determinao dos esforos de flexo biaxial, toro, deformao axial e cisalhamento, a partir da integrao das tenses ao longo da seo. Estes esforos so determinados nas extremidades de cada elemento e em sees ao longo do comprimento escolhidas pelo usurio. No modelo em EF da laje nervurada em estudo, que pode ser visualizado na Figura 8, a capa, as nervuras, as vigas de borda e os macios foram modelados com elementos de casca, totalizando 23.000 elementos aproximadamente, e os pilares com elementos de barra. Foi realizada uma anlise no linear da estrutura, aplicando-se trs tipos de carregamento, em seqncia: peso prprio, carga de alvenaria sobre as vigas e carga experimental. Os resultados numricos correspondentes ao carregamento experimental foram considerados como sendo a diferena entre os valores obtidos na etapa inicial e final da carga experimental (Figura 9).
Figura 10 Grelha gerada pelo sistema computacional TQS
3.2 Sistema computacional TQS

O Sistema Computacional TQS um programa desenvolvido pela TQS Informtica Ltda. para projeto de estruturas de concreto armado, que utiliza o mtodo da anlise matricial de grelhas. Este
Figura 11 Carregamento aplicado na laje durante o programa experimental
388
Figura 12 Linha deformada SAP + TQS + EXPERIMENTAL: plano de corte A referente Figura 6
Figura 13 Linha deformada SAP + TQS + EXPERIMENTAL: plano de corte B referente Figura 6
sistema, formado por um conjunto de mdulos, monta uma base de dados de toda a edificao, agrupando geometrias e carregamentos, gerenciando a anlise da estrutura, a transferncia e o clculo de esforos e o detalhamento de todos os elementos constituintes da estrutura[9]. No caso de lajes nervuradas, o Sistema Computacional TQS permite que a anlise da grelha seja linear ou no-linear. Na anlise linear, as armaduras no so consideradas na determinao da rigidez das sees transversais. A anlise no linear calcula os deslocamentos verticais, bem como a modificao das solicitaes, considerando a armadura adotada e a no linearidade fsica do concreto por conta da fissurao, com base nas recomendaes da norma brasileira NBR 6118:2003 (ABNT, 2004[10]). Este programa computacional permite que o projetista da estrutura defina os valores de alguns critrios de projeto utilizados na determinao de esforos e detalhamento de armaduras. No caso de lajes nervuradas, o usurio pode modificar a inrcia flexo e/ ou toro das barras, assim como definir a rigidez das ligaes entre vigas e pilares. Neste caso de estudo, entretanto, nenhuma reduo de inrcia flexo e/ou toro foi considerada entre as nervuras e as vigas de borda. Alm disso, a rigidez das ligaes entre as vigas de borda e os pilares tambm no sofreu reduo. Esta deciso foi tomada a partir da anlise paramtrica realizada em uma laje re-
tangular, onde foram variadas as rigidezes dos elementos, tendo como balizador o programa de elementos finitos[7]. Fornecidos os dados, o programa gerou automaticamente o modelo numrico da laje, como pode ser visualizado na Figura 10. Aqui, foi tambm realizada uma anlise no linear da estrutura. Esse programa, porm, no caso da anlise no linear, somente permite a considerao nica da soma de todas as cargas aplicadas na estrutura, incluindo, neste caso, o peso prprio. Essa carga global pode, entretanto, ser divida em um nmero de etapas a ser definido pelo usurio. Neste estudo, os resultados provenientes apenas da considerao do carregamento experimental foram obtidos pela diferena entre duas diferentes anlises: uma aplicando a soma dos carregamentos do peso prprio, da carga permanente e da carga de alvenaria aplicada diretamente sobre as vigas de borda e outra adotando apenas o peso prprio e carga de alvenaria sobre as vigas.
4. Anlise dos resultados

Neste trabalho, so apresentados os resultados obtidos nas anlises numricas e medidos experimentalmente referentes carga total aplicada, como indicado na Figura 11. As linhas deformadas e diagramas de momento fletor so formados por pontos pertencentes aos planos de corte A e B indicados na Figura 6.
Figura 14 Deformaes especficas adotadas para (a) Experimental Estdio I s inferido, (b) Experimental Estdio I s medido e (c) Experimental Estdio II
389
4.1 Deslocamentos verticais

As Figuras 11 e 12 apresentam linhas deformadas, obtidas a partir das anlises no lineares da estrutura, bem como os valores obtidos experimentalmente. Verifica-se um comportamento semelhante da estrutura entre os resultados numricos dos dois modelos entre si e com os experimentais, indicando que ambos os modelos numricos representam, adequadamente, o comportamento deformado da laje.
4.2 Momentos fletores

Para obter-se a comparao de momentos fletores, numricos e experimentais, as deformaes medidas deram subsdio ao clculo de momentos, atravs do equilbrio interno de foras nas sees instrumentadas [11]. Foram calculados trs diferentes valores para o momento fletor em cada seo, ilustrados na Figura 14 e denominados como segue: n Experimental Estdio I s inferido: Considera a seo no fissurada (Estdio I), com a deformao especfica da armadura inferida a partir do valor medido no concreto. n Experimental Estdio I s medido: Considera a seo no fissurada (Estdio I), porm com a deformao efetivamente medida na armadura. Este presume a possibilidade de um princpio de fissurao da estrutura no entorno da seo e considera a contribuio do concreto tracionado entre fissuras (efeito tension stiffening). n Experimental Estdio II: Considera a seo fissurada (Estdio II). A Figura 15 apresenta as distribuies de momentos fletores, obtidas a partir dos resultados dos dois modelos numricos, comparados com os valores obtidos experimentalmente para o ponto E1. Os dois modelos numricos apresentam uma forma de distribuies de momentos fletores muito similares, indicando flexo positiva mxima na rea de carregamento e momento negativo mximo na regio do macio. No ponto E1, zona de mxima solicitao, o momento obtido experimentalmente - Experimental Estdio I s medido coincide com os valores previstos pelos modelos numricos, indicando um possvel incio de fissurao nas cercanias desta seo, diminuindo a capacidade resistente do concreto, sendo os esforos transferidos para a armadura.
Os dois modelos indicaram momento negativo na ligao das nervuras com as vigas de borda. O valor indicado pelo modelo em EF, entretanto, superior. Esta diferena deve-se, provavelmente, ao fato de que esse modelo considera a capa da laje como uma placa continua, gerando um maior efeito de engastamento na viga de borda, em comparao ao modelo de grelhas. Alm disso, o modelo em EF apresenta picos de momentos negativos, nas regies dos macios, bem mais pronunciados. Esta divergncia pode estar associada diferente modelagem dos macios existente nos dois programas. Nestas posies, infelizmente, no foram coletados dados experimentais que permitissem a verificao de qual dos programas possui uma modelagem mais adequada representao dessas ligaes e da influncia dos macios.
5. Concluses
Este trabalho apresentou uma anlise comparativa entre resultados numricos e experimentais obtidos para uma laje nervurada de concreto armado, projetada para servir de base para quadras de tnis. Os resultados indicam que, apesar dos mtodos numricos utilizados, adotando a anlise matricial de grelhas e o mtodo dos elementos finitos, apresentarem bons resultados, existem questes que ainda precisam ser melhor estudadas. Os resultados mostram, ainda, a importncia da realizao de estudos experimentais com o objetivo de verificar e ajustar os modelos numricos, principalmente quando so adotados novos elementos. As principais concluses verificadas neste trabalho so: 1. Os resultados das anlises numricas do sistema nervurado, obtidos a partir da analogia de grelha, aqui representado pelo Sistema Computacional TQS, mostraram que a modelagem atravs deste mtodo bastante eficiente, desde que adotadas as inrcias reais e considerado o vnculo de engastamento total entre os elementos estruturais. 2. O modelo em elementos finitos, concebido no programa computacional SAP2000, mostrou-se igualmente adequado. Todavia, comparado simplicidade de uma analogia a grelhas, o modelo em EF revela-se muito trabalhoso, demonstrando ser de difcil utilizao no dia a dia dos escritrios de projeto. 3. Ressalve-se, entretanto, que essa analogia a grelhas resulta em momentos negativos, tanto nas ligaes ao vigamento de borda como nas regies de macios, inferiores aos indicados pelo modelo em elementos finitos. Essa divergncia de resultados pode estar associada s diferentes modelagens utilizadas pelos dois programas. Torna-se, portanto, importante uma coleta de dados experimentais que permita definir qual das duas modelagens melhor representa o comportamento de um sistema nervurado.
Figura 15 Diagrama de momento fletor TQS + SAP + EXPERIMENTAL: plano de corte A referente Figura 6
6. Agradecimentos
Os autores agradecem ao CNPq e s empresas Projetak - Tavares Eng. Associados S/C Ltda e Constrututora Tedesco pelo apoio na realizao desta pesquisa.
7. Referncias bibliogrficas
[01] DIAS, R. H. Anlise Numrica de Pavimentos de Edifcios em Lajes Nervuradas. Dissertao (Mestrado
390
em Engenharia) Escola de Engenharia de So Carlos, Universidade de So Paulo (USP), So Carlos, 2003. [02] AJDUKIEWICZ, A.; STAROLSKI, W. Reinforced- concrete slab-column stuctures. New York: Elsevier Science Publishers, 1986. [03] SELISTRE, S. L.C. Anlise Terico-Experimental de uma Laje Nervurada de Microconcreto Submetida a um Carregamento Uniformemente Distribudo. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, Universidade Federal do Rio Grande do Sul, Porto Alegre, 2000. [04] ABDUL-WAHAB, H. M. S.; KHALIL, M. H. Rigidity and Strength of Orthotropic Reinforced Concrete Waffle Slabs. Journal of Structural Engineering, v. 126, n. 2, Feb., p. 219-227, 2000. [05] SOARES, Y.V. Anlise Experimental de Lajes Cogumelo Nervuradas de Concreto Armado com Pilares Metlicos. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, UFSC, Florianpolis, 2003. [06] SCHWETZ, P. F. Anlise terico-experimental de uma laje nervurada modelo reduzido sujeita a um carregamento linear. Dissertao (Mestrado em Engenharia) Curso de Ps-Graduao em Engenharia Civil, UFRGS, Porto Alegre, 2005. [07] SCHWETZ, P.F. Anlise Numrico-Experimental de Lajes Nervuradas Sujeitas a Cargas Estticas de Servio. Tese (Doutorado em Engenharia) Escola de Engenharia, Universidade Federal do Rio Grande do Sul, Porto Alegre, 2011. [08] COMPUTERS AND STRUCTURES, INC. CSI Analysis Reference Manual. Berkeley: University Avenue, 2010. [09] TQS Informtica Ltda. (a) . Sistemas Computacionais Engenharia Estrutural: CAD/Formas. So Paulo, s.d. vrias paginaes. [10] ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. NBR 6118:2003 : projeto e execuo de obras de concreto armado. Rio de Janeiro, 2004. [11] PARK R.; PAULAY T. Reinforced Concrete Structures. University of Canterbury, Christchurch, New Zeland, 1975.
391
Nonconventional concrete hollow blocks evaluation by destructive and non-destructive testing Avaliao de blocos de concreto vazados no convencionais por ensaios no destrutivos e destrutivos
M.S. RODRIGUES a michatcsr@yahoo.com.br G.C.S. FERREIRA b gisleiva@ft.unicamp.br L. SHIROMA c leandro@recicon.com.br A.L. BERALDO d beraldo@feagri.unicamp.br
Abstract
The aim of this study was to evaluate cementitious matrices properties by partial replacement of Portland cement by silica fume (SF) or by rice husk ash (RHA), and their application in nonbearing hollow blocks, tested by destructive and non-destructive methods. The following mixtures were produced: reference (100% of Portland cement) and Portland cement replacement (10% by mass) with SF or RHA. The non-destructive testing showed that the highest values of UPV were obtained for SF-based blocks and RHA-based blocks. The destructive test showed better results for SF-based blocks, but there was no statistical difference between the RHA-based and control ones. Keywords: concrete hollow blocks, rice husk ash, pozzolan, non-destructive testing.
Resumo
O objetivo desse estudo foi o de avaliar as propriedades de matrizes cimentcias com a substituio parcial do cimento Portland por slica ativa (SA) ou cinza de casca de arroz (CCA) e suas aplicaes em blocos de concreto no estruturais, avaliados por mtodos destrutivos e no destrutivos. As seguintes misturas foram produzidas: referencia (100% de cimento Portland), e mistura de cimento com SA ou CCA (10% de substituio, em massa). O ensaio no destrutivo mostrou que maiores valores de velocidade do pulso ultra-snico (VPU) foram dos blocos com SA e com CCA. No ensaio destrutivo, os melhores resultados foram obtidos para os blocos com SA, mas no h diferena estatisticamente significativa entre os blocos com CCA e os blocos controle. Palavras-chave: blocos de concreto vazados, cinza de casca de arroz, pozolana, ensaio no destrutivo
Agricultural Engineering, State University of Campinas, michatcsr@yahoo.com.br. Av. Candido Rondon, 501, Cidade Universitria Zeferino Vaz, zip code 13083-875, Baro Geraldo, Campinas, So Paulo, Brazil b Faculty of Technology, State University of Campinas, gisleiva@ft.unicamp.br. Rua Paschoal Marmo, 1888, zip code 13484-332, Limeira, So Paulo, Brazil c Agricultural Engineering, State University of Campinas, leandro@recicon.com.br. Av. Candido Rondon, 501, Cidade Universitria Zeferino Vaz, zip code 13083-875, Baro Geraldo, Campinas, So Paulo, Brazil d Agricultural Engineering, State University of Campinas, beraldo@feagri.unicamp.br. Av. Candido Rondon, 501, Cidade Universitria Zeferino Vaz, zip code 13083-875, Baro Geraldo, Campinas, So Paulo, Brazil
a
Received:01 Nov 2012 Accepted: 07 Mar 2013 Available Online: 12 Jun 2013
2013 IBRACON
M.S. RODRIGUES | G.C.S. FERREIRA | L. SHIROMA | A.L. BERALDO
1. Introduction
Concrete hollow blocks applications are common in many countries. However, the manufacture of these blocks does not provide satisfactory technology control due to the need for quick release of the lots. Brazilian and international standards (ABNT NBR6136 [1], ASTM C55-11 [2], BS 6073/81 [3]) describe hollow blocks compressive test at 28 days old, but this period can be considered long for practical purposes. Therefore, the need for more rapid testing enables the application of non-destructive testing (ultrasound) for materials characterization. Prassianakis & Prassianakis [4] evaluated the ability of non-destructive testing to determine the integrity of concrete specimens, and they concluded that this methodology is feasible for blocks evaluation. Hamidetal.[5] performedultrasoundtestsonhigh performance concrete (HPC), with the addition ofsilica fume,on cubic specimens of150mmside.After that, the specimens weretested in compression. Results confirmed that ultrasonic pulse velocity (UPV)change was more evident between 1 and14 days and that UPV increases only slightly at older ages(28,56 and90 days old). Moreover, there is an actualneed toadopt practices tominimize environmental impacts withthe useof admixturesinPortland cement(SNIC [6]).The use of ash from agro-industrial and pozzolanic materials allow reducing the Portland cement consumption, thus making cement industry more sustainable. Additionally, cement replacement by ashes also improves the mechanical characteristics and the durability of cementitious matrices. According to Mehta and Monteiro [7], the highly reactive pozzolan, even when present in small contents (30%) in a cementitious matrix, is able to consume almost completely calcium hydroxide, thus improving not only the mechanical strength but also the resistance to acid and sulfate attacks. Researchers studied the use of silica fume in concrete, comparing the results of destructive and non-destructive testing (ultrasound). They observed that the mixtures with higher percentage of silica fume (30%) had smaller UPV when compared with mixtures with 10% and 20% silica fume. This can be justified because, with silica fume utilization, concrete becomes less dense, despite showing the same compressive strength (Hamid [5]). Cement-based materials are heterogeneous mixtures, and the evolution of its strength over time is not properly understood. However, the use of non-destructive methods, such as the ultrasound monitoring of structural changes, could be a quick and efficient way to predict material behavior (Beraldo [8], Teixeira [9], Beraldo [10]). The aim of this study was to evaluate cementitious matrices properties by partial replacement of Portland cement with SF or RHA, and to analyze their applications in nonbearing hollow blocks, tested by destructive and non-destructive methods.
2. Materials and experimental program

2.1 Raw materials
It was employed a residual rice husk ash (RHA), obtained by noncontrolled burning process in rice industry located in So Paulo State, Brazil. Silica fume (SF) was used as reference. These residues were used without any treatment. Concrete hollow blocks were manufactured with ordinary Portland cement (CPV-ARI-40, according to Brazilian standard NBR 5733 [11]), natural sand, gravel and a plasticizer additive.
2.2 Chemical and physical characterization of RHA and SF

The chemical composition of residues was determined by X-ray fluorescence (XRF), and the loss on ignition (LOI) was determined by ASTM C-114 [12]. Mineralogical characterization of the ashes was carried out by X-ray diffractions. Granulometric size distributions of ashes were performed using a Malvern Mastersizer 2000 apparatus, which allows an analysis of particles by laser diffraction from 0.02 to 2000 m, in liquid mode as dispersant, with 10% - 15% of obscuration and ultrasonic agitation during 60 s.
2.3 Pozzolanic activity evaluation by electrical conductivity

The development of the reaction of pozzolan/calcium hydroxide produces insoluble products, therefore the concentration of CH [Ca(OH)2] in the solution decreases.The decrease of Ca2+ ions due to the increase phase of CSH (calcium silicate hydrate) produces a decrease in conductivity, which rate depends if the reactivity of the pozzolan is higher or lower (Villar-Cocia [13]).Several authors (Luxn [14], Tashiro [15], Pay [16]) applied the electrical conductivity test to evaluate the pozzolanic activity of a material.This test involves taking measurements of the electrical conductivity of a solution of calcium hydroxide and pozzolan, or those of RHA. Thus, a qualitative analysis is obtained by means of an electrical conductivity curve as a function of the time for RHA and for SF. It was added 200 mg of Ca (OH)2 in deionized water to obtain an unsaturated solution, at 60 C.After stabilization of the solutions conductivity, it was placed 5.25 g of ash into the solution, and electrical conductivity was obtained.Conductivity values were collected every 20 s, and the test duration was 24 h.
2.4 Concrete hollow blocks manufacturing

For the hollow blocks production, three difference mixtures were
Table 1 Ratios used in the mixtures (by mass)

Mixtures Control Cement/RHA Cement/SF Portland cement 1 0.9 0.9 Waste 0.1 0.1 Sand 7.6 Gravel 6.1 Water/binder 0.5 0.6 0.7 Plasticizer additive 0.03
393
Nonconventional concrete hollow blocks evaluation by destructive and non-destructive testing
Figure 1 Directions of the measurements: height, length and width
of age, according to Brazilian standard NBR 12118 [17]. Blocks were soaked in water at 23+5 C during 24 h, and then, they were weighed in saturated condition. Blocks absorption was calculated according to equation 1.
mh ms .100 ms
(1)
w (%) = water absorption; mh = block saturated mass (kg); ms = block dried mass (kg).
2.7 Compressive strength

The test was conducted in a universal testing machine model DL30000, with device displacement of 2 mm.min-1. The results of the compressive strength (MPa) were obtained by dividing the maximum load (N) by the area of section (mm2), obtaining the average of repetitions, in accordance with standard. Data were analyzed by statistical software (Statgraphics). Average comparison was performed using the Tukey test at 5% of significance level.
employed: a control (100% of Portland cement), blocks with RHA and blocks with SF. The percentage of replacement of Portland cement by ashes was 10%. The ratios used in mixtures are presented in Table 1. The materials were placed in a mixer, and after complete mixtures homogenization the mixture was pressed in a pneumatic machine Permac model MB0100. The pressure of the piston was 50 MPa. Hollow blocks were produced with dimensions of 390 x 190 x 140 mm, classified as M-15 block, according to Brazilian standard NBR 6136 [1]. The cure was performed in a laboratory environment with temperature range of 23+5 C.

3.1 Residues characterization
Chemical composition and LOI of the residues are shown in Table 2. RHA shows high content of silica, thus corroborating the analysis presented by others researchers (Krishnaro [18], Bui [19]). Potassium content was low for both samples, which is a favorable aspect, as the presence of potassium may cause problems with alkali-aggregate reaction in the cement matrix. However, LOI for RHA was high. According to the Brazilian standard (NBR 12653 [20]), pozzolan maximum content must be 6% of LOI. XRD patterns of SF and RHA are presented in Figures 2 and 3, respectively. As expected, crystalline phases were not identified for SF, and the amorphous material is evidenced. Peaks of cristobalite and quartz were observed at RHA diffractogram, showing the crystalline phases formation, probably as a consequence of non-controlled burning. The more crystalline ashes, the lower its reactivity (John [21]). Table 3 and Figure 4 show the particle size distribution and particle size curve, respectively. It is observed that SF has the smallest particles size when compared to RHA; however, RHA has 90% of its particles below 80 m, and it presents small average diameter. Materials with a high fineness may show high reactivity in contact with calcium hydroxide (NBR 12118 [17]). Moreover,
2.5 Ultrasonic pulse velocity (UPV)

The same hollow blocks were tested by nondestructive test and then by the destructive one. UPV was recorded once a week during 28 days before the destructive test. An Ultrasonic Tester Steinkamp, model BP7, with exponential transducers with a resonance frequency of 45 kHz was employed. Measurements were carried out in three directions of the blocks, height, length and width, as showed in Figure 1. Statistical analysis was conducted by the Statgraphics Software, version Centurion XV.I. UPVs data were evaluated statistically aiming to obtain correlations with the compressive strength allowing applications for the control of material quality.
2.6 Water absorption test

The blocks were subjected to the water absorption test at 28 days
Table 2 Chemical composition of RHA and SF

Na2O SF RHA 0.17 0.09 MgO 0.37 0.37 Al2O3 0.08 0.19 SiO2 95.2 86.8 P2O5 0.12 0.66 SO3 0.28 0.09 Cl 0.21 0.03 K2O 1.22 1.67 CaO 0.48 0.65 MnO 0.04 0.21 Fe2O3 0.10 0.09 LOI 1.68 9.12
394
Figure 2 XRD pattern of SF

1000
Figure 3 XRD pattern of RHA

5000
C
4000
3000
Counts
500
Counts
Q - quartz C - cristobalite
2000
1000 Q
0 0 10 20 30 40 50 60 70
CC
30
C C CC
40
C
50
0 0 10 20
C C CC C
60 70
Position (2 theta degrees)
Position (2 theta degrees)
Table 3 Particle size distribution

Samples RHA SF D10 (m) 4.725 3.945 D50 (m) 26.964 18.983 D90 (m) 79.686 42.466 Average (m) 35.489 21.315
the fine material fills the voids and presents the effect of a filler. With the combination of these two positive effects, a material might partially replace the cement with advantage, as it does not imply a large increase in water/binder ratio (Cordeiro [22]). In this work, it was observed a small increase of water/binder ratio in blocks produced with replacement (by SF or by RHA), due to the fineness of the materials.
3.2 Pozzolanic evaluation

In order to evaluate the reactivity of RHA, the electrical conductivity of a solution of CH/ash was measured, which was also carried out for SF. The curves are shown in Figure 5. It was observed that the CH/SF solution is highly reactive, with large decreases in conductivity in the first hours. This behavior was already expected, as SF is well known as pozzolan. The
Figure 4 Particle size distribution of samples Figure 5 Electrical conductivity curve of samples
395
Figure 6 UPV of concrete hollow blocks at the transverse direction (width)

Control RHA SF
Figure 7 UPV of concrete hollow blocks at the longitudinal direction (length)

Control RHA SF
2,6
2,6
2,4
2,4
UPV (Km/s)
UPV (Km/s)
2,2
2,2
2,0
2,0
1,8
1,8 0 7 14 21 28
14
21
28
Age (days)
Age (days)
CH/RHA curve shows the reactivity of this ash. It is observed that RHA reacts slower than SF; however, RHA presented low electrical conductivity, thus showing that it may indicate the formation of CSH products. After 12 h, for both samples, it was observed the electrical conductivity stabilization, with values close to 0.5 mS.cm-1.
3.3 Ultrasonic pulse velocity (UPV)

Figures 6, 7 and 8 show the ultrasonic pulse velocity (UPV) values across the concrete hollow blocks at three different directions, respectively width, length and height. At the width direction (Figure 6), UPV across the hollow blocks with SF was higher than for other hollow blocks, which was expected, as SF is commercially known as a pozzolan, which presents high reactivity. It was observed that, as age increases, there was a small UPV variation across the blocks. There was a statistically significant difference among the UPV for all of the blocks. At the length direction (Figure 7), it was observed a similar behavior with those from Figure 6. However, control hollow blocks and RHA-based hollow blocks showed no statistical difference between them, but both were statistically different when compared to the SF-based hollow blocks, which showed the highest values of UPV.
The same behavior was observed for the UPV at the height direction. UPV was higher across the SF-based hollow blocks. RHA-based and SF-based hollow blocks present higher values of UPV when compared to the control ones due to the filler physical effect, with the filling of voids in concrete produced by the different particles size. Statistical analysiscomparingthe UPV values at 7 days among the hollow blocks directions shows that there wasa statistical significant difference.For length and width, there is nostatistical difference; however, theyshowed statistical difference when comparedwith theheight direction.The same behavior was observed when comparing UPV at28 days. For the length and width directions, there is the effect of confinement of the wall mold, which does not happen to the height direction.
3.4 Water absorption

According to the Brazilian standard (NBR 12118 [17]), the maximum water absorption for concrete hollow blocks must be 15%. It is observed in Table 4 that the hollow blocks manufactured with SF or RHA showed low water absorption, when compared to the control ones. If pozzolanic reaction occurs, the products reaction formed are very effective in filling the capillary spaces, thus improving the mechanical strength and impermeability of the matrix (ASTM C55-11 [2]), as it was observed for the replacements tested.
Table 4 Water absorption of hollow blocks

Concrete hollow blocks Control Cement/RHA Cement/SF Water absorption (%) 11.56 8.97 8.67 Standard deviation 1,81 Coefficient of variation 17,22
396
4. Conclusions
Figure 8 UPV of concrete hollow blocks at the perpendicular direction (height)
Control RHA SF
2,6
2,4
2,2
2,0
1,8 0 7 14 21 28
Age (days)
The characterization of RHA showed high content of silica, low content of potassium and high fineness, thus denoting its reactivity. However, this ash presents high LOI content, and XRD showed a crystalline material with peaks of cristobalite and quartz. The pozzolanic assessment of RHA showed pozzolanic reactivity when evaluated by electrical conductivity in a CH /ash solution. Non-destructive evaluation of the concrete hollow blocks showed that the highest values of UPV were obtained across the SF-based hollow blocks for three different directions. Statistical analysis showed that, for length and height directions, the UPV for these hollow blocks presents statistically significant difference when compared to the others ones. RHA-based hollow blocks and SFbased hollow blocks presented lower water absorption when compared to the control ones. Despite SF-based hollow blocks showing the highest values of compressive strength, statistical analysis showed no statistically significant difference among the types of hollow blocks. Data correlation between UPV and compressive strength showed a small determination coefficient (0.38).
UPV (Km/s)
3.5 Compressive strength of concrete hollow blocks

As expected, the compressive strength of the SF-based hollow blocks was higher than the other ones (Figure 9). However, statistical analysis showed no statistically difference among the compressive strength of the hollow blocks from the three different mixtures employed (control, cement/SF, cement/RHA). Therefore, according to this test, it is possible to produce hollow blocks with partial replacement of Portland cement by RHA, without decreasing compressive strength. Based on the results of physical and mechanical tests, it was observed that, in addition to the residues acting as pozzolans, they may also perform as fillers, thus contributing to lower water absorption and, therefore, increasing the durability of the hollow blocks.
5. References
[01] Brazilian Technical Standard Association Rio de Janeiro. NBR 6136 Hollow concrete blocks for concrete masonry. Rio de Janeiro, Brazil, 2006. [02] American Society for Testing and Materials C55-11, Standard Specification for Concrete Building Brick. 2011. [03] BS 6073/81 Part 1 Precast concrete masonry units Specification for precast for concrete masonary units, 2011. [04] Prassianakis, I.N.; Prassianakis, N.I. Ultrasonic testing of non-metallic materials: concrete and marble. Theoretical and Applied Fracture Mechanics. Amsterdam, 42 (2004) 191-198.
3.6 Correlation between UPV and compressive strength

In order to obtain a correlation between UPV and compressive strength, a simple regression was performed. UPV data are related to height direction that was also considered for destructive test. The software used for regression analysis was Statgraphics. Despite a positive tendency of the compressive strength with respect to the UPV, the determination coefficient was only 0.38, thus denoting a small dependence of the variables. The linear model presented by the software is presented by equation 2.
Figure 9 Compressive strength of blocks
Rc 1,1571 2,2177 UPV

Rc = compressive strength (MPa); UPV = ultrasonic pulse velocity (km.s-1)
(2)
397
[05] Hamid, R.; Yosof, K.M.; Zain, M.F.M. A combined ultrasound method applied to high performance concrete with silica fume. Construction and Building Materials, vol. 24, 2010, p.94-98. [06] Information on http://www.snic.org.br [07] Mehta, P.K.; Monteiro, P.J.M. Concrete, structures, properties and materials. 1st .Ed., Pini, So Paulo, 1994, pp. 574. [08] Beraldo, A.L. Ultrasonic speed wave: cure type influence on wood-cement composite compression strength. In: Proceedings of the NDTISS99 International Symposium on Non Destructive Testing Contribution to the Infrastructure Safety Systems in the 21st Century, UFSM, Torres, RS, 1999, pp. 86-90, 1999. [09] Teixeira, D.E.; Molesmi, A. Assessing modulus of elasticity of wood-fiber cement (WFC) sheets using nondestructive evaluation (NDE). Bioresource Technology, 24 (2001) 193-198. [10] Beraldo, A.L.; Martins, S.C.F. Destructive and non-destructive evaluation of bamboo chips Portland cement composite. Journal of Bamboo and Rattan, 6 (2007) 61-70. [11] Brazilian Technical Standard Association Rio de Janeiro. NBR 5733 High early strength Portland cement. Specification. Rio de Janeiro. 1991. 5p. (In Portuguese). [12] American Society for Testing And Materials, Standard Test Method for Chemical Analysis of Hydraulic Cement: ASTM C114-03. 2003. Philadelphia. [13] Villar-Cocia, E.; Valencia-Morales, E.; Gonzlez-Rodrguez, R.; Hernndez-Ruz, J. Kinetics of the pozzolanic reaction between lime and sugar cane straw ash by electrical conductivity measurement: A kineticdiffusive model. Cement Concrete Research, 33 (2003) 517524. [14] Luxn, M.P.; Madruga,F.; Saavedra, J. Rapid evaluation of pozzolanic activity of natural products by conductivity measurement. Cement and Concrete Research, 19 (1989) 63-68 [15] Tashiro, C.; Ikeda, K.; Inoue, Y. Evaluation of pozzolanic activity by electric resistance measurement method. Cement and Concrete Research, 24 (1994) 1133-1139. [16] Pay, P.; Borrachero, M.V. ; Monz, J. Peris-Mora, E.; Amahjour, F. Determination of amorphous silica in rice husk ash by a rapid analytical method. Cement and Concrete Research, 31 (2001) 227-231. [17] Brazilian Technical Standard Association Rio de Janeiro. NBR 12118 Hollow concrete blocks for concrete masonry. Test method. Rio de Janeiro. 2007. 12p. (In Portuguese). [18] Krishnarao, R.V.; Subrahmanyam, J.E.; Kumar, T.J. Studies on the formation of black particles in rice husk silica ash. Journal of the European Ceramic Society, 21 (2001) 99-104. [19] Bui, D.D.; Hu, J.; Stroeven, P. Particle size effect on the strength of rice husk ash blended gap-graded Portland cement concrete. Cement & Concrete Composites; 27(2005) 357-366.
[20] [21] [22]
Brazilian Technical Standard Association Rio de Janeiro. NBR 12653 Pozzolans Specification. Rio de Janeiro. 1992. 4p. (In Portuguese). John, V.M.; Cincotto, M.A.; Silva, M.G.; Ashes and Non Conventional Binders. In: Tecnology and Non Conventional Materials for Constructions. FREIRE, W. J. and BERALDO, A. L. (Coord.). Campinas, SP: UNICAMP Press, 2003. 145-190. (In Portuguese). Cordeiro, G.C. Ultrafine ashes from sugar-cane bagasse and rice husk as mineral additive for concrete. D.Sc. Thesis, COPPE, Universidade Federal do Rio de Janeiro UFRJ, Rio de Janeiro, Brazil. 2006. (In Portuguese).
398
Study of the influential factors on the rheological behavior of adhesive mortar available in the market Estudo dos fatores influentes no comportamento reolgico de argamassas colantes de mercado
M. R. M. M. COSTA a mariennecosta@uol.com.br E. PEREIRA b engenheiroeduardopereira@gmail.com R. G. PILEGGI c rafael.pileggi@poli.usp.br M. A. CINCOTTO d cincotto@poli.usp.br
Abstract
Rheology studies the deformation and flow of matter and seeks to describe the deformations of material depending on the time they are exposed to mechanical actions. This paper proposes to understand the behavior of fresh mortar adhesives in Squeeze Flow tests based on rheological characterization of different commercial compositions. In the tests, flow of material results from the application of a compressive load on the sample in the fresh state which causes displacement due to stresses generated during radial shear flow. We performed further physical characterizations of mortars and viscosity on material passing the through a sieve with a 0.075 mm sieve opening with the objective of basing the analysis of the results obtained from Squeeze Flow tests. It was observed that the mortars studied showed differences in rheological behavior, probably resulting from synergistic action of some of the composition parameters, especially the particle size distribution and particle morphology. Keywords: rheology; squeeze flow test; particle size distribution; morphology; adhesive mortar.
Resumo
A reologia estuda a deformao e o escoamento da matria buscando descrever as deformaes dos materiais dependentes do tempo, quando os mesmos so expostos a aes mecnicas. O presente trabalho prope o entendimento do comportamento no estado fresco de argamassas colantes, com base na caracterizao reolgica de diferentes composies comerciais, a partir do ensaio Squeeze Flow. Neste ensaio, o escoamento do material decorre da aplicao de uma carga de compresso sobre a amostra no estado fresco, a qual ocasiona deslocamentos no seu interior, devido a esforos de cisalhamento radiais originados durante o fluxo. Foi realizada ainda a caracterizao fsica das argamassas e a viscosidade da frao passante na peneira da frao passante na peneira de abertura 0,075 mm, com o objetivo de embasar a anlise dos resultados obtidos com o Squeeze Flow. Observou-se que as argamassas estudadas apresentam diferenas de comportamento reolgico, decorrente provavelmente, de ao sinrgica de alguns parmetros da composio, com destaque para a distribuio granulomtrica e morfologia das partculas. Palavras-chave: reologia; squeeze flow teste; distribuio granulomtrica; morfologia; argamassa colante.
Departamento de Construo Civil da Universidade Federal do Paran UFPR, mariennecosta@uol.com.br, Curitiba, Brasil; Departamento de Construo Civil da Universidade Federal do Paran UFPR, engenheiroeduardopereira@gmail.com, Curitiba, Brasil; c Departamento de Engenharia Civil Escola Politcnicna da Universidade de So Paulo - USP, rafael.pileggi@poli.usp.br, So Paulo, Brasil; d Departamento de Engenharia Civil Escola Politcnicna da Universidade de So Paulo - USP, cincotto@poli.usp.br, So Paulo, Brasil.
a b
Received: 04 May 2012 Accepted: 24 Feb 2013 Available Online: 12 Jun 2013
2013 IBRACON
Study of the influential factors on the rheological behavior of adhesive mortar available in the market
1. Introduction
Adhesive mortars consist of cement, aggregate and other additives. The most common industrially produced additives are cellulose ethers (HEC e MHEC) and latex polymers (PVA). Each perform specific functions in fresh and hardened mortars [1]. Latex additives improve the rheological behavior of fresh mortars, but mainly provides flexibility and tensile strength for hardened mortars [2].Ethers are water soluble polymers and small amounts of polymer increases water retention and viscosity in mortars [3, 4]. Quartz sand, with dimensions less than 0.50 mm, is predominantly used in mortar. Characteristics, such as shape, texture and grain size, have a significant influence on the workability and adhesion of mortar [5]. The shape and texture of the grains are primarily responsible for changes in the friction coefficient. Mendes [6] demonstrated that the more rounded the particles, the lower the friction coefficient compared to those that are irregularly shaped. Pereira [7] found that the shape of particles larger than 0.090 mm significantly influence the flow phenomena of mortar and the more rough the particles, the greater the shear required for the flow of the mixture. Another issue related to the influence of particle size distribution is the degree of compaction. The ideal condition is to gradually fill the larger voids with smaller grains [8]. It should be noted, however, that you should not utilize an excess of fines as they may increase the surface area, hence increasing the amount of water needed to coat the particles. Sand with elevated fineness increases the water consumption in the mixture, which may cause significant shrinkage in the mortar after water evaporation, and fine sand may fill the pores of the substrates occupying the spaces of the hydration products [5, 9]. Moreover, an increase in the proportion of sand, maintaining the same amount of water in the mixture, decreases the workability of the mortar, thus influencing the development of adhesive strength of the coatings. It should also be considered that increased coarseness in the sand could cause inefficient distribution of the grains in the cement mixture leading to difficulty in mortar flow and compromising adhesion [9]. However, it should be emphasized that an adequate proportion of sand is essential for the formation of the undeformed skeletal structure of the mixture and to reduce mortar shrinkage. The fineness of cement has a significant influence in adhesive mortar. Different cements exhibit different physical characteristics. However, fineness is the most significant parameter [10]. The finer
the particles of cement, the more water and additives are required for the same workability. From the perspective of rheological behavior, it is assumed that mortar adhesives are concentrated suspensions of solid particles immersed in a viscous liquid which has the function of ensuring the cohesion of the system and providing lubrication and space for the movement of aggregates [8]. If the volume of the matrix is not sufficient, the rheological characteristics and properties of the mortar is harmed [11, 6]. With increased matrix content, fluid flows more easily and fluidity of the system will be governed by the viscosity of the matrix. Thus, to understand the rheology of the fluid and have a global view of rheological behavior, it is necessary to analyze the phenomena that occur in the matrix, such as those derived from aggregate [12]. To define the behavior of mortar in its fresh state we frequently use consistency parameter data obtained from tests Flow Table and the Dropping Ball tests [13]; the so-called monopoint tests. However, as easy as these tests are to perform, they present conceptual limitations in the characterization of mortars. It is not possible to describe the behavior of material by a single measurement. But, it should be measured by a rheological profile test that, preferably, simulates actual conditions since behavior can vary depending on the characteristics of the actual application [14]. To fill this limitation gap, the Squeeze Flow test was developed. It provides for the characterization of rheological behavior through viscosity and yield parameters and is the defined as standard test NBR 15839 [15]. Despite the aforementioned standard, it is not specific for the analysis of adhesive mortars. But, this methodology is being used successfully by various researchers [8,7,16] along with adaptations and adjustments to the test apparatus. Within the above-mentioned context, the objective of this work is to understand the behavior of fresh mortar adhesives based on rheological characterization of different Brazilian commercial compositions. An understanding of the rheological behavior of adhesive mortar is the basis for a technological breakthrough, considering that their application is in a fluid state.
2. Experimental program and materials

The experimental program was designed with the objective of correlating the physical characteristics of mortar adhesives in the market, such as particle size, particle morphology and air content, with their rheological behavior determined from Squeeze Flow test parameter characteristics.
Table 1 Detail of the preliminary data of the samples (Slip and water/dry materials ratio in the pastes and mortars)
A( G ) Amount of water indicated by the manufacturer for adhesive mortar (ml/ Kg) Proportion water/dry mix materials for paste (ml/g) Slip test results of mortar (mm)
(G) Good quality; (M) Medium quality; (B) Bad Quality
B( G ) 220 1.1 0.5
Adhesive mortar and paste C( M ) D( B ) E( M ) 200 0.9 0.7 230 0.9 3.4 230 1.2 1.1
F( B ) 200 0.9 5.9
230 1.1 0.5
400
M. R. M. M. COSTA | E. PEREIRA | R. G. PILEGGI | M. A. CINCOTTO
Figure 1 Schematic representation of the components of the Squeeze Flow Test
the No. 200 sieve (0.075 mm opening). Industrialized materials were also utilized in a reconstruction approximating the amounts of components in procedures recommended by Quarcioni [19]. The aggregate particles retained in sieve openings of 0.6, 0.3 and 0.15 mm were analyzed for their morphology in order to determine texture differences and form irregularities between the particles. The particles measuring above the 0.6 mm were not analyzed due to the small amount within this range. The particle study was performed using images acquired by a stereoscopic Zeiss, Stemi model 2000-C. For each mortar, a minimum of 100 grains were photographed in each sieve. Morphological analysis of the aggregate was performed utilizing an IMAGO program, version 2.2.4 loaned to us by the business ESSS[20]. The technique could not be applied to the portion of material passing through the 0.075 mm sieve opening because of the natural agglomeration of the grains. Their small size prevented them from appearing individually in the images. 2.2.2 Fresh state In the fresh state mortar was rheologically evaluated by the Squeeze Flow test, in addition to determining densities and the amount of incorporated air. In characterizing the rheological behavior of adhesive mortar in the Squeeze Flow test, the flow of material resulting from the application of a compressive load on the fresh state sample caused shifts in its interior due to the radial shear forces created during the flow. To execute this test the equipment used was a universal test machine commonly existing in building construction material laboratories for measuring application load sample as described in NBR15839 [15]. The details of the test employed are indicated in Figure 1. The movable punch has the same diameter as the sample in order to ensure the load application on the entire surface. In general, the rheological behavior expected for adhesive mortar, throughout the test, was increased compression load as it increases the displacement of the movable punch due to the approximation of the fraction larger than 0.075 mm in the mortar. In practice, comparative analysis between the mortar enables the identification of the ease of application by the mason and how it forms over the toothed trowel. The viscosity of the portion passing through the 0.075 mm sieve opening (i.e., pulp) was measured in order to determine the influence of the fine fraction (paste) on the rheological behavior of the mortar. This test was performed with a Brookfield Programmable Viscometer DV-II+ Version 5.0. To determine the amount of air incorporated in the mixtures, density tests were conducted in the fresh state, in addition to bulk density test in the dry state as previously mentioned. Bulk density in the fresh state was determined following the procedure proposed by NBR 13278 [21].
2.1 Definition of the sample universe

Initially, one bag each of thirty different brands of type AC-I adhesive mortar [17] was acquired from building material dealers throughout Brazil. Classification rules, based on manufacturers specifications, were considered for the type AC-I mortar. In order to reduce the initial sample size and provide a parameter for selecting the mortars used in the research, each mortar was subjected to a slip test, as determined by NBR 14085 [18]. In this test, the mortar was applied on a substrate pattern and the strings were formed in transverse direction to the substrate. Ceramic tiles were positioned within the strings and the substrate was placed in an upright position in order to observe for sliding of the tiles. Based upon the sliding results, six brands were chosen for the sample universe. Two brands were below the limit specified by NBR 14081 [17]. Two were at the limit. And, two were above the specified limit. The maximum sliding was specified at 0.7 mm. Table 1 reflects the sliding results of the mortar chosen. The relation of water/fine fraction dry material was obtained by mixing with the water content specified by the mortar manufacturer with respect to the proportion of fine fraction of each mortar. For example, if the amount of water indicated for mortar A was 230 ml/kg and the fine fraction was 21.8% of the mortar mass, then the ratio of water/dry material for mortar A would be equal to 1.1, or 230 ml/218 g. The water content used was specified by the manufacturer product packaging. Table 1 also reflects the amount of water recommended by the mortar manufacturers selected, in addition to the amount used.
2.2 Physical characteristics of the adhesive mortars

2.2.1 Anhydrous state Mortar was physically characterized in the anhydrous state by grain size and grain morphology. Bulk density was measured to assist in determining the amount of incorporated air. We determined the complete granulometric curves by using screening test, with the normal series of ABNT sieves, on the mortar and a laser particle size analyzer for the fraction passing through

Table 2 reflects the proportion of coarse and fine fractions (paste) adhesive mortar, or of material content larger than 75 micron and size of less than 75 micron, respectively. The fine fraction is represented mostly by cement grains and fillers and the coarse fraction by sand grains. Note that Sample D had a higher content of paste than Sample E.
401
Table 2 Reconstitution of the aproximation formulation of adhesive mortars

Adhesive mortar C D 74.67 17.77 7.55 0.16 70.67 17.90 11.44 0.15
A Aggregate (%) Cement (%) Limestone filler (%) Moisture (%) 74.75 15.86 9.39 0.13
B 75.80 20.70 3.50 0.10
E 80.27 17.86 1.87 0.16
F 74.31 19.52 6.17 0.15
Figure 2 Particle size distribution of adhesive mortars
Figure 3 Particle size distribution for particles finer than 75 m
Figure 2 reflects a diverse particle distribution in the mortars and Figure 3 illustrates a specific particle size distribution of fine fraction (particles smaller than 75 m). By analyzing the profile of the size distribution curves of the mortar (Figure 2), mortars A, C and D had narrow distributions, which tends to reduce the mean distance of separation between the particles, influencing the rheological behavior of the adhesive mortar. Opened distribution increases the flow benefits, since the smaller grains are more separated from the larger grains, thereby reducing the friction between them. From the data obtained from the sieve analysis, it can be stated that the mortar adhesives have their greatest variations when the particle size distribution is in the range 100 a 1000 m, repre-
sented by grains larger than 0.075 mm. Table 3 reflects the percentages retained and passed through the sieves with openings of 0.075, 0.15, 0.3 and 0.6 mm. It is possible to verify that there are considerable differences in the percentages retained in each sieve when compared to different mortars in sieves with openings of over 0.075 mm (coarse fraction). The fine fraction, in turn, showed a similar distribution, without significant differences between them, the only exception being Sample D which differed from the others by at least 5%. Table 4 reflects the morphological results of particles retained in the sieve openings of 0.15, 0.30 and 0.60 mm and their coefficients of roundness and the elongation of the grains. Table 5 reflects the bulk density results in the anhydrous and fresh mortar states along
Table 3 Percent of material retained and Percent of material pass on sieve

Sieve mesh 200 200 100 50 30 Adhesive mortar C D 22.8 12.2 47.0 13.0 5.0 27.2 15.6 51.2 5.4 0.6
mm % fine fraction (passing the sieve) % fine fraction (retained in the sieve) 0.075 0.075 0.15 0.30 0.60
A 21.8 31.2 42.2 4.3 0.5
B 20.9 12.7 35.8 28.1 2.5
E 19.4 31.8 24.8 22.9 1.1
F 21.8 9.5 33.1 32.8 2.8
402
Table 4 Morphological characterization of the coarse sand fraction in the adhesive mortars
Sieve mm mesh 0.15 mm 100 0.3 mm 50 0.6 mm 30 0.15 mm 100 0.3 mm 50 0.6 mm 30 Adhesive mortar C D 0.508 0.492 0.502 0.522 0.474 0.516 0.730 0.678 0.700 0.729 0.662 0.720
Roundness Sphericity
A 0.518 0.521 0.500 0.728 0.738 0.729
B 0.509 0.525 0.504 0.706 0.734 0.699
E 0.492 0.490 0.521 0.717 0.685 0.733
F 0.506 0.792 0.496 0.735 0.755 0.696
Table 5 Specific gravity and the air entrained content in the adhesive mortar
Adhesive mortar C D 2.60 2.65 1.55 1.75 22.9 15.7
Specific gravity Air entrained content
Dry Fresh Fresh
A 2.70 1.45 28.2
B 2.70 1.60 21.9
E 2.60 1.70 15.5
F 2.60 1.50 26.6
with the incorporated air content in fresh state mortar when using the manufacturer designated water content. In the results it was observed that the mortars had similar rounded particles (in the range of 0.474 to 0.521 mm) except for Sample F (fraction 0.3 mm sieve) which had superior rounding (0.792) compared to the other samples. It is also possible to verify that mortar clusters have similar elongation coefficients in the range of 0.662 to 0.755. In regard to the bulk density of an anhydrous mixture, the mortar exhibited variations in a very narrow range between 2.63 and 2.67 g/cm3. This can be attributed to the inherent variability of the test method (on the order of 1.5%). However, for the mortar analyzed, the differences resided in the fresh state bulk density which was in the range between 1.46 and 1.71 g/cm3. These differences may arise in chemical and physical changes caused by the reactivity of cement, additives and the mixing procedure, resulting in a distribution of constituents other than that present in the anhydrous state. The cement determines hydrated components in the hardened mixture. Additives, and the mixing procedure itself, will incorporate air into the fresh mixture. Throughout the tests, there was a strong correlation between fresh state bulk density and incorporated air content (R=0.9693). A variation in the level and type of additives can have a significant impact on the behavior of fresh mortar. It was not the purpose of this research to determine the quantitative data associated with these variations. Figure 4 reflects the viscosity of a mixture as a function of shear rate. The data obtained in the tests indicated that all mixtures had pseudoplastic behavior, i.e., there was a decrease in viscosity with a corresponding increase in shear rate. Although, with different mortar viscosities between 0 a 50 S-1, the data indicates a tendency of similar viscosities at higher shear rates similar to those in the Squeeze Flow tests. Thus, it was not considered
pertinent in this research to indicate viscosity as a variable in the assessment of the rheological behavior of mortar. The profile curves obtained in the Squeeze Flow test for adhesive mortar are reflected in Figure 5. The curves in Figure 5 show that adhesive mortar exhibits different rheological characteristics as indicated by each mortar having distinct compression loading absorption capacity when imposed by the movable punch equipment. Sample C had a higher compressive load as compared to the other samples which resulted in greater spreading difficulty and deformation of the strings. Sample F had greater fluidity as compared to the others which would result in easier spreading and formation of the strings but would probably result in excessive slippage of the ceramic tile. As it was not possible to obtain a parameter of effort
Figure 4 Relation between viscosities and shear rate of pastes
403
Figure 5 Curve shape of load versus displacement of adhesive mortars, obtained with the Squeeze Flow test
applied by the mason for determining spreading and bead formation, it cannot be stated that Sample C does not provide satisfactory spreading and bead formation. Probably the six studied mortars are acceptable for use in coatings. The analysis that was performed was comparative and the results obtained were restricted to the water/ dry material relationship indicated by the manufacturer. Table 6 highlights the analyzed parameters of this research and demonstrates the significant influence that particle size has on the rheological behavior of the adhesive mortars studied compared to the results obtained by the Squeeze Flow test. The numerous influential parameters are indicated in this table as high (a), intermediate (a) or low (a). Morphology is classified as either irregular or regular. When the mortar exceeds a parameter, it is indicated with a + sign. Comparing the Squeeze Flow test results to the data obtained from mortar property analysis, it was concluded that the primary influential parameter on the performance of the mortar used in this study was particle size distribution of grains larger than 0.075 mm. Calculating correlations between various properties analyzed, the percentage of aggregate retained between sieves with aperture 0.3 to 0.15 mm was shown to be more significant. This behavior might best be explained by the phenomenon of imbrication of grains, i.e., rapprochement between the aggregates and the consequent friction between them during flow, which directly influence the yield strength of mortar.
Samples A, C and D had a narrow distribution, which reduces the average distance of separation between the particles and hinders the flow. This effect was observed in the tests where mortar had the highest compressive loads. It was determined in Sample C that narrow particle size distribution was decisive in achieving superior maximum load compression. It was determined in Sample A that narrow particle size distribution overcomes the effects of high incorporated air levels, resulting in a compressive load inferior to that of Sample C. Sample D exhibited low levels of incorporated air and narrow particle size distribution resulting in a high compressive load. Sample D also had a higher content of fines than the other samples. It is possible that the water content of the mixture was not sufficient to ensure adequate mixture flow. Despite the similarities in particle size distribution between the samples, the maximum compression loads were different. This is probably due to the combined effect of high levels of incorporated air in the mortar. The lowest compression loads were represented by Samples E and F. Both of these samples had open size distribution but differed in the content of incorporated air. In Sample E, grain size was decisive in compressive loading in that it had more significant effect than a low level of incorporated air. The low load compression level in Sample F resulted from high levels of incorporated air, open particle size distribution and in the morphological characteristics of particles highlighted by lower level of roughness. The observed results of the samples confirm the conclusions obtained by Cardoso [14] in his doctoral thesis where he indicated that open particle size resulted in smaller voids in the mortar and lower loads in the Squeeze Flow test while narrow particle size distribution resulted in larger voids and higher loads in the Squeeze Flow test. The morphology of the aggregates allowed the verification of its influence on rheological behavior. Specifically in Sample F, rounded grains acted together with incorporated air, particle size distribution and reduced compression load. After analyzing the results, it was determined that it is likely that there is a synergistic relationship between particle size distribution, incorporated air content and morphology on the load compression obtained in the Squeeze Flow test. However, it is noteworthy to consider that, even with mortar additives, a portion of these observed behaviors may have been due to these synergistic relationships, thus creating the necessity for further investigation on the subject.
Table 6 Summary of the adhesive mortars analysis

Adhesive mortar A B C D E F Compressive load (N) High (1.26) Intermediate (1.20) + High (1.54) High (1.21) Low (0.62) + Low (0.48) Paste viscosity (mPa.s) Intermediate (64) Intermediate (66) Intermed/Low (45) Low (21) Intermediate (66) High (87) Particle size distribution Narrow Open + Narrow Narrow + Open Open Air entrained content (%) + High (28.2) Intermediate (21.9) High (22.9) Low (15.7) + Low (15.5) High (26.6) Roundness Irregular (0.518) Irregular (0.515) Irregular (0.504) Irregular (0.495) Irregular (0.492) Regular (0.616)
404
4. Conclusions
The experimental program proposed in this research yielded the following conclusions: n The profile and load level data curves obtained from the tests are sufficient for understanding the rheological behavior of mortar; n The adhesive mortar utilized in this study reflected different rheological characteristics, each mixture indicating distinct compression loading absorption capacity imposed by the mobile punch equipment; n It was observed that there is a synergistic relationship between particle size distribution, the amount of incorporated air and grain morphology on the compressive load of the mortar; n Size distribution and the morphology of the grains, within the size range of 0.3 - 0.15 mm, were indicated as the most influential parameters for the studied adhesive mortars.
5. References
[01] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Changes in microstructures and physical properties of polymer-modified mortars during wet storage. Cement and Concrete Research. Volume 36. 2005a. [02] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Influence of polymerer on microstructure and adhesive strength of cementitiuos tile adhesive mortars. Cement and Concrete Research. Volume 36. 2005b. [03] SANTOS, N. Avaliao de micro e nanoporosidade de argamassas colantes. Dissertao de Mestrado, Programa de Ps Graduao em Engenharia Civil da Universidade Federal de Santa Catarina - UFSC, Florianpolis (SC), 2006. [04] SENNA, D. O. Avaliao do comportamento das argamassas modificada por polmeros base de ter de celulose em ambientes agressivos. Dissertao de Mestrado. Programa de Ps Graduao em Engenharia Civil e Ambiental. Departamento de Tecnologia. Universidade Estadual de Feira de Santana. Feira de Santana - Bahia. 2008. [05] MANSUR, A. A. P. Mecanismo fsico-qumico de aderncia na interface argamassa modificada com polmeros/cermica de revestimento. Tese (Doutorado), Universidade Federal de Minas Gerais. Belo Horizonte. 2007. [06] MENDES, T. M. Influncia do coeficiente de atrito entre os agregados e da viscosidade da matriz no comportamento reolgico de suspenses concentradas heterogneas. Dissertao de Mestrado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2008. [07] PEREIRA, E. Estudo da influncia das propriedades de argamassas colantes na resistncia de aderncia de revestimentos cermicos aplicados no assentamento de piso sobre piso. Dissertao (Mestrado). Programa de Ps-Graduao em Engenharia de Construo Civil. Universidade Federal do Paran. Curitiba, 2012.
[08] [09] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
COSTA, M. R. M. M. Anlise comparativa de argamassas colantes de mercado atravs de parmetros reolgicos. Tese de Doutorado, Escola Politcnica da Universidade de So Paulo - USP, So Paulo (SP), 2006. ANTUNES, G. R. Estudo de Manifestaes Patolgicas em Revestimentos de Fachada em Braslia Sistematizao da Incidncia de Casos. Dissertao de Mestrado. Universidade de Braslia. Distrito Federal. 2010. MANSUR, A. A. P; NASCIMENTO, O. L; MANSUR, H. S. Physico-chemical characterization of EVA-modified mortar and porcelain tiles interfaces. Cement and Concrete Research. Volume 39. pp 1199-1208. 2009. OLIVEIRA, I.R. ; STUDART, A.R. ; PILLEGI, R.G. ; PANDOLFELLI, V.C. Disperso e empacotamento de partculas : princpios e aplicaes em processamento cermico. 18 ed. Fazendo arte editorial, 195 p, 2000. PILEGGI, R. G. Ferramentas para o estudo e desenvolvimento de concretos refratrios. 2001. 187p. Tese de doutorado. Universidade de So Carlos, 2001. BAUER, E. Revestimentos de argamassa caractersticas e peculiaridades. 1. ed. BRASLIA: LEM-UnB - SINDUSCON/DF. 2005. CARDOSO, F. A. Mtodo de formulao de argamassas de revestimento baseado em distribuio granulomtrica e comportamento reolgico. Tese de Doutorado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2009. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa de assentamento e revestimento de paredes e tetos Caracterizao reolgica pelo mtodo squeeze-flow - NBR15839. Rio de Janeiro, 2010 KUDO, E. K; CARDOSO, F. A; PILEGGI, R. G. Squeeze Flow aplicado a argamassas colantes: influncia de parmetros experimentais de configurao e taxa de deslocamento. In: IV Simpsio brasileiro de tecnologia de argamassas. Anais do IV SBTA. Belo Horizonte. 2011. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Especificaes - NBR 14081. Rio de Janeiro, 2004. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Determinao do deslizamento - NBR 14085. Rio de Janeiro, 2004. QUARCIONI, V. A. Reconstituio de trao de argamassas: atualizao do Mtodo IPT. Dissertao (Mestrado) Escola Politcnica, Universidade de So Paulo. So Paulo, 1998. ESSS - Engineering Simulation and Scientific Software: Fatores de forma Descrio e usabilidade dos fatores de forma na plataforma IMAGO. Florianpolis, 2005. 19 p. Relatrio. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa para assentamento de paredes e revestimento de paredes e tetos Determinao da densidade da massa e do teor de ar incorporado NBR 13278. Rio de Janeiro, 2004.
405
Abstract
Resumo
a b
2013 IBRACON
1. Introduction
Adhesive mortars consist of cement, aggregate and other additives. The most common industrially produced additives are cellulose ethers (HEC e MHEC) and latex polymers (PVA). Each perform specific functions in fresh and hardened mortars [1]. Latex additives improve the rheological behavior of fresh mortars, but mainly provides flexibility and tensile strength for hardened mortars [2].Ethers are water soluble polymers and small amounts of polymer increases water retention and viscosity in mortars [3, 4]. Quartz sand, with dimensions less than 0.50 mm, is predominantly used in mortar. Characteristics, such as shape, texture and grain size, have a significant influence on the workability and adhesion of mortar [5]. The shape and texture of the grains are primarily responsible for changes in the friction coefficient. Mendes [6] demonstrated that the more rounded the particles, the lower the friction coefficient compared to those that are irregularly shaped. Pereira [7] found that the shape of particles larger than 0.090 mm significantly influence the flow phenomena of mortar and the more rough the particles, the greater the shear required for the flow of the mixture. Another issue related to the influence of particle size distribution is the degree of compaction. The ideal condition is to gradually fill the larger voids with smaller grains [8]. It should be noted, however, that you should not utilize an excess of fines as they may increase the surface area, hence increasing the amount of water needed to coat the particles. Sand with elevated fineness increases the water consumption in the mixture, which may cause significant shrinkage in the mortar after water evaporation, and fine sand may fill the pores of the substrates occupying the spaces of the hydration products [5, 9]. Moreover, an increase in the proportion of sand, maintaining the same amount of water in the mixture, decreases the workability of the mortar, thus influencing the development of adhesive strength of the coatings. It should also be considered that increased coarseness in the sand could cause inefficient distribution of the grains in the cement mixture leading to difficulty in mortar flow and compromising adhesion [9]. However, it should be emphasized that an adequate proportion of sand is essential for the formation of the undeformed skeletal structure of the mixture and to reduce mortar shrinkage. The fineness of cement has a significant influence in adhesive mortar. Different cements exhibit different physical characteristics. However, fineness is the most significant parameter [10]. The finer
the particles of cement, the more water and additives are required for the same workability. From the perspective of rheological behavior, it is assumed that mortar adhesives are concentrated suspensions of solid particles immersed in a viscous liquid which has the function of ensuring the cohesion of the system and providing lubrication and space for the movement of aggregates [8]. If the volume of the matrix is not sufficient, the rheological characteristics and properties of the mortar is harmed [11, 6]. With increased matrix content, fluid flows more easily and fluidity of the system will be governed by the viscosity of the matrix. Thus, to understand the rheology of the fluid and have a global view of rheological behavior, it is necessary to analyze the phenomena that occur in the matrix, such as those derived from aggregate [12]. To define the behavior of mortar in its fresh state we frequently use consistency parameter data obtained from tests Flow Table and the Dropping Ball tests [13]; the so-called monopoint tests. However, as easy as these tests are to perform, they present conceptual limitations in the characterization of mortars. It is not possible to describe the behavior of material by a single measurement. But, it should be measured by a rheological profile test that, preferably, simulates actual conditions since behavior can vary depending on the characteristics of the actual application [14]. To fill this limitation gap, the Squeeze Flow test was developed. It provides for the characterization of rheological behavior through viscosity and yield parameters and is the defined as standard test NBR 15839 [15]. Despite the aforementioned standard, it is not specific for the analysis of adhesive mortars. But, this methodology is being used successfully by various researchers [8,7,16] along with adaptations and adjustments to the test apparatus. Within the above-mentioned context, the objective of this work is to understand the behavior of fresh mortar adhesives based on rheological characterization of different Brazilian commercial compositions. An understanding of the rheological behavior of adhesive mortar is the basis for a technological breakthrough, considering that their application is in a fluid state.
2. Experimental program and materials

The experimental program was designed with the objective of correlating the physical characteristics of mortar adhesives in the market, such as particle size, particle morphology and air content, with their rheological behavior determined from Squeeze Flow test parameter characteristics.
Table 1 Detail of the preliminary data of the samples (Slip and water/dry materials ratio in the pastes and mortars)
A( G ) Amount of water indicated by the manufacturer for adhesive mortar (ml/ Kg) Proportion water/dry mix materials for paste (ml/g) Slip test results of mortar (mm)
(G) Good quality; (M) Medium quality; (B) Bad Quality
B( G ) 220 1.1 0.5
Adhesive mortar and paste C( M ) D( B ) E( M ) 200 0.9 0.7 230 0.9 3.4 230 1.2 1.1
F( B ) 200 0.9 5.9
230 1.1 0.5
400
Figure 1 Schematic representation of the components of the Squeeze Flow Test
the No. 200 sieve (0.075 mm opening). Industrialized materials were also utilized in a reconstruction approximating the amounts of components in procedures recommended by Quarcioni [19]. The aggregate particles retained in sieve openings of 0.6, 0.3 and 0.15 mm were analyzed for their morphology in order to determine texture differences and form irregularities between the particles. The particles measuring above the 0.6 mm were not analyzed due to the small amount within this range. The particle study was performed using images acquired by a stereoscopic Zeiss, Stemi model 2000-C. For each mortar, a minimum of 100 grains were photographed in each sieve. Morphological analysis of the aggregate was performed utilizing an IMAGO program, version 2.2.4 loaned to us by the business ESSS[20]. The technique could not be applied to the portion of material passing through the 0.075 mm sieve opening because of the natural agglomeration of the grains. Their small size prevented them from appearing individually in the images. 2.2.2 Fresh state In the fresh state mortar was rheologically evaluated by the Squeeze Flow test, in addition to determining densities and the amount of incorporated air. In characterizing the rheological behavior of adhesive mortar in the Squeeze Flow test, the flow of material resulting from the application of a compressive load on the fresh state sample caused shifts in its interior due to the radial shear forces created during the flow. To execute this test the equipment used was a universal test machine commonly existing in building construction material laboratories for measuring application load sample as described in NBR15839 [15]. The details of the test employed are indicated in Figure 1. The movable punch has the same diameter as the sample in order to ensure the load application on the entire surface. In general, the rheological behavior expected for adhesive mortar, throughout the test, was increased compression load as it increases the displacement of the movable punch due to the approximation of the fraction larger than 0.075 mm in the mortar. In practice, comparative analysis between the mortar enables the identification of the ease of application by the mason and how it forms over the toothed trowel. The viscosity of the portion passing through the 0.075 mm sieve opening (i.e., pulp) was measured in order to determine the influence of the fine fraction (paste) on the rheological behavior of the mortar. This test was performed with a Brookfield Programmable Viscometer DV-II+ Version 5.0. To determine the amount of air incorporated in the mixtures, density tests were conducted in the fresh state, in addition to bulk density test in the dry state as previously mentioned. Bulk density in the fresh state was determined following the procedure proposed by NBR 13278 [21].
2.1 Definition of the sample universe

Initially, one bag each of thirty different brands of type AC-I adhesive mortar [17] was acquired from building material dealers throughout Brazil. Classification rules, based on manufacturers specifications, were considered for the type AC-I mortar. In order to reduce the initial sample size and provide a parameter for selecting the mortars used in the research, each mortar was subjected to a slip test, as determined by NBR 14085 [18]. In this test, the mortar was applied on a substrate pattern and the strings were formed in transverse direction to the substrate. Ceramic tiles were positioned within the strings and the substrate was placed in an upright position in order to observe for sliding of the tiles. Based upon the sliding results, six brands were chosen for the sample universe. Two brands were below the limit specified by NBR 14081 [17]. Two were at the limit. And, two were above the specified limit. The maximum sliding was specified at 0.7 mm. Table 1 reflects the sliding results of the mortar chosen. The relation of water/fine fraction dry material was obtained by mixing with the water content specified by the mortar manufacturer with respect to the proportion of fine fraction of each mortar. For example, if the amount of water indicated for mortar A was 230 ml/kg and the fine fraction was 21.8% of the mortar mass, then the ratio of water/dry material for mortar A would be equal to 1.1, or 230 ml/218 g. The water content used was specified by the manufacturer product packaging. Table 1 also reflects the amount of water recommended by the mortar manufacturers selected, in addition to the amount used.
2.2 Physical characteristics of the adhesive mortars

2.2.1 Anhydrous state Mortar was physically characterized in the anhydrous state by grain size and grain morphology. Bulk density was measured to assist in determining the amount of incorporated air. We determined the complete granulometric curves by using screening test, with the normal series of ABNT sieves, on the mortar and a laser particle size analyzer for the fraction passing through

Table 2 reflects the proportion of coarse and fine fractions (paste) adhesive mortar, or of material content larger than 75 micron and size of less than 75 micron, respectively. The fine fraction is represented mostly by cement grains and fillers and the coarse fraction by sand grains. Note that Sample D had a higher content of paste than Sample E.
401
Table 2 Reconstitution of the aproximation formulation of adhesive mortars

Adhesive mortar C D 74.67 17.77 7.55 0.16 70.67 17.90 11.44 0.15
A Aggregate (%) Cement (%) Limestone filler (%) Moisture (%) 74.75 15.86 9.39 0.13
B 75.80 20.70 3.50 0.10
E 80.27 17.86 1.87 0.16
F 74.31 19.52 6.17 0.15
Figure 2 Particle size distribution of adhesive mortars
Figure 3 Particle size distribution for particles finer than 75 m
Figure 2 reflects a diverse particle distribution in the mortars and Figure 3 illustrates a specific particle size distribution of fine fraction (particles smaller than 75 m). By analyzing the profile of the size distribution curves of the mortar (Figure 2), mortars A, C and D had narrow distributions, which tends to reduce the mean distance of separation between the particles, influencing the rheological behavior of the adhesive mortar. Opened distribution increases the flow benefits, since the smaller grains are more separated from the larger grains, thereby reducing the friction between them. From the data obtained from the sieve analysis, it can be stated that the mortar adhesives have their greatest variations when the particle size distribution is in the range 100 a 1000 m, repre-
sented by grains larger than 0.075 mm. Table 3 reflects the percentages retained and passed through the sieves with openings of 0.075, 0.15, 0.3 and 0.6 mm. It is possible to verify that there are considerable differences in the percentages retained in each sieve when compared to different mortars in sieves with openings of over 0.075 mm (coarse fraction). The fine fraction, in turn, showed a similar distribution, without significant differences between them, the only exception being Sample D which differed from the others by at least 5%. Table 4 reflects the morphological results of particles retained in the sieve openings of 0.15, 0.30 and 0.60 mm and their coefficients of roundness and the elongation of the grains. Table 5 reflects the bulk density results in the anhydrous and fresh mortar states along
Table 3 Percent of material retained and Percent of material pass on sieve

Sieve mesh 200 200 100 50 30 Adhesive mortar C D 22.8 12.2 47.0 13.0 5.0 27.2 15.6 51.2 5.4 0.6
mm % fine fraction (passing the sieve) % fine fraction (retained in the sieve) 0.075 0.075 0.15 0.30 0.60
A 21.8 31.2 42.2 4.3 0.5
B 20.9 12.7 35.8 28.1 2.5
E 19.4 31.8 24.8 22.9 1.1
F 21.8 9.5 33.1 32.8 2.8
402
Table 4 Morphological characterization of the coarse sand fraction in the adhesive mortars
Sieve mm mesh 0.15 mm 100 0.3 mm 50 0.6 mm 30 0.15 mm 100 0.3 mm 50 0.6 mm 30 Adhesive mortar C D 0.508 0.492 0.502 0.522 0.474 0.516 0.730 0.678 0.700 0.729 0.662 0.720
Roundness Sphericity
A 0.518 0.521 0.500 0.728 0.738 0.729
B 0.509 0.525 0.504 0.706 0.734 0.699
E 0.492 0.490 0.521 0.717 0.685 0.733
F 0.506 0.792 0.496 0.735 0.755 0.696
Table 5 Specific gravity and the air entrained content in the adhesive mortar
Adhesive mortar C D 2.60 2.65 1.55 1.75 22.9 15.7
Specific gravity Air entrained content
Dry Fresh Fresh
A 2.70 1.45 28.2
B 2.70 1.60 21.9
E 2.60 1.70 15.5
F 2.60 1.50 26.6
with the incorporated air content in fresh state mortar when using the manufacturer designated water content. In the results it was observed that the mortars had similar rounded particles (in the range of 0.474 to 0.521 mm) except for Sample F (fraction 0.3 mm sieve) which had superior rounding (0.792) compared to the other samples. It is also possible to verify that mortar clusters have similar elongation coefficients in the range of 0.662 to 0.755. In regard to the bulk density of an anhydrous mixture, the mortar exhibited variations in a very narrow range between 2.63 and 2.67 g/cm3. This can be attributed to the inherent variability of the test method (on the order of 1.5%). However, for the mortar analyzed, the differences resided in the fresh state bulk density which was in the range between 1.46 and 1.71 g/cm3. These differences may arise in chemical and physical changes caused by the reactivity of cement, additives and the mixing procedure, resulting in a distribution of constituents other than that present in the anhydrous state. The cement determines hydrated components in the hardened mixture. Additives, and the mixing procedure itself, will incorporate air into the fresh mixture. Throughout the tests, there was a strong correlation between fresh state bulk density and incorporated air content (R=0.9693). A variation in the level and type of additives can have a significant impact on the behavior of fresh mortar. It was not the purpose of this research to determine the quantitative data associated with these variations. Figure 4 reflects the viscosity of a mixture as a function of shear rate. The data obtained in the tests indicated that all mixtures had pseudoplastic behavior, i.e., there was a decrease in viscosity with a corresponding increase in shear rate. Although, with different mortar viscosities between 0 a 50 S-1, the data indicates a tendency of similar viscosities at higher shear rates similar to those in the Squeeze Flow tests. Thus, it was not considered
pertinent in this research to indicate viscosity as a variable in the assessment of the rheological behavior of mortar. The profile curves obtained in the Squeeze Flow test for adhesive mortar are reflected in Figure 5. The curves in Figure 5 show that adhesive mortar exhibits different rheological characteristics as indicated by each mortar having distinct compression loading absorption capacity when imposed by the movable punch equipment. Sample C had a higher compressive load as compared to the other samples which resulted in greater spreading difficulty and deformation of the strings. Sample F had greater fluidity as compared to the others which would result in easier spreading and formation of the strings but would probably result in excessive slippage of the ceramic tile. As it was not possible to obtain a parameter of effort
Figure 4 Relation between viscosities and shear rate of pastes
403
Figure 5 Curve shape of load versus displacement of adhesive mortars, obtained with the Squeeze Flow test
applied by the mason for determining spreading and bead formation, it cannot be stated that Sample C does not provide satisfactory spreading and bead formation. Probably the six studied mortars are acceptable for use in coatings. The analysis that was performed was comparative and the results obtained were restricted to the water/ dry material relationship indicated by the manufacturer. Table 6 highlights the analyzed parameters of this research and demonstrates the significant influence that particle size has on the rheological behavior of the adhesive mortars studied compared to the results obtained by the Squeeze Flow test. The numerous influential parameters are indicated in this table as high (a), intermediate (a) or low (a). Morphology is classified as either irregular or regular. When the mortar exceeds a parameter, it is indicated with a + sign. Comparing the Squeeze Flow test results to the data obtained from mortar property analysis, it was concluded that the primary influential parameter on the performance of the mortar used in this study was particle size distribution of grains larger than 0.075 mm. Calculating correlations between various properties analyzed, the percentage of aggregate retained between sieves with aperture 0.3 to 0.15 mm was shown to be more significant. This behavior might best be explained by the phenomenon of imbrication of grains, i.e., rapprochement between the aggregates and the consequent friction between them during flow, which directly influence the yield strength of mortar.
Samples A, C and D had a narrow distribution, which reduces the average distance of separation between the particles and hinders the flow. This effect was observed in the tests where mortar had the highest compressive loads. It was determined in Sample C that narrow particle size distribution was decisive in achieving superior maximum load compression. It was determined in Sample A that narrow particle size distribution overcomes the effects of high incorporated air levels, resulting in a compressive load inferior to that of Sample C. Sample D exhibited low levels of incorporated air and narrow particle size distribution resulting in a high compressive load. Sample D also had a higher content of fines than the other samples. It is possible that the water content of the mixture was not sufficient to ensure adequate mixture flow. Despite the similarities in particle size distribution between the samples, the maximum compression loads were different. This is probably due to the combined effect of high levels of incorporated air in the mortar. The lowest compression loads were represented by Samples E and F. Both of these samples had open size distribution but differed in the content of incorporated air. In Sample E, grain size was decisive in compressive loading in that it had more significant effect than a low level of incorporated air. The low load compression level in Sample F resulted from high levels of incorporated air, open particle size distribution and in the morphological characteristics of particles highlighted by lower level of roughness. The observed results of the samples confirm the conclusions obtained by Cardoso [14] in his doctoral thesis where he indicated that open particle size resulted in smaller voids in the mortar and lower loads in the Squeeze Flow test while narrow particle size distribution resulted in larger voids and higher loads in the Squeeze Flow test. The morphology of the aggregates allowed the verification of its influence on rheological behavior. Specifically in Sample F, rounded grains acted together with incorporated air, particle size distribution and reduced compression load. After analyzing the results, it was determined that it is likely that there is a synergistic relationship between particle size distribution, incorporated air content and morphology on the load compression obtained in the Squeeze Flow test. However, it is noteworthy to consider that, even with mortar additives, a portion of these observed behaviors may have been due to these synergistic relationships, thus creating the necessity for further investigation on the subject.
Table 6 Summary of the adhesive mortars analysis

Adhesive mortar A B C D E F Compressive load (N) High (1.26) Intermediate (1.20) + High (1.54) High (1.21) Low (0.62) + Low (0.48) Paste viscosity (mPa.s) Intermediate (64) Intermediate (66) Intermed/Low (45) Low (21) Intermediate (66) High (87) Particle size distribution Narrow Open + Narrow Narrow + Open Open Air entrained content (%) + High (28.2) Intermediate (21.9) High (22.9) Low (15.7) + Low (15.5) High (26.6) Roundness Irregular (0.518) Irregular (0.515) Irregular (0.504) Irregular (0.495) Irregular (0.492) Regular (0.616)
404
4. Conclusions
The experimental program proposed in this research yielded the following conclusions: n The profile and load level data curves obtained from the tests are sufficient for understanding the rheological behavior of mortar; n The adhesive mortar utilized in this study reflected different rheological characteristics, each mixture indicating distinct compression loading absorption capacity imposed by the mobile punch equipment; n It was observed that there is a synergistic relationship between particle size distribution, the amount of incorporated air and grain morphology on the compressive load of the mortar; n Size distribution and the morphology of the grains, within the size range of 0.3 - 0.15 mm, were indicated as the most influential parameters for the studied adhesive mortars.
5. References
[01] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Changes in microstructures and physical properties of polymer-modified mortars during wet storage. Cement and Concrete Research. Volume 36. 2005a. [02] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Influence of polymerer on microstructure and adhesive strength of cementitiuos tile adhesive mortars. Cement and Concrete Research. Volume 36. 2005b. [03] SANTOS, N. Avaliao de micro e nanoporosidade de argamassas colantes. Dissertao de Mestrado, Programa de Ps Graduao em Engenharia Civil da Universidade Federal de Santa Catarina - UFSC, Florianpolis (SC), 2006. [04] SENNA, D. O. Avaliao do comportamento das argamassas modificada por polmeros base de ter de celulose em ambientes agressivos. Dissertao de Mestrado. Programa de Ps Graduao em Engenharia Civil e Ambiental. Departamento de Tecnologia. Universidade Estadual de Feira de Santana. Feira de Santana - Bahia. 2008. [05] MANSUR, A. A. P. Mecanismo fsico-qumico de aderncia na interface argamassa modificada com polmeros/cermica de revestimento. Tese (Doutorado), Universidade Federal de Minas Gerais. Belo Horizonte. 2007. [06] MENDES, T. M. Influncia do coeficiente de atrito entre os agregados e da viscosidade da matriz no comportamento reolgico de suspenses concentradas heterogneas. Dissertao de Mestrado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2008. [07] PEREIRA, E. Estudo da influncia das propriedades de argamassas colantes na resistncia de aderncia de revestimentos cermicos aplicados no assentamento de piso sobre piso. Dissertao (Mestrado). Programa de Ps-Graduao em Engenharia de Construo Civil. Universidade Federal do Paran. Curitiba, 2012.
[08] [09] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
COSTA, M. R. M. M. Anlise comparativa de argamassas colantes de mercado atravs de parmetros reolgicos. Tese de Doutorado, Escola Politcnica da Universidade de So Paulo - USP, So Paulo (SP), 2006. ANTUNES, G. R. Estudo de Manifestaes Patolgicas em Revestimentos de Fachada em Braslia Sistematizao da Incidncia de Casos. Dissertao de Mestrado. Universidade de Braslia. Distrito Federal. 2010. MANSUR, A. A. P; NASCIMENTO, O. L; MANSUR, H. S. Physico-chemical characterization of EVA-modified mortar and porcelain tiles interfaces. Cement and Concrete Research. Volume 39. pp 1199-1208. 2009. OLIVEIRA, I.R. ; STUDART, A.R. ; PILLEGI, R.G. ; PANDOLFELLI, V.C. Disperso e empacotamento de partculas : princpios e aplicaes em processamento cermico. 18 ed. Fazendo arte editorial, 195 p, 2000. PILEGGI, R. G. Ferramentas para o estudo e desenvolvimento de concretos refratrios. 2001. 187p. Tese de doutorado. Universidade de So Carlos, 2001. BAUER, E. Revestimentos de argamassa caractersticas e peculiaridades. 1. ed. BRASLIA: LEM-UnB - SINDUSCON/DF. 2005. CARDOSO, F. A. Mtodo de formulao de argamassas de revestimento baseado em distribuio granulomtrica e comportamento reolgico. Tese de Doutorado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2009. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa de assentamento e revestimento de paredes e tetos Caracterizao reolgica pelo mtodo squeeze-flow - NBR15839. Rio de Janeiro, 2010 KUDO, E. K; CARDOSO, F. A; PILEGGI, R. G. Squeeze Flow aplicado a argamassas colantes: influncia de parmetros experimentais de configurao e taxa de deslocamento. In: IV Simpsio brasileiro de tecnologia de argamassas. Anais do IV SBTA. Belo Horizonte. 2011. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Especificaes - NBR 14081. Rio de Janeiro, 2004. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Determinao do deslizamento - NBR 14085. Rio de Janeiro, 2004. QUARCIONI, V. A. Reconstituio de trao de argamassas: atualizao do Mtodo IPT. Dissertao (Mestrado) Escola Politcnica, Universidade de So Paulo. So Paulo, 1998. ESSS - Engineering Simulation and Scientific Software: Fatores de forma Descrio e usabilidade dos fatores de forma na plataforma IMAGO. Florianpolis, 2005. 19 p. Relatrio. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa para assentamento de paredes e revestimento de paredes e tetos Determinao da densidade da massa e do teor de ar incorporado NBR 13278. Rio de Janeiro, 2004.
405
Abstract
Resumo
a b
2013 IBRACON
1. Introduo
As argamassas colantes so compostas por cimento, agregado, aditivos e eventualmente adies. Produzidas industrialmente, os aditivos mais comuns so os teres de celulose (HEC e MHEC) e os ltex polimricos (PVA). Cada um desempenha funes especificas no estado fresco e endurecido da argamassa [1]. O ltex melhora o comportamento reolgico da argamassa fresca, mas, principalmente, proporcionam flexibilidade e resistncia trao para a argamassa endurecida [2].Os teres so polmeros solveis em gua e pequenas quantidades deste polmero levam elevao da reteno de gua e da viscosidade em argamassas [3, 4]. A areia usada nas dosagens de argamassas so predominantemente quartzosas com dimenses menores que 0,50 mm. Caractersticas como forma, textura e granulometria exercem grande influncia na trabalhabilidade e aderncia das argamassas [5]. A forma e textura dos gros so as maiores responsveis por alteraes no coeficiente de atrito. Mendes [6] demonstrou que partculas mais arredondadas possuem um coeficiente de atrito menor se comparado quelas com forma irregular. Pereira [7] verificou que a forma das partculas maiores que 0,090 mm influenciam significativamente os fenmenos de escoamento da argamassas, sendo quanto mais rugosas as partculas maior o cisalhamento necessrias para escoamento das misturas. Outra questo relacionada com a influncia da distribuio granulomtrica das partculas o seu grau de empacotamento. A condio ideal seria aquela onde h um preenchimento gradual entre os vazios deixados pelas partculas, de forma que vazios maiores sejam preenchidos por partculas menores [8]. Deve-se, no entanto garantir que no haja um excesso de finos, pois estes podero acarretar um elevado aumento na rea superficial e, consequentemente, maior quantidade de gua necessria para recobrir as partculas. Areias com uma finura elevada aumentam o consumo de gua da mistura o que pode provocar uma retrao acentuada nas argamassas, aps a evaporao da gua, alm de que os finos da areia podem preencher os poros dos substratos ocupando espaos dos produtos de hidratao [5, 9]. Por outro lado, um aumento na proporo de areia, mantida a mesma quantidade de gua na mistura, diminui a trabalhabilidade das argamassas, influenciando no desenvolvimento da resistncia de aderncia dos revestimentos. Verifica-se tambm que areias mais grossas provocam um ineficiente envolvimento dos gros pela pasta de cimento, devido a dificuldade de escoamento da argamassa, o que tende a comprometer a extenso da aderncia
[9]. No entanto, deve-se ressaltar que um adequado porporcionamento na quantidade de areia imprescindvel para formao do esqueleto indeformvel da mistura, reduzindo-se assim a retrao nas argamassas. Quanto especificao do tipo de cimento para argamassas colantes, a finura do material exerce influncia significativa. Diferentes cimentos apresentam caractersticas fsicas dsitintas, sendo, no entanto, a finura um dos parmetros mais significativos [10]. Cimentos mais finos demandam mais gua ou a utilizao de aditivos para uma mesma trabalhabilidade. Sob o ponto de vista do comportamento reolgico, assume-se que as argamasssas colantes so suspenses concentradas de partculas slidas imersas em um lquido viscoso, o qual tem como funo assegurar a coeso do sistema, alm de proporcionar lubrificao e espao para a movimentao dos agregados [8]. Se o volume de matriz no for suficiente, as caractersticas reolgicas e as propriedades da argamassa so prejudicadas [11, 6]. Com o aumento no teor de matriz, o fluido escoa com maior facilidade e a fluidez do sistema passa a ser governada pela viscosidade da matriz. Com isso, para entender a reologia do fluido, necessria a anlise tanto dos fenmenos que ocorrem na matriz, como os que derivam dos agregados, para que se tenha uma viso global do seu comportamento reolgico [12]. Para definir o comportamento das argamassas no estado fresco frequentemente utilizam-se relaes com sua consistncia, sendo utilizado para tal dados obtidos atravs de ensaios como o Flow Table e o Dropping Ball [13], os chamados ensaios monopontos. No entanto, mesmo estes ensaios sendo de fcil execuo, apresentam limitaes conceituais na caracterizao das argamassas. No possvel que o comportamento complexo de um material seja descrito por um nico valor de medida, mas deve ser sim medido por um perfil reolgico e, preferencialmente, simulando as solicitaes prticas reais, visto que o comportamento pode variar em funo das caractersticas da solicitao [14]. Para preencher esta lacuna desenvolveu-se o ensaio Squeeze flow que permite caracterizar o comportamento reolgico das argamassas atravs dos parmetros caractersticos: viscosidade e tenso de escoamento, sendo este ensaio normalizado pela NBR 15839 [15]. Apesar da norma citada no ser especfica para anlise de argamassas colantes, a metodologia j utilizada com sucesso por diversos pesquisadores [8, 7, 16], com ajustes e adaptaes nos aparatos de ensaio. Dentro deste contexto, o objetivo deste trabalho compreender o comportamento no estado fresco de argamassas colantes, com
Tabela 1 Resultados de deslizamento e teor de gua utilizado nas argamassas e pastas

Argamassas colantes e pastas B( G ) C( M ) D( B ) E( M ) 220 1,1 0,5 200 0,9 0,7 230 0,9 3,4 230 1,2 1,1
A( G ) Quantidade de gua indicada pelo fabricante para as argamassas (ml/ Kg) Proporo gua/materiais secos para ensaio de viscosidade das pastas (ml/g) Deslizamento das argamassas (mm) 230 1,1 0,5
F( B ) 200 0,9 5,9
Obs: ( B ) Boa qualidade; (M) Qualidade mdia; (R) Ruim avaliao pelo deslizamento
407
Figura 1 Representao esquemtica dos componentes do ensaio Squeeze Flow
argamassa, logo a proporo gua/materiais secos para a pasta A foi igual a 1,1, ou seja, 230 ml / 218 g. Para as argamassas, o teor de gua empregado foi aquele indicado na embalagem do produto pelo fabricante. A tabela 1 apresenta ainda a quantidade de gua recomendada pelos fabricantes para adio nas argamassas selecionadas, alm do teor de gua utilizado para as pastas.
2.2 Caracterizao fsica das argamassas colantes

2.2.1 Estado anidro As argamassas foram caracterizadas fisicamente no estado anidro atravs da granulometria e morfologia dos gros. A densidade de massa aparente foi tambm determinada para auxiliar na determinao do teor de ar incorporado. Foram determinadas as curvas granulomtricas completas das argamassas, sendo utilizado o ensaio de peneiramento com a srie normal de peneiras da ABNT para a argamassa e o granulmetro laser para a frao passante na peneira n 200 (abertura 0,075 mm). Sendo o material industrializado, foi ainda executada uma reconstruo aproximada dos traos das argamassas colantes atravs dos procedimentos recomendados por Quarcioni [19]. As partculas de agregado retidas nas peneiras de abertura 0,6, 0,3 e 0,15 mm foram analisadas quanto sua morfologia, com o objetivo de determinar diferenas de textura e de irregularidades de forma entre as partculas. As fraes acima da peneira 0,6 mm no foram analisadas devido a pequena quantidade de gros nesta faixa, impossibilitando a amostragem. O estudo foi realizado nas imagens adquiridas com o estereoscpio marca Zeiss, modelo Stemi 2000-C. Para cada argamassa, foram fotografados pelo menos 100 gros em cada peneira. A anlise morfolgica do agregado foi feita com o auxilio do programa IMAGO verso 2.2.4 cedido pela empresa ESSS [20]. A tcnica empregada no pde ser aplicada frao passante na peneira de abertura 0,075 mm, pois a aglomerao natural desses gros, devido sua pequena dimenso, impede que os mesmos apaream individualizados nas imagens. 2.2.2 Estado fresco No estado fresco as argamassas foram avaliadas reologicamente a partir do ensaio Squeeze Flow, alm da determinao do teor de ar incorporado e da densidade. Na caracterizao do comportamento reolgico das argamassas colantes pelo ensaio Squeeze Flow, o escoamento do material decorre da aplicao de uma carga de compresso sobre a amostra no estado fresco, a qual ocasiona deslocamentos no seu interior devido aos esforos de cisalhamento radiais originados durante o fluxo. Para execuo deste ensaio, o equipamento utilizado uma mquina universal de ensaios comumente existente nos laboratrios de materiais de construo para a aplicao de carga em corpos-de-prova e descrito na NBR15839 [15]. A geometria de ensaio empregada est detalhada na Figura 1. O puno mvel possui o mesmo dimetro da amostra, visando garantir a aplicao de carga sobre toda a superfcie da mesma. De um modo geral, o comportamento reolgico que se espera para as argamassas colantes ao longo do ensaio, o aumento da carga de compresso medida que se aumenta o deslocamento do puno mvel, devido aproximao das partculas da
base na caracterizao reolgica de diferentes composies comerciais brasileiras. O entendimento do comportamento reolgico das argamassas colantes a base para um avano tecnolgico, tendo em vista que a sua aplicao ocorre no estado fluido.
2. Materiais e programa experimental

O programa experimental foi planejado com o objetivo de correlacionar caractersticas fsicas de argamassas colantes de mercado, tais como: granulometria, morfologia das partculas e teor de ar incorporado, com o seu comportamento reolgico, a partir de parmetros caractersticos determinados pelo mtodo Squeeze flow.
2.1 Definio do universo amostral

Foram inicialmente adquiridas trinta marcas de argamassa colante tipo AC-I [17] nas revendas de materiais de construo de todo o Brasil, pela compra de um saco de cada marca. A classificao das argamassas nesta categoria normativa (AC-I) foi considerada com base na especificao do fabricante. Cada argamassa foi submetida ao ensaio de deslizamento determinado pela NBR 14085 [18], tendo sido tomado como parmetro para a escolha das argamassas utilizadas na pesquisa, visando-se com isso a reduo do universo amostral inicial. Nesse ensaio, a argamassa aplicada sobre um substrato padro e os cordes so formados no sentido transversal do substrato. Placas cermicas so posicionadas sobre os cordes e o substrato deslocado na vertical para que seja observado o deslizamento das placas. Com base nos resultados de deslizamento, foram escolhidas 6 marcas, sendo duas com resultado abaixo do limite especificado pela NBR 14081 [17], duas com resultado no limite e outras duas com resultado acima do mesmo, compondo assim o universo amostral do programa experimental. O deslizamento mximo especificado foi de 0,7 mm. A tabela 1 apresenta os resultados de deslizamento das argamassas escolhidas. A relao gua/ materiais secos da frao fina foi obtida pela mistura com o teor de gua indicado pelo fabricante para a argamassa, respeitando a proporo de frao fina de cada argamassa. Por exemplo, a quantidade de gua indicada pelo fabricante para a argamassa A foi 230 ml/kg e a frao fina representa 21,8% da massa desta
408
Tabela 2 Proporo aproximada de cada material nas argamassas colantes

Argamassa colante C D 74,.67 17,77 7,55 0,16 70,67 17,90 11,44 0,15
A Agregado (%) Cimento (%) Fler Calcrio (%) Umidade (%) 74,75 15,86 9,39 0,13
B 75,80 20,70 3,50 0,10
E 80,27 17,86 1,87 0,16
F 74,31 19,52 6,17 0,15
Figura 2 Distribuio granulomtrica completa das argamassas colantes
Figura 3 Distribuio granulomtrica das fraes finas partculas com dimenso inferior a 75 m
frao grada da argamassa. Na prtica, a anlise comparativa entre as argamassas possibilita identificar as de maior facilidade de aplicao por parte do oficial pedreiro e conformao atravs da desempenadeira denteada. A viscosidade da frao passante na peneira de abertura 0,075 mm (ou seja, da pasta) foi determinada, tendo-se em vista a influncia da condio reolgica da frao fina (pasta) sobre o comportamento reolgico da argamassa. O ensaio foi realizado no equipamento Programmable Viscometer Brookfield DV-II+ Version 5.0. Para a determinao do teor de ar incorporado das misturas, realizaram-se os ensaios de densidade de massa especfica no estado fresco, alm da densidade de massa aparente no estado anidro j
comentado. A densidade de massa no estado fresco foi determinada seguindo procedimento proposto pela NBR 13278 [21].
3. Resultados e discusses
A Tabela 2 apresenta a proporo das fraes grossa e fina (pasta) das argamassas colantes, ou seja, do teor de material com dimenses superiores a 75 m e o teor de material com dimenses inferiores a 75 m, respectivamente. A frao fina representada em sua maioria pelos gros de cimento e tambm pelo fler da areia e a frao grossa pela areia. Observa-se um maior teor de pasta na argamassa D e um menor teor na argamassa E. Na Figura 2, est apresentada a distribuio granulomtrica com-
Tabela 3 Porcentagem de material retido e passante em cada peneira

Peneira mm malha % Frao fina (passante na peneira) % Frao grossa (retida na peneira) 0,075 0,075 0,15 0,30 0,60 200 200 100 50 30 Argamassa colante C D 22,8 12,2 47,0 13,0 5,0 27,2 15,6 51,2 5,4 0,6
A 21,8 31,2 42,2 4,3 0,5
B 20,9 12,7 35,8 28,1 2,5
E 19,4 31,8 24,8 22,9 1,1
F 21,8 9,5 33,1 32,8 2,8
409
Tabela 4 Caracterizao morfolgica da frao grossa das argamassas

Caracterizao da morfologia Coeficiente de arredondamento Coeficiente de alongamento Peneira mm malha #0,15 mm 100 #0,3 mm 50 #0,6 mm 30 #0,15 mm 100 #0,3 mm 50 #0,6 mm 30 Argamassa colante C D 0,508 0,492 0,502 0,522 0,474 0,516 0,730 0,678 0,700 0,729 0,662 0,720
A 0,518 0,521 0,500 0,728 0,738 0,729
B 0,509 0,525 0,504 0,706 0,734 0,699
E 0,492 0,490 0,521 0,717 0,685 0,733
F 0,506 0,792 0,496 0,735 0,755 0,696
Tabela 5 Densidade de massa e teor de ar incorporado nas argamassas colantes

Caracterizao das argamassas Densidade de massa Ar incorporado Anidra Fresca Fresca Argamassa colante C D 2,60 2,65 1,55 1,75 22,9 15,7
A 2,70 1,45 28,2
B 2,70 1,60 21,9
E 2,60 1,70 15,5
F 2,60 1,50 26,6
pleta das argamassas e na Figura 3, a distribuio granulomtrica especfica da frao fina (dimenso inferior a 75 m). Analisando-se o perfil das curvas granulomtricas das argamassas (figura 2), as argamassas A, C e D apresentaram uma distribuio estreita, o que tende a reduzir a distncia mdia de separao entre as partculas, influenciando o comportamento reolgico das argamassas colantes. A distribuio granulomtrica aberta beneficia o fluxo, pois gros menores distanciam os maiores, reduzindo com isso o atrito entre os mesmos. A partir dos dados obtidos da anlise granulomtrica pode-se afirmar que as argamassas colantes, caracterizadas neste trabalho, apresentam suas maiores variaes, quanto distribuio granulomtrica, na faixa de 100 a 1000 m, representada pela frao grada. Na Tabela 3 apresentam-se os resultados de porcentagens retidas e passantes nas peneiras de abertura 0,075, 0,15, 0,3 e 0,6 mm. possvel verificar que existem diferenas considerveis nas porcentagens retidas em cada peneira quando comparada as diferentes argamassas nas peneiras acima de 0,075 mm (frao grossa). A frao fina, por sua vez, apresentou distribuio semelhante, no apresentando diferenas significativas entre elas, sendo a nica exceo a argamassa D diferindo das demais em no mnimo 5%. A Tabela 4 apresenta os resultados morfolgicos das partculas retidas nas peneiras de abertura de 0,15, 0,30 e 0,60 mm, sendo determinados os coeficientes de arredondamento e alongamento dos gros. Na Tabela 5 esto apresentados os resultados de densidade de massa no estado anidro e fresco das argamassas, bem como o teor de ar incorporado nas argamassas no estado fresco para o teor de gua determinado pelos fabricantes. Pelos resultados, observa-se que as argamassas apresentam as partculas com arredondamento semelhante (na faixa de 0,474 a 0,521), com exceo da frao 0,3 mm da argamassa F que apresentou arredondamento bem superior s demais (0,792). possvel verificar tambm que os agregados das argamassas apre-
sentam coeficiente de alongamento semelhante, com resultados variando de 0,662 a 0,755. Quanto densidade de massa anidra as argamassas apresentam variaes numa faixa muito estreita compreendida entre 2,63 e 2,67 g/cm3, cuja diferena pode ser atribuda variabilidade intrnseca ao mtodo de ensaio (da ordem de 1,5%). No entanto, para as argamassas analisadas, as diferenas residem na densidade de massa no estado fresco, cujos resultados esto compreendidos entre 1,46 e 1,71 g/cm3. Estas diferenas podem ter origem nas transformaes qumicas e fsicas originadas pela reatividade do cimento e aditivos e pelo procedimento de mistura, resultando em outra distribuio de constituintes diferente daquela presente no estado anidro. O cimento origina compostos hidratados, e os aditivos e o procedimento de
Figura 4 Relao entre viscosidade e a taxa de cisalhamentoda frao passante na peneira de abertura 0,075 mm
410
Figura 5 Perfil da curva carga vs deslocamento das argamassas colantes, obtidas com o ensaio Squeeze Flow
mistura originam a incorporao de ar na mistura fresca. Atravs dos resultados verificou-se uma forte correlao entre densidade no estado fresco e teor de ar incorporado (R = 0,9693). Uma variao no teor e tipo de aditivo pode ter uma impacto significativo sobre o comportamento da argamassa fresca, no tendo sido possvel determinar nesta pesquisa estes percentuais adicionados. A Figura 4 apresenta a viscosidade das pastas em funo da taxa de cisalhamento. Os dados obtidos no ensaio mostram que todas as pastas apresentaram comportamento pseudoplstico, ou seja, decrscimo da viscosidade com acrscimo da taxa de cisalhamento. Apesar de as argamassas apresentarem viscosidades diferentes na faixa de 0 a 50 S-1 as curvas indicam uma tendncia de aproximao da viscosidade a partir de taxas de cisalhamento superiores, semelhantes s taxas de cisalhamento impostas no ensaio Squeeze Flow para as argamassas. Desta forma, no foi considerado nesse trabalho a viscosidade da pasta como uma varivel na avaliao do comportamento reolgico das argamassas. O perfil das curvas obtido no ensaio Squeeze Flow para as argamassas colantes est apresentado na Figura 5. As curvas da Figura 5 demonstram que as argamassas colantes apresentam caractersticas reolgicas diferentes, pois cada argamassa mostrou capacidade distinta de absoro do carregamento de compresso imposto pelo puno mvel do equipamento. A argamassa C, que apresentou maior carga mxima de com-
presso que as demais, apresentar maior dificuldade ao espalhamento e deformao dos cordes. A argamassa F, por ter se apresentado muito fluida frente s demais, mostrar facilidade ao espalhamento e formao dos cordes, provavelmente at em excesso, com possvel deslizamento da placa. Como no foi possvel obter um parmetro do esforo aplicado pelo oficial pedreiro para o espalhamento e formao dos cordes de argamassa, no cabe afirmar que a argamassa C no proporcionar um espalhamento satisfatrio e formao de cordes; provavelmente as seis argamassas estudadas so aplicveis. A anlise que est sendo colocada comparativa e os resultados obtidos esto restritos para a relao gua/materiais secos indicadas pelos fabricantes. Na Tabela 6 esto destacados os parmetros analisados neste trabalho, demostrando indicativos de maior influncia da granulometria no comportamento reolgico das argamassas colantes estudadas, comparativamente aos resultados obtidos pelo Squeeze Flow. Os parmetros influentes, cujos resultados so numricos, esto indicados nessa tabela em alto (a), intermedirio (a) ou baixo (a). A morfologia est classificada em irregular e regular. Quando a argamassa se destacou em algum parmetro utilizado o sinal de +. Comparando-se os resultados do ensaio Squeeze Flow com os dados obtidos na anlise das propriedades das argamassas, conclui-se que o principal parmetro influente no desempenho das argamassas em estudo foi a distribuio granulomtrica da frao grada. Calculando-se correlaes entre as diversas propriedades analisadas, a porcentagem de agregado retido entre as peneiras com abertura de 0,3 a 0,15 mm foi a que se mostrou mais significativa. Este comportamento talvez possa ser explicado pelo fenmeno de embricamento dos gros, ou seja, aproximao entre os agregados e consequente atrito entre estes durante o escoamento, que influenciam diretamente na tenso de escoamento das argamassas. As argamassas A, C e D apresentaram distribuio estreita, o que reduz a distncia mdia de separao entre as partculas e dificulta o fluxo. Esse efeito foi observado no ensaio, onde essas argamassas apresentaram as maiores cargas de compresso. Para a argamassa C, cuja carga mxima de compresso superou as demais, a distribuio granulomtrica estreita foi decisiva nesse desempenho. Na argamassa A, a distribuio granulomtrica estreita supera o efeito do alto teor de ar incorporado, resultando numa carga de compresso inferior C. Na argamassa D, o baixo teor de ar incorporado, juntamente com a distribuio granulom-
Tabela 6 Resumo da anlise das argamassas

Argamassa A B C D E F Carga de compresso (N) Alta (1,26) Interm. (1,20) + Alta (1,54) Alta (1,21) Baixa (0,62) + Baixa (0,48) Viscosidade da pasta (mPa.s) Intermediria (64) Intermediria (66) Interm/Baixa (45) Baixa (21) Intermediria (66) Alta (87) Distribuio granulomtrica Estreita Aberta + Estreita Estreita + Aberta Aberta Teor de ar Morfologia das incorporado (%) partculas (coef.arred.) + Alto (28,2) Intermedirio (21,9) Alto (22,9) Baixo (15,7) + Baixo (15, 5) Alto (26,6) Irregular (0,518) Irregular (0,515) Irregular (0,504) Irregular (0,495) Irregular (0,492) Regular (0,616)
411
trica estreita, resultou em alta carga de compresso. A argamassa D possui ainda um teor de finos maior que as demais. Possivelmente o teor de gua na mistura no foi suficiente para garantir o escoamento das misturas. Apesar das similaridades de distribuio granulomtrica entre essas argamassas, as cargas mximas de compresso foram diferentes, provavelmente devido ao efeito conjunto do alto teor de ar incorporado na argamassa. As menores cargas de compresso foram apresentadas pelas argamassas E e F. Ambas as argamassas apresentaram distribuio granulomtrica aberta, mas se diferenciaram quanto ao teor de ar incorporado. Para a argamassa E, a granulometria foi decisiva na carga de compresso, ou seja, teve seu efeito prevalecido sobre o baixo teor de ar incorporado. A baixa carga de compresso apresentada pela argamassa F foi resultante do alto teor de ar incorporado, da distribuio granulomtrica aberta e das caractersticas morfolgicas dos gros que se destacaram pela baixa rugosidade. Os resultados verificados para as argamassas confirmam os resultados obtidos por Cardoso [14] que em sua tese de doutorado enfatiza que a granulometria aberta resulta em um menor ndice de vazios nas argamassas e em menor carga no Squeeze Flow enquanto uma distribuio granulomtrica fechada resulta em maior ndice de vazios e maiores cargas no Squeeze Flow. A morfologia dos agregados permitiu a verificao da sua influncia sobre o comportamento reolgico, principalmente na argamassa F, cujos gros arredondados atuaram juntamente com o ar incorporado e com a distribuio granulomtrica e reduziram a carga de compresso da argamassa. Com a anlise dos resultados foi possvel observar que provvel que tenha ocorrido uma ao sinrgica entre distribuio granulomtrica, teor de ar incorporado e morfologia, sobre a carga mxima de compresso obtida no ensaio Squeeze flow. Porm, ressalta-se ainda que sendo estas argamassas aditivadas, parte do comportamento observado pode ter sido devido a ao destes sobre as misturas, demandando maiores investigaes sobre o tema.
[01] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Changes in microstructures and physical properties of polymer-modified mortars during wet storage. Cement and Concrete Research. Volume 36. 2005a. [02] JENNI, A; ZURBRIGGEN, R; HOLZER, L; HERWEGH, M. Influence of polymerer on microstructure and adhesive strength of cementitiuos tile adhesive mortars. Cement and Concrete Research. Volume 36. 2005b. [03] SANTOS, N. Avaliao de micro e nanoporosidade de argamassas colantes. Dissertao de Mestrado, Programa de Ps Graduao em Engenharia Civil da Universidade Federal de Santa Catarina - UFSC, Florianpolis (SC), 2006. [04] SENNA, D. O. Avaliao do comportamento das argamassas modificada por polmeros base de ter de celulose em ambientes agressivos. Dissertao de Mestrado. Programa de Ps Graduao em Engenharia Civil e Ambiental. Departamento de Tecnologia. Universidade Estadual de Feira de Santana. Feira de Santana - Bahia. 2008. [05] MANSUR, A. A. P. Mecanismo fsico-qumico de aderncia na interface argamassa modificada com polmeros/cermica de revestimento. Tese (Doutorado), Universidade Federal de Minas Gerais. Belo Horizonte. 2007. [06] MENDES, T. M. Influncia do coeficiente de atrito entre os agregados e da viscosidade da matriz no comportamento reolgico de suspenses concentradas heterogneas. Dissertao de Mestrado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2008. [07] PEREIRA, E. Estudo da influncia das propriedades de argamassas colantes na resistncia de aderncia de revestimentos cermicos aplicados no assentamento de piso sobre piso. Dissertao (Mestrado). Programa de Ps-Graduao em Engenharia de Construo Civil. Universidade Federal do Paran. Curitiba, 2012. [08] COSTA, M. R. M. M. Anlise comparativa de argamassas colantes de mercado atravs de parmetros reolgicos. Tese de Doutorado, Escola Politcnica da Universidade de So Paulo - USP, So Paulo (SP), 2006. [09] ANTUNES, G. R. Estudo de Manifestaes Patolgicas em Revestimentos de Fachada em Braslia Sistematizao da Incidncia de Casos. Dissertao de Mestrado. Universidade de Braslia. Distrito Federal. 2010. [10] MANSUR, A. A. P; NASCIMENTO, O. L; MANSUR, H. S. Physico-chemical characterization of EVA-modified mortar and porcelain tiles interfaces. Cement and Concrete Research. Volume 39. pp 1199-1208. 2009. [11] OLIVEIRA, I.R. ; STUDART, A.R. ; PILLEGI, R.G. ; PANDOLFELLI, V.C. Disperso e empacotamento de partculas : princpios e aplicaes em processamento cermico. 18 ed. Fazendo arte editorial, 195 p, 2000.
4. Concluses
O programa experimental proposto neste trabalho permitiu obter as seguintes concluses: n As curvas obtidas pelo ensaio so suficientes para a compreenso do comportamento reolgico das argamassas, atravs do seu perfil e dos nveis de carga atingidos; n As argamassas colantes estudadas apresentam caractersticas reolgicas diferentes, pois cada mistura mostrou capacidade distinta de absoro do carregamento de compresso imposto pelo puno mvel do equipamento; n Quanto influncia da composio, com a anlise dos resultados foi observado que h uma ao sinrgica entre a distribuio granulomtrica, teor de ar incorporado e morfologia dos gros, sobre a carga mxima de compresso da argamassa. n A distribuio granulomtrica e a morfologia dos gros se apresentaram como parmetros de maior influncia para as argamassas estudadas, principalmente nas faixas entre 0,3 e 0,15 mm.
412
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
PILEGGI, R. G. Ferramentas para o estudo e desenvolvimento de concretos refratrios. 2001. 187p. Tese de doutorado. Universidade de So Carlos, 2001. BAUER, E. Revestimentos de argamassa caractersticas e peculiaridades. 1. ed. BRASLIA: LEM-UnB - SINDUSCON/DF. 2005. CARDOSO, F. A. Mtodo de formulao de argamassas de revestimento baseado em distribuio granulomtrica e comportamento reolgico. Tese de Doutorado. Escola Politcnica da Universidade de So Paulo USP. So Paulo (SP). 2009. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa de assentamento e revestimento de paredes e tetos Caracterizao reolgica pelo mtodo squeeze-flow - NBR15839. Rio de Janeiro, 2010 KUDO, E. K; CARDOSO, F. A; PILEGGI, R. G. Squeeze Flow aplicado a argamassas colantes: influncia de parmetros experimentais de configurao e taxa de deslocamento. In: IV Simpsio brasileiro de tecnologia de argamassas. Anais do IV SBTA. Belo Horizonte. 2011. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Especificaes - NBR 14081. Rio de Janeiro, 2004. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa colante industrializada para assentamento de placas cermicas Determinao do deslizamento - NBR 14085. Rio de Janeiro, 2004. QUARCIONI, V. A. Reconstituio de trao de argamassas: atualizao do Mtodo IPT. Dissertao (Mestrado) Escola Politcnica, Universidade de So Paulo. So Paulo, 1998. ESSS - Engineering Simulation and Scientific Software: Fatores de forma Descrio e usabilidade dos fatores de forma na plataforma IMAGO. Florianpolis, 2005. 19 p. Relatrio. ASSOCIAO BRASILEIRA DE NORMAS TCNICAS. Argamassa para assentamento de paredes e revestimento de paredes e tetos Determinao da densidade da massa e do teor de ar incorporado NBR 13278. Rio de Janeiro, 2004.
413
Image-based method for monitoring of crack opening on masonry and concrete using Mobile Platform Mtodo com base em imagem para o monitoramento da abertura de fissura em alvenaria e concreto usando plataforma mvel
A. P. Martins a apms6592@gmail.com.br J. C. Pizolato Junior b jcpizolato@yahoo.com.br V. L. Belini c belini@ufscar.br
Abstract
This paper proposes an automatic method based on the computing vision, implemented in a mobile platform, to inspect cracks in masonry and concrete. The developed algorithm for image processing performs this task from images of the cracks evolution. The contribution of this paper is the development of a mobile tool with quick response aiming to assist technicians in periodic visits when monitoring the crack opening in masonry and concrete. The obtained results show, successfully, the dimensional alterations of cracks detected by mobile phone in a faster and accurate way compared with the conventional measurement technique. Regardless the irregular shape of the cracks, the proposed method has the advantage of producing results statistically significant in measurement repetition by decreasing the subjectivity inherent to manual measurement technique. Keywords: computational vision, image processing, mobile platform, cracks.
Resumo
Este artigo prope um mtodo automtico com base em viso computacional, implementado em uma plataforma mvel, para inspecionar fissuras em alvenaria e concreto. O algoritmo desenvolvido para processamento de imagem realiza esta tarefa a partir de imagens da evoluo de fissuras. A contribuio deste trabalho o desenvolvimento de uma ferramenta mvel, com resposta rpida visando auxiliar os tcnicos em visitas peridicas no monitorameto da abertura de fissuras em alvenaria e concreto. Os resultados obtidos mostram, com sucesso, as alteraes dimensionais de fissuras detectadas pelo telefone celular de uma maneira mais rpida e precisa em comparao com a tcnica de medio convencional. Independente da forma irregular das fissuras, o mtodo proposto tem a vantagem de produzir resultados estatisticamente significativos na repetio de medio, diminuindo a subjetividade inerente tcnica de medio manual. Palavras-chave: viso computacional, processamento de imagem, plataforma mvel, fissuras.
Alecsander Pereira Martins, Universidade Estadual de Londrina (UEL) - Mestrado em Engenharia Eltrica na Universidade Estadual de Londrina (UEL), apms6592@gmail.com.br, Londrina, PR, Brasil; Jos Carlos Pizolato Jr,Universidade Federal de So Carlos (UFSCar) -Professor Adjunto do Departamento de Engenharia Eltrica da Universidade Federal de So Carlos (UFSCar), jcpizolato@yahoo.com.br, So Carlos, SP, Brasil; c Valdinei Lus Belini, Universidade Federal de So Carlos (UFSCar), Professor Adjunto do Departamento de Engenharia Eltrica da Universidade Federal de So Carlos (UFSCar), belini@ufscar.br, So Carlos, SP, Brasil.
a
2013 IBRACON
A. P. Martins | J. C. Pizolato Junior | V. L. Belini
Introduction
On the last years there was an elevated increase on the civil construction field (a 42.41% increase between 2004 and 2010), being able to keep the extension between 4% and 5% for the next years [1]. In the same rate grows the number of edifications that present structural alterations, for project upgrades and environmental changes reasons. In this context, there is a growing concern with the inspection of the masonry and concrete structures that may present modification or visible damages along the time. The cracks analysis is an important security issue to be valued on civil engineering structures inspections [2], [3] and [4]. The masonry constructions inspection allows observing the stability or the increase of the opening of crack that can indicate the necessity for structural repairs. The crack opening term involves the opening of one or more cracks on masonry and concrete. The crack opening alteration is a referential item on the technicians analysis on periodic structural inspections. The inspection on the damaged structure is accomplished on different periods to the identification of possible cracks alterations. A crack opening constant monitoring method is the use of the mechanical or electrical strain gages, installed between two damaged structures. On another method, the conventional measurement is obtained by an inspection technician, with a dimensional metrology instrument manually placed on the damaged structure. On the reading of a crack opening, the technician observes on the instrument the measuring trait and the scale value correspondent to the crack opening. The direct observation on conventional measuring is a very thorough and exhausting task, and it depends on the technicians experience. Besides that, factors as tiredness, fatigue and distractions can contribute to different crack dimensions registers from the actual dimension. Although this manual technique is known and low priced, the time requested for measurement and the imprecision on the results have been motivating the researches for automatic methods. The present papers objective is to contribute with a tool for comparative analysis of the crack opening on masonry and concrete. The purpose is to reduce the inspection technicians workload, facilitating and presenting more exact and consistent results (smaller repeatability error). Many works have been proposing the automation of the identification and valuation of the increase rate of cracks on masonry. Some works have utilized video images, applying color contrast techniques between concrete and grout aggregates [5], and on the detection of crack on pavement artificial neural networks were used [6]. However, other works have used image processing routines in order to identify cracks on masonry, with Fourier analysis application [7], and on cracks edges detection applying image binarization techniques [8], [9] and [10]. Images of surfaces on masonry prepared with fluorescence marker were analyzed for micro-cracks detection [11]. Spatial filters on colored image were applied to perform the crack region separation from the rest of the image [12]. The Sobel and Canny edge detectors, associated to the Fast Fourier transform [13], were used to identify the region occupied by masonry cracks. With an image processing algorithm applying the Hough transform,
the verification of the cracks length and width (opening) were obtained [14]. Cannys edge detector and Otsus thresholding [15] were utilized to identify cracks on concrete tubes. Cameras optical measuring and Green-Lagranges expression [16] are tools that were applied to measurement on cracks. Texture filters on the interest surface area [17] were used for crack identification. With webcams and images tools from Matlab software [18], it is possible to find and measure the masonrys cracks. On the math algorithm applying [19] occurs cracks detection by the difference of reflections of thermographic light radiation from the masonry intact area. The cracks are also identified by an algorithm applying a median filter on the image [20]. The digital photogrammetry techniques [21] and [22] are utilized to measure the cracks dimensions and to estimate civil engineering structural deformations. Despite the accuracy and consistency of the results presented by the discussed proposals, the performance of most of the automatic systems is presented utilizing a fixed platform to accomplish the measurement. Specifically, the fixed platform is composed of a video camera to the image acquisition and a personal computer (PC) to execute the processing algorithms and image analysis. This requirement contributes to the partial restriction of the use of these methods in research laboratories or specifics environments. The increase of the data processing and data storage capacity, associated to the integration of high resolution video cameras on the modern mobile phones with low prices, have been allowed the substantial enlargement of these sophisticated mobile devices applications. Accordingly, this article proposes the automation of the masonry and concrete cracks dimensional comparing (structural masonry or closure masonry), starting with the hardware and software integration in only one mobile platform. To reach such objective, this proposal used as mobile platform the mobile phone Samsung model Omnia WSGH-i677 UG, 1.4 GHz processor and 512 MB (RAM) and Windows Phone 7.5 operational system. In this paper an important contribution is the reduction of the technician subjectivity on the monitoring of the crack opening and the uncertain statistics, when compared to the conventional method. Although the mobile platforms processing capacity has increased during the last years, the image processing algorithm originally developed on PC software are still computationally more complex to be executed directly on mobile phones. In this context, the optimization of an image processing algorithm, typically utilized on PC, in a way that it may be executed on a mobile phone, represents another important contribution of this work. The article is organized as follows. The Section 2 presents the necessary procedures on the conventional measurement on the evaluation of the masonry cracks growth. The proposed monitoring system is described on Section 3. The Section 4 presents the image digital processing and the Section 5 discusses the obtained results. The conclusions are on Section 6.
2.
Conventional method
On a conventional procedure, the inspection technician verifies the growth of the crack opening through a period of time, by means of manual measurement. This kind of manual measuring on
415
Monitoring of crack opening in masonry and concrete using mobile platform
Figure 1 (a). Graduated scale manually put on mansory structure with crack, on conventional measuring procedure. Figure 1(b). Crack comparator put on a crack of a faulty structure on masonry. Figure 1(c). Scale detail of the comparator instrument, presenting the 0.70 mm indication of a crack opening
A
inspections is conducted by the technician that uses a measuring tool, such as a graduated scale or cracks comparator. Figure 1 shows one of the most used tools of the conventional method for measuring cracks opening. The main problems about those measuring by comparing methods could be enumerated as low accuracy and traceability, subjectivity on reading and the difficulties on recording data [23]. Another dimensional instrument with better resolution applied on the conventional measuring procedures is the digital pachymeter, illustrated in the Figure 2(a). On other conventional measuring procedure, the technician utilizes blades calibrator with thickness between 0.1 mm and 2.0 mm. In this procedure the technician must choose and insert the metallic blade in the crack opening, as shown in the Figure 2(b). Another inconvenient on the blades calibrator procedure is the constant insertion of metal blades. Besides the time spent on changing the blades for the measuring, the insertion of the blades may lead to further degradation on the crack opening area, which could cause both mistaken measures and compromising the inspection process. The conventional dimensional metrology instruments present subjectivity on the manual positioning during the measuring of
the masonry crack opening, which may change the quantitative analysis. Other factor that could impose some difficult on the conventional measuring is the irregular forms of the cracks on masonry. Besides that, with the quoted conventional instruments it is not possible to detect rapidly small variations on the crack opening, due to its resolution and to the used technique. As a result, it becomes necessary to make several measures in order to obtain a conclusive measure. On the conventional measuring procedure there is an uncertainness associated to the handling of the instrument by the technician. This way, the conventional measuring methods may generate different readings between inspection technicians, for the same crack on masonry or concrete region. This inconsistency demands a higher number of repetitive measures to have more accurate values. On the continuous crack monitoring method, the electric or mechanical strain gages instrument is installed between the damaged structures [21], but demands the permanency of it between the periodic inspections. With the purpose of reducing the inspection technician work load and the generation of more accurate results this proposal aims to automate the identification of the crack opening on masonry and
Figure 2 (a). Digital pachymeter on a conventional procedure of a crack on masonry conventional measuring. Figure 2(b). Calibrator instrument formed by blades. This instrument on a conventional measuring allows a comparative verification between the crack openingon mansory and the blade's thickness
416
Figure 3 Block diagram of the proposed method for the identification of crack growth
concrete. The next section presents the proposed method for the crack monitoring and the image acquisition.
3. Proposed methodology
This paper work fundaments itself on the application of digital processing techniques on images in the spatial domain using a mobile phone. As a result, the implemented algorithm works directly mathematical morphology of the pixels on images previously converted to binary matrix.
The proposed method objective is to achieve the percentual index of the crack area on the structure on masonry or concrete image. The block diagram in the Figure 3 shows the proposed method. The entry of the proposed method(block1 in Figure3) corresponds to amasonryor concretestructurewithcrackas shown inFigure 1 (masonrystructure). The block 2 on Figure 3 represents the capture of a cracked structures image using a mobile phone, and this image is digitally processed on block 3. The digital processing final result, whose algorithm will be detailed on Section 4, consists on the identification of the pixels that represent a crack region on the image. Starting from this information it is possible to indentify a crack region on the image. The block 4 of Figure 3 has as function calculating the percentual index of occupation of the crack on the monitored structure image. However, the proposed method is sensitive to sideways displacement and to the distance of the mobile phone to the structure with the crack being analyzed. Thereby, it was proposed the application of a support base for the positioning and alignment of the mobile phones photographic camera in relation to the crack region, as shown in Figure 4(a). Figure 4(b) illustrates in detail the structures directional guide and the acquisition systems support base from Figure 4(a). This guide was proposed with the objective of aligning correctly the image acquisition system to the masonry structure. Thereby, the acquisition systems support base allows keeping a constant positioning of the mobile phone camera in relation to the monitored structure on periodic inspections.
Figure 4 Image acquisition of the proposed vision system. (a). mobile phone positioned on the support base on a masonry structure. In this case a 200 mm focal length was adjusted between the mobile phone's camera and the masonry structure. (b). Masonry structure directional guide put on the support base of the proposed image acquisition system. (c). Internal view of the proposed system's support base and the detail of the opening for the device's camera localized behind the illumination
417
A comparative test for different distances between the masonry structure and the mobile phone camera (focal length) was made. The test allowed finding that, for distances bellow 120 mm occurs the distortion of the image seen on the mobile phones screen. This image distortion is noticed by the user, because it shows the cracks with bigger ends compared to the real dimension. There was also verified that with focal length above 280 mm occurs the absence of the visualization of the crack beginning appearances, and bigger repeatability errors on measurements. In this case, the optimized focal length, in order to avoid distortions or crack identifications losses on the image, corresponds to 200 mm. In Figure 4 it is illustrated the proposed acquisition system that uses the mobile phone on the crack on masonry image acquisition and processing. Another characteristic of the proposed acquisition system is the illumination system with a lamp directed to the cracks region, as illustrated in Figure 4(c). The implemented illumination system has as objective avoiding luminosity interferences, for example, the alteration of the environment light, shadows or reflexes during the image capture process. The illumination system avoids bigger errors on the repeatability of reading of the proposed method. In Figure 4(c) it is possible to observe the details of the circular illumination device positioned at the image acquisition system.
The mobile phone can easily be removed from the acquisition system in order to be utilized to its usual communicational applications, as can be seen in Figure 4(a).
4. Digital image processing

The proposed digital processing is accomplished at the spatial plan of the image obtained from the structure on masonry or concrete. This processing acts directly on the pixels morphology of image. Figure 5 illustrates the blocks diagram of the proposed image processing. The block 1 in Figure 5 represents the cracked structure on masonry or concrete image obtained by the acquisition system described on Section 3. A crack on masonry image illustrated in Figure 6(a) is quantized on 256 gray levels. On the block 2 of the block diagram from Figure 5, the image with 256 gray levels from Figure 6(a) is represented on the matrix form by the function f(x, y), when x and y represent, respectively, the coordinates for row and column in each pixel position. The block 3 in Figure 5 normalizes the pixels intensity values (0255) for the interval between 0 and 1. This calculation is accomplished based on the equation (1) and the new values are stored on the matrix g(x, y).
Figure 5 Block Diagram of the image digital process of the proposed method
g (x, y)
f (x, y) max( f (x, y))
(1)
Where: f(x, y) is the original intensity value(0 f(x, y) 255); max(f(x, y)): is the highest value for the matrix f(x, y) and g(x, y) is the resultant matrix with normalized values (0 g(x, y) 1). The function of the block 4 in Figure 5 is to implement the binarization process of the matrix g(x, y). In this case, was applied the segmentation technique for image thresholding as described on [24] and mathematically described by the equation (2).
1 se g ( x , y ) L, h ( x, y ) 0 se g ( x , y ) L
(2)
On the proposed method the images threshold L on equation (2) is adjusted by the user in the beginning of the crack detection process. The value is determined aiming to identify the beginning of the cracks appearance in reason of the tonality on the structure on masonry or concrete. The surface of the structure that will be monitored must be prepared, and will be described on the next section. In the obtained results, the threshold value was equal 130 on all of the proposed method monitoring. Figure 6(b) illustrates the binary image resultant of the process accomplished by the block 4. The measuring of the interest pixels quantity is accomplished by the block 5 (Figure 5) that implements the equation (3).
N 1 M 1 n 0 m 0
n pm q
(3)
418
Figure 6 (a). Image of crack on masonry represented in Gray-level with 256 levels of grey. 6(b). Image after binarization process
Where N and M represent the quantity of rows and columns on the image, and Q represents the quantity of interest pixels on the image. Finally, the block 6 in Figure 5 calculates the percentual of the region occupied by crack on the input image. This percentual is obtained at the comparison between the quantity of crack pixels and the total pixels quantity on the image. This pixels comparison is accomplished independent of the opening, length and index (quantity) of the cracks found on the input image. The processing algorithm output information, which corresponds to the percentual of crack occupation of the structure on masonry, is exhibited directly on the mobile phones screen.
5. Results and discussions

The performance and efficiency evaluation of the proposed method was conducted by means of comparisons between measuring recorded by the mobile phone and those from the conventional method, on different crack openings. The comparisons between the proposed and the conventional method were accomplished on concrete specimen with standard shape, on a compression test machine. The proposed method may be applied on masonry and concrete structures. The objective of the experiment with specimen and testing machine was to obtain cracks for dimensional analysis in a faster way, in comparison to the collect on periodic inspections on masonry. Different types of specimen composition may be used, once the experiments do not take into account the mechanical characteristics of the specimen. Figure 7 illustrates a specimen in the compression test machine. The experiment with the testing machine was carried out on six cylindrical specimens, with a diameter of 100 mm and length of 200 mm with perpendicular faces. The surface monitored with the proposed system (masonry or concrete) was prepared by adding white acrylic paint on the specimen surface in order to facilitate the subsequent crack observation. In the experiment were recovered three specimens that presented bigger cracks focused on the fron-
tal region of the testing machine. The specimens were chosen aiming to facilitate the measuring by the conventional method and the positioning of the acquisition system, for the image capture on the proposed method. After the breaking beginning of the specimen, an advancement of compressions axis of the testing machine, was applied gradually, which promoted proportional and observables crack increases on the specimen. The conventional measuring of the crack opening was performed by two technicians, using initially a 0.5 mm resolution graduated scale. Due to the low scale resolution, other conventional measuring was accomplished utilizing a 0.01 mm resolution pachymeter. On the comparison experiment were accomplished in each method (proposed and conventional) four measures, resulting in 22 average measures to be presented in this section. The technicians that performed the monitoring in a conventional method also used the proposed system on the same conditions and specimens.
Figure 7 Testing machine used in the civil engineering to verify behavior and r esistance to compression specimen
419
Figure 8 Area % occupied by cracks on the specimen obtained in the proposed method with the mobile phone device. There were 22 average measurements on each one of three specimen during the comparison experiment
On the measuring comparison, the values of the proposed method, that resulted the following graphics were collected, processed and presented to the user directly on the mobile phones screen. Figure 8 presents a graphic illustration external to the mobile phone, with the increase rates of the crack opening registered by the proposed system on three different specimens. The measuring values were obtained on 22 gaps of advancement application axis of compression of the testing machine. The measuring was accomplished on the three specimens after the appearance of the first crack. The other results analysis in this section refers to the specimen number three. This specimen was chosen due to its best visual positioning of the cracks disposed in the testing machine, because the specimen is not withdrawn after the beginning of the compression test. Another factor was the improvement of the comparison with the proposed method against the conventional method, due to the bigger amplitude of the crack opening observed. To validate the proposed method, the captured images were submitted to a comparative test with off-line processing on PC. In that comparison it was employed a toolbox image processing (Matlab) and a pixels counting algorithm. Figure 9 illustrates the results comparision on PC image processing measuring and the proposed method. The proposed method followed the referential measuring of the processed algorithm on the PC software and the difference between the methods represents a value below 1.4%. On a periodic inspection, the technician verifies the increase or stabilization of the crack in relation to the period of time between monitoring, without the metric relation between the crack on masonry and the image. It is only considered the difference of occupation of the crack during the observed im-
ages. Figure 10 presents the real mobile phones screens on the proposed method. Figure 10 (c) presents a captured image of the masonry structure with different dimensions on the crack opening. Figure 10 (d) presents the binary image with a total result of 307200 pixels, and 16.198% of crack occupation (black pixels) on the monitored image. Regardless the length, index and opening dimension, all the visible by eye cracks were processed by the proposed method. Figure 11 presents a comparative analysis between measures presented by the use of a graduated scale, a digital pachymeter and the proposed method. Due to the difference between the proposed method and the measures obtained with the use of a graduated scale (9% and 34%, respectively), the values adopted as reference on the conventional method were obtained with a digital pachymeter, with measuring variations between 4.2% and 4.9%. The same region of the specimen presenting crack opening was utilized on all measures, regardless the measuring method. The use of the scale and pachymeter demanded more time to obtain the conventional measures, due to the two measures of the two technicians on each instrument. The successive measures were necessary to obtain the corresponding averages for each phase of the cracks evolution, and are illustrated in Figure 11. In total seven cracks were observed during the period of the specimen compression on the testing machine. Along that procedure of compression, significant differences on the dimensions of the crack opening were observed. Different from the conventional technique, that demands extra calculations from the manual readings, the proposed method the percentual of crack occupations as the images were captured. The 0.5 mm resolution of the graduated scale, associated to the difficulty of preview (scale reading and positioning on the crack) contributed to produce measures with higher er-
Figure 9 Measurement comparison between the images in the proposed method and the images processing in PC. Total were 22 average measurements
420
rors. The pachymeter, in turn, presented readings directly on the instrument screen, with the disadvantage of the necessity of manual positioning over the crack on the specimen. The repeatability errors on the measuring of the pachymeter and on
the proposed system, for each phase of the crack opening are illustrated in the Figure 12. The outcomes of Figure 12 indicates that the repeatability errors of the pachymeter were higher than the ones presented by
Figure 10 Processing of an image in the mobile phone with Windows Phone System: (a) system selection of image capture to be processed (wPhoneFissuras); (b) System selection (Home screen) and of threshold to the image processing to be captured; (c) Captured image Processing (Monitoring); (d) Binary image of crack and the introduction of outcomes of monitored image pixels
421
Figure 11 Measurement comparison between the proposed method, conventional method with graduated scale and digital pachymeter in 22 measurements in the specimen
the automatical method independent of the amount of measures accomplished. The pachymeter presented a repeatability error between 4.2% and 4.9%. In particular, these errors were higher during the beginning of the only crack opening observed, on each the crack dimensions are close to the pachymeter resolution. Starting from the 9th measuring with the pachymeter there is an increase on the repeatability error due to a bigger index of cracks on
the evaluated specimen. On the other hand, the proposed method gathered better results due to its capacity to detect and count minors quantities of pixels for the same phase of increase of the cracks on the specimen, presenting a 0.42% and 0.77% repeatability error. A performance evaluation for the proposed method (Figure 13) shows that it was necessary nine seconds to capture and process an image with up to seven cracks on the specimen against an average of 235 seconds for the conventional method. Those durations correspond to 22 measures collected on the same specimen. It is observed in Figure 13 that the time demanded by the conventional measuring until the 8th measure is contrary to the increase of the only crack opening observed on the specimen. It is also observed that the time demanded by the measuring of the 9th and 22th measures is directly proportional to the increase of the index of cracks identified on the specimen. The comparison of the repeatability error in relation to the dimensions of the cracks is illustrated in Figure 14. It is noticed that the errors presented by the digital pachymeter outweigh the ones found in the automatical method. The outcomes in Figure 14 indicate that, as the only observed crack grows in dimension until the 8th measure, smaller are the errors registered by the conventional method. That behavior occurred due to the higher facility to view the crack increase, along the compression test. In Figure 14, as the new cracks arise in initial phase, the error becomes higher. In these cases, the error for the conventional method increased from 4.2% to 4.9% between the 9th and 22th measures. On the other hand, the error exhibited by the proposed method increased slightly at the end of measures, due to the higher quantity of pixels that the crack occupied. However, that imprecision remained below 1% during all measures.
6. Conclusions
The proposed monitoring method in this work identifies the percentual index of the region of the crack on masonry or concrete,
Figure 12 Repeatability errors of crack measurements in the specimen. Total were performed 22 comparisons between the digital pachymeter and the proposed method
Figure 13 Measurements time during 22 stages of the crack opening for a total of seven observed cracks
422
by means of an image processing algorithm in a mobile platform. The comparative experiments were accomplished between the vision system proposed and the conventional crack opening measuring method. In the comparing, it was used the compression test on specimen, for the faster acquisition of different cracks dimensions. The results demonstrated that the proposed dimensional crack monitoring method overcame in various aspects the conventional technique. The vision system presented in this work, which uses a mobile phone for the image acquisition and processing, produced the following results: n Smaller subjectivity on the identification of alterations on the crack region, without depend of the technicians experience on reading and interpreting procedures of the metrology instruments; n More accurate and reliable results, regardless the index of the cracks found and of the observed crack opening dimensions; n The percentual index of the cracks area is obtained directly on the mobile phones screen as the images are captured and processed, differently of the conventional method that demands exhaustive measuring and off-line calculations to achieve the results; n The shorter measuring time, constant and independent of the length, form and index of crack found, represented other attractive results; n With the proposed method is possible to share the data obtained on the cracks monitoring by means of mobile communication. The use of a mobile phone with the support base offers repeatability on the positioning of the proposed method in the crack area analysis. An advantage of the support portable base (light and small) is the mobility, especially interesting when different analysis must be conducted on different environment with masonry structures.
The cracks monitoring on this present work may also be applied on the analysis of the behavior of different concrete composts submitted to mechanical trials, with a crack evolution observation.
7. Acknowledgements
The authors would like to thank the Department of Electrical Engineering of both the State University of Londrina (UEL) and the Federal University of So Carlos (UFSCar) for their technical support for conducting the present research and experiments.
8. References
[01] CBIC - CMARA BRASILEIRA DA INDSTRIA DA CONSTRUO, Informativo Econmico Construo Civil: Desempenho e Perspectivas, Banco de Dados, Braslia DF, disponvel em: http://www.cbicdados.com.br/files/textos/064.pdf, Dezembro 2011, p.6. [02] PARK, H. S.; LEE, H. M.; ADELI, H.; Lee, I. A new approach for health monitoring of structures: terrestrial laser scanning, Computer-Aided Civil and Infrastructure Engineering, v. 22, n. 1, 2007, p.1930. [03] YAMAGUCHI, T.; HASMUTO S. Practical Image Measurement of Crack Width for Real Concrete Structure. Eletronics and Communications in Japan, v. 92, n.10, 2009, p. 112. [04] CHEN, B.; Liu, W. Mobile agent computing paradigm for building a flexible structural health monitoring sensor network, Computer-Aided Civil and Infrastructure Engineering, 2010, v. 25, n.7, p. 50416. [05] GUDMUNDSSON, H.; CHATTERJI, S.; JENSEN, A. D.; THULOW, N.; CHRISTENSEN, P. The Measurement of paste content in hardened concrete using automatic image analyzing technique, Concrete Research. 1979, v.9, p. 607612. [06] KASEKO, M.S.; LO, Z.P.; RITCHIE, S.G. Comparison of traditional and Neural Classifier for Pavement-Crack Detection, Journal of Transportation Engineering ASCE, 1994, v.120, n.4, p. 552569. [07] ISSA, M.A.; HAMMAD, A.M. Assessment and Evaluation of Fractal Dimension of Concrete Fracture Surface Digitized Images. Cement and Concrete Research, 1994, v.24, n.2, p. 325334. [08] TILLOTSON, H. T.; SNAITH M. S.; HUANG, Y. Detecting cracks by image analysis on a parallel computer, in Proceedings of the International Computing Congress, ASCE, Boston, 1998, p.1123. [09] CHEN, L. C.; JAN, H. H.; HAUNG C. W. Mensuration of Concrete Cracks Using Digital Close-range Photographs, in Proceedings of the 22th Asian Conference on Remote Sensing, Singapore, 2001, p.12481253. [10] DARE, P.M.; HANLEY, H. B.; FRASER, C.S.; RIEDEL, B.; NIEMEIER W. An Operational Application of Automatic Feature Extraction The Measurement of Cracks in Concrete Structures. Photogrammetric Record, 2002, v.99, n.17, p.453464.
Figure 14 Repeatability error comparison between the proposed method and the pachymeter conventional method, in relation to the crack opening physical dimensions
423
[11] AMMOUCHE, A.; RISS, J.; BREYSSE, D.; MARCH AND, J. Image analysis for the automated study of microcracks in concrete, Cement and Concrete Composites, 2001, v.23, p.267278. [12] SCHUTTER, G. D. Advanced monitoring of cracked structures using video microscope and automated image analysis. NDT&E International, 2002, v.35, p.209212. [13] QADER, A.L.; ABUDAYYEH, O.; KELLY, M.E. Analysis of edge-detection techniques for crack identification in bridges, Journal of Computing In Civil Engineering, v. 17, n. 4, p. 255263, 2003. [14] SOHN, H.G.; Lim, Y.M.; YUM, K.H.; KIM, G.H. Monitoring Crack Changes in Concrete Structures, Computer- Aided Civil and Infrastructure Engineering, 2005, v.20 p.5261. [15] SINHA, S. K.; FIEGUTH P. W. Automated detection of cracks in buried concrete pipe images. Automation in Construction, 2006, v.15 p.5872. [16] LECOMPTE, D.;VANTOMME, J.; SOL, H. Crack Detection in a Concrete Beam using Two Different Camera Techniques, Structural Health Monitoring, 2006, v.5, p.5968. [17] YAMAGUCHI, T.; NAKAMURA, S.; SAEGUSA, R.; HASHIMOTO, S. Image-based Crack Detection for Real Concrete Surfaces, IEEJ Transactions On Electrical and Electronic Engineering, 2008, v.3, n.1, p.128135. [18] BERNSTONE, C.; HEYDEN, A. Image analysis for monitoring of crack growth in hydropower concrete structures, Measurement, 2009, v.42, p.878893. [19] SHAM, F. C.; XU, W. T.; LO, T. Application of flash thermography for crack identification in concrete materials, Insight - Non-Destructive Testing and Condition Monitoring, 2010,v.52, n.9, p.494497. [20] FUJITA, Y.; HAMAMOTO, Y. A robust automatic crack detection method from noisy concrete surfaces, Machine Vision and Applications, 2011, v.22, p.245254. [21] BARAZZETTI, L.; SCAIONI, M. Crack Measurement: Development, Testing and Applications of an Automatic Image-Based Algorithm, ISPRS Journal of Photogrammetry and Remote Sensing, Lecco, Italy, 2009, p.285296. [22] BARAZZETTI, L.; SCAIONI, M. Development and Implementation of Image-based Algorithms for Measurement of Deformations in Material Testing, Sensors, 2010, v.10, p.7469-7495. [23] AVAK, R.; SCHWUCHOW R.; MARCEL, F. Concrete- crack monitoring using digital image processing techniques, Bautechnik, 2007, v.84, n.6, p.403-408. [24] GONZALEZ, R. C.; WOODS, R. Digital Image Processing. Prentice Hall, 2 ed. New Jersey, 2002.
424
Image-based method for monitoring of crack opening on masonry and concrete using Mobile Platform Mtodo com base em imagem para o monitoramento da abertura de fissura em alvenaria e concreto usando plataforma mvel
A. P. Martins a apms6592@gmail.com.br J. C. Pizolato Junior b jcpizolato@yahoo.com.br V. L. Belini c belini@ufscar.br
Abstract
This paper proposes an automatic method based on the computing vision, implemented in a mobile platform, to inspect cracks in masonry and concrete. The developed algorithm for image processing performs this task from images of the cracks evolution. The contribution of this paper is the development of a mobile tool with quick response aiming to assist technicians in periodic visits when monitoring the crack opening in masonry and concrete. The obtained results show, successfully, the dimensional alterations of cracks detected by mobile phone in a faster and accurate way compared with the conventional measurement technique. Regardless the irregular shape of the cracks, the proposed method has the advantage of producing results statistically significant in measurement repetition by decreasing the subjectivity inherent to manual measurement technique. Keywords: computational vision, image processing, mobile platform, cracks.
Resumo
Este artigo prope um mtodo automtico com base em viso computacional, implementado em uma plataforma mvel, para inspecionar fissuras em alvenaria e concreto. O algoritmo desenvolvido para processamento de imagem realiza esta tarefa a partir de imagens da evoluo de fissuras. A contribuio deste trabalho o desenvolvimento de uma ferramenta mvel, com resposta rpida visando auxiliar os tcnicos em visitas peridicas no monitorameto da abertura de fissuras em alvenaria e concreto. Os resultados obtidos mostram, com sucesso, as alteraes dimensionais de fissuras detectadas pelo telefone celular de uma maneira mais rpida e precisa em comparao com a tcnica de medio convencional. Independente da forma irregular das fissuras, o mtodo proposto tem a vantagem de produzir resultados estatisticamente significativos na repetio de medio, diminuindo a subjetividade inerente tcnica de medio manual. Palavras-chave: viso computacional, processamento de imagem, plataforma mvel, fissuras.
Alecsander Pereira Martins, Universidade Estadual de Londrina (UEL) - Mestrado em Engenharia Eltrica na Universidade Estadual de Londrina (UEL), apms6592@gmail.com.br, Londrina, PR, Brasil; Jos Carlos Pizolato Jr,Universidade Federal de So Carlos (UFSCar) -Professor Adjunto do Departamento de Engenharia Eltrica da Universidade Federal de So Carlos (UFSCar), jcpizolato@yahoo.com.br, So Carlos, SP, Brasil; c Valdinei Lus Belini, Universidade Federal de So Carlos (UFSCar), Professor Adjunto do Departamento de Engenharia Eltrica da Universidade Federal de So Carlos (UFSCar), belini@ufscar.br, So Carlos, SP, Brasil.
a
2013 IBRACON
1. Introduo
Nos ltimos anos houve um elevado crescimento no segmento de Construo Civil (aumento de 42,41% entre 2004 e 2010), podendo continuar a manter a expanso entre 4 e 5% para os prximos anos [1]. Dessa forma cresce tambm o nmero de edificaes que apresentam alteraes estruturais, em razo de atualizaes de projeto e mudanas ambientais locais. Neste contexto, existe cada vez mais a preocupao com a inspeo de estruturas em alvenaria e concreto, que podem apresentar modificaes ou avarias visveis ao longo do tempo. A anlise de fissuras ou trincas um importante item de segurana a ser avaliado nas vistorias ou inspees de estruturas na Engenharia Civil [2], [3] e [4]. A vistoria das construes em alvenaria permite observar a estabilidade ou aumento da abertura de fissura, que pode indicar a necessidade de reparos estruturais. O termo abertura de fissura envolve a abertura de uma ou mais fissuras em alvenaria e concreto. A alterao na abertura de fissura um item referencial na anlise do tcnico nas vistorias peridicas estruturais. A vistoria realizada na estrutura avariada em perodos diferentes, para identificar possveis alteraes de fissura. Um mtodo de monitoramento constante da abertura de fissura o uso do instrumento extensmetro mecnico ou eltrico, instalado entre as estruturas avariadas. Em outro mtodo, a medio convencional obtida por um tcnico de inspeo, com um instrumento de metrologia dimensional colocado manualmente sobre a estrutura avariada. Na leitura da abertura de uma fissura, o tcnico observa no instrumento o trao de medida e o valor da escala correspondente abertura de fissura. A observao direta na medio convencional uma tarefa muito minuciosa e exaustiva e depende da experincia do tcnico. Alm disso, fatores como cansao, fadiga e distraes podem contribuir para conduzir a registros de dimenses de fissuras diferentes das dimenses verdadeiras. Embora essa tcnica manual seja conhecida e de baixo custo, o tempo demandado nas medies e as imprecises nos resultados tm motivado a pesquisa por mtodos automticos. O objetivo do presente trabalho contribuir com uma ferramenta para a anlise comparativa da abertura de fissura em alvenaria e concreto. O propsito reduzir a carga de trabalho do tcnico de inspeo, facilitando e produzindo resultados mais exatos e consistentes (menor erro de repetibilidade). Muitos trabalhos tm proposto a automatizao dos processos de identificao e avaliao da taxa de aumento de fissuras em alvenaria. Alguns trabalhos utilizaram imagens de vdeo, aplicando tcnicas de contraste de cor entre agregados do concreto e a argamassa [5], e na deteco de fissuras em pavimento [6] utilizou-se redes neurais artificiais. Todavia, outros trabalhos empregaram rotinas de processamento de imagem para identificar fissuras em alvenaria, com aplicao da anlise de Fourier [7], e na deteco de bordas de fissuras aplicando tcnicas de binarizao da imagem [8], [9] e [10]. Imagens de superfcies em alvenaria preparada com fluorescncia foram analisadas para a deteco de microfissuras [11]. Filtros espaciais em imagem colorida foram aplicados para realizar a separao da regio da fissura do restante da imagem [12]. Os detectores de borda Sobel e Canny, associados com a transformada rpida de Fourier [13], foram utilizados para identificar a regio ocupada por fissuras em alvenaria. Com um algoritmo de processamento de imagem aplicando a
transformada Hough [14], obteve-se a verificao do comprimento e largura (abertura) de fissuras. O detector de borda de Canny e limiarizao de Otsu [15] foi utilizado para identificar trincas em tubos de concreto. A medio ptica de cmeras e a expresso de Green-Lagrange [16] so ferramentas que foram aplicadas para medies em fissuras. Filtros de textura na superfcie da regio de interesse [17] foram utilizados para a identificao de fissuras. Com webcam e ferramentas de imagens do software Matlab [18], torna-se possvel encontrar e medir as fissuras em alvenaria. Na aplicao de algoritmo matemtico [19] ocorre deteco de fissuras pela diferena de reflexes de radiao de luz termogrfica da regio intacta da alvenaria. As fissuras so tambm identificadas por meio de um algoritmo [20] aplicando um filtro mediano na imagem. As tcnicas de fotogrametria digital [21] e [22] so utilizadas para medir as dimenses de fissuras e estimar deformaes em estruturas da Eng. Civil. Apesar da exatido e da consistncia dos resultados demonstrados por essas propostas, o desempenho da maioria dos sistemas automticos dependente de uma plataforma fixa para realizar a medio. A plataforma fixa composta por uma cmera de vdeo para a aquisio de imagens e um computador pessoal (PC) para executar os algoritmos de processamento e de anlise das imagens. Essa exigncia contribui para restringir parcialmente o uso de tais mtodos em laboratrios de pesquisa ou ambientes restritos. O aumento da capacidade de processamento e de armazenamento de dados, associados integrao de cmeras de vdeo de alta resoluo aos modernos telefones celulares de baixo custo, tm permitido ampliar substancialmente o elenco de aplicaes desses sofisticados equipamentos portteis. Nesse sentido, esse artigo prope a automao da comparao dimensional de fissuras em alvenaria (alvenaria estrutural ou alvenaria de fechamento) e em concreto, a partir da integrao do hardware e do software em uma nica plataforma mvel. Para atingir tal objetivo, essa proposta usou como plataforma mvel o aparelho celular Samsung modelo Omnia WSGH-i677 UG, processador de 1,4 GHz e 512 MB (RAM) e sistema operacional Windows Phone 7.5. Pretende-se contribuir para a reduo da subjetividade do tcnico no monitoramento da abertura de fissura e das incertezas estatsticas, quando comparadas com o mtodo convencional. Embora a capacidade de processamento dessas plataformas mveis tenha aumentado, os algoritmos de processamento de imagem originalmente desenvolvidos para PC ainda mostram-se computacionalmente mais complexos, para serem executados diretamente em aparelhos celulares. Nesse contexto, a otimizao de um algoritmo de processamento de imagem, tipicamente utilizado em PC, de tal maneira que ele possa ser executado em aparelho celular, representa outra contribuio desse trabalho. O artigo organizado conforme a seguir. A Seo 2 apresenta os procedimentos necessrios da medio convencional na avaliao do aumento da abertura de fissura em alvenaria. O sistema de monitoramento proposto descrito na Seo 3. A Seo 4 apresenta o processamento digital da imagem e a Seo 5 discute os resultados obtidos. As concluses esto na Seo 6.
2. Mtodo convencional
Em um procedimento convencional, o tcnico de inspeo verifica o aumento da abertura de fissura ao longo de um perodo de tempo, por meio da medio manual. Esse tipo de medio ma-
426
Figura 1 (a). Escala graduada colocada manualmente sobre a estrutura em alvenaria com fissura, em procedimento de medio convencional. Figura 1(b). Comparador de fissuras colocado sobre a fissura de uma estrutura em alvenaria avariada. Figura 1(c). Detalhe da escala do instrumento comparador, apresentando a indicao de 0,70 mm de abertura de fissura
Fissura Escala graduada
A
nual em vistorias conduzido pelo tcnico que utiliza algum instrumento de medida, como uma escala graduada ou comparador de fissuras. A Figura 1 ilustra instrumentos utilizados no mtodo convencional para a medio da abertura de fissura. Os principais problemas desses mtodos de medio por comparao so: baixa preciso e rastreabilidade, subjetividade na leitura e as dificuldades de registros dos dados [23]. Outro instrumento dimensional com melhor resoluo aplicado no procedimento de medio convencional o paqumetro digital, ilustrado na Figura 2(a). Em outro procedimento de medio convencional, o tcnico de inspeo utiliza um calibrador de lminas com espessuras entre 0,1 e 2,0 mm. Nesse procedimento o tcnico deve escolher e inserir a lmina metlica na abertura da fissura, como ilustrado na Figura 2(b). Outro inconveniente na medio com o instrumento calibrador de lminas a insero constante das lminas de metal. Alm do tempo dispendido na troca de lminas para a medio, a insero das lminas pode levar a degradao na regio da abertura da fissura, ocasionando medies errneas e comprometendo o processo de vistoria. Os instrumentos convencionais de metrologia dimensional apresentam subjetividade no posicionamento manual durante a medi-
o da abertura da fissura em alvenaria, podendo alterar a anlise quantitativa. Outro fator que dificulta o mtodo convencional de medio a forma irregular das fissuras em alvenaria. Alm disso, com os instrumentos convencionais citados no possvel detectar rapidamente pequenas variaes da abertura de fissura, devido sua resoluo e a tcnica empregada. Como resultado, torna-se necessrio realizar vrias medies, para a obteno de uma medida conclusiva. No procedimento de medio convencional com instrumentos de metrologia h uma incerteza associada ao manuseio do instrumento pelo tcnico. Desta forma, o mtodo convencional de medio pode gerar diferentes leituras entre tcnicos de inspeo, para a mesma regio da fissura em alvenaria ou concreto. Esta inconsistncia demanda maior quantidade de medies repetidas para produzir valores mais exatos. No mtodo de monitoramento contnuo em fissuras, o instrumento extensmetro mecnico ou eltrico instalado entre as estruturas avariadas [21], mas exige a permanncia do mesmo entre as vistorias peridicas. Com o propsito de reduzir o tempo de trabalho do tcnico de inspeo e produzir resultados com maior exatido, essa proposta visa automatizar a identificao da abertura de fissura em alvenaria e
Figura 2 (a) Instrumento paqumetro digital num procedimento de medio convencional da fissura em alvenaria. Figura 2 (b). Instrumento calibrador formado por lminas. Este instrumento numa medio convencional permite uma verificao comparativa entre a abertura da fissura em alvenaria e a espessura da lmina
427
Figura 3 Diagrama em blocos do mtodo proposto para a identificao do aumento da abertura de fissura
concreto. A prxima Seo apresenta o mtodo proposto para o monitoramento de fissura e o sistema de aquisio da imagem.
3. Metodologia proposta
Esse trabalho fundamenta-se na aplicao de tcnicas de processamento digital no domnio espacial da imagem em um aparelho
celular. Como resultado, o algoritmo implementado trabalha diretamente a morfologia matemtica dos pixels em imagens previamente convertidas para uma matriz binria. O objetivo do mtodo proposto obter o ndice % da regio de fissura na imagem da estrutura em alvenaria ou concreto. O diagrama em blocos da Figura 3 ilustra o mtodo proposto. A entrada do mtodo proposto (bloco 1 da Figura 3) corresponde a uma estrutura em alvenaria ou concreto com fissura, como mostrado na Figura 1 (estrutura em alvenaria). O bloco 2 da Figura 3 representa a captura de uma imagem da estrutura com fissura utilizando o aparelho celular, e esta imagem processada digitalmente no bloco 3. O resultado final do processamento digital, cujo algoritmo ser detalhado na Seo 4, consiste na identificao dos pixels que representam a regio de fissura na imagem. A partir desta informao possvel identificar a regio de fissura na imagem. O bloco 4 da Figura 3 tem como funo calcular o ndice % de ocupao de fissura na imagem da estrutura monitorada. Entretanto, o mtodo proposto sensvel a deslocamentos laterais e de distncia do aparelho celular em relao estrutura com fissura a ser analisada. Desta forma, foi proposto a aplicao de uma base suporte para o posicionamento e o alinhamento da cmera fotogrfica do aparelho celular em relao regio de fissura, como ilustrado na Figura 4(a). A Figura 4(b) ilustra em detalhes o guia direcional da estrutura e a base suporte do sistema de aquisio da Figura 4(a). Este guia foi proposto com o objetivo de alinhar corretamente o sistema de aquisio da imagem estrutura em
Figura 4 Sistema de aquisio de imagem do sistema de viso proposto. (a). Aparelho celular posicionado na base suporte sobre uma estrutura em alvenaria. Neste caso foi ajustada uma distncia focal de 200 mm entre a cmera do aparelho celular e a estrutura em alvenaria. (b). Guia direcional da estrutura em alvenaria colocado na base suporte do sistema de aquisio de imagem proposto. (c). Vista interna da base suporte do sistema proposto e o detalhe da abertura para a cmera do aparelho celular localizado atrs da iluminao
428
alvenaria. Desta forma, a base suporte do sistema de aquisio permite manter o posicionamento constante da cmera do celular em relao estrutura monitorada em vistorias peridicas. Teste comparativo para diferentes distncias entre a estru tura em alvenaria e a cmera do aparelho celular (distncia focal) foi realizado. O teste permitiu a constatao de que, para distncias inferiores a 120 mm ocorre distoro na ima gem visualizada na tela do aparelho celular. Essa distoro da imagem percebida pelo usurio, pois apresenta a fissura com extremidades maiores comparadas dimenso real. Tambm verificou-se que distncia focal superior a 280 mm ocorre a ausncia de visualizao do incio do surgimento de fissura, e maior erro de repetibilidade nas medies. Neste caso, a distncia focal otimizada, para evitar distores ou perdas de identificao de fissura na imagem corresponde a 200 mm. Na Figura 4 est ilustrado o sistema de aquisio proposto que utiliza o aparelho celular na obteno e processamento da imagem de fissura em alvenaria. Outra caracterstica do sistema de aquisio proposto o sistema de iluminao com uma lmpada direcionada para a regio da fissura, conforme ilustrado na Figura 4(c). O sistema de iluminao
implementado tem por objetivo evitar interferncias de luminosidade, como por exemplo, a alterao de luz ambiente, sombras ou reflexos durante o processo de captura da imagem. O sistema de iluminao evita maiores erros na repetibilidade de leitura do mtodo proposto. Na Figura 4(c) possvel observar os detalhes do dispositivo de iluminao circular posicionado no sistema de aquisio de imagem. O aparelho celular pode ser facilmente removido do sistema de aquisio para continuar sendo utilizado nas aplicaes habituais de comunicao, como pode ser observado na Figura 4 (a).
4. Processamento digital da imagem

O processamento digital proposto realizado no plano espacial da imagem obtida da estrutura em alvenaria ou concreto. Esse processamento atua diretamente na morfologia dos pixels da imagem. A Figura 5 ilustra o diagrama em blocos do processamento da imagem proposto. O bloco 1 da Figura 5 representa a imagem da estrutura em alvenaria ou concreto com fissura obtida pelo sistema de aquisio descrito na Seo 3. A imagem de uma fissura em alvenaria ilustrada na Figura 6(a) quantizada em 256 nveis de cinza. No bloco 2 do diagrama em blocos da Figura 5, a imagem com 256 nveis de cinza da Figura 6(a) representada na forma matricial pela funo f(x, y), sendo que x e y representam, respectivamente, as coordenadas de linha e de coluna de cada posio do pixel. O bloco 3 da Figura 5 normaliza os valores de intensidade de pixels (0-255) para o intervalo compreendido entre 0 e 1. Este clculo realizado segundo a equao (1) e os novos valores so armazenados na matriz g(x, y).
Figura 5 Diagrama em blocos do processamento digital da imagem do mtodo proposto
g (x, y)
f (x, y) max( f (x, y))
(1)
Onde: f(x, y) o valor de intensidade original (0 f(x, y) 255); max(f(x, y)): o maior valor da matriz f(x, y) e g(x, y) a matriz resultante com valores normalizados (0 g(x, y) 1). A funo do bloco 4 da Figura 5 implementar o processo de binarizao da matriz g(x, y). Neste caso, foi aplicada a tcnica de segmentao da imagem por limiarizao como descrito por [24] e descrita matematicamente pela equao (2).
1 se g ( x , y ) L, h ( x, y ) 0 se g ( x , y ) L
(2)
No mtodo proposto o limiar L da imagem na equao (2) ajustado pelo usurio no incio do processo de deteco de fissura. O valor determinado visando identificar o incio do surgimento de fissura em razo da tonalidade da estrutura em alvenaria ou concreto. A superfcie da estrutura a ser monitorada deve ser preparada, e ser descrita na prxima Seo. Nos resultados obtidos, o valor do limiar foi igual a 130 em todos os monitoramentos do mtodo proposto. A Figura 6(b) ilustra a imagem binria resultante do processo rea-
429
Figura 6 (a). Imagem de fissura em alvenaria representada em escala de cinza com 256 nveis de cinza. 6(b). Imagem aps processo de binarizao
lizado pelo bloco 4. A medio da quantidade de pixels realizada pelo bloco 5 (Figura 5) que implementa a equao 3.
N 1 M 1 n 0 m 0
n pm q
(3)
Onde M e N representam a quantidade de linhas e colunas na imagem, e Q representa a quantidade de pixels de interesse na imagem. Finalmente, o bloco 6 da Figura 5 calcula a % da regio ocupada de fissura na imagem de entrada. Essa % obtida na comparao entre a quantidade de pixels de fissura a quantidade de pixels total da imagem. Essa comparao de pixels realizada independente a abertura, comprimento e ndice (quantidade) das fissuras encontradas na imagem de entrada. A informao de sada do algoritmo de processamento, que corresponde % de ocupao de fissura na estrutura em alvenaria exibida diretamente na tela do aparelho celular.
corpo de prova na mquina de ensaio de compresso. O experimento com a mquina de ensaio foi realizado utilizando seis corpos de prova cilndricos, com dimenses de 100 mm de dimetro por 200 mm de comprimento e faces perpendiculares. A superfcie a ser monitorada pelo sistema proposto (alvenaria ou concreto) preparada, e consiste em adicionar tinta acrlica da cor branca na superfcie do corpo de prova, para facilitar a posterior observao das fissuras. No experimento foram aproveitados trs corpos de prova que apresentaram maiores fissuras voltadas regio frontal da mquina de ensaio. Os corpos de prova foram escolhidos visando facilitar a medio pelo mtodo convencional e o posicionamento do sistema de aquisio, para a captura de imagem no mtodo proposto. Aps a ruptura inicial do corpo de prova, aplicou-se gradativamente um avano do eixo de compresso da mquina de ensaio, o qual promoveu aumentos proporcionais e observveis de fissuras no corpo de prova.
5. Resultados e discusses
A avaliao do desempenho e da eficincia do sistema proposto foi conduzida por meio de comparaes entre medies registradas pelo aparelho celular e aquelas provenientes do mtodo convencional, em diferentes aberturas de fissuras. As comparaes entre o mtodo proposto e o convencional foram realizadas em corpo de prova de concreto em formato padro, numa mquina de ensaio de compresso. O mtodo proposto pode ser aplicado em estruturas em alvenaria e tambm em concreto. O objetivo do experimento com corpo de prova e mquina de ensaio foi obter fissuras para anlise dimensional de maneira mais rpida, comparado com a coleta nas vistorias peridicas em alvenaria. Diferentes tipos e composies do corpo de prova podem ser utilizados, uma vez que os experimentos desconsideraram as caractersticas mecnicas do corpo de prova. A Figura 7 ilustra um
Figura 7 Mquina de ensaio utilizada na Engenharia Civil para verificar comportamento e resistncia compresso em corpo de prova
Corpo de prova Mquina de ensaio
430
Figura 8 Regio % ocupada pelas fissuras no corpo de prova obtido no mtodo proposto com o aparelho celular. Foram 22 medies mdias em cada um dos trs corpos de prova durante o experimento de comparao
As medies convencionais da abertura de fissura foram realizadas por dois tcnicos, utilizando inicialmente uma escala graduada com resoluo de 0,5 mm. Devido baixa resoluo desta escala, outras medies convencionais foram realizadas utilizando um paqumetro digital de 0,01 mm de resoluo. No experimento de comparao foram realizadas em cada mtodo (proposto e convencional) quatro medies, obtendo-se 22 medies mdias que sero apresentadas nesta Seo. Os tcnicos que realizaram o monitoramento no mtodo convencional tambm utilizaram o sistema proposto nas mesmas condies e corpos de prova. Na comparao de medio, os valores do mtodo proposto que resultaram os grficos a seguir, foram coletados, processados e apresentados diretamente na tela do aparelho celular ao usurio. A Figura 8 apresenta uma ilustrao grfica externa ao aparelho celular, com as taxas de aumento das aberturas de fissuras registradas pelo sistema proposto em trs diferentes corpos de prova. Os valores das medies foram obtidos em 22 intervalos de aplicao de avano do eixo de compresso da mquina de ensaio. As medies foram realizadas nos trs corpos de prova aps o surgimento da 1 fissura. As demais anlises de resultados desta Seo referem-se ao corpo de prova nmero trs. Esse corpo de prova foi escolhido devido melhor posio visual das aberturas de fissuras dispostas na mquina de ensaio, pois o corpo de prova no retirado aps o incio do ensaio de compresso. Outro fator foi melhorar a comparao do mtodo proposto contra o mtodo convencional, devido maior amplitude da abertura das fissuras observadas. Para validar o mtodo proposto, as imagens capturadas foram submetidas a um teste comparativo com o processamento off-line em PC. Nessa comparao utilizou-se o software Matlab com toolbox de processamento de imagens e algoritmo de contagem de pixels. A Figura 9 ilustra a comparao dos resultados de medio
do processamento das imagens em PC e o mtodo proposto. O mtodo proposto acompanhou a medio de referncia do algoritmo processado no software em PC. As diferenas entre os mtodos representam um valor abaixo de 1,4 %. Em uma vistoria peridica, o tcnico verifica o aumento ou estabilizao de fissura em relao ao perodo entre os monitoramentos, sem a relao mtrica entre a fissura em alvenaria e a imagem. Apenas considerada a diferena de ocupao de fissura no decorrer das imagens observadas. A Figura 10 apresenta as telas reais do aparelho celular no mtodo proposto. A Figura 10 (c) apresenta uma imagem capturada da estrutura em alvenaria com diferentes dimenses na abertura de fissura. A Figura 10 (d) apresenta a imagem binria com o resultado total de 307200 pixels, e 16,198% de ocupao de fissura (pixels pretos) na imagem monitorada. Independente do comprimento, ndice e dimenso de abertura, todas as fissuras visveis a olho nu foram processadas pelo mtodo proposto. A Figura 11 apresenta uma anlise comparativa entre medidas apresentadas pelo uso de uma escala graduada, um paqumetro digital e o mtodo proposto. Devido s diferenas entre o mtodo proposto e as medidas obtidas com o uso de uma escala graduada (9 e 34%), os valores adotados como referncias no mtodo convencional foram obtidos com um paqumetro digital, com variaes de medida entre 4,2 e 4,9 %. A mesma regio do corpo de prova apresentando abertura de fissura foi utilizada em todas as medies, independente do mtodo de medida. A utilizao da escala e do paqumetro digital demandou maior tempo na obteno da medio convencional, devido s duas medidas dos dois tcnicos em cada instrumento. As medies sucessivas foram necessrias para a obteno das correspondentes mdias para cada fase de evoluo das fissuras, e so ilustradas na Figura 11. No total sete fissuras foram observadas durante o perodo de compresso do corpo de prova na mquina de ensaio. Ao longo desse procedimento de ensaio de
Figura 9 Comparao de medio entre as imagens no mtodo proposto e o processamento das imagens em PC. No total foram 22 medies mdias
431
compresso, diferenas significativas nas dimenses das aberturas de fissuras puderam ser observadas. Diferente da tcnica convencional, a qual demanda clculos extras a partir das leituras manuais, o mtodo proposto apre-
sentou o % de ocupao das fissuras medida que as imagens eram capturadas. A resoluo de 0,5 mm da escala graduada, associada dificuldade de visualizao (leitura da escala e posicionamento na fissura) contribuiu para produzir medies com
Figura 10 Processamento de uma imagem no aparelho celular com o sistema Windows Phone: (a) Seleo do sistema de captura de imagem a ser processada (wPhoneFissuras); (b) Seleo do sistema (tela inicial) e do limiar para o processamento da imagem a ser capturada; (c) Processamento (monitoramento) da imagem capturada; (d) Imagem binria da fissura e a apresentao do resultado dos pixels monitorados da imagem
432
Figura 11 Comparao de medio entre o mtodo proposto, mtodo convencional com escala graduada e paqumetro digital em 22 medies no corpo de prova
maiores erros. O paqumetro digital, por sua vez, apresentou leituras diretamente no mostrador do instrumento, com a desvantagem da necessidade de realizar o posicionamento manual sobre a fissura no corpo de prova. Os erros de repetibilidade das medies obtidas no paqumetro digital e no sistema proposto, para cada fase de abertura das fissuras so ilustrados na Figura 12.
Os resultados da Figura 12 indicam que os erros na repetibilidade do paqumetro digital superaram aqueles apresentados pelo mtodo automtico independente da quantidade de medies realizadas. O paqumetro digital apresentou erro de repetibilidade entre 4,2 e 4,9 %. Em particular, esses erros foram maiores durante o incio de abertura da nica fissura observada, na qual as dimenses de fissura aproximam-se da resoluo do paqumetro. A partir da 9 medio com o paqumetro digital h um aumento no erro de repetibilidade devido ao maior ndice de fissuras no corpo de prova avaliado. Em contrapartida, o mtodo proposto obteve melhores resultados em razo da sua capacidade em detectar e contar menores quantidades de pixels para a mesma fase de aumento das fissuras no corpo de prova, apresentando erro de repetibilidade entre 0,42 e 0,77%. Uma avaliao do desempenho do mtodo proposto (Figura 13) demonstra que so necessrios nove segundos para capturar e processar uma imagem com at sete fissuras no corpo de prova, contra uma mdia de 235 segundos para o mtodo convencional. Essas duraes correspondem a 22 medies coletadas no mesmo corpo de prova. Observa-se na Figura 13 que o tempo demandado pela medio convencional at a 8 medio inversamente proporcional ao aumento da abertura da nica fissura observada no corpo de prova. Tambm observa-se que o tempo demandado pela medio convencional entre a 9 e 22 medio diretamente proporcional ao aumento do ndice de fissuras identificadas no corpo de prova. A comparao de erro de repetibilidade em relao s dimenses das fissuras observadas ilustrada na Figura 14. Nota-se que os erros apresentados pelo paqumetro digital superam aqueles encontrados pelo mtodo automtico. Os resultados da Figura 14 indicam que, medida que a nica fissura observada aumenta em dimenso at a 8 medio, menores so os erros registrados pelo mtodo convencional. Esse comportamento ocorreu em razo da maior facilidade na visualizao do aumento da fissura ao longo do ensaio de compresso.
Figura 12 Erros de repetibilidade de medies de fissura no corpo de prova. No total, foram realizadas 22 comparaes entre o paqumetro digital e o mtodo proposto
Figura 13 Tempo de medies durante 22 estgios de abertura de fissura para um total de sete fissuras observadas
433
Na Figura 14 medida que surgem novas fissuras em fase inicial, o erro se torna maior. Para esses casos, o erro do mtodo convencional foi de 4,2% para 4,9 % entre a 9 e a 22 medio. Em contrapartida, o erro exibido pelo mtodo proposto aumentou ligeiramente ao final das medies, devido maior quantidade de pixels que as fissuras ocuparam. Todavia, essa impreciso permaneceu abaixo de 1% durante todas as medies.
6. Concluses
O mtodo de monitoramento proposto neste trabalho identifica o ndice % da regio de fissura em alvenaria ou concreto, por meio de algoritmo de processamento de imagens em plataforma mvel. Experimento comparativo foi realizado entre o sistema de viso proposto e o mtodo convencional de medio de abertura de fissura. Na comparao, utilizou-se o ensaio de compresso em corpo de prova, para obteno mais rpida de diferentes dimenses de fissuras. Os resultados demonstraram que o mtodo proposto de monitoramento dimensional de fissuras superou a tcnica convencional em diversos aspectos. O sistema de viso apresentado neste trabalho, o qual utiliza um aparelho celular para aquisio de imagens e processamento, produziu os seguintes resultados: n Menor subjetividade na identificao de alteraes da regio da fissura, sem depender de experincia do tcnico em procedimentos de leitura e interpretao de instrumentos de metrologia; n Resultados mais exatos e consistentes, independente do ndice de fissuras encontradas e das dimenses das aberturas de fissuras observadas; n O ndice % da regio das fissuras obtido diretamente na tela do aparelho celular medida que as imagens eram capturadas
e processadas, diferentemente do mtodo convencional que demanda medies exaustivas e clculos externos para a obteno de resultados; n O menor tempo de medio, constante e independente do comprimento, forma e ndice de fissuras encontradas, representaram outro resultado atrativo; n Com o mtodo proposto possvel compartilhar os dados obtidos no monitoramento de fissuras por meio da comunicao mvel. O uso de um aparelho mvel com a base suporte oferece repetibilidade no posicionamento do mtodo proposto na anlise da regio de fissura. Uma vantagem da base suporte porttil (leve e pequena) a mobilidade, especialmente interessante quando diferentes anlises devem ser conduzidas em diferentes ambientes com estruturas em alvenaria. O monitoramento de fissuras do presente trabalho tambm pode ser aplicado na anlise do comportamento de diferentes compostos do concreto submetidos a ensaios mecnicos, com a observao da evoluo das fissuras..
7. Agradecimentos
Os autores gostariam de agradecer ao Departamento de Engenharia Eltrica, tanto da Universidade Estadual de Londrina (UEL) e da Universidade Federal de So Carlos (UFSCar), pelo apoio tcnico para a realizao da presente pesquisa e experimentos.
[01] [02] [03] [04] [05] [06] [07] CBIC - CMARA BRASILEIRA DA INDSTRIA DA CONSTRUO, Informativo Econmico Construo Civil: Desempenho e Perspectivas, Banco de Dados, Braslia DF, disponvel em: http://www.cbicdados.com.br/files/textos/064.pdf, Dezembro 2011, p.6. PARK, H. S.; LEE, H. M.; ADELI, H.; Lee, I. A new approach for health monitoring of structures: terrestrial laser scanning, Computer-Aided Civil and Infrastructure Engineering, v. 22, n. 1, 2007, p.1930. YAMAGUCHI, T.; HASMUTO S. Practical Image Measurement of Crack Width for Real Concrete Structure. Eletronics and Communications in Japan, v. 92, n.10, 2009, p. 112. CHEN, B.; Liu, W. Mobile agent computing paradigm for building a flexible structural health monitoring sensor network, Computer-Aided Civil and Infrastructure Engineering, 2010, v. 25, n.7, p. 50416. GUDMUNDSSON, H.; CHATTERJI, S.; JENSEN, A. D.; THULOW, N.; CHRISTENSEN, P. The Measurement of paste content in hardened concrete using automatic image analyzing technique, Concrete Research. 1979, v.9, p. 607612. KASEKO, M.S.; LO, Z.P.; RITCHIE, S.G. Comparison of traditional and Neural Classifier for Pavement-Crack Detection, Journal of Transportation Engineering ASCE, 1994, v.120, n.4, p. 552569. ISSA, M.A.; HAMMAD, A.M. Assessment and Evaluation of Fractal Dimension of Concrete Fracture Surface Digitized Images. Cement and Concrete Research,
Figura 14 Comparao de erro de repetibilidade entre o mtodo proposto e o mtodo convencional com o paqumetro digital, em relao s dimenses fsicas das aberturas de fissuras
434
1994, v.24, n.2, p. 325334. [08] TILLOTSON, H. T.; SNAITH M. S.; HUANG, Y. Detecting cracks by image analysis on a parallel computer, in Proceedings of the International Computing Congress, ASCE, Boston, 1998, p.1123. [09] CHEN, L. C.; JAN, H. H.; HAUNG C. W. Mensuration of Concrete Cracks Using Digital Close-range Photographs, in Proceedings of the 22th Asian Conference on Remote Sensing, Singapore, 2001, p.12481253. [10] DARE, P.M.; HANLEY, H. B.; FRASER, C.S.; RIEDEL, B.; NIEMEIER W. An Operational Application of Automatic Feature Extraction The Measurement of Cracks in Concrete Structures. Photogrammetric Record, 2002, v.99, n.17, p.453464. [11] AMMOUCHE, A.; RISS, J.; BREYSSE, D.; MARCH AND, J. Image analysis for the automated study of microcracks in concrete, Cement and Concrete Composites, 2001, v.23, p.267278. [12] SCHUTTER, G. D. Advanced monitoring of cracked structures using video microscope and automated image analysis. NDT&E International, 2002, v.35, p.209212. [13] QADER, A.L.; ABUDAYYEH, O.; KELLY, M.E. Analysis of edge-detection techniques for crack identification in bridges, Journal of Computing In Civil Engineering, v. 17, n. 4, p. 255263, 2003. [14] SOHN, H.G.; Lim, Y.M.; YUM, K.H.; KIM, G.H. Monitoring Crack Changes in Concrete Structures, Computer- Aided Civil and Infrastructure Engineering, 2005, v.20 p.5261. [15] SINHA, S. K.; FIEGUTH P. W. Automated detection of cracks in buried concrete pipe images. Automation in Construction, 2006, v.15 p.5872. [16] LECOMPTE, D.;VANTOMME, J.; SOL, H. Crack Detection in a Concrete Beam using Two Different Camera Techniques, Structural Health Monitoring, 2006, v.5, p.5968. [17] YAMAGUCHI, T.; NAKAMURA, S.; SAEGUSA, R.; HASHIMOTO, S. Image-based Crack Detection for Real Concrete Surfaces, IEEJ Transactions On Electrical and Electronic Engineering, 2008, v.3, n.1, p.128135. [18] BERNSTONE, C.; HEYDEN, A. Image analysis for monitoring of crack growth in hydropower concrete structures, Measurement, 2009, v.42, p.878893. [19] SHAM, F. C.; XU, W. T.; LO, T. Application of flash thermography for crack identification in concrete materials, Insight - Non-Destructive Testing and Condition Monitoring, 2010,v.52, n.9, p.494497. [20] FUJITA, Y.; HAMAMOTO, Y. A robust automatic crack detection method from noisy concrete surfaces, Machine Vision and Applications, 2011, v.22, p.245254. [21] BARAZZETTI, L.; SCAIONI, M. Crack Measurement: Development, Testing and Applications of an Automatic Image-Based Algorithm, ISPRS Journal of Photogrammetry and Remote Sensing, Lecco, Italy, 2009, p.285296.
[22] [23] [24]
BARAZZETTI, L.; SCAIONI, M. Development and Implementation of Image-based Algorithms for Measurement of Deformations in Material Testing, Sensors, 2010, v.10, p.7469-7495. AVAK, R.; SCHWUCHOW R.; MARCEL, F. Concrete- crack monitoring using digital image processing techniques, Bautechnik, 2007, v.84, n.6, p.403-408. GONZALEZ, R. C.; WOODS, R. Digital Image Processing. Prentice Hall, 2 ed. New Jersey, 2002.
435
Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load Anlise numrica de blocos sobre duas estacas com clice embutido, submetido ao de fora de compresso excntrica
R. G. DELALIBERA a delalibera@pq.cnpq.br J. S. GIONGO b jsgiongo@sc.usp.br
Abstract
The structural behavior of pile caps with sockets embedded is influenced by interface of column-socket, which can be smooth or rough. With intent to analyze the behavior of two pile caps with embedded socket, considering the friction between the column and the socket, with eccentric normal load, the numerical simulations were carried out, using a program based on the Finite Element Methods (FEM). In the numerical analysis the non-linear behavior of materials was considered, also the friction between the column and the socket. It was considered perfect bond between the reinforcement and the concrete around. It was observed that the embedded length is preponderant factor in the structural behavior of the analyzed element. Keywords: pile caps, foudations, precast concrete.
Resumo
O comportamento estrutural dos blocos sobre estacas com clice embutido influenciado pelo tipo de interface pilar-clice, a qual pode ser lisa ou rugosa. Com o objetivo de analisar o comportamento de blocos sobre duas estacas com clice embutido, considerando a rugosidade da interface pilar-clice, submetido ao de fora normal excntrica, realizaram-se simulaes numricas, desenvolvidas em programa de computador baseado no mtodo dos elementos finitos MEF. Na anlise numrica as no-linearidades fsicas dos materiais concreto e ao foram consideradas, como tambm o atrito nas interfaces entre as paredes do pilar e do clice. Considerou-se aderncia perfeita entre as barras de ao das armaduras com o concreto do entorno. Observou-se que o comprimento de embutimento do pilar no interior do bloco fator preponderante no comportamento estrutural do elemento em anlise. Palavras-chave: blocos sobre estacas; fundaes; concreto pr-moldado.
Departamento de Engenharia Civil, Universidade Federal de Gois Campus Catalo, dellacivil@pq.cnpq.br, Av. Lamartine Pinto de Avelar, 1120, CEP: 75704-020, Catalo GO, Brasil. b Departamento de Engenharia de Estruturas, Escola de Engenharia de So Carlos, Universidade de So Paulo, jsgiongo@sc.usp.br, Av. Trabalhador Socarlense, 400, CEP: 13566-590, So Carlos SP, Brasil.
a
Received: 04 Sep 2012 Accepted: 07 Mar 2013 Available Online: 12 Jun 2013
2013 IBRACON
1. Introduction
The choice of the kind of foundation is made after analysis which considers the technical and economic conditions of the building, the vicinity of the neighboring buildings, the nature and characteristic of the soil, the magnitude of the actions and of the kinds of foundations available in the market. The foundations in piles are adopted when the soil in its superficial layers is not able to stand actions from the superstructure, being necessary, therefore, to consider strength in deep layers. When using solution in deep foundation by piles, it is necessary the construction of another structural element, the pile caps. According to NBR 6118:2007[1], pile caps are important structural elements which function is to transfer the actions of the superstructure to a group of piles. These structural elements, in spite of being essential to the safety of the superstructure, usually does not allow the visual inspection while working, thus, it is important the knowledge of its real behavior in the Limit States of Service and Last . Figure [1] illustrates this structural element. When the superstructure is precast it is necessary that the column is built-in on the pile. The length of the built-in of the column within the pile, so that it is considered clamped, it is function of the internal forces (bending, normal force and shear force) and the kind of superficial shape of the walls (rough or smooth) of the precast and column concrete. The lengths of embed are definite by NBR 9062:2006[2]. There are some variations in the types of pile with precast (external precast, partially embed precast and totally embed precast). Figure [2] shows the variations of the blocks on piles for precast columns. The technical mean uses more intensely the blocks with external precast and partially embed. The European code (EUROCODE 2) EN 1992-1-1 [3] presents recommendations regarding the precast project for linking columnfoundation, considering the monolithic behavior of the set when using shear key. For the smooth conformation of the precast and the column walls, it indicates that the friction coefficient between the column faces and the precast faces is higher than 0.3 and the embed length higher or equal to 1.2 times the biggest column dimension.
Figure 1 Pile cap with socket embedded
As there is no consensus between the scientific means in relation to the section of the pile cap where it forms the stanchion (or connecting rod), a numeric analysis was performed, aiming at observing the structural behavior of the pile caps. Thereunto, eighteen two piles caps were analyzed, nine with rough interface and nine with smooth interface. The embed length of the column ( emb) and the thickness of the bottom slab of the pile cap (h s) were assorted. Thus, the embed lengths of the precast columns were equal to 80 cm, 60 cm and 50 cm. The thicknesses of the bottom slab were equal to 30 cm, 20 cm and 10 cm. Bottom slab, it is not the best term to express the dimension h s because there is not such a slab in the bottom of the block. Although, it is consid ered that this term is the one which presents better understanding to the readers.
Figure 2 Pile caps with external socket, partly embedded e fully embedded
437
Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load
The purpose of this work is to analyze the structural behavior of two piles caps with embed precast, with smooth and rough interfaces of the walls of the column and the precast, subject to the action of normal eccentric compression force and, verify by variance analysis, which variable presents the biggest relevance in their behavior. The method used was the one related to works with emphasis in numeric analysis. Definite the models which were studied, it was developed a statistical analysis, using values obtained by numeric simulations. The pile caps were numerically analyzed by a computer program based on the methods of the finite elements MEF. The physical and geometrical nonlinearities were considered. The cracking of the concrete and the reinforcement of the pile caps were also taken into account. For the numerical analysis of the pile caps via method of the finite elements, it was used the program ANSYS [23] (ANalyser SYStem).
2. Geometric properties of the numerically analyzed models

The models were dimensioned with the criteria of Blvot & Frmy [4]. The rates of reinforcement were fixed, because they were not analysis objects. The distance between the axles of the piles was
also fixed, varying the length of the embed of the column (emb) and the thickness of the pile cap bottom slab (hs). The dimensioning of the main reinforcement of the pile cap (tie) and the verifications of the crushing of the concrete of the compressed diagonal with the inferior nodal zone (pile/pile cap) and with the superior nodal zone (column/pile cap) were performed according to the recommendations of the model proposed by Blvot & Frmy [4]. It was used as limits to verify the stress with the superior and inferior nodal zones the values observed by the French researchers, in other words, 1,40fck near the column and fck near the pile. We opted for using straight anchoring, because it was verified through the recommendations of the NBR 6118:2007[1] that the hook was not necessary (anchoring criteria in extreme supports). By admitting this hypothesis the researcher did not consider (project decision) the recommendation of NBR 6118:2007[1] which demands the use of hooks in pile caps. We remember the results obtained by Delalibera [5], Miguel [6] e Adebar et al. [7], for rigid pile caps, where the stress on the tip of the rod was negligible, corroborating this decision. For the dimensioning of the pile caps, we used a pattern pile cap, in other words, all pile caps had the same dimensions in plant of the pattern pile cap, varying only the embed length of the column (emb) and the thickness of the pile cap bottom slab (hs). It is understood,
Figure 3 Geometric properties of standard pile cap
A Dimension
Shear key
438
5 5 5 5 5
R.. G. DELALIBERA | J. S. GIONGO
so, that the total heights of the models and the inclination angles of the compression rod also suffered variations. For the pattern pile cap we adopted piles of squared transversal section equal to 30 cm side, column also with squared transversal section, however with 40 cm side. The embed length of the column in the pattern pile cap was equal to 60 cm. This value represents the minimal embed length for smooth walls (precast and column) specified in NBR 9062:2006 [2], when the relation [Md/(Ndh)] 0,15. Figure [3] presents the geometric properties of the pattern pile cap. The distance between the axles of the piles was determined in function of the inferior limit angle established by the French researchers, in other words, 45. Therefore, the total height of the pattern block was equal to 90 cm, and the dimensions in plants were equal to 240 cm for 84 cm. The distance between the axles of the piles of all pile caps was equal to a hundred eight centimeters. The embed length of the piles in the inferior face of the pile caps followed the suggestion of the Montoya et al. [8], namely, it was inlaid ten centimeters of the pile shaft inside the pile cap. In relation to the columns and the piles the compression strength concrete was equal to 50 MPa seeking to avoid, thus, ruining these elements. The reinforcement of the piles were composed by eight steel bars with a 20 mm diameter with strength equal to 500 MPa, totalizing an steel area equal to 32.7 cm2. The external length of the column was equal to 100 cm. For the filling material, we adopted compression strength concrete equal to 50 MPa (value equal compression strengths concrete of the column). Figure [4] shows the factors which present variations in the numerical analysis. Table [1] presents the properties of the analyzed pile caps in relation to the conformation of the smooth walls and to the shear key (rough), requested by the compression force supposedly centered in the column and by a horizontal force applied on the top of the column. The nomenclature used in Table [1] is described: L, pile with smooth conformation of the walls of the precast and the column; R, rough conformation of the precast and the column; e80, embed of the column equal to eighty centimeters (more numbers are analogue); hs30, thickness of the bottom slab equal to thirty centimeters (more numbers are analogue); NM means that eccentric compression force was applied. In the same table, Ast represents the area of the transversal section of the reinforcement class CA-50, Blx the length of the pile cap, Bly the breadth of the pile cap, emb the embed length of the precast column and hs, the thickness of the pile cap bottom slab. Using the indications of Blvot & Frmy [4], it was performed a forecast of the resistant capacities of the pile caps.
Figure [5] presents the finite elements net used to discretization of the elements of the volume (concrete) and the bar elements (steel bars of the reinforcement). The normal force was considered through a pressure of 20 MPa applied on the top of the column. The moment was considered through a horizontal force, also applied on the column top, which value was equal to 100 kN, generating a moment in the column base of 100 kNm. See Figure [04]. As the models were symmetric, it was used the resource of symmetry. Therefore, it was analyzed only half of the structural element. The translations were restricted (in the three directions, x, y and z) of the element nodes placed on the tips of the piles. Figure [6] presents the conditions of shape used in the models. The length of a meter of the piles and the column, was chosen in function of the principle of Saint Venant, thus, the influences were eliminated from the disturbances of stress in the areas of application of the forces and in the areas of translation restrictions.
3.1 Finite element used

For the modeling of the concrete material, we used the finite element Solid 65. This element has eight nodes with three degrees of freedom per node translations in the directions x, y and z. The element presents plastic deformations, cracking and crushing in three orthogonal directions. In the element Solid 65, the cracking occurs when the main stress of the traction in any direction reaches the rupture superficies. After the cracking, the elasticity module of the concrete has value equal to zero in the considered direction. The
Figure 4 Variation of factors chosen
3. Numerical analysis
The goal of the numerical analysis was to provide results for the application of a statistical analysis named ANOVA (variance analysis), thus, two piles caps with embed precast, with conformation of the smooth and rough walls were analyzed. The numerical analysis did not aim at calibrating bends of experimental results, but presenting behavior trends of the analyzed models. The geometry of all models were created in the computer program AutoCad and exported to the computer program ANSYS[23], through the extension SAT.
439
Table 1 Pile caps properties analyzed for smooth and rough walls
Total heigth h (cm) 110 100 90 90 80 70 80 70 60 Measure the Measure the sides of sides of the column the piles (cm) (cm) 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30
Pile caps Le80hs30NM Re80hs30NM Le80hs20NM Re80hs20NM Le80hs10NM Re80hs10NM Le60hs30NM Re60hs30NM Le60hs20NM Re60hs20NM Le60hs10NM Re60hs10NM Le50hs30NM Re50hs30NM Le50hs20NM Re50hs20NM Le50hs10NM Re50hs10NM
Ast (cm2) 22 22 22 22 22 22 22 22 22
Blx (cm) 240 240 240 240 240 240 240 240 240
Bly (cm) 84 84 84 84 84 84 84 84 84
emb (cm) 80 80 80 60 60 60 50 50 50
hs Classification ABNT (cm) NBR 6118:2007 30 20 10 30 20 10 30 20 10 Rigid Rigid Rigid Rigid Rigid Rigid Rigid Rigid Rigid
Note: h, total height of the pile cap; Ast, sum of the areas of the cross sections of steel bars; Blx, the total size of the pile cap in the longitudinal direction; Bly, the total size of the pile cap in the transverse direction; emb, length embed column in pile cap; hs, thickness of the bottom slab.
crushing occurs when all compression tensions reach the rupture superficies, subsequently, the elasticity module has value equal to zero in all directions. Figure [7] presents the element Solid 65. In the modeling of the steel bars of the reinforcement it was used the finite element Link 8. This element has two nodes, so that each
node has three degrees of freedom translations in the directions x, y and z. Figure [8] shows this element. We chose this element, because the armors in the models were discrete. In the models, it was not considered the phenomenon of the adherence between the steel bars and the concrete. In spite, the results
Figure 5 Discretization used in the models
440
Figure 6 Boundary conditions and finite element
Figure 8 Link 8, Ansys
The finite elements of contact were used only in the models with smooth conformation of the walls of the precast and the column, because, due to researches already performed by several researchers, it can be considered that the link column-foundation through precast with shear key have monolithic behavior. of the comparisons among experimental and numerical results were satisfactory, as presented in the item 3.3. To represent the friction between the walls of the precast and the filling material (grout) and the faces of the precast column, elements of contact were used, being definite contact superficies between the materials (contact superficies and target superficies). The contact superficies between the materials were represented by tow finite elements, named contact pair. For the contact superficies, the finite element Contact 173 was used and for the target superficies, the finite element Target 170 was used. These elements have three freedom degrees in each node and the geometric properties are the same as in the faces of the solid elements to which they are linked, which may have triangular or quadrangular geometry. Figure [9] presents the contact pairs (element Contact 173 and Target 170).
3.2 Materials properties

Developing a model able to represent the concrete behavior as close as the real behavior is a challenge. The reinforced concrete is an almost fragile material and has different behaviors in the compression and in the traction. In the compression, the curve tension vs. deformation of the concrete is elastic and linear until nearly 30% of the last compression force. After this point, the concrete loses rigidity and follows elevating the tension values until rupture force. Thereafter, there is no increase of the tension suffering softening. In the traction, the curve tension vs. deformation of the concrete is nearly elastic and linear until the tension of the maximal traction. After this point, the concrete cracks and its strength is not considering the softening in the traction. To model the concrete material, it is necessary to provide the program Ansys[23] the following input data: longitudinal elasticity module of the concrete; ultimate strength of the concrete to compression and traction; Poisson coefficient; and transfer coefficient of shear. Ansys[23] also allows as input data, the inclusion of a tension stress vs. deformation to represent the mechanical properties of concrete. This is normally done, when by convergence problems, the processing is abruptly interrupted by early crushing of the concrete. Kachlakev et al. [9] bring bigger information on this phenomenon. In the analyzed models, this problem did not occur. The longitudinal elasticity module of the concrete, Ec, as well as, the concrete traction strength, ftk, were determined based in the recommendations of NBR 6118:2007 [1]. Poisson coefficient, n, adopted to the concrete was equal to 0.2 and the shear transference coefficients, b adopted were equal to 1 to open and closed cracking. This value for the coefficient b was used, because tests performed showed bigger efficiency in the convergence of the processing when used the mentioned value, see Delalibera [5]. Concrete rupture criterion provided by Ansys was used. For the
Figure 7 Solid 65, Ansys
441
Figure 9 Finite Element Contact, Ansys
definition of the rupture superficies it is necessary only two parameters: the strengths to the last compression and traction of the concrete. Concrete rupture criterion is analogue to William-Warnke rupture criterion. Figure [10] presents the rupture superficies. In all pile caps was adopted compressive strength of concrete (fck) equal to 25 MPa. For the steel bars, it was adopted perfect elastic-plastic behavior. The elasticity module used was equal to 210 GPa, Poisson coefficient equal to 0.3 and the tension steel equal to 500 MPa. Through the tests performed, we verified that Newton-Raphson criterion was the one which presented the best results regarding
Figure 10 Failure Surface in Principal Stress Space, Concrete, Ansys
the convergence of the models, thus, in all analyses this criterion was used. For the properties of the contact elements, it was used the Coulomb friction model, being necessary to define the value of the friction coefficient, m, the maximal shear stress, tmaximum and two constants, FKN and FTOLN, FKN represent a normal rigidity coefficient of the contact element and FTOLN constant is a tolerance factor to be applied in the sense of the normal vector of the superficies. This factor is used to determine the penetration compatibility. The contact compatibility is satisfied if the penetration is in permissible tolerance (FTOLN measures the deepness of the underlying elements). The deepness is definite by the average deepness of each individual element of the contact in the pair. If the computer program Ansys[23] detects any penetration bigger than this tolerance, the global solution does not converge, even if the residual forces and the displacement increments are found in the criteria of the adopted convergence. For FKN coefficient we used value equal to 1 and for FTOL value equal to 0.1. The choice of the correct value of the friction coefficient is a hard task, as it depends on several factors: type of superficies, intensity of actions, mechanic properties of the materials which compose the link column-foundation. There are, in the technical literatures several indications for the value of the friction coefficient concreteconcrete. According to Nielsen [10] the value to be used is 0.6, EN 1992-1-1 [03] indicates that the friction coefficient value for the situation where the link column-foundation by half precast with smooth walls, must be higher than 0,3. Canha [11] and Ebeling [12] analyzed the influence of the friction coefficient in links columnfoundation by half precast, varying the value of 0.60, 0.45 and 0.30. Osani et al. [13] suggests that the values of friction coefficient have values equal to 0,5 and 1, function of the embed length and of the king of the precast and column walls conformation. In this work, suggesting the recommendations of Canha & El Debs [14] used the friction coefficient equal to 0,6. It is important to remember that this numerical analysis aims at presenting the behavior trend of the link column-foundation behavior through embed precast in blocks on two piles, with main end of analyzing the relevance of the analyzed factors.
442
In relation to the shear tension, tmaximum, we adopted the value used by the computer program, [y/(3)], in which y is the strength to the outflow of the rupture criterion of Von Mises, of the underlying material to the contact superficies. For y we used values indicated by Canha [11], where this tension values thirty per cent of the compressive strength of concrete of minor strength belonging to the contact. Therefore, y value adopted was equal to 7.5 MPa, resulting 2.5 MPa for the value of tmaximum.
3.3 Comparative analysis

In order to verify if the adopted models for the numerical analysis of the two pile caps show correct behavior trend, comparative analysis of experimental tests and numerical simulations of several researchers. In this work, it will be presented the results of the comparative numerical analyses of the tests performed by Mautoni [15] and Adebar et al. [16] and of the numerical simulation developed by Sam & Iyer [17]. In Delalibera [5], it can be observed the rest of the comparative analyses performed and more details of the numerical analyses developed. In the comparative analysis, we also used the same finite elements used in the analyses of the two pile caps, i.e., Solid 65 discretizing the concrete material and Link 8 modeling the steel bars. All mechanic and geometric properties adopted in the comparative analysis were the same as in the experimental tests. The contour conditions used in the tests with higher possible reality degree, the same thing happening to the loading. The first comparative analysis is of the pile cap B1-A tested by Mautoni [15]. The pile cap had rupture by shear with last force equal to 800 kN. The first crack appeared with nearly thirty per cent of the last force, about 240 kN. The pile caps was twenty five centimeters high, the column was squared with a 225 cm area and the piles had transversal sections equal to 10 cm x 15 cm. The compressive strength of concrete was equal to 32.30 MPa and the steel bars of the tie presented strength of equal to 720MPa. As there was no information on the elasticity module and on steel and concrete Poisson coefficients, we adopted the recommendations of NBR 6118:2007 [1].
We applied the numerical model three hundred force increments, whereas to each increment the applied force value was 2,67 kN. Figure [11] shows the final configuration obtained in the experiment and in the simulation of the block B1-A. The force value which originated the first crack in the numerical model was 312,33 kN. It occurred difference of 23.15% in relation to the force value which originated the first crack in the experimental model. It occurred because in the experimental model, the force that provoked the first crack was determined as a function of visual observation, i.e, the first crack visible to the human eye. In relation to the last force, the numerical model presented last force of 799,98 kN, practically, did not occur difference with the value experimentally obtained. The second comparative analysis is two blocks tested by Adebar et al. [16]. Pile caps A, B, C, D and F were simulated. All pile caps were sixty centimeters tall and Poisson coefficients equal to 0,3 and 0,2 for steel and concrete respectively. In all pile five hundred force increments were applied. Pile cap A was constituted by four piles of twenty centimeters and column with squared transversal section with thirty centimeters side. The average compressive strength of concrete (fcm) obtained in the tests was equal to 27,10 MPa. The steel bars had tensile strength equal to 479 MPa. In the test, the first crack emerged with force equal to 1186 kN and the rupture force was equal to 1781 kN. In the numerical model, the first crack occurred with force equal to 1403,86 and the ruin force obtained was equal to 1781,10 kN. Pile cap B had the same geometric properties of Pile cap A, however, the reinforcement was distributed on piles, while in Pile A, the reinforcement was distributed in mesh. The compressive strength of concrete was 24,80 MPa and the tensile strength the steel bars of the tie was the same as in Pile cap A. In the test, the first crack emerged for a 1679 kN force, now, in the numerical model, the first crack occurred with 1505,71 kN. The experimental ruin force was registered with value equal to 2189 kN and in the numerical simulation, the force obtained was 2186 kN. Pile cap C had six piles with diameters of twenty centimeters and column with the same transversal section of the other models. The
Figure 11 Final configuration of the pile cap B1-A, tested by Mautoni (1972)
443
Figure 12 Pile cap A, Adebar et al. [7]
average compressive strength of concrete was 27.10 MPa while the tensile strength the tie was equal 479 MPa. The first crack in the test emerged to a force of 1780 kN and the rupture force observed was 2892 kN. In the numerical model these forces were 1588.30 kN and 2647.70 kN respectively. Pile cap D was similar to Block B, changing only the area of the steel bars of the reinforcement. The compressive strength of concrete was equal to 30.30 MPa and the tensile strength of the bars was equal to 486 MPa. The first crack presented for the force of 1122 kN and the pile cap rupture occurred for the force of 3222 kN. In the numerical simulation, the force that originated the first crack was 1097.74 kN and the ruin force observed was 3212.17 kN. In spite of obtaining good correlation between the cracking and last forces, the numerical simulation presented substantial differences in relation to the rigidity of Pile cap D experimentally tested, see Figure [15]. Pile cap F had the same mechanical properties of Pile cap B, however, changed only its geometric form. The force that provoked the
Figure 13 Pile cap B, Adebar et al. [7]
Figure 15 Pile Cap D, Adebar et al. [7]
Figure 14 Pile cap C, Adebar et al. [7]
Figure 16 Pile Cap F, Adebar et al. [7]
444
Table 2 Relations between numerical results and experimental pile caps tested by Adebar et al. [7]
Pile caps Bloco A Bloco B Bloco C Bloco D Bloco F Fu,exp (kN) 1781 2189 2892 3222 3026 Fu,num (kN) 1781,10 2186 2647,7 3212,70 2589,74 Fu,exp/Fu,num 0,99 1,00 1,09 1,01 1,17
Figure 18 Reinforcement distributed over the piles, Sam & Iyer [17]
first crack was equal to 650 kN and the rupture force observed was equal to 3026 kN. In the numerical simulation, these forces were 501,94 kN and 2589.74 kN. Figures [12], [13], [14], [15] e [16] showed the correlations obtained between the numerical and experimental results and Table [2] relations between the last experimental and numerical forces. The third comparative analysis refers to the blocks numerically simulated by Iyer & Sam [17]. The piles caps were simulated with arrangement reinforcement distributed in mesh and on the piles. The pile cap were twenty-two centimeters and five millimeters of height and were composed by four piles with square transversal section of 10 cm x 10 cm. The columns also had squared section with 14.14 cm x 14.14 cm side. The compressive strength of concrete was equal to 19 MPa and the tensile strength of the steel bars equal to 300 MPa. Poisson coefficient adopted to steel and concrete was 0.3 and 0.2 respectively. In the numerical simulations were adopted 345 force increments. The last forces obtained in the analyses developed by Iyer & Sam (1995) were: 600 kN, for reinforced on mesh and 560 kN, for reinforced on piles. In the numerical simulation developed in this text, the forces found were: 582.17 kN for the pile cap with reinforcement arrangement distributed on mesh and 594.59 kN for the pile caps with reinforcement arrangement distributed on piles. Figures [17] and [18] presented the correlations among the results obtained.
We observe in Figures [17] and [18] that there is great correlation among the results, indicating that the model adopted in the numerical analysis is consistent.
3.4 Influence of the length of the pile and the soil

As the experimental test of the pile caps with real length piles are of difficult execution, we simulated pile caps with real length piles, aiming to observe the behavior of the main stress outflow of compression and the length influence of the piles on the block. Thereunto, four pile caps were modeled three with piles with real lengths and
Figure 19 Result of SPT, Senna Jnior [18]
Figure 17 Main reinforcement mesh, Sam & Iyer [17]
5 2 4 3 4 4 7 9 9 7 7 9 11 14 12 15 13 14 18 13
brown argillaceous sand cenozoic sediment E = 46 MPa c = 15 KPa attrition angle = 22 pebble line N.W.: -10m Red argillaceous sand Soil residual - Bauru group E = 70 MPa c = 26 KPa attrition angle = 28
End of the puncture: -20 m
445
Figure 20 Pile caps analyzed actual length of the pile

Pile Cap A
35 117,5
Figure 22 Flow principal stress compression, Pile cap B
Pile Cap B
117,5
Pile Cap C
55 117,5 55
Pile Cap D
55 117,5 55
0m
Soil 1 Soil 2 Soil 3 Soil 4 Soil 5

-9 m
-18 m
-22,5 m
one with small height piles. In the model was used one pile cap on two piles with angle of inclination of the strut equal to 45. The length of the piles was determined according to the result of the survey executed by Geotechnical Department, in Campus I of USP in So Carlos. Thus, the length calculated of the piles shaft was equal to 9.0 m. Figure [19] shows the result of the survey the percussion (S.P.T.) and Figure [20] the pile caps numerically analyzed. The friction between the pile and the soil was not taken into account. For the existing soil around the pile shaft, we adopted a plastic behavior (rupture criterion Drucker-Prager). The result of the survey the percussion and mechanical properties of the soil were obtained in Senna Jnior [18]. The force was applied through
one hundred steps, admitting it centered. In the models where there was no contribution of the soil (Pile caps A and B Figure [20], it applied force until the pile cap rupture. In the other models, the action applied corresponded to the piles loading capacity, i.e., 600 kN. In the soil modeling, we used the model of elastic Continuum model (Soil 1, 4 and 5 Figure [20]). The soil could be modeled using Winkler hypotheses (contact pressures are proportional to settlements), however, according to Velloso [19], the Medium Continuum model represents with higher accuracy the phenomenon of the interaction soil-structure. To avoid localized disturbance stress on the force application point, it was modeled a steel plate on the head of the column five centimeters thick and elastic and linear material, the same for all the numerical analysis developed.
Figure 21 Flow principal stress compression, Pile cap A
Figure 23 Flow principal stress compression, Pile cap C
446
As it was intended to analyze the behavior of the pile caps, it was adopted in the piles and in the columns, compressive strength of concrete equal to 50 MPa. The coefficients of shear transference also had value equal to 1. The elasticity model of the steel and concrete, as well as their respective Poisson coefficients, were determined with recommendations of NBR 6118:2007 [1]. The reinforcement rates were equal in all models, whereas, the areas of
reinforcement of piles, of columns and ties are presented in Table [3]. For more information, see Delalibera [5]. By Figures [21] to [24], it can be observed that there is more concentration of compression stress in the pile caps sections just below the column and in the section of the piles farther from the pile cap board, sections F and G Figure [20]. It was verified that the piles not requested uniformly, i.e., the sections farther from the pile
Table 3 Criteria used in the pile cap modeling, considering the influence of the piles and soil
Estructural element Finite element Real constants Material properties Additional informaticon
Column
Solid 65
Pile caps
Solid 65
Piles
Solid 65
As, tirante As, pilar e estacas Steel plate Stirrups
Link 8 Link 8 Solid 65 Link 8
As = 3,15 cm i = 0
2
As = 1,25 cm i = 0
2
As = 0,50 cm2 i = 0
Soil 1
Solid 65
Soil 2 Soil 3 Soil 4 Soil 5
Solid 65 Solid 65 Solid 65 Solid 65
Ec = 33658 MPa = 0,3 Plastic behavior concrete fck = 50 MPa ftk = 4,07 MPa =1 Ec = 21287 MPa = 0,3 Plastic behavior concrete fck = 20 MPa ftk = 2,21 MPa =1 Ec = 33658 MPa = 0,3 Plastic behavior concrete fck = 50 MPa ftk = 4,07 MPa =1 Es = 210 MPa fy = 500 MPa Elastic-plastic perfect behavior = 0,3 Es = 210 MPa fy = 500 MPa Elastic-plastic perfect behavior = 0,3 Es = 210 MPa Elastic-plastic perfect behavior = 0,3 Es = 210 MPa fy = 500 MPa Elastic behavior = 0,3 Esolo = 46 MPa c = 15 KPa Plastic behavior Drucker-Prager solo = 22 = 0,3 Esolo = 70 MPa c = 26 KPa Plastic behavior Drucker-Prager solo = 28 = 0,3 Esolo = 46 MPa Elastic behavior = 0,3 Esolo = 70 MPa Elastic behavior = 0,3 Esolo = 1000 MPa Elastic behavior = 0,3
447
Figure 24 Flow principal stress compression, Pile cap D
Figure 25 Nodal sections investigated
caps boards (sections F and G Figure [25]) were more solicited than the sections closer to the board (sections D and I Figure [25]). This phenomenon occurred in all models analyzed. As a function of the difference between the intensity of forces applied in the pile caps without the presence of the soil (blocks A and B Figure [20]), which were bigger, in relation to the blocks modeled with the soil (blocks C and D Figure [20]), occurred differences in the intensities of the main compression stress, what was foreseen, once the rupture of the set soil-structure of the pile caps C and D occurred by the soil. Through the values presented in Table [4] it is possible to conclude that the sections F and G, in the inferior nodal zone (near the piles) were more requested, because with the distribution of the stress deriving from the strut does not happen uniformly on the head of the piles, the regions closer to the column were the ones which presented higher stress. These results are similar to the experimental results obtained by Delalibera & Giongo [25]. According to these results, all analyses developed were performed using short piles, because it was found that the distribution of the flows of main compression stress is little influenced by the piles rigidity. An interesting result observed during these analyses was in rela-
tion to standardization of the main stress of compression on the piles. In blocks B, C and D (see Figure [20]), which had long length piles, it was checked that the stress were uniformed nearly to 1/3 of the height of the pile shaft.
4. Analysis of variance
The analysis of variance is a statistics test widespread among the statistical professionals and fundamentally aims at verifying if there is substantial difference among the averages and if the factors exercise influence on any dependent variable. The proposed values can be of qualitative or quantitative origin, but the dependent variable must be necessarily continuous. The main application of ANOVA (analysis of variance) is the comparison of averages coming from different groups, also called treatments. There are two kinds of problems to be solved by ANOVA: fixed factors or random factors. The randomness determines the question of the problem. In most of the cases they are risk factors, after all, the second kind of problem (random) will only emerge when a study involving random choice of factors is performed. The variance analysis is broadly used in several areas. In Civil
Table 4 Intensities compressive principal stress, considering the influence of the piles and soil
Section A B C D=I E=H F=G Tenses principais de compresso (MPa) Pile cap B Pile cap C 11,5 7,5 11,5 3,5 7,5 19,5 13,3 22 13,3 4,2 13,3 30,6
Pile cap A 16,7 11,9 16,7 5,10 10,9 28,3
Pile cap D 11,7 21,3 11,7 2,10 11,7 21,3
448
Table 5 Analysis of Variance, addressing general, Montgomery [24]

Factor M N MxN Error Total Sum of squares SSM SSN SSMN SSE SST Freedom degrees a 1 b1 (a 1) (b 1) abc (n 1) abcn 1 Square average MSM = SSM / (a 1) MSN = SSN / (b 1) MSMN = SSMN / [(a 1) (b 1)] MSE = SSE / [abc (n 1)] F0
F0
F0
F0
MSM MSE
MSN MS E
MSMN MS E
Engineering, its use is still restricted, but, there are already works developed using the technique, as: Lima Jnior [20], Delalibera [5] and Pituba et al. [21]. In the variance analysis developed in this work, we used fixed factors, choosing three study variables: the embed length of the column in the sockets (emb); the thickness of the pile caps bottom slab (hs); and the wall conformation of the embed chalice and the precast column. The variables chosen totalized eighteen cases of combinations. The models were divided in two groups (smooth walls L and rough walls R).
(2)
(3)
To verify the relevance of a determined main variable fixed or combined, it is applied test F. Through tabulated values of Fcritical, provided by Montgomery [24], it is compared the value calculated of F0 with the value of Fcritical. If the calculated value of F0 is higher than the tabulated value of Fcritical it means that this factor is relevant, otherwise, it implies that the factor does not have substantial importance. The values of Fcritical are a function of the number of freedom degrees and of each variable and of the total freedom degrees number.
4.1 Formulation of the variance analysis

Being N and M the main fixed factors of the variance analysis, a, b and c, the variations of these factors and n the number of replicas. In general there will be abcn possible combinations. If all the experiment factors are fixed, the problems can be easily formulated, obtaining results that indicate which of the analyzed factors are important as well as their combinations. Table [5] presents a variance analysis with two factors. To verify the relevance of a determined fixed principal factor or combinations among the main factors, it occurs the relation between the average of the squares of each main factor or combination of the main factors by the average of the squares of the mistakes. The division between the average of the squares of each main factor or combination of main factors by the average of mistakes is called F0. The number of freedom degrees of each main factor is equal to the number of variations of each factor less the unity. The number of freedom degrees of the main factors combined is the product between the main factors which were combined. The total sum of the squares is calculated through Equation [1]. The sum of the squares of the combination N x M is expressed through Equation [2]. The sum of the squares of the mistake is defined by Equation [3].
5. Results obtained
5.1 Analysis of two pile caps normal force, moment and smooth walls
Nine pile caps solicited by action of normal force of compression and moment were analyzed (the moment applied to the pile cap was obtained through application of a horizontal force applied on the top of the column). The pile caps presented variations in the embed length of the column (emb) and in the thickness of the slab of pile cap bottom (hs). The variation of the factors analyzed modified substantially the distribution of the main stress of compression and the panorama of cracks in the last force increment applied to the models. Table [6] presents the results of the numerical analyses performed. Considering the results of Table [6], it can be observed that for minor embed lengths of the column and minor thicknesses of the bottom slab (in case of the models Le60hs10NM and Le50hs10NM) occurred substantial differences in relation to the analytical values.
(1)
449
Table 6 Analysis of variance, ANOVA, blocks with smooth conformation and action of compressive force eccentric
Measure the sides of the column (cm) 20 x 20 20 x 20 20 x 20 20 x 30 20 x 30 20 x 30 20 x 40 20 x 40 20 x 40 Measure Fu the sides h e hs emb (kN) Fblvot / Fnum of the (cm) (cm) (graus) (cm) (cm) piles (cm) Blvot Num. 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 30 20 10 30 20 10 30 20 10 80 80 80 60 60 60 50 50 50 54,0 51,3 48,4 48,4 45,0 41,2 45,0 41,2 36,9 120 110 100 100 100 80 90 80 70 3663 3415 3128 3128 2800 2428 2800 2428 2016 3663 3415 3128 3128 2800 2428 2800 2428 2016 1,27 1,17 1,28 1,18 1,22 1,91 1,24 1,96 3,16 2,44 3,00 2,84 2,90 4,35 5,96 3,59 4,01 4,44 Ab,est (cm2) Blvot Num. 323,6 269,6 487,8 365,8 673 639,4 299,1 299,1 441,9 397,7 592,8 497,9 282,8 207,3 411,7 288,2 355,9 177,9
Pile caps Le80hs30NM Le80hs20NM Le80hs10NM Le60hs30NM Le60hs20NM Le60hs10NM Le50hs30NM Le50hs20NM Le50hs10NM
Note: Num., value obtained by numerical simulation Blvot, value calculated using the criteria Blvot & Fremy [4] Fu, Rupture force applied in the column, and eccentricity of the normal force, Ab,est, area of the strut near the column.
This shows that the resistant capacity of the block cannot be determined in function of the rupture of the compressed diagonal, but by the puncturing effect of the bottom slab. It is interesting to observe that, the puncturing effect will only occur if, and only if, the force on the column is transmitted (almost all) to the bottom slab and, this will only happen if the embed length is small, not being possible the formation of the connection (or stanchion), being the element rupture characterized by punching shear. Figure [26] presents bends reaction in the most solicited pile vs. displacement in the middle of the pile cap span for the models numerically analyzed.
Figure [27] shows the distribution of the main stress of compressions inside one of the numerically analyzed pile caps and Figure [28] shows the panorama of cracking of the same pile cap presented in Figure [27], referring to the last force increment. For the pile caps models with conformation of the sockets and column walls, with performance of normal force and moment in the pile cap, factor emb is the main relevant factor, followed by factor hs. Table [7] shows the results of the analysis of variance of the pile caps models with conformation of the walls of the sockets and col-
Figure 26 Curve Force vs. displacement in the middle of the span, pile caps with smooth conformation, normal force and moment
Figura 27 Principal stress compression, pile cap with smooth conformation, normal force and moment
Le60hs20NM
450
Figure 28 Panorama cracking, pile caps with smooth conformation, normal force and moment
Figura 29 Principal stress compression, pile cap with rough conformation, normal force and moment
Le60hs20NM
umn, with performance of normal force and moment in the block. Through numerical analysis, it was observed that the existing differences in the areas of the struts with the piles, when compared to the areas of the strut calculated using the criteria by Blvot & Frmy [4] also occurred, corroborating with the experimental results found by Delalibera & Giongo [25]. In Table [6] are presented the results of the areas of strut calculated by Blvot criterion (considering a slant of q) and the areas of the strut obtained numerically (for the calculation of the areas of the strut obtained through numerical results, it was used the values of the main compression stress with the piles), as well as the eccentricities of the normal force of each model.
Le60hs20NM
slab (hs). The variation of the factors analyzed modified, also in this case, the distribution of the main compression stress and the crack panorama in the last forces increment (horizontal and vertical) applied to the models. Table [8] presents results of the numerical analyses, Figure [29] shows the distribution of the main compression stress and Figure [30] the crack panorama of the pile caps numerically analyzed. The cause of the precocious rupture of the model Re60hs10NM is probably related to the punching shear of the bottom slab of the pile cap and the rupture of the model Re50hs20NM, due to the small embed length of the column. Through Table [8], it can be observed the differences among the areas of the strut with the piles, calculated by analytic criterion developed by Blvot & Frmy [4] and using the numerical models. In the same table, are presented the values of the eccentricities of the normal force of each model numerically analyzed. It is possible to observe through Figures [29] and [30] the formation of the compressed diagonals and the panorama of cracking of one of the blocks numerically analyzed. It is noted that the strut and tie
5.2 Analysis of the two pile caps normal force, moment and rough walls
Nine pile caps solicited by action of normal compression force and moment were analyzed, with variations in the embed length of the column (emb) and in the thickness of the bottom pile cap
Table 7 Analysis of variance, ANOVA, pile caps with smooth conformation and action of compressive force eccentric
Factor emb hs emb x hs Erro Total Sum of squares 2805913 2026374 302545 194219 5329056 Freedom degrees 2 2 4 3 8 Square average 1402956 1013187 75637,5 64734 666132 F0 21,67 15,65 1,16 Significance F0,01 8,65 8,65 7,01
Note: emb x hs, coupling between the embedded length and the thickness of the column slab background.
451
Table 8 Results obtained by numerical analysis for blocks with rough conformation and action of compressive force eccentric
Measure the sides of the column (cm) 20 x 20 20 x 20 20 x 20 20 x 30 20 x 30 20 x 30 20 x 40 20 x 40 20 x 40 Measure Fu the sides h e hs emb (kN) Fblvot / Fnum of the (cm) (cm) (graus) (cm) (cm) piles (cm) Blvot Num. 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 30 20 10 30 20 10 30 20 10 80 80 80 60 60 60 50 50 50 54,0 51,3 48,4 48,4 45,0 41,2 45,0 41,2 36,9 120 110 100 100 100 80 90 80 70 3663 3415 3128 3128 2800 2428 2800 2428 2016 2832 2660 2460 2360 2368 1103 2260 1253 1828 1,29 1,28 1,27 1,33 1,18 2,20 1,24 1,94 1,10 3,18 3,76 3,48 3,93 4,74 7,13 3,91 5,03 4,47 Ab,est (cm2) Blvot Num. 323,6 487,8 673 299,1 441,9 592,8 282,8 411,7 355,9 223,8 404,9 639,4 299,1 371,2 462,4 243,2 308,8 153,1
Pile caps Re80hs30NM Re80hs20NM Re80hs10NM Re60hs30NM Re60hs20NM Re60hs10NM Re50hs30NM Re50hs20NM Re50hs10NM
Note: Num., value obtained by numerical simulation Blvot, value calculated using the criteria Blvot & Fremy [4] Fu, Rupture force applied in the column, and eccentricity of the normal force, Ab,est, area of the strut near the column.
model must be modified according to the external actions acting on the pile cap, i.e., idealize a model that contemplates the actions of moment, normal force and horizontal force. Figure [31] shows bends reaction in the most solicited pile vs. displacement in the middle of the pile cap span. The behavior of the pile cap submitted to the action of moment and compression force, indicate that the pile caps with more rigidity present higher bearing capacity and the presence of moment in them, reduce it. Delalibera [5] experimentally proves this statement.
Delalibera & Giongo [26], analyzed the same pile caps of this work, but it was applied only centered compression force. It was observed that the bearing capacity of the pile caps with only centered force was superior to the pile caps with eccentric force. Table [9] presents the results of the variance analysis of the pile caps submitted to eccentric compression force and rough conformation of the column-sockets interface. Another very interesting result consistent with statements of other researchers can be observed through the values of Table [9]. As the link with the shear key grants the link column-foundation monolithic behavior, it is expected that the rupture of the pile caps is associated to other factors. It was verified, because, it is noted that
Figura 30 Panorama cracking, pile caps with rough conformation, normal force and moment
Figura 31 Curve maximum reaction vs. displacement in the middle of the span, pile caps with rough conformation, normal force and moment
Le60hs20NM
452
6. Conclusion
Figure 32 Strength of reinforcement, Le60hs20NM, strengths expressed in MPa
The numerical simulations proved to be coherent and pointed tendency of the structural behavior of two pile caps with embed socket, with smooth and rough conformation of the socket walls and of the columns. It was observed for the pile caps with smooth interface between the socket and column walls, the following factors: embed length of the column - emb and thickness of the bottom slab hs, has got relevant importance in the bearing capacity of the models. Because, the models numerically analyzed with bigger embed lengths of the column and bigger thicknesses of the bottom slab, presented higher bearing capacity. Furthermore, it was evident that in the situations in which there was embed length of the column diminished and small thickness of the bottom slab, the possible ruin occurred by punching shear of the pile caps in the region of contact between the column base and the pile cap base. For the models simulated with rough interface between the socket and column (shear key), the factors analyzed ( emb and h s) were not relevant in the determination of the bearing capacity of the pile caps. These results are interesting, because it is possible to project pile caps with embed socket and shear key, with strut and tie model, only getting worried with the punching shear of the bottom slab of the pile cap, during the assembling phase (i.e., before the application of the filling material), since respecting the minimal embed length of the column in the socket, suggested by NBR 9062:2006. It is necessary a higher number of experimental tests, so that it can be verified the possibility of the reduction of the embed length of the column in the socket. It was also observed that the rigidity of the pile cap influences in the bearing capacity of it. This can be stated as a function of the results obtained, because the models with higher heights were the ones that presented higher resistant capacity. It was found, and it was expected, that the models numerically analyzed with smooth interface, presented minor bearing capacity in relation to the models numerically analyzed with rough interface.
the factors previously chosen for this analysis were not relevant, i.e., did not present substantial values. Such results corroborate with the results presented by Canha & El Debs [14], where it was found that respecting the minimal embed length of the column established by NBR 9062:2006, it can be considered monolithic the behavior of the link column-sockets.
5.3 Stress on the steel bars

Analyzing Figure [32], it can be noticed that reinforcement of the models presented consistent values and in some cases, occurred yield. It was observed that in some reinforcement of the piles, occurred traction tensions. These results are coherent as a function of the external actions applied. The results presented in Figure [32] corroborate with the results presented by Adebar et al. [7], Miguel [6] and Delalibera [5], i.e., the stress in the reinforcement is not constant, having null values (or even negative) on the tip of the rods and maximum values in the middle of the span. It can also be observed that, the reinforcement that of the piles present stress with different values, what indicates flexion-compression on the piles.
7. Acknowledgments
To: CNPq Conselho Nacional de Desenvolvimento Cientfico e Tecnolgico and to FAPESP Fundao de Amparo Pesquisa do Estado de So Paulo for the financial support to perform the research that made it possible to write this paper.
Table 9 Analysis of variance, ANOVA, blocks with rough conformation and action of compressive force eccentric
Factor emb hs emb x hs Erro Total Sum of squares 1697377 875380 43964 1606137 4223308 Freedom degrees 2 2 4 3 8 Square average 848688 437915 10991 535379 527913 F0 1,58 0,82 0,02 Significance F0,01 8,65 8,65 7,01
Note: emb x hs, coupling between the embedded length and the thickness of the column slab background.
453
8. Bibliographical references
[01] ASSOCIAO BRASILEIRA DE NORMA TCNICAS, ABNT NBR 6118:2003 Projeto de estruturas de concreto, 2007. Rio de Janeiro. [02] ASSOCIAO BRASILEIRA DE NORMA TCNICAS, ABNT NBR 9062:2005 Projeto e execuo de estruturas de concreto pr-moldado, 2006. Rio de Janeiro. [03] EN 1992-1-1, Eurocode 2: Design of concrete structures Part 1-1: General rules and rules for buildings, 2004. European standard. [04] BLVOT, J.; FRMY, R.. Semelles sur piex. Analles dInstitut Techique du Btiment et des Travaux Publics, Paris,1967, v. 20, n. 230, p. 223-295, fev. [05] DELALIBERA, R. G.. Anlise terica e experimental de blocos de concreto armado sobre duas estacas submetidos a ao de fora centrada e excntrica. Tese (doutorado) Escola de Engenharia de So Carlos, Universidade de So Paulo, 2006, So Carlos. [06] MIGUEL, G. M.. Anlise experimental e numrica de blocos sobre trs estacas. Tese (doutorado) Escola de Engenharia de So Carlos, Universidade de So Paulo, 2000, So Carlos. [07] ADEBAR, P.; KUCHMA, D.; COLLINS, M. P.. Strut-and-tie models for design of pile caps: an experimental study. ACI Journal, 1990, v. 87, p. 81-91, Jan/Feb. [08] MONTOYA, P.J.; MESEGUER, A.; CABRE, M. Hormigon Armado 14.a Edicin Basada em EHE ajustada al Cdigo Modelo y al Eurocdig, 2000. Barcelona, Gustavo Gili. [09] KACHLAKEV, D.; MILLER, T; YIM, S.; CHANSAWAT, K.; POTISUK, T.. Finite element modeling of reinforced concrete structures strengthened with frp laminates. Final Report, 2001. Oregon Department of Transportation, May. [10] NIESEN, M. P.. Limit analysis and concrete plasticity. Prentice-Hall series in Civil Engineering, Englewood Cliffs, 1984. New Jersey, 420 pg. [11] CANHA, R. M. F.. Estudo terico-experimental da ligao pilar-fundao por meio de clice em estruturas de concreto pr-moldado. Tese (Doutorado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 2004. So Carlos. [12] EBELING, E. B.. Anlise da base de pilares pr-moldados na ligao com clice de fundao. Dissertao (Mestrado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 2006. So Carlos. [13] OSANAI, Y.; WATANABE, F.; OKAMOTO, S.. Stress transfer mechanism of socket base connections with precast concrete columns. ACI Structural journal, 1996, v. 93, n. 3, p. 226-276, May/June. [14] CANHA, R. M. F.; EL DEBS, M. K.. Proposta de modelo de prometo para a ligao pilar-fundao por meio de clice em estruturas de concreto pr-moldado. Revista Ibracon de Estruturas, 2006. Vol. 2, n. 2, p. 137-166, junho.
[15] MAUTONI, M.. Blocos sobre dois apoios. Grmio Politcnico, 1972. So Paulo. [16] ABEBAR, P.. One-way shear strength of large footings. Can. J. Civ. Eng, 2000. 27: 553562. [17] SAM, C., IYER, P. K.. Nonlinear finite element analysis of reinforced concrete four-pile caps. International Journal of Structures, 1995, v. 15, n. 1, p. 18-34, Jan/Jun. [18] SENNA JNIOR, R. S.. Distribuio de carga em grupos de estacas escavadas de pequeno dimetro. Dissertao (Mestrado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 1993. So Carlos. [19] VELLOSO, D. A.; LOPES, F. R.. Fundaes. COPPE/ UFRJ, 1996. Rio de Janeiro. [20] LIMA JNIOR, H. C.. Avaliao da ductilidade de pilares de concreto armado, submetidos a flexo-compresso reta com e sem adio de fibras metlicas. Tese (Doutorado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 200., So Carlos. [21] PITUBA, J. J. C. ; DELALIBERA, R. G.; RODRIGUES, F. S. . Numerical and Statistical Analysis about Displacements in Reinforced Concrete Beams using Damage Mechanics, 2012. Computers and Concrete, an International Journal (Print). [22] DELALIBERA, R. G.. Anlise numrica de blocos sobre duas estacas, com clice embutido, submetido ao de fora normal e excntrica. Relatrio cientfico, 2012. Universidade Federal de Gois / CNPq Conselho Nacional de desenvolvimento cientfico e tecnolgico. [23] ANSYS Users Manual. Theory Manual, 2011. ANSYS revision 5.5. [24] MONTGOMERY, D. C.. Design and Analysis of Experiments Fourth Edition, 1996. Arizona State University, John Wiley & Sons. [25] DELALIBERA , R. G.; GIONGO, J. S.. Deformations in the strut of two pile caps. Ibracon Structures and Materials Journal - RIEM, 2008, v. 1, p. 121-138. [26] DELA LIBERA , R. G. ; GIONGO, Jos Samuel. Anlise de varincia de blocos sobre duas estacas com clice embutido,submetido ao de fora de compresso centrada. Revista Eletrnica de Engenharia Civil, 2012, v. 1, p. 1-16.
454
Numerical analysis of two pile caps with sockets embedded, subject the eccentric compression load Anlise numrica de blocos sobre duas estacas com clice embutido, submetido ao de fora de compresso excntrica
R. G. DELALIBERA a delalibera@pq.cnpq.br J. S. GIONGO b jsgiongo@sc.usp.br
Abstract
The structural behavior of pile caps with sockets embedded is influenced by interface of column-socket, which can be smooth or rough. With intent to analyze the behavior of two pile caps with embedded socket, considering the friction between the column and the socket, with eccentric normal load, the numerical simulations were carried out, using a program based on the Finite Element Methods (FEM). In the numerical analysis the non-linear behavior of materials was considered, also the friction between the column and the socket. It was considered perfect bond between the reinforcement and the concrete around. It was observed that the embedded length is preponderant factor in the structural behavior of the analyzed element. Keywords: pile caps, foudations, precast concrete.
Resumo
O comportamento estrutural dos blocos sobre estacas com clice embutido influenciado pelo tipo de interface pilar-clice, a qual pode ser lisa ou rugosa. Com o objetivo de analisar o comportamento de blocos sobre duas estacas com clice embutido, considerando a rugosidade da interface pilar-clice, submetido ao de fora normal excntrica, realizaram-se simulaes numricas, desenvolvidas em programa de computador baseado no mtodo dos elementos finitos MEF. Na anlise numrica as no-linearidades fsicas dos materiais concreto e ao foram consideradas, como tambm o atrito nas interfaces entre as paredes do pilar e do clice. Considerou-se aderncia perfeita entre as barras de ao das armaduras com o concreto do entorno. Observou-se que o comprimento de embutimento do pilar no interior do bloco fator preponderante no comportamento estrutural do elemento em anlise. Palavras-chave: blocos sobre estacas; fundaes; concreto pr-moldado.
Departamento de Engenharia Civil, Universidade Federal de Gois Campus Catalo, dellacivil@pq.cnpq.br, Av. Lamartine Pinto de Avelar, 1120, CEP: 75704-020, Catalo GO, Brasil. b Departamento de Engenharia de Estruturas, Escola de Engenharia de So Carlos, Universidade de So Paulo, jsgiongo@sc.usp.br, Av. Trabalhador Socarlense, 400, CEP: 13566-590, So Carlos SP, Brasil.
a
Received: 04 Sep 2012 Accepted: 07 Mar 2013 Available Online: 12 Jun 2013
2013 IBRACON
1. Introduo
A escolha do tipo de fundao feita aps anlise que considere as condies tcnicas e econmicas da obra, as proximidades dos edfcios limtrofes, a natureza e caracterstica do subsolo, as magnitudes das aes e dos tipos de fundaes disponveis no mercado. As fundaes em estacas so adotadas quando o solo em suas camadas superficiais no capaz de suportar aes oriundas da superestrutura, sendo necessrio, portanto, considerar resistncia em camadas profundas. Ao se utilizar soluo em fundao profunda por estacas (ou tubules), faz-se necessrio a construo de outro elemento estrutural, o bloco de coroamento, tambm denominado bloco sobre estacas. Segundo a NBR 6118:2007[1], blocos sobre estacas so importantes elementos estruturais cuja funo transferir as aes da superestrutura para um grupo de estacas. Esses elementos estruturais, apesar de serem fundamentais para a segurana da superestrutura, geralmente no permitem a inspeo visual quando em servio, sendo assim, importante o conhecimento de seu real comportamento nos Estados Limites de Servio e ltimo. A Figura [1] ilustra esse elemento estrutural. Quando a superestrutura pr-moldada faz-se necessrio que o pilar seja embutido no bloco. O comprimento de embutimento do pilar no interior do bloco, para que o mesmo seja considerado engastado, funo dos esforos solicitantes (momento fletor, fora normal e fora cortante) e do tipo de conformao superficial das paredes (rugosa ou lisa) do clice e do pilar pr-moldado. Os comprimentos de embutimento so definidos pela NBR 9062:2006[2]. Existem algumas variaes nos tipos de blocos com clice (clice externo, clice parcialmente embutido e clice totalmente embutido). A Figura [2] mostra as variaes dos blocos sobre estacas para pilares pr-moldados. O meio tcnico utiliza com maior intensidade os blocos com clice externo e parcialmente embutido. O cdigo europeu (EUROCODE 2) EN 1992-1-1 [3] apresenta recomendaes a respeito do projeto de clice para ligao pilar-fundao, considerando o comportamento monoltico do conjunto quando se utiliza chave de cisalhamento. Para conformao lisa das paredes
Figura 1 Blocos sobre estacas com clice embutido
do clice e do pilar, indica que o coeficiente de atrito entre as faces do pilar e as faces do clice seja maior que 0,3 e o comprimento de embutimento do pilar maior ou igual a 1,2 vez a maior dimenso do pilar. Como no h consenso junto aos meios tcnico e cientfico com relao seo do bloco onde se forma a escora (ou biela), realizou-se anlise numrica, com intuito de observar o comportamento estrutural dos blocos. Para isso foram analisados numericamente dezoito blocos sobre duas estacas, nove com interface rugosa e nove com interface lisa. O comprimento de embutimento do pilar (emb) e a espessura da laje de fundo do bloco (hs) foram variados. Assim, os comprimentos de embutimento dos pilares pr-moldados foram iguais a 80 cm, 60 cm e 50 cm. As espessuras da laje de fundo foram iguais a 30 cm, 20 cm e 10 cm. Laje de fundo do bloco, no o melhor termo para expressar a dimenso hs, em funo de no existir uma
Figura 2 Blocos sobre estacas com clice externo, parcialmente embutido e totalmente embutido
456
laje propriamente dita no fundo do bloco. Porm, julga-se que esse termo o que apresenta melhor compreenso para os leitores. O objetivo deste trabalho analisar o comportamento estrutural de blocos sobre duas estacas com clice embutido, com interfaces lisa e rugosa das paredes dos pilar e do clice, submetido ao de fora normal de compresso excntrica e, verificar por meio de anlise de varincia, qual a varivel apresenta maior relevncia no comportamento dos mesmos. O mtodo empregado foi o relativo a trabalhos com nfase em anlise numrica. Definidos os modelos que foram estudados, anlise estatstica foi desenvolvida, utilizando valores obtidos por meio de simulaes numricas. Os blocos de concreto armado sobre estacas foram analisados numericamente por meio de programa de computador baseado nos mtodos dos elementos finitos MEF. As no-linearidades fsica e geomtrica foram consideradas. A fissurao do concreto e a armadura do bloco tambm foram levadas em considerao. Para a anlise numrica dos blocos via mtodo dos elementos finitos, utilizou-se o programa ANSYS [23] (ANalyser SYStem).
2. Propriedades geomtricas dos modelos analisados numericamente

Os modelos foram dimensionados com os critrios de Blvot &
Frmy [4]. As taxas de armadura foram fixadas, pois no eram objetos de anlise. A distncia entre os eixos das estacas tambm foi fixada, variando-se o comprimento de embutido do pilar (emb) e a espessura da laje de fundo do bloco (hs). O dimensionamento das barras de ao da armadura principal de trao do bloco e as verificaes do esmagamento do concreto da diagonal comprimida junto zona nodal inferior (bloco-estaca) e junto zona nodal superior (bloco-pilar) foram feitos de acordo com as recomendaes do modelo proposto por Blvot & Frmy [4]. Utilizaram-se como limites para verificao das tenses junto s zonas nodais superior e inferior os valores observados pelos pesquisadores franceses, ou seja, 1,40fck junto ao pilar e fck junto estaca. Optou-se por utilizar ancoragem reta, pois verificou-se por meio das recomendaes da NBR 6118:2007[1] que o gancho no era necessrio (critrios de ancoragens em apoios extremos). Ao admitir essa hiptese o pesquisador no considerou (deciso de projeto) a recomendao da NBR 6118:2007[1] que exige a utilizao de ganchos em blocos sobre estacas. Lembram-se os resultados obtidos por Delalibera [5], Miguel [6] e Adebar et al.[7], para blocos rgidos sobre estacas, onde a tenso na ponta do tirante foi despresvel, corroborando com esta desciso. Para o dimensionamento dos blocos, utilizou-se um bloco padro, ou seja, todos os blocos tiveram as mesmas dimenses em planta do blo-
Figura 3 Propriedades geomtricas do bloco padro
457
co padro, variando-se apenas os comprimento de embutimento do pilar (emb) e a espessura da laje de fundo do bloco (hs). Subentende-se, ento, que as alturas totais dos modelos e os ngulos de inclinao da biela de compresso tambm sofreram variaes. Para o bloco padro adotaram-se estacas de seo transversal quadrada igual a 30 cm de lado, pilar tambm com seo transversal quadrada, porm com 40 cm de lado. O comprimento de embutimento do pilar no bloco padro foi igual a 60 cm. Esse valor representa o comprimento mnimo de embutimento para paredes lisas (clice e pilar) especificado na NBR 9062:2006 [2], quando a relao [Md/(Ndh)] 0,15. A Figura [3] apresenta as propriedades geomtricas do bloco padro. A distncia entre os eixos das estacas foi determinada em funo do ngulo limite inferior estabelecido pelos pesquisadores franceses, ou seja, 45. Portanto, a altura total do bloco padro foi igual a 90 cm, e as dimenses em plantas foram iguais a 240 cm por 84 cm. A distncia entre os eixos das estacas de todos os blocos foi igual a cento e oitenta centmetros. O comprimento de embutimento das estacas na face inferior do bloco seguiu sugesto de Montoya et al. [8], isto , embutiu-se dez centmetros do fuste da estaca no interior do bloco. Com relao aos pilares e as estacas adotaram-se resistncias caractersticas do concreto compresso igual a 50 MPa procurando evitar assim, runa desses elementos. As armaduras das estacas foram compostas por oito barras de ao com dimetro de 20 mm e resistncia caracterstica de escoamento igual a 500 MPa, totalizando rea de ao igual a 25,1 cm2 (8 barras de 20 mm). Nos pilares utilizaram-se doze barras de 20 mm com resistncia caracterstica ao escamento igual a 500 MPa, totalizando rea de ao igual a 32,7 cm2. O comprimento externo do pilar foi igual a 100 cm. Para o material de preenchimento, adotou-se resistncia caracterstica compresso igual a 50 MPa (valor igual resistncia caracterstica do concreto do pilar). A Figura [4] mostra os fatores que apresentam variaes na anlise numrica. A Tabela [1] apresenta as propriedades dos blocos analisados com relao conformao das paredes lisas e com chave de cisalhamento (rugosa), solicitada pela fora de compresso supostamente centrada no pilar e por uma fora horizontal aplicada no topo do pilar. A nomenclatura utilizada na Tabela [1] descrita: L, blocos com conformao lisa das paredes do clice e do pilar; R, conformao rugosa das paredes do clice e do pilar; e80, embutimento do pilar igual a oitenta centmetros (demais nmeros so anlogos); hs30, espessura da laje de fundo igual a trinta centmetros (demais nmeros so anlogos); NM significa que foi aplicada fora de compresso excentrica. Na mesma tabela Ast representa a rea da seo transversal das barras de ao classe CA-50, Blx o comprimento do bloco, Bly a largura do bloco, emb o comprimento de embutimento do pilar pr-moldado e hs a espessura da laje de fundo do bloco. Utilizando as indicaes de Blvot & Frmy [4], fez-se uma previso das capacidades resistentes dos blocos.
analisados. A anlise numrica no tinha como objetivo calibrar curvas de resultados experimentais, mas sim, apresentar tendncias de comportamento dos modelos analisados. A geometria de todos os modelos foram criadas no programa de computador AutoCad e exportadas para o programa de computador ANSYS[23], por meio de extenso SAT. A Figura [5] apresenta a rede de elementos finitos utilizadas para a discretizao dos elementos de volume (concreto) e os elementos de barra (barras de ao das armaduras). A fora normal foi considerada por meio de uma presso de 20 MPa aplicada no topo do pilar. O momento foi considerado por meio de uma fora horizontal, tambm aplicada no topo do pilar, cujo valor foi igual a 100 kN, gerando um momento na base pilar de 100 kNm. Ver Figura [04]. Como os modelos eram simtricos, utilizou-se o recurso de simetria. Portanto, analisou-se apenas metade do elemento estrutural. Restringiu-se as translaes (nas trs direes, x, y e z) dos ns dos elementos posicionados nas pontas das estacas. A Figura [6] apresenta as condies de contorno empregadas nos modelos. O comprimento de um metro das estacas e do pilar, foi escolhido em funo do princpio de Saint Venant, assim, eliminaram-se as influncias das perturbaes de tenses nas regies de aplicao das foras e nas regies de restries de translaes.
3.1 Elementos finitos utilizados

Para a modelagem do material concreto, utilizou-se o elemento finito Solid 65. Esse elemento possui oito ns com trs graus de
Figura 4 Variao dos fatores escolhidos
3. Anlise numrica
O objetivo da anlise numrica foi fornecer resultados para a aplicao de uma anlise estatstica denominada ANOVA (anlise de varincia), sendo assim, blocos sobre duas estacas com clice embutido, com conformao das paredes lisas e rugosas foram
458
Tabela 1 Propriedades dos blocos analisados, para paredes lisas e rugosas

Altura total h (cm) 110 100 90 90 80 70 80 70 60 Medida dos lados do pilar (cm) 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 40 x 40 Medida dos lados da estacas (cm) 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30 30 x 30
Blocos Le80hs30NM Re80hs30NM Le80hs20NM Re80hs20NM Le80hs10NM Re80hs10NM Le60hs30NM Re60hs30NM Le60hs20NM Re60hs20NM Le60hs10NM Re60hs10NM Le50hs30NM Re50hs30NM Le50hs20NM Re50hs20NM Le50hs10NM Re50hs10NM
Ast (cm2) 22 22 22 22 22 22 22 22 22
Blx (cm) 240 240 240 240 240 240 240 240 240
Bly (cm) 84 84 84 84 84 84 84 84 84
emb (cm) 80 80 80 60 60 60 50 50 50
hs Classificao ABNT (cm) NBR 6118:2007 30 20 10 30 20 10 30 20 10 Rgido Rgido Rgido Rgido Rgido Rgido Rgido Rgido Rgido
Nota: h, altura total do bloco; Ast, soma das reas das sees transversais das barras de ao; Blx, dimenso total do bloco na direo longitudinal; Bly, dimenso total do bloco na direo transversal; emb, comprimento de embutimendo do pilar no bloco; hs, espessura da laje de fundo.
liberdade por n translaes nas direes x, y e z. O elemento apresenta deformaes plsticas, fissurao e esmagamento em trs direes ortogonais. No elemento Solid 65, a fissurao ocorre quando a tenso principal de trao em qualquer direo atinge a superfcie de ruptura. Depois da fissurao, o mdulo de elasti-
cidade do concreto tem valor igual a zero na direo considerada. O esmagamento ocorre quando todas as tenses de compresso atingem a superfcie de ruptura, subseqentemente, o mdulo de elasticidade tem valor igual a zero em todas as direes. A Figura [7] apresenta o elemento Solid 65.
Figura 5 Discretizao utilizada nos modelos
459
Figura 6 Condies de contorno e rede de elementos finitos
Figura 8 Link 8, Ansys
Na modelagem das barras de ao da armadura foi utilizado o elemento finito Link 8. Esse elemento tem dois ns, sendo que cada n possui trs graus de liberdade translaes nas direes x, y e z. A Figura [8] mostra esse elemento. Optou-se por esse elemento, pois as armaduras nos modelos foram discretas. Nos modelos, no foi considerado o fenmeno da aderncia entre as barras de ao e o concreto. Apesar disso, os resultados das comparaes entre resultados experimentais e numricos mostraram-se satisfatrios, como apresentado no item 3.3. Para representar o atrito entre as paredes do clice e o material de preenchimento (graute) e as faces do pilar pr-moldados, utilizaram-se elementos de contato, definindo-se superfcies de contato entre os materiais (superfcie de contato e superfcie alvo). As superfcies de contato entre os materiais foram representadas por
dois elementos finitos, denominado par de contato. Para a superfcie de contato, utilizou-se o elemento finito Contact 173 e para a superfcie alvo, utilizou-se o elemento finito Target 170. Esses elementos possuem trs graus de liberdade em cada n e as propriedades geomtricas so as mesmas das faces dos elementos slidos aos quais esto ligados, podendo ter geometria triangular ou quadrangular. A Figura [9] apresenta os pares de contado (elemento Contact 173 e Target 170). Os elementos finitos de contato foram utilizados apenas nos modelos com conformao das paredes do clice e o pilar lisa, pois, em virtude de pesquisas j realizadas por diversos pesquisadores, pode-se considerar que a ligao pilar-fundao por meio de clice com chave de cisalhamento tenha comportamento monoltico.
3.2 Propriedades dos materiais

Desenvolver um modelo capaz de representar o comportamento do concreto o mais prximo do comportamento real um desafio. O concreto armado um material quase-frgil e tem diferentes comportamentos na compresso e na trao. Na compresso, a curva tenso vs. deformao do concreto elstica e linear at aproximadamente 30% da fora ltima de compresso. Aps esse ponto, o concreto perde rigidez e segue elevando os valores das tenses at a fora de ruptura. Depois disso, no h aumento da tenso sofrendo amolecimento. Na trao, a curva tenso vs. deformao do concreto aproximadamente elstica e linear at a tenso de trao mxima. Aps esse ponto, o concreto fissura e sua resistncia no considerando o amolecimento na trao. Para modelar o material concreto, necessrio fornecer ao programa Ansys[23] os seguintes dados de entrada: mdulo de elasticidade longitudinal do concreto; resistncia ltima do concreto compresso e trao; coeficiente de Poisson; e coeficientes de transferncia de cisalhamento. O Ansys[23] tambm permite como dado de entrada, a incluso de uma curva tenso vs. deformao para representar as propriedades mecnicas do concreto. Isso normalmente feito, quando por problemas de convergncia, o processamento interrompido bruscamente por esmagamento precoce do concreto. Kachlakev et al. [9] trazem maiores informaes sobre esse fenmeno. Nos modelos analisados, no ocorreu este problema.
Figura 7 Solid 65, Ansys
460
Figura 9 Elementos finitos de contato, Ansys
O mdulo de elasticidade longitudinal do concreto, Ec, como tambm, a resistncia caracterstica trao, ftk, foram determinadas com base nas recomendaes da NBR 6118:2007 [1]. O coeficiente de Poisson, n, adotado para o concreto foi igual a 0,2 e os coeficientes de transferncia cisalhamento, b adotados foram iguais a 1 para fissuras abertas e fechadas. Esse valor para os coeficientes b foi utilizado, pois testes feitos demonstraram maior eficincia na convergncia do processamento quando utilizou-se o valor mencionado, ver Delalibera [5]. Foi utilizado o critrio de ruptura Concrete fornecido pelo Ansys. Para a definio da superfcie de ruptura se fazem necessrios
Figura 10 Superfcie de ruptura, Concrete, Ansys
apenas dois parmetros: as resistncias compresso e trao ltimas do concreto. O critrio de ruptura Concrete anlogo ao critrio de ruptura de Willam-Warnke. A Figura [10] apresenta a superfcie de ruptura. Em todos os blocos adotaram-se resistncia caracterstica do concreto compresso (fck) igual a 25 MPa. Para as barras de ao, adotou-se comportamento elastoplstico perfeito. O mdulo de elasticidade utilizado foi igual a 210 GPa, o coeficiente de Poisson igual a 0,3 e a resistncia caracterstica ao escoamento igual a 500 MPa. Por meio de testes realizados, constatou-se que o critrio de Newton-Raphson foi o que apresentou melhores resultados quanto convergncia dos modelos, sendo assim, em todas as anlises utilizou-se esse critrio. Para as propriedades dos elementos de contato, utilizou-se o modelo de atrito de Coulomb, sendo necessrio definir o valor do coeficiente de atrito m, as tenses mximas de cisalhamento, tmx e duas constantes, FKN e FTOLN. FKN representa um coeficiente de rigidez normal do elemento de contato e a constante FTOLN um fator de tolerncia a ser aplicado no sentido do vetor normal da superfcie. Esse fator utilizado para determinar a compatibilidade da penetrao. A compatibilidade do contato satisfeita se a penetrao estiver dentro de uma tolerncia permissvel (FTOLN mede a profundidade de elementos subjacentes). A profundidade definida pela profundidade mdia de cada elemento individual do contato no par. Se o programa de computador Ansys[23] detectar qualquer penetrao maior do que esta tolerncia, a soluo global no converge, mesmo que as foras residuais e os incrementos de deslocamento se encontrem dentro dos critrios da convergncia adotada. Para o coeficiente FKN utilizou-se valor igual a 1 e para FTOL valor igual 0,1. A escolha do valor correto do coeficiente de atrito uma tarefa difcil, pois o mesmo depende de vrios fatores: tipo de superfcie, intensidades das aes, propriedades mecnicas dos materiais que compem a ligao pilar-fundao. Existem nas literaturas tcnicas vrias indicaes para o valor do coeficiente de atrito concreto-concreto. Segundo Nielsen [10] o valor a ser utilizado 0,6, a EN 1992-1-1 [03] indica que o valor do coeficiente de atrito
461
para a situao onde a ligao pilar-fundao por meio clice com paredes lisas, deve ser maior que 0,3. Canha [11] e Ebeling [12] analisaram a influncia do coeficiente de atrito em ligaes pilar-fundao por meio de clice, variando-se o valor de 0,60, 0,45 e 0,30. Osanai et al. [13] sugere que os valores dos coeficientes de atrito tenham valores iguais a 0,5 e 1, funo do comprimento de embutimento do pilar e do tipo de conformao das paredes do clice e do pilar. Neste trabalho, sugerindo as recomendaes de Canha & El Debs [14] utilizou-se coeficiente de atrito igual a 0,6. importante lembrar que esta anlise numrica tem como objetivo apresentar tendncia do comportamento da ligao pilar-fundao por meio de clice embutido em blocos sobre duas estacas, com fim principal de analisar a relevncia dos fatores analisados. Com relao tenso de cisalhamento, tmx, adotou-se o valor utilizado pelo programa de computador, [y/(3)], em que y a resistncia ao escoamento do critrio de ruptura de Von Mises, do material adjacente superfcie de contato. Para y utilizou-se valores indicados por Canha [11], onde essa tenso vale trinta por cento da resistncia compresso do concreto de menor resistncia pertencente ao contato. Portanto, o valor de y adotado foi igual a 7,5 MPa, resultando 2,5 MPa para o valor de tmx.
3.3 Anlise comparativa

Visando verificar se os modelos adotados para a anlise numrica dos blocos sobre duas estacas mostram tendncia de comportamento correta, fizeram-se anlises comparativas de ensaios experimentais e simulaes numricas de vrios pesquisadores. Neste trabalho sero apresentados os resultados das anlises numricas comparativas dos ensaios realizados por Mautoni [15] e Adebar et al.[16] e da simulao numrica desenvolvida por Sam & Iyer [17]. Em Delalibera [5], pode-se observar os restantes das anlises comparativas realizadas e maiores detalhes das anlises numricas desenvolvidas. Na anlise comparativa, tambm utilizaram-se os mesmos elementos finitos empregados nas anlises dos blocos sobre duas estacas, ou seja, o Solid 65 discretizando o material concreto
e o Link 8 modelando as barras de ao. Todas as propriedades mecnicas e geomtricas adotadas na anlise comparativa foram s mesmas dos ensaios experimentais. As condies de contorno empregadas nos modelos numricos tentaram representar as condies de contorno empregadas nos experimentos com maior grau de realidade possvel, o mesmo acontecendo com o carregamento. A primeira anlise comparativa trata-se do bloco B1-A ensaiado por Mautoni [15]. O bloco teve runa por cisalhamento com fora ltima igual a 800 kN. A primeira fissura surgiu com aproximadamente trinta por cento da fora ltima, cerca de 240 kN. O bloco tinha altura de vinte e cinco centmetros, o pilar era quadrado com rea igual a 225 cm2 e as estacas tinham sees transversais iguais a 10 cm x 15 cm. O concreto apresentou resistncia compresso igual a 32,30 MPa e as barras de ao do tirante apresentavam resistncia ao escoamento igual a 720 MPa. Como no existiam informaes sobre o mdulo de elasticidade e sobre os coeficientes de Poisson do ao e concreto, adotaram-se as recomendaes da NBR 6118:2007[1]. Aplicou-se no modelo numrico trezentos incrementos de fora, sendo que, para cada incremento o valor da fora aplicada era de 2,67 kN. A Figura [11] mostra a configurao final obtida no experimento e na simulao do bloco B1-A. O valor da fora que originou a primeira fissura no modelo numrico foi de 312,33 kN. Ocorreu diferena de 23,15% com relao ao valor da fora que originou a primeira fissura no modelo experimental. Isso ocorreu, pois, no modelo experimental, a fora que provocou a primeira fissura foi determinada em funo de observao visual, ou seja, a primeira fissura visvel ao olho humano. Com relao fora ltima, o modelo numrico, apresentou fora ltima de 799,98 kN, praticamente, no ocorrendo diferena com o valor obtido experimentalmente. A segunda anlise comparativa trata-se dos blocos ensaiados por Adebar et al. [16]. Foram simulados os blocos A, B, C, D e F. Todos os blocos tinham altura de sessenta centmetros e coeficientes de Poisson iguais 0,3 e 0,2 para ao e concreto respectivamente. Em todos os blocos foram aplicados quinhentos incrementos de fora.
Figura 11 Configurao final do bloco B1-A, ensaiado por Mautoni (1972)
462
Figura 12 Bloco A, Adebar et al. [7]
O Bloco A foi constitudo por quatro estacas com dimetro de vinte centmetros e pilar com seo transversal quadrada com trinta centmetros de lado. A resistncia mdia compresso do concreto (fcm) obtida nos ensaios foi igual a 27,10 MPa. As barras de ao possuam resistncia ao escoamento igual a 479 MPa. No ensaio, a primeira fissura surgiu com fora igual a 1186 kN e a fora de ruptura foi igual a 1781 kN. No modelo numrico, a primeira fissura ocorreu com fora igual a 1403,86 e a fora de runa obtida foi igual a 1781,10 kN. O Bloco B possua as mesmas propriedades geomtricas do Bloco A, porm, a armadura era distribuda sobre as estacas, enquanto que no Bloco A, a armadura era distribuda em malha. A resistncia do concreto compresso foi de 24,80 MPa e a resistncia ao escoamento das barras de ao dos tirantes foi a mesma do Bloco A. No ensaio, a primeira fissura surgiu para uma fora de 1679 kN, j, no modelo numrico, a primeira fissura ocorreu com 1505,71 kN. A fora de runa experimental foi registrada com valor igual a 2189 kN e na simulao numrica, a fora obtida foi de 2186 kN.
Figura 13 Bloco B, Adebar et al. [7]
Figura 15 Bloco D, Adebar et al. [7]
Figura 14 Bloco C, Adebar et al. [7]
Figura 16 Bloco F, Adebar et al. [7]
463
Tabela 2 Relaes entre resultados numricos e experimentais dos blocos de Adebar et al. [7]
Blocos Bloco A Bloco B Bloco C Bloco D Bloco F Fu,exp (kN) 1781 2189 2892 3222 3026 Fu,num (kN) 1781,10 2186 2647,7 3212,70 2589,74 Fu,exp/Fu,num 0,99 1,00 1,09 1,01 1,17
Figura 18 Armadura distribuda sobre as estacas, Sam & Iyer [17]
O Bloco C tinha seis estacas com dimetros de vinte centmetros e pilar com a mesma seo transversal dos demais modelos. A resistncia mdia do concreto compresso era de 27,10 MPa enquanto que a resistncia ao escoamento das barras de ao dos tirantes tinha valor igual 479 MPa. A primeira fissura no ensaio surgiu para uma fora de 1780 kN e a fora de ruptura observada foi de 2892 kN. No modelo numrico essas foras foram de 1588,30 kN e 2647,70 kN respectivamente. O Bloco D era semelhante ao Bloco B, mudando-se apenas a rea das barras de ao da armadura do tirante. A resistncia do concreto compresso foi igual a 30,30 MPa e a resistncia ao escoamento das barras igual a de 486 MPa. A primeira fissura apresentou-se para a fora de 1122 kN e a ruptura do bloco se deu para a fora de 3222 kN. Na simulao numrica, a fora que originou a primeira fissura foi de 1097,74 kN e a fora de runa observada foi de 3212,17 kN. Apesar de obter boa correlao entre as foras de fissurao e ltima, a simulao numrica apresentou diferenas significativas em relao rigidez do bloco D ensaiado experimentalmente, ver Figura [15]. O Bloco F tinha as mesmas propriedades mecnicas do Bloco B, porm, mudou-se apenas a forma geomtrica do mesmo. A fora que provocou a primeira fissura no ensaio foi igual a 650 kN e a fora de runa observada foi igual a 3026 kN. Na simulao numrica, essas foras foram de 501,94 kN e 2589,74 kN.
As Figuras [12], [13], [14], [15] e [16] mostram as correlaes obtidas entre os resultados experimentais e numricos e a Tabela [2] as relaes entre as foras ltimas experimentais e numricas. A terceira anlise comparativa refere-se aos blocos simulados numericamente por Iyer & Sam [17]. Foram simulados os blocos com arranjo de armadura distribuda em malha e sobre as estacas. Os blocos tinham altura de vinte e dois centmetros e cinco milmetros e eram compostos por quatros estacas com seo transversal quadrada de 10 cm x 10 cm. Os pilares tambm tinham seo quadrada com 14,14 cm x 14,14 cm de lado. A resistncia caracterstica do concreto era igual a 19 MPa e a resistncia caracterstica
Figura 19 Resultado do ensaio de SPT, Senna Jnior [18]
Figura 17 Armadura distribuda em malha, Sam & Iyer [17]
464
Figura 20 Blocos analisados com comprimento real da estaca
Figura 22 Fluxo das tenses principais de compresso, Bloco B
entre os resultados, indicando que o modelo adotado na anlise numrica coerente. ao escoamento das barras de ao igual a 300 MPa. O coeficiente de Poisson adotado para ao e concreto foi 0,3 e 0,2, respectivamente. Nas simulaes numricas adotaram-se 345 incrementos de fora. As foras ltimas obtidas nas anlises desenvolvidas por Iyer & Sam (1995) foram de: 600 kN, para armadura em malha e 560 kN, para armadura sobre as estacas. Na simulao numrica desenvolvida neste texto, as foras encontradas foram de: 582,17 kN para os blocos com arranjo de armadura distribuda em malha e 594,59 kN para os blocos com arranjo de armadura distribuda sobre as estacas. As Figuras [17] e [18] apresentam as correlaes entre os resultados obtidos. Observa-se nas Figuras [17] e [18], que existe tima correlao
3.4 Influncia do comprimento da estaca e do solo

Como o ensaio experimental de blocos com estacas de comprimentos reais so de difcil execuo, simularam-se blocos com estacas de comprimentos reais, com objetivo de observar o comportamento do fluxo de tenses principais de compresso e a influncia do comprimento das estacas no bloco. Para isso, quatro blocos foram modelados trs com estacas com comprimentos reais e um com estacas de pequena altura. Na modelagem utilizou-se um bloco sobre duas estacas com ngulo de inclinao da biela de compresso igual a 45. O comprimento das estacas foi determinado em funo do resultado da sondagem executada pelo Departamento de Geo-
Figura 21 Fluxo das tenses principais de compresso, Bloco A
Figura 23 Fluxo das tenses principais de compresso, Bloco C
465
tecnia, no Campus I da USP em So Carlos. Assim, o comprimento calculado do fuste das estacas foi igual a 9,0 m. A Figura [19] mostra o resultado da sondagem a percusso (S.P.T.) e a Figura [20] os blocos analisados numericamente. O atrito entre a estaca e o solo no foi levado em considerao. Para o solo existente em torno do fuste da estaca, adotou-se comportamento plstico (critrio de ruptura Drucker-Prager). O resulta-
do da sondagem a percusso e as propriedades mecnicas do solo foram obtidas em Senna Jnior [18]. A fora foi aplicada por meio de cem incrementos, admitindo-a centrada. Nos modelos onde no existia a contribuio do solo (Bloco A e B Figura [20]), aplicou-se fora at a ruptura do bloco. Nos demais modelos, a ao aplicada correspondeu capacidade de carga das estacas, ou seja, 600 kN. Na modelagem do solo, utilizou-se o modelo do Meio Contnuo
Tabela 3 Critrios utilizados na modelagem de blocos, considerando a influncia das estacas e do solo
Elemento estrutural Elemento finito Constantes reais Propriedades dos materiais Informaes adicionais
Pilar
Solid 65
Bloco
Solid 65
Estaca
Solid 65
As, tirante As, pilar e estacas Placa de ao Estribos
Link 8 Link 8 Solid 65 Link 8
As = 3,15 cm i = 0
2
As = 1,25 cm i = 0
2
As = 0,50 cm2 i = 0
Solo 1
Solid 65
Solo 2 Solo 3 Solo 4 Solo 5
Solid 65 Solid 65 Solid 65 Solid 65
Ec = 33658 MPa = 0,3 Comportamento fck = 50 MPa plstico Concrete ftk = 4,07 MPa =1 Ec = 21287 MPa = 0,3 Comportamento fck = 20 MPa plstico Concrete ftk = 2,21 MPa =1 Ec = 33658 MPa = 0,3 Comportamento fck = 50 MPa plstico Concrete ftk = 4,07 MPa =1 Es = 210 MPa Comportamento fy = 500 MPa elastoplstico perfeito = 0,3 Es = 210 MPa Comportamento fy = 500 MPa elastoplstico perfeito = 0,3 Comportamento Es = 210 MPa elastoplstico perfeito = 0,3 Es = 210 MPa Comportamento fy = 500 MPa elstico = 0,3 Esolo = 46 MPa Comportamento c = 15 KPa plstico Drucker-Prager solo = 22 = 0,3 Esolo = 70 MPa c = 26 KPa Plastic behavior Drucker-Prager solo = 28 = 0,3 Esolo = 46 MPa Elastic behavior = 0,3 Esolo = 70 MPa Elastic behavior = 0,3 Esolo = 1000 MPa Elastic behavior = 0,3
466
Figura 24 Fluxo das tenses principais de compresso, Bloco D
Figura 25 Sees nodais investigadas
elstico (Solo 1, 4 e 5 Figura [20]) e elastoplstico (Solo 1 e 2 Figura [20]). O solo poderia ser modelado utilizando as hipteses de Winkler (presses de contato so proporcionais aos recalques), porm, segundo Velloso [19], o modelo do Meio Contnuo representa com maior preciso o fenmeno da interao solo-estrutura. Para evitar que ocorresse perturbao localizada de tenses no ponto de aplicao da fora, modelou-se uma placa de ao na cabea do pilar com espessura de cinco centmetros e material elstico e linear, o mesmo ocorrendo em toda a anlise numrica desenvolvida. Como se pretendia analisar o comportamento dos blocos, adotaram-se nas estacas e nos pilares, concreto com resistncia caracterstica compresso igual a 50 MPa. Os coeficientes de transferncia de cisalhamento tambm tiveram valor igual a 1. O mdulo de elasticidade do ao e do concreto, como tambm os seus respectivos coeficientes de Poisson, foram determinados com auxlio da NBR 6118:2007[1]. As taxas de armadura foram iguais em todos os modelos, sendo que, as reas de armadura das estacas, dos pilares e dos tirantes so apresentadas na Tabela [3]. Para maiores informaes, ver Delalibera [5]. Por meio das Figuras 21 a 24, observa-se que existe maior concentrao de tenses de compresso nas sees do bloco logo abaixo do pilar e nas sees das estacas mais afastadas da borda do bloco, sees F e G Figura [20]. Verificou-se que as estacas no so
solicitadas de maneira uniforme, ou seja, as sees mais afastadas das bordas do bloco (sees F e G Figura [25]) foram mais solicitadas que as sees mais prximas da borda (sees D e I Figura [25]). Este fenmeno ocorreu em todos os modelos analisados. Em funo da diferena entre as intensidades das foras aplicadas nos blocos sem a presena do solo (blocos A e B Figura [20]), que foram maiores, em relao aos blocos modelados com o solo (blocos C e D Figura [20]), ocorreram diferenas nas intensidades das tenses principais de compresso, o que era previsto, uma vez que a ruptura do conjunto solo-estrutura dos blocos C e D se deram pelo solo. Por meio dos valores apresentados na Tabela [4] possvel concluir que as sees F e G, na zona nodal inferior (junto s estacas) foram mais solicitadas, pois como a distribuio das tenses oriundas da escora (ou biela) no se d de maneira uniforme na cabea das estacas, as regies mais prximas do pilar foram as que apresentaram maiores tenses. Esses resultados so similares aos resultados experimentais obtidos por Delalibera & Giongo [25]. Em funo destes resultados, todas as anlises desenvolvidas foram feitas utilizando estacas curtas, pois constatou-se, que a distribuio dos fluxos das tenses principais de compresso pouco influenciada pela rigidez das estacas.
Tabela 4 Intensidades das tenses principais de compresso, considerando a influncia das estacas e do solo
Sees A B C D=I E=H F=G Tenses principais de compresso (MPa) Bloco B Bloco C 11,5 7,5 11,5 3,5 7,5 19,5 13,3 22 13,3 4,2 13,3 30,6
Bloco A 16,7 11,9 16,7 5,10 10,9 28,3
Bloco D 11,7 21,3 11,7 2,10 11,7 21,3
467
Tabela 5 Anlise de Varincia, equacionamento geral, Montgomery [24]

Fatores M N MxN Erro Total Soma dos quadrados SSM SSN SSMN SSE SST Graus de liberdade a 1 b1 (a 1) (b 1) abc (n 1) abcn 1 Mdia dos quadrados MSM = SSM / (a 1) MSN = SSN / (b 1) MSMN = SSMN / [(a 1) (b 1)] MSE = SSE / [abc (n 1)] F0
F0
F0
F0
MSM MSE
MSN MS E
MSMN MS E
Um resultado interessante constatado durante estas anlises foi com relao uniformizao das tenses principais de compresso nas estacas. Nos blocos B, C e D (ver Figura [20]), que tinham estacas com comprimentos longos, verificou-se que as tenses se uniformizaram aproximadamente a 1/3 da altura do fuste da estaca.
4. Anlise de varincia
A anlise de varincia um teste estatstico amplamente difundido entre os estatsticos e visa fundamentalmente verificar se existe diferena significativa entre as mdias e se os fatores exercem influncia em alguma varivel dependente. Os fatores propostos podem ser de origem qualitativa ou quantitativa, mas a varivel dependente necessariamente dever ser contnua. A principal aplicao da ANOVA (analise of variance) a comparao de mdias oriundas de grupos diferentes, tambm chamados tratamentos. Existem dois tipos de problemas a serem resolvidos por meio da ANOVA: fatores fixos ou fatores aleatrios. A aleatoriedade determina a questo do problema. Na grande maioria dos casos trata-se de fatores fixos, afinal o segundo tipo de problema (aleatrio) somente surgir quando ocorrer um estudo envolvendo uma escolha aleatria de fatores. A anlise de varincia amplamente utilizada em diversas reas. Na Engenharia Civil, sua utilizao ainda restrita, porm, j existem trabalhos desenvolvidos utilizando a tcnica, como por exemplo: Lima Jnior [20], Delalibera [5] e Pituba et al. [21]. Na anlise de varincia desenvolvida neste trabalho, utilizaram-se fatores fixos, escolhendo-se trs variveis de estudo: o comprimento de embutimento do pilar no clice (emb); a espessura da laje de fundo do bloco (hs); e a conformao das paredes do clice embutido e do pilar pr-moldado. As variveis escolhidas totalizaram dezoito casos de combinaes. Os modelos foram divididos em dois grupos (paredes lisas L e paredes rugosas R).
do experimento forem fixados, pode-se facilmente formular o problema, obtendo resultados que indicam quais dos fatores analisados so importantes como tambm suas combinaes. A Tabela [5] apresenta uma anlise de varincia com dois fatores. Para verificar a relevncia de um determinado fator principal fixo ou combinaes entre os fatores principais, faz-se a relao entre a mdia dos quadrados de cada fator principal ou combinao dos fatores principais pela mdia dos quadrados dos erros. A diviso entre a mdia dos quadrados de cada fator principal ou combinao dos fatores principais pela mdia dos erros chamada de F0. O nmero de graus de liberdade de cada fator principal igual ao nmero de variaes de cada fator menos a unidade. O nmero de graus de liberdade dos fatores principais combinados o produto entre os fatores principais que foram combinados. A soma total dos quadrados calculada por meio da Equaao [1]. A soma dos quadrados da combinao N x M expressa por meio da Equao 2. A soma dos quadrados do erro definida pela Equao 3.
(1)
(2)
(3)
Para verificao da relevncia de uma determinada varivel principal fixa ou combinada, aplica-se o teste F. Por meio de valores tabelados de Fcrtico, fornecidos por Montgomery [24], compara-se o valor calculado de F0 com o valor de Fcrtico. Se o valor calculado de F0 for maior que o valor tabelado de Fcrtico significa que esse fator relevante, caso contrrio, implica que o fator no possui im-
4.1 Formulao da anlise de varincia

Sejam N e M os fatores principais fixos da anlise de varincia, a, b e c, as variaes desses fatores e n o nmero de rplicas. Em geral existir abc...n combinaes possveis. Se todos os fatores
468
Tabela 6 Resultados obtidos por meio da anlise numrica para os blocos do com conformao lisa e ao de fora de compresso excntrica (fora normal e momento)
Seo das estacas (cm) 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 Fu (kN) Blvot Num. 30 20 10 30 20 10 30 20 10 80 80 80 60 60 60 50 50 50 54,0 51,3 48,4 48,4 45,0 41,2 45,0 41,2 36,9 120 110 100 100 100 80 90 80 70 3663 3415 3128 3128 2800 2428 2800 2428 2016 3663 3415 3128 3128 2800 2428 2800 2428 2016 1,27 1,17 1,28 1,18 1,22 1,91 1,24 1,96 3,16 2,44 3,00 2,84 2,90 4,35 5,96 3,59 4,01 4,44 Ab,est (cm2) Blvot Num. 323,6 269,6 487,8 365,8 673 639,4 299,1 299,1 441,9 397,7 592,8 497,9 282,8 207,3 411,7 288,2 355,9 177,9
Blocos
Seo do pilar (cm) 20 x 20 20 x 20 20 x 20 20 x 30 20 x 30 20 x 30 20 x 40 20 x 40 20 x 40
hs emb h (cm) (cm) (graus) (cm)
Fblvot / Fnum
e (cm)
Le80hs30NM Le80hs20NM Le80hs10NM Le60hs30NM Le60hs20NM Le60hs10NM Le50hs30NM Le50hs20NM Le50hs10NM
Nota: Num., valor obtido por meio da simulao numrica; Blvot, valor calculado por meio dos critrios de Blvot & Fremy [4]; Fu, fora ltima aplicada no pilar; e, excentricidade da fora normal; Ab,est, rea da biela (ou escora) junto estaca.
portncia significativa. Os valores de Fcrtico so funo do nmero de graus de liberdade de cada varivel e do nmero de graus de liberdade total.
5. Resultados obtidos
5.1 Anlise dos blocos sobre duas estacas fora normal, momento e paredes lisas
Foram analisados nove blocos solicitados por ao de fora de normal de compresso e momento (o momento aplicado ao bloco foi
obtido por meio de aplicao de uma fora horizontal aplicada no topo do pilar). Os blocos apresentavam variaes no comprimento de embutimento do pilar (emb) e na espessura da laje de fundo do bloco (hs). A variao dos fatores analisados modificou de modo significativo a distribuio das tenses principais de compresso e o panorama de fissuras no ltimo incremento de fora aplicado aos modelos. A Tabela [6] apresenta resultados das anlises numricas realizadas. Considerando os resultados da Tabela [6], observa-se que para menores comprimentos de embutimento do pilar e menores es-
Figura 26 Curvas fora vs. deslocamento no meio do vo, blocos com conformao lisa, fora normal e momento
Figura 27 Tenses principais de compresso, bloco com conformao lisa, fora normal e momento
Le60hs20NM
469
Figura 28 Panorama de fissurao, blocos com conformao lisa, fora normal e momento
Figura 29 Tenses principais de compresso, bloco com conformao rugosa, fora normal e momento
Le60hs20NM
pessuras da laje de fundo (no caso dos modelos Le60hs10NM e Le50hs10NM) ocorreram diferenas significativas com relao aos valores analticos. Isso demonstra que a capacidade resistente do bloco no pode ser determinada em funo da ruptura da diagonal comprimida, mas sim, pelo efeito de puno da laje de fundo. interessante observar, que, o efeito da puno s ocorrer se, e somente se, a fora no pilar for transmitida (quase toda) para a laje de fundo e, isso s ocorrer se o comprimento de embutimento do pilar for pequeno, no sendo possvel a formao da biela (ou escora), sendo a ruptura do elemento caracterizada por puno. A Figura [26] apresenta curvas reao na estaca mais solicitada vs. deslocamento no meio do vo do bloco para os modelos analisados numericamente. A Figura [27] mostra a distribuio das tenses principais de compresso no interior de um dos blocos analisados numericamente e a Figura [28] mostra o panorama de fissurao do mesmo bloco apresentado na Figura [27], referente ao ltimo incremento de fora.
Le60hs20NM
Para os modelos de blocos com conformao das paredes do clice e do pilar, com atuao de fora normal e momento no bloco, o fator emb o principal fator relevante, seguido do fator hs. A Tabela [7] mostra os resultados da anlise de varincia dos modelos de blocos com conformao das paredes do clice e do pilar, com atuao de fora normal e momento no bloco. Por meio da anlise numrica, observou-se que as diferenas existentes nas reas das bielas de compresso junto s estacas, quando comparadas com as reas das bielas calculadas utilizando os critrios de Blvot & Frmy [4] tambm ocorreram, corroborando com os resultados experimentais encontrados por Delalibera & Giongo [25]. Na Tabela [6] apresentam-se os resultados das reas das bielas calcudas pelo critrio de Blevt (considerando um plano inclincado de q ) e as reas das bielas obtidas numericamente (para o clculo das reas das bielas obtidas por meio dos resultados numricos, utilizou-se os valores das tenses principais de compresso junto s estacas), como tambm as excentricidades da fora normal de cada modelo.
Tabela 7 Anlise de varincia, Anova, blocos do com conformao lisa e ao de fora de compresso excntrica
Fator emb hs emb x hs Erro Total Soma dos quadrados 2805913 2026374 302545 194219 5329056 Graus de liberdade 2 2 4 3 8 Mdia dos quadrados 1402956 1013187 75637,5 64734 666132 F0 21,67 15,65 1,16 Significncia F0,01 8,65 8,65 7,01
Nota: emb x hs, acoplamento entre o comprimento de embutimento do pilar e a espessura da laje de fundo.
470
Tabela 8 Resultados obtidos por meio da anlise numrica para os blocos com conformao rugosa e ao de fora de compresso excntrica
Seo do pilar (cm) 20 x 20 20 x 20 20 x 20 20 x 30 20 x 30 20 x 30 20 x 40 20 x 40 20 x 40 Seo das estacas (cm) 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 20 x 20 25 x 25 30 x 30 Fu (kN) Blvot Num. 30 20 10 30 20 10 30 20 10 80 80 80 60 60 60 50 50 50 54,0 51,3 48,4 48,4 45,0 41,2 45,0 41,2 36,9 120 110 100 100 100 80 90 80 70 3663 3415 3128 3128 2800 2428 2800 2428 2016 2832 2660 2460 2360 2368 1103 2260 1253 1828 1,29 1,28 1,27 1,33 1,18 2,20 1,24 1,94 1,10 3,18 3,76 3,48 3,93 4,74 7,13 3,91 5,03 4,47 Ab,est (cm2) Blvot Num. 323,6 487,8 673 299,1 441,9 592,8 282,8 411,7 355,9 223,8 404,9 639,4 299,1 371,2 462,4 243,2 308,8 153,1
Blocos
hs emb h (cm) (cm) (graus) (cm)
Fblvot / Fnum
e (cm)
Re80hs30NM Re80hs20NM Re80hs10NM Re60hs30NM Re60hs20NM Re60hs10NM Re50hs30NM Re50hs20NM Re50hs10NM
Nota: Num., valor obtido por meio da simulao numrica; Blvot, valor calculado por meio dos critrios de Blvot & Fremy [4]; Fu, fora ltima aplicada no pilar; e, excentricidade da fora normal; Ab,est, rea da biela (ou escora) junto estaca.
5.3 Anlise dos blocos sobre duas estacas fora normal, momento e paredes rugosas
Foram analisados nove blocos solicitados por ao de fora normal de compresso e momento, com variaes no comprimento de embutimento do pilar (emb) e na espessura da laje de fundo do bloco (hs). A variao dos fatores analisados modificou, tambm neste caso, a distribuio das tenses principais de compresso e o panorama de fissura no ltimo incremento de foras (horizontal e
vertical) aplicadas aos modelos. A Tabela [8] apresenta resultados das anlises numricas, a Figura [29] mostra a distribuio das tenses principais de compresso e a Figura [30] o panorama de fissurao dos blocos analisados numericamente. A causa da runa precoce do modelo Re60hs10NM provavelmente est relacionada com a puno da laje de fundo do bloco e do modelo Re50hs20NM, em virtude do pequeno comprimento do embutimento do pilar. Por meio da Tabela [8], observam-se as diferenas entre as reas das bielas de compresso junto s estacas, calculadas pelo critrio analtico desenvolvido por Blvot & Frmy [4] e utilizando os modelos numricos. Na mesma tabela, apresentam-se os valores
Figura 30 Panorama de fissurao, blocos com conformao rugosa, fora normal e momento
Figura 31 Curvas reao na estaca mais comprimida vs. deslocamento no meio do vo, blocos com conformao rugosa, fora normal e momento
Le60hs20NM
471
Figura 32 Tenses nas barras de ao, Le60hs20NM, tenses expressas em MPa
Outro resultado muito interessante que condiz com afirmaes de outros pesquisadores pode ser observado por meio dos valores da Tabela [9]. Como a ligao com chave de cisalhamento, confere ligao pilar-fundao comportamento monoltico, de esperar que a runa dos blocos esteja associada com outros fatores. Isso foi constatado, pois, nota-se que os fatores previamente escolhidos para esta anlise no foram relevantes, ou seja, no apresentaram valores significantes. Tais resultados corroboram com os resultados apresentados por Canha & El Debs[14], onde constatou-se que se respeitando o comprimento de embutimento mnimo do pilar estabelecido pela NBR 9062:2006, pode-se consdiderar monontico o comportamento da ligao pilar-clice.
5.3 Tenses nas barras de ao

Analisando-se a Figura [32], percebe-se que as barras de ao da armadura dos modelos apresentaram valores consistentes e em alguns casos, ocorreu escoamento. Observou-se que em algumas barras de ao das estacas, ocorreram tenses de trao. Esses resultados so coerentes em funo das aes externas aplicadas. Os resultados apresentados na Figura [32] corroboram com os resultados apresentados por Adebar. et al. [7], Miguel [6] e Delalibera [5], ou seja, a tenso nas barras da armadura principal de trao no constante, tendo valores nulos (ou at negativo) na ponta dos tirantes e valores mximos no meio de vo. Observa-se tambm que, as barras de ao que compem a armadura das estacas apresentam tenses com valores diferentes, o que indica flexo-compresso nas estacas.
das excentricidades da fora normal de cada modelo analisado numericamente. possvel observar por meio das Figuras [29] e [30] a formao das diagonais comprimidas e o panorama de fissurao de um dos blocos analisados numericamnete. Constata-se que os modelos de bielas de tirantes devem ser modificados em funo das aes externas atuantes no bloco, ou seja, idealizar um modelo que contemple as aes de momento, fora normal e fora horizontal. A Figura [31] mostra curvas reao na estaca mais solicitada vs. deslocamento no meio do vo do bloco. O comportamento dos blocos submeticos a ao de momento e fora de compresso, indicam que os blocos com maior rigidez apresentam maior capacidade portante e a presena de momento nos mesmos, a reduz. Delalibera [5] comprova experimentalmente essa afirmao. Delalibera & Giongo [26], analisaram os mesmos blocos deste trabalho, porm aplicou-se apenas fora de compresso centrada. Observou-se, que a capacidade portante dos blocos com apenas fora centrada foi superior aos blocos com fora excntrica. A Tabela [9] apresenta os resultados da anlise de varincia dos blocos submetidos fora de compresso excntrica e conformao rugosa da interface pilar-clice.
6. Concluso
As simulaes numricas se mostraram coerentes e apontaram tendncia do comportamento estrutural de blocos sobre duas estacas com clice embutido, com conformao das paredes do clice e do pilar lisas e rugosas. Observaram-se para os blocos com interface lisa entre as paredes do clice e do pilar, os seguintes fatores: comprimento de embutimento do pilar emb e espessura da laje de fundo hs, tem importncia relevante na capacidade portante dos modelos. Pois, os modelos analisados numericamente com maiores comprimentos de embutimento do pilar e maiores espessuras da laje de fundo, apresentaram maior capacidade portante. Alm disso, ficou evidente que nas situaes em que se tinha comprimento de embutimento do pilar diminudo e pequena espessura da laje de fundo,
Tabela 9 Anlise de varincia, Anova, blocos do com conformao rugosa e ao de fora de compresso excntrica
Fator emb hs emb x hs Erro Total Soma dos quadrados 1697377 875380 43964 1606137 4223308 Graus de liberdade 2 2 4 3 8 Mdia dos quadrados 848688 437915 10991 535379 527913 F0 1,58 0,82 0,02 Significncia F0,01 8,65 8,65 7,01
Nota: emb x hs, acoplamento entre o comprimento de embutimento do pilar e a espessura da laje de fundo.
472
a possvel runa ocorreu por puno do bloco na regio de contato entre a base do pilar e a base do bloco. Para os modelos simulados com interface rugosa entre as paredes do clice e do pilar (chave de cisalhamento), os fatores analisados (emb e hs) no foram relevantes na determinao da capacidade portante dos blocos. Esses resultados so interessantes, pois possvel projetar blocos sobre estacas com clice embutidos e chave de cisalhamento, com modelos de bielas e tirantes, preocupando-se apenas com a puno da laje de fundo do bloco, durante a fase de montagem (ou seja, antes da aplicao do material de preenchimento), desde que respeitado o comprimento mnimo de embutimento do pilar no clice, sugerido pela NBR 9062:2006. Faz-se necessrio um maior nmero de ensaios experimentais, para que possa verificar a possiblidade da diminuio do comprimento de embutimento do pilar no clice. Observou-se tambm que a rigidez do bloco influencia na capacidade portante do mesmo. Pode-se se afirmar isso, em funo dos resultados obtidos, pois os modelos com maiores alturas foram os que apresentaram maior capacidade resistente. Constatou-se, e isso j era esperado, que os modelos analisados numericamente com interface lisa, apresentaram menor capacidade portante com relao aos modelos analisados numericamente com interface rugosa.
7. Agradecimentos
Ao CNPq Conselho Nacional de Desenvolvimento Cientfico e Tecnolgico e FAPESP Fundao de Amparo Pesquisa do Estado de So Paulo pelo apoio financeiro para fazer a pesquisa que permitiu escrever este artigo.
[01] [02] [03] [04] [05] [06] [07] ASSOCIAO BRASILEIRA DE NORMA TCNICAS, ABNT NBR 6118:2003 Projeto de estruturas de concreto, 2007. Rio de Janeiro. ASSOCIAO BRASILEIRA DE NORMA TCNICAS, ABNT NBR 9062:2005 Projeto e execuo de estruturas de concreto pr-moldado, 2006. Rio de Janeiro. EN 1992-1-1, Eurocode 2: Design of concrete structures Part 1-1: General rules and rules for buildings, 2004. European standard. BLVOT, J.; FRMY, R.. Semelles sur piex. Analles dInstitut Techique du Btiment et des Travaux Publics, Paris,1967, v. 20, n. 230, p. 223-295, fev. DELALIBERA, R. G.. Anlise terica e experimental de blocos de concreto armado sobre duas estacas submetidos a ao de fora centrada e excntrica. Tese (doutorado) Escola de Engenharia de So Carlos, Universidade de So Paulo, 2006, So Carlos. MIGUEL, G. M.. Anlise experimental e numrica de blocos sobre trs estacas. Tese (doutorado) Escola de Engenharia de So Carlos, Universidade de So Paulo, 2000, So Carlos. ADEBAR, P.; KUCHMA, D.; COLLINS, M. P.. Strut-and-tie models for design of pile caps: an experimental study. ACI Journal, 1990, v. 87, p. 81-91, Jan/Feb.
[08] MONTOYA, P.J.; MESEGUER, A.; CABRE, M. Hormigon Armado 14.a Edicin Basada em EHE ajustada al Cdigo Modelo y al Eurocdig, 2000. Barcelona, Gustavo Gili. [09] KACHLAKEV, D.; MILLER, T; YIM, S.; CHANSAWAT, K.; POTISUK, T.. Finite element modeling of reinforced concrete structures strengthened with frp laminates. Final Report, 2001. Oregon Department of Transportation, May. [10] NIESEN, M. P.. Limit analysis and concrete plasticity. Prentice-Hall series in Civil Engineering, Englewood Cliffs, 1984. New Jersey, 420 pg. [11] CANHA, R. M. F.. Estudo terico-experimental da ligao pilar-fundao por meio de clice em estruturas de concreto pr-moldado. Tese (Doutorado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 2004. So Carlos. [12] EBELING, E. B.. Anlise da base de pilares pr-moldados na ligao com clice de fundao. Dissertao (Mestrado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 2006. So Carlos. [13] OSANAI, Y.; WATANABE, F.; OKAMOTO, S.. Stress transfer mechanism of socket base connections with precast concrete columns. ACI Structural journal, 1996, v. 93, n. 3, p. 226-276, May/June. [14] CANHA, R. M. F.; EL DEBS, M. K.. Proposta de modelo de prometo para a ligao pilar-fundao por meio de clice em estruturas de concreto pr-moldado. Revista Ibracon de Estruturas, 2006. Vol. 2, n. 2, p. 137-166, junho. [15] MAUTONI, M.. Blocos sobre dois apoios. Grmio Politcnico, 1972. So Paulo. [16] ABEBAR, P.. One-way shear strength of large footings. Can. J. Civ. Eng, 2000. 27: 553562. [17] SAM, C., IYER, P. K.. Nonlinear finite element analysis of reinforced concrete four-pile caps. International Journal of Structures, 1995, v. 15, n. 1, p. 18-34, Jan/Jun. [18] SENNA JNIOR, R. S.. Distribuio de carga em grupos de estacas escavadas de pequeno dimetro. Dissertao (Mestrado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 1993. So Carlos. [19] VELLOSO, D. A.; LOPES, F. R.. Fundaes. COPPE/ UFRJ, 1996. Rio de Janeiro. [20] LIMA JNIOR, H. C.. Avaliao da ductilidade de pilares de concreto armado, submetidos a flexo-compresso reta com e sem adio de fibras metlicas. Tese (Doutorado), Escola de Engenharia de So Carlos, Universidade de So Paulo, 200., So Carlos. [21] PITUBA, J. J. C. ; DELALIBERA, R. G.; RODRIGUES, F. S. . Numerical and Statistical Analysis about Displacements in Reinforced Concrete Beams using Damage Mechanics, 2012. Computers and Concrete, an International Journal (Print). [22] DELALIBERA, R. G.. Anlise numrica de blocos sobre duas estacas, com clice embutido, submetido
473
ao de fora normal e excntrica. Relatrio cientfico, 2012. Universidade Federal de Gois / CNPq Conselho Nacional de desenvolvimento cientfico e tecnolgico. [23] ANSYS Users Manual. Theory Manual, 2011. ANSYS revision 5.5. [24] MONTGOMERY, D. C.. Design and Analysis of Experiments Fourth Edition, 1996. Arizona State University, John Wiley & Sons. [25] DELALIBERA , R. G.; GIONGO, J. S.. Deformations in the strut of two pile caps. Ibracon Structures and Materials Journal - RIEM, 2008, v. 1, p. 121-138. [26] DELA LIBERA , R. G. ; GIONGO, Jos Samuel. Anlise de varincia de blocos sobre duas estacas com clice embutido,submetido ao de fora de compresso centrada. Revista Eletrnica de Engenharia Civil, 2012, v. 1, p. 1-16.
474
Plastic rotation and tension stiffening effect analysis in beams using photogrammetry Anlise da rotao plstica e da contribuio do beto entre fendas em vigas usando fotogrametria
R. N. F. do Carmo a carmo@isec.pt J. Valena b jonatas@isec.pt D. Dias-da-Costa c dias-da-costa@dec.uc.pt
Abstract
Innovative procedures for monitoring experimental tests using photogrammetry and image processing have been recently proposed. This manuscript aims at providing a practical demonstration of the advantages of using these new techniques in experimental tests up to failure. In this scope, focus is given to the curvature, rotation and to the tension stiffening effect, i.e., the contribution of the concrete between cracks to increase the bending stiffness. These parameters are crucial to describe the structural behavior of reinforced concrete beams both in serviceability and in ultimate limit states. These new techniques allow monitoring a high number of points, enhancing the traditional monitoring methods and providing more information which would be impossible to obtain using the traditional methods. Keywords: curvature, plastic rotation, experimental tests, photogrammetry, image processing.
Resumo
A monitorizao de ensaios experimentais tem registado desenvolvimentos significativos nos ltimos anos mediante o recurso fotogrametria e ao processamento de imagem. Neste artigo pretende-se demostrar a aplicabilidade prtica destas novas tcnicas na monitorizao de ensaios at rotura. Neste trabalho dada especial ateno curvatura, rotao plstica e contribuio do beto entre fendas para o aumento da rigidez flexo, i.e., o tension stiffening effect, os quais so fundamentais para caracterizar o comportamento estrutural de vigas de beto armado, tanto em servio como em estado limite ltimo. Verifica-se que estas novas tcnicas permitem monitorizar um nmero muito elevado de pontos, complementando os mtodos tradicionais de monotorizao, e possibilitando a determinao de informao impossvel de obter atravs dos mtodos tradicionais. Palavras-chave: rotao plstica, ensaios experimentais, fotogrametria, processamento de imagem.
CIEC, Instituto Superior de Engenharia de Coimbra, carmo@isec.pt, Rua Pedro Nunes - Quinta da Nora. 3030-199, Coimbra, Portugal; ICIST, Instituto Superior de Engenharia de Coimbra, jonatas@isec.pt, Rua Pedro Nunes - Quinta da Nora. 3030-199, Coimbra, Portugal; c INESC, University of Coimbra, dias-da-costa@dec.uc.pt, Rua Lus Reis Santos, 3030-788 Coimbra, Portugal. School of Civil Engineering, The University of Sydney, NSW2006, Australia.
a b
Received: 29 Nov 2012 Accepted: 14 Apr 2013 Available Online: 12 Jun 2013
2013 IBRACON
Plastic rotation and tension stiffening effect analysis in beams using photogrammetry
1. Introduction
Monitoring experimental tests performed on structural elements is crucial to quantify the loading effects. Measuring forces, curvatures, displacements and strains in key sections allow characterizing the structural behavior of the element. The observation of the failure mechanism in reinforced concrete elements is also important simultaneously with the identification of the crack pattern. The above quantities are usually measured by traditional monitoring instruments, including: load cells, mechanical strain gauges, demecs, strain gauges and LVDTs. Recently, new monitoring tools using photogrammetry and image processing were developed to determine some of the above parameters [1-8]. These tools allow to assess a large amount of data which is difficult, or even not possible, to measure with traditional methods. A detailed curvature evolution along the concrete beam axis is one example. This study aims at demonstrating how photogrammetry and image processing can be applied to study the flexural behavior of concrete beams and how these techniques go beyond the limitations of traditional methods. Furthermore, the analysis of the tension stiffening effect, combining the information obtained by photogrammetry and image processing, is presented.
Figure 1 Moment-curvature relation in a reinforced concrete section under pure bending before reinforcement yielding
1/r I - curvature at an uncracked section
Bending moment M
[(1/r)II - (1/r)I ]
Tension stiffening effect 1/r II - curvature at a fully cracked section
Cracking moment 1/rm= .(1/r)II + (1- ).(1/r) I
Curvature 1/r
2. Plastic rotation and tension stiffening effect in reinforced concrete beams

The possibility of achieving plastic analysis and linear analysis with moment redistribution requires a certain amount of plastic rotation in critical sections. It is important to ensure that critical sections can reach the foreseen failure type. In uncertain situations it becomes necessary to make an explicit verification of this capability. Thus, it is
crucial to know the moment versus curvature relationship or, alternatively, plastic rotation capacity versus x/d parameter (see EC2 [9]). The plastic rotation capacity is defined as the difference between the rotation at the ultimate load and at the steel bars yielding onset. Therefore, the plastic rotation can be defined as the integral of the curvature after steel yielding in the plastified area (Eq. 1).
Figure 2 Curvature evolution and tension stiffening effect (example of a region under support)
476
R. N. F. do Carmo | J. Valena | D. Dias-da-Costa
(1)
pl - plastic rotation capacity lpl - length of the plastic hinge 1/r - total curvature 1/ry - yielding curvature es - total strain of steel reinforcement esy - yielding strain of steel reinforcement d - effective depth of a cross-section x - neutral axis depth Computing the rotation in critical regions can be a difficult task since the curvature has a discontinuous development along the beam axis due to bending stiffness difference between cracked sections and uncracked sections. On the other hand, in the plastic hinge region and near failure, the assumption of plane sections is not valid. This makes difficult determining the rotation by integrating the curvature along the beam axis. The contribution of the concrete between cracks on the tensile strength originates a significant variation of the bending stiffness, known as tension stiffening effect. Disregarding this effect may lead to unrealistic predictions, i.e., if only the curvature in a fully cracked section is considered, a value greater than the actual rotation would be obtained (Figure 1). According to EC2 [9] and others codes, the mean curvature must be computed by considering both uncracked and entirely cracked states, i.e., by applying Equation 2.
(1/r)II - curvature at a fully cracked section z - distribution coefficient which take into account the tension stiffening effect In a reinforced concrete member, the tensile reinforcement strain is variable along the beam axis. Consequently, the plastic rotation has also a discontinuous variation, depending essentially on the curvature of the cracked sections and, in less extent, on the curvature of the sections between cracks. In Figure 2, the diagonally dashed area corresponds to the integral of the plastic curvature along the plastic hinge length, i.e., to the plastic rotation. In Figure 2 it is also observed that the plastic curvature tends to localize in cracked sections. Based on this, Bachmann, 1967 [10] proposed a straightforward method for computing the rotation in certain regions of beams due to cracks (rotation between the two opposite surfaces of the crack). The main characteristic of this method is the fact of not obtaining the rotation from the curvature. To apply this model it is necessary to know the number of cracks in the region of interest and the width and depth of the neutral axis at each crack [11-12]. In this case, photogrammetry and image processing are excellent techniques to obtaining all required data. The model is based on a discrete analysis of the reinforced concrete member, being the tangent to the deformed beam discontinuous at each crack. Figure 3 exemplifies this procedure for a region with negative moments, where the total rotation is equal to the sum of the rotations in the n existing cracks.
(3)
(1/r)m = .(1/r)II + (1 - ).(1/r)I

(1/r)m - mean curvature (1/r)I - curvature at an uncracked section
(2)
(4)
Figure 3 Bachmann's method for computing the rotation (example of a region over a support) [11-12]
477
wi - i-th crack width d - effective depth of the cross-section xi - neutral axis depth - rotation
3. Photogrammetry and image processing

3.1 Photogrammetry
Photogrammetry allows measuring the displacement field at targets placed at the surface of the specimen [1-4]. Alternatively, advanced image correlation techniques can be applied to avoid the use of pre-defined targets. In both cases, after obtaining the displacement field, the strain field is computed applying standard finite element method (MEF) procedures [5]. All operations performed after image acquisition to obtain the strain field, are described in this section, i.e.: i) target detection, where the positions of all targets are identified in all images; ii)homography, which allows scaling and orientating all images and thus obtaining the coordinates of the targets in the surface of the specimen; and iii)mesh generation, where a Delaunay triangulation is applied to define the post-processing mesh [13]. The first step, target detection, is performed using the Hough transform [14] to identify the geometrical center of the targets, at all stages. In summary, the average pixel radius allows computing a parametric transform, which results in a map of peaks coincident with the geometrical center of the targets [14, 15]. Secondly, after detecting all targets, their position is obtained in the world coordinate system. Since all displacements are expected to occur within a plane, a simple homography is established to match image coordinates with real plane coordinates. The concept is accomplished by solving the following system of equations for the reference stage:
where is a scale factor. Since only the ratio of the homography parameters is relevant, the number of unknowns to be solved is eight. Therefore, an exact solution is obtained using four targets. Usually, there are many more targets available and the resulting system is over determined. In this case, the solution is obtained by minimizing the norm Ah of Equation5 for n points. The eigenvector corresponding to the least eigenvalue of A T A allows to directly obtaining the solution. It should be noted that the homography define a map between two planes, independently of their position and orientation. The homography parameters can then be used to compute the real plane coordinates for all the targets and at any stage evaluated. The differences in the coordinates in relation to the reference stage allow to directly computing the corresponding displacement field. Then, an auxiliary mesh is assembled by means of a Delaunay triangulation and using the reference targets [13]. Lastly, the strain field is directly computed by a strain-nodal displacement matrix associated to the auxiliary mesh (see [3] for more details). It should be mention that all acquired images were orientated and scaled to 1:5 using the homography parameters. This value is used since the mean resolution of the original image frames was circa 0.2mm/pixel. Later, the value is also used to measure the crack width (see Section5.3).
3.2 Image processing

Digital image processing is a technique which allows detecting discontinuities in the image, i.e, points where sudden changes in the intensity level of the pixels occur. Therefore, detecting and measuring cracks on concrete surfaces is enabled. In the most cases, the detection of discontinuities in an image was performed by applying edge detectors [6-8]. This allows obtaining a binary image which enhances the crack pattern to be characterized. However, the method presents results strongly dependent on the surface conditions (other source of discontinuities). Thus, surfaces need to be carefully prepared and to adequate lighting conditions have to be assured. Therefore, combined approaches have emerged, in which the strain field is used to define critical regions where image processing is performed [15]. In this section, a brief description of the digital image processing operations required to enhance and measure cracking is presented. The surface of the specimen was painted white in order to obtain a homogeneous background, thus further enhancing cracking appearing during the experimental test. This method was developed for monitoring cracking by combining digital image processing and mathematical morphology operations, and a complete description of the procedure, including experimental examples, can be found in [7]. After orientating and scaling all the images according to the previous Section, the procedure comprises the following main steps: i) binarization of the images using the Otsus method; ii) mathematical morphology operations (cleaning, linking and filling) in order to prevent other sources of discontinuities (e.g. surface imperfections) to be misleadingly taken as cracks; and finally iii) measuring any selected crack on the scaled image.
(5)
where Xi and Yi are the real plane coordinates for each target i provided by the reference grid size painted at the surface of the specimen (in this case 20 x 20 mm2), xi and yi are the corresponding coordinates in the image, and h1 to h9 are the homography parameters. Thus, the following relation can be written, for any point belonging to the surface of the specimen [16]:
X h Y = h4 1 h7
h2 h5 h8
h3 x h6 y 1 h9
(6)
4. Experimental program
4.1 Overview
The method was applied to monitoring an experimental test un-
478
Figure 4 Cross section, structural scheme, shear force and bending moment diagrams
til failure of a reinforced concrete beam with 3.0m long, 12cm width and 27cm height. The beam was produced with lightweight concrete with a density of 1900 kg/m3, an average compressive strength of 51 MPa at 28 days and 54 MPa at the test day (166 days after casting) [17-18]. The concrete Youngs modulus of 24 GPa was also experimentally measured. For the longitudinal and transversal reinforcement hot rolled and ribbed steel bars of
S500NR-SD class were used. Tensile tests were performed to measure the steel yielding stress, fsym, and the tensile strength, fsum, 545 and 645 MPa, respectively. The beam was simply supported at both edges, 2.8 m span, and was subject to symmetrical loading, one concentrated load P which was divided into two equal loads distanced 1.0 m from the supports. Thus, the central region of the beam was theoretically under pure bending
Figure 5 General view of the test set-up
479
Figure 6 Force vs. testing time duration

80
support (sections where loading was applied) and the third was placed exactly at mid-span. Two LVDTs were also placed to measure the horizontal stretching of the beam at the pure bending region (Figure 5).
4.2 Preparation and image acquisition

Previously works using photogrammetry and image processing in structural monitoring have shown that the precision obtained in the displacement field is typically better than 0.050mm. Thus, in order to measure a uniform strain field with 0.5% precision (value suitable for characterizing the crack formation) a minimum distance between targets of 20 mm is required (Dias-da-Costa et al. [3]). Therefore, the circular targets were painted in the surface of the specimen composing a square grid with sides of 20 mm length, oppositely to the surface used for placing the traditional instruments. All images were captured by a digital camera installed in front of the beam, at a distance of 1.75m. A tripod and a remote shutter were used to trigger the camera without compromising the stability. The images were acquired at full resolution of 46083072 pixel. Before start loading, ten images were sequentially acquired and later on used for scaling and orientating all images used to monitoring the experimental test and also for estimate the error in the results. Four stages of the test are herein analyzed (Figure6).
60
Force P (kN)
40
20
photogrammetry
0 0:00
0:28
0:57
1:26
1:55
2:24
(Figure 4). The main load was applied by a 300 kN capacity hydraulic servo-actuator, attached to the metallic support frame. The test was performed using displacement control at a constant rate of 0.01 mm/s. Several parameters were measured during the test to characterize the beam behavior, namely, support reactions and displacements at selected critical sections. The test was also monitored using the image processing methods described in Section 3 and traditional instrumentation, in particular, load cells under the supports and displacement transducers (LVDTs) placed in the pure bending region. Three LVDTs were used to measure vertical displacements, two placed at 1.0 m from each
4.3 Error analysis

As previously mentioned the ten images acquired before start loading were used to estimate the error of the results, namely: i) the error of the homography; and ii) the precision of the coordinates at each target. In the first case, the error was 0.236mm and
Figure 7 Error map of target detection in the monitored surface
Table 1 Vertical displacements in mm using LVDTs and photogrammetry (units in mm)

Section 1.0 m at the left support Photog. LVDT-1 9.60 8.68 17.14 15.97 33.55 32.57 60.49 67.27 Section at mid-span Photog. LVDT-2 10.62 9.69 19.65 18.31 40.09 39.21 78.77 83.13 Section 1.0 m at the right support Photog. LVDT-3 8.98 9.62 17.12 18.05 34.89 35.81 74.18 72.10
Stage 1 Stage 2 Stage 3 Stage 4
480
Figure 8 Cracks pattern and strain map (stage 3)
A Cracks pattern
0.315mm in average and at a 95% confidence level, respectively. The precision in target detection was 0.031mm and 0.046mm in average and at a 95% confidence level, respectively (Figure 7). It is important to highlight that the error of the homography is a systematic error, i.e, the error on the displacement is assumed to correspond to the error in the target detection. Finally, the error in the strain field was circa 0.3% for the defined grid spacing.
First principal strain map
5.2 Curvature
The mean curvature at the pure bending region can be defined using the parabolic equation that best fits the vertical displacements. The curvature is the directly computed by the 2nd derivative of the equation. For this purpose, photogrammetry presents a significant advantage because it allows recording the displacement at a large number of cross sections which can be used to obtain a reliable approximation of the parabolic equation. The average curvature can also be assessed by measuring the horizontal stretching at the pure bending region. As mentioned above, two horizontal LVDTs were used to measure these deformations. Knowing the vertical distance between the LVDTs axis and the horizontal distance between reading points, the average strains and corresponding curvature, can be computed. In Figure 9, the moment versus curvature relation for the pure bending region (800 mm) is shown, identifying the onset of reinforcement yielding, stage 3 and stage 4. The curvature for a specific section and at a particular stage is determined based on strains measured along the section height. Initially, it is necessary to define the best fit straight line to the strains. Figure 10 presents the strain evolution and the corresponding curvature comput-
5. Results analysis
5.1 Preliminary results
The vertical displacements obtained using photogrammetry were compared with those obtained by traditional monitoring. The obtained differences were relatively small, 6% on average (Table 1). From the displacements measured at each target it is possible to compute the strain field (see Section 3.1). In Figure 8, the first principal strain field at stage 3 is shown (see Section 3.2). It is highlighted that it is consistent with the existing cracking pattern. Additionally, image processing can be used to measure the crack width at each test stage. This information is later used to determine the rotation in Section 5.3.
Figure 9 Bending moment vs. average curvature measured using horizontal LVDTs
40
32 (1/r)3 - (1/r)y Stage 3 (1/r)4 - (1/r)y 16

Start of steel yielding
M (kN.m)
24
Stage 4
8 800 mm 0 0 32 64 96 1/rm (10-3 m-1) 128 160
481
Figure 10 Curvature at a section placed at 50 mm of the left edge in the pure bending span
Figure 11 Curvature along the pure bending span (800 mm)
A Stage 1
Stage 2
C Stage 3
Stage 4
482
Figure 12 Plastic rotation using method 1 with shaded area representing the plastic rotation
A Stage 3
ed for the four stages considered in this analysis. The selected section was placed at 50 mm from the left edge of the pure bending region. A perfect linearity between strains was not found because, during the deformation, the sections do not remain plain due to concrete cracking. It is reminded that mechanical strain gauges and demcs only allow obtaining the average curvature in lengths of 100 or 200 mm (values limited by the mechanical strain gauge length) and not curvature in a specific section (Carmo and Lopes [19]). Another disadvantage of using mechanical strain gauges compared with photogrammetry is the time required to perform all readings. Photogrammetry provides a detailed curvature analysis because the spacing between targets can be significantly reduced, in this case 20 mm was adopted. The detailed evolution of curvature along the beam axis allows analyzing the influence of the concrete between the cracks and cracked sections in the beam deformation. Figure 11 presents the curvature evolution along the beam axis for the four analyzed stages, where it can be seen that the curvature is not constant. In stages 1 and 2, despite the cracks width not being very significant, a larger curvature was already detected in cracked sections. Between stages 2 and 3, the curvature increases in cracked sections, being approximately constant in sections between cracks. In stage 4, it can be seen that sections with maximum cur-
Stage 4
vatures tend to concentrate on a limited length, about 300-350 mm.
5.3 Plastic rotation

The reinforced concrete members have a nonlinear behavior when subjected to higher loads. Before reinforcement reaches the yield strength, the nonlinear behavior is caused by concrete cracking. After reaching the steel yielding, the influence of the nonlinearity of steel stress-deformation must also be taken into account. At this moment, it is considered that a plastic hinge is formed. To quantify the ductility of the member, the evaluation of the plastic rotation capacity at certain zones of the structure is generally used [20]. The plastic rotation capacity of a beam is the maximum plastic rotation supported immediately before collapsing. To compute the plastic rotation three methods are presented: i) curvature integration after steel yielding in the plastified area (Eq. 1 and Figure 12), ii) multiplying the difference of the average curvatures, at the steel yielding onset and at the analyzed stage, by 800 mm length (curvatures determined by the horizontal LVDTs) (Figure 9), iii) Bachmanns method, based on the sum of rotations which occur between the two sides of the crack (Eqs. 3 and 4 and Figure 13). The latter is easily applied using photogrammetry and image pro-
Figure 13 Crack width and neutral axis depth in mm (stage 3)
483
Table 2 Rotation computed in each identified crack

Stage 3 Crack 1 2 3 4 5 6 7 8 xi (mm) wi (mm) i (10 rad)
-3
86 1.60 92 1.20 100 1.00 87 1.00 99 1.20 91 0.80 109 0.80 120 1.00 i (10-3 rad) =
1.60 1.20 1.00 1.00 1.20 0.80 0.80 1.00 63.2
to the reinforcement yielding. Theoretically, flexural stiffness in a specific section can be determined in state I and II: in state I the whole cross-section of concrete and steel is considered, whereas in state II only the concrete under compression and, obviously, the steel area are considered. The flexural stiffness experimentally measured should be between these two limits due to the tension stiffening effect. Figure 14 presents the evolution of flexural stiffness (EI) with the applied load, P. The latter parameter is written using the dimensionless ratio P/Py (where Py is the load required to the reinforcement yielding). The stiffness was measured using the relation between bending moment applied and mean curvature, in this case measured using the horizontal LVDTs. Initially, stiffness assumes high values because the curvature is very low and therefore very sensitive to any reading variation from the LVDTs. The tension stiffening effect is particularly important in the analysis of deformations in concrete structures under serviceability conditions, as recommended in design codes for concrete structures. As mentioned in Section 2, the distribution coefficient z considers the contribution of the tension stiffening effect. An analysis of this effect is provided below to stage 1, where the applied load is ap-
Table 3 Total rotation obtained by Bachmann method

Stage 1 2 3 4 i (10 rad)
-3
Figure 14 Flexural stiffness vs. applied load
19.5 63.2 123.5
Table 4 Plastic rotation (x 10-3 rad)

Stage 3 4 Method 1 31.7 90.1 Method 2 30.8 95 Method 3 43.7 104
cessing, since all required data can be easily measured after the test, thus not requiring to stop the test to perform readings (Tables 2 and 3). However, this method has a drawback: in stages before steel yielding, when cracks are relatively small, is not possible to achieve the required accuracy to measure the cracks width. Table 4 summarizes the plastic rotations in pure bending region obtained by the three methods mentioned. It is found that the values obtained using methods 1 and 2 are very similar, with differences lower than 5%, and with the values smaller than those obtained using method 3, particularly at step 3, where the differences can reach 40%.
Figure 15 Average curvature in the monitored region (stage 1)
5.4 Tension stiffening effect evaluation

The beam flexural stiffness (EI) decreases as the applied load increases, initially due to concrete cracking and, at a later stage, due
484
proximately 70% of the maximum load, which corresponds roughly to serviceability conditions. Considering the flexural stiffness value experimentally obtained, i.e. 2679 kN.m2, and applying Equation 2 is obtained a value for the distribution coefficient z of 0.883. This coefficient can also be determined using the curvature computed by photogrammetry and image processing. At this stage the bending moment is 23.5 kN.m. Knowing the stiffness in states I and II, the curvature is: (1/r) I = 3.9 x10-3 m-1 and (1/r)II = 10.5 x10-3 m-1 . Since the average curvature, in the analyzed region, is 9.85 x10-3 m-1, the distribution coefficient z is 0.902 (Figure 15). The values experimentally measured can be compared with the EC2 [9], z = 1 - b.(Mcr/M)2. In this case, b is 1 (short-term loading) and the cracking moment Mcr is 6 kN.m, considering the concrete tensile strength and all homogenized section. Taking into account this data, and according the EC2, the distribution coefficient value is 0.935, very close to the values obtained experimentally.
of Civil Engineering of the Polytechnic of Coimbra for providing the conditions to carry out this study and the colleagues Hugo Costa, Tiago Simes, Ctia Loureno and Diogo Andrade, for their collaboration in the experimental programme.
8. References
[01] Hegger J, Sherif A, Grtz S (2004) Investigation of pre-and postcracking shear behavior of prestressed concrete beams using innovative measuring techniques. ACI Structural Journal 101 (2):183-192. [02] Lange J, Benning W, Siering K Crack detection at concrete construction units from photogrammetric data using image processing procedures. In: ISPRS Commission VII Mid-term Symposium Remote Sensing: From Pixels to Processes, Enschede, Netherlands, 8-11 May 2006. pp 493-496. [03] Dias-da-Costa D, Valena J, Jlio E (2011) Laboratorial test monitoring applying photogrammetric post-processing procedures to surface displacements. Measurement 44 (3):527-538. doi:10.1016/j.measurement.2010.11.014. [04] Valena J, Jlio E, Arajo H (2012) Application of photogrammetry to structural assessment. Experimental Techniques 36 (5). doi:10.1111/j.1747-1567.2011.00731.x. [05] Hoffman ME, Manevitz LM, Wong EK, Geers MGG, De Borst R, Brekelmans WAM (1996) Computing strain fields from discrete displacement fields in 2D-solids. International Journal of Solids and Structures 33 (29):4293-4307. doi:10.1016/0020-7683(95)00240-5 [06] Abdel-Quarter I, Abudayyeh O, Kelly M (2003) Analysis of edge detection techniques for crack identification in bridges. Journal of Computing in Civil Engineering 17 (3):255-263. doi:10.1061/~ASCE!0887-3801~2003!17:4~255! [07] Valena J, Dias-da-Costa D, Jlio ENBS (2012) Characterisation of concrete cracking during laboratorial tests using image processing. Construction and Building Materials 28 (1):607-615. doi:10.1016/j.conbuildmat.2011.08.082 [08] Hutchinson TC, Chen Z (2006) Improved Image Analysis for Evaluating Concrete Damage. Journal of Computing in Civil Engineering 20 (3):210-216 [09] CEN (2004) EN 1992-1-1: Eurocode 2: Design of Concrete Structures - Part 1-1: General Rules and Rules for Buildings, in: European Committee for Standardization (CEN). [10] Bachmann H (1967) Zur plastizittstheoretischen Berechnung statisch unbestimmter Stahlbetonbalken. Technische Hochschule, Zrich [11] Park R, Paulay T (1975) Reinforced concrete structures. Wiley, New York. [12] Buchaim R (2001) A influncia da no-linearidade fsica do concreto armado na rigidez flexo e na capacidade de rotao plstica. So Paulo University. [13] Barber CB, Dobkin DP, Huhdanpaa H (1996) The quickhull algorithm for convex hulls. ACM
6. Conclusions
Recently, innovative methods for monitoring structures using photogrammetry and image processing have been developed. This paper presents a practical example of these techniques, namely, in the characterization of the structural behavior of reinforced concrete beams, with particular attention in plastic rotation and tension stiffening effect. Photogrammetry and image processing were used to monitor a large number of points on a concrete surface until failure. Four stages during the experimental test were analyzed, highlighting the interest in the developed tools, which allow the computation of the strain field on concrete surfaces and characterization of the cracking pattern. Throughout the test, similar values of the vertical displacements between photogrammetry and LVDTs were obtained. The detailed information of the strain field in the concrete surface allowed computing the curvature in specific sections and evaluating their progress along the beam axis. This information is extremely difficult to obtain with traditional methods, since the use of mechanical strain gauges is required. Usually, the mean curvature along regions can only be determined with 100 or 200 mm resolution due to the limitation of the equipment size. Photogrammetry allows obtaining the curvature along smaller regions, in this case with 20 mm length. The detailed evolution of the curvature along the beam axis was used to study the plastic rotation and the tension stiffening effect. The method allowed to experimentally verify that the plastic curvature was localized in cracked sections and in a restricted length of the beam, in this case around 350 mm. Photogrammetry was also used to determine rotation between the cracks boundaries. These values have also been used to compute the plastic rotation in the pure bending zone. In summary, the presented study aims validating and demonstrating that photogrammetry and image processing techniques can provide significant and relevant data relatively to traditional methods of monitoring, both in serviceability and in ultimate states. Furthermore, due to the amount of data recorded, is possible to perform analyses that would otherwise be unfeasible.
7. Acknowledgments
The authors would like to express their gratitude to the Department
485
Transactions on Mathematical Software 22 (4):469-483. doi:10.1145/235815.235821. [14] Ballard D (1981) Generalizing the Hough Transform to Find Arbitrary Shapes. Pattern Recognition 13:111122 [15] Valena J, Dias-da-Costa D, Jlio E, Arajo H, Costa H (2012) Automatic crack monitoring using photogrammetry and image processing. Measurement (in press). doi:10.1016/j.measurement.2012.07.019. [16] Criminisi A, Reid I, Zisserman A (2000) Single view metrology. Int J Comput Vision 40 (2):123-148. doi:10.1023/a:1026598000963. [17] Carmo RNF, Costa H, Loureno C, Andrade D, Simes T (2012) Influence of both concrete strength and transverse confinement on bending behaviour of reinforced LWAC beams. Engineering Structures (in press). doi:10.1016/j.engstruct.2012.09.030. [18] EN 12390, Testing hardened concrete (2009). European Committee for Standardisation, Brussels, Belgium. [19] Carmo RNF, Lopes SM (2005) Influence of the shear force and transverse reinforcement ratio on plastic rotation capacity. Structural Concrete, Journal of the fib (Thomas Telford Journal) 6 (3):107-117 [20] CEB (1998) Ductility of reinforced concrete structures. Bulletin dInformation n 242. Lausanne, Switzerland.
486
Plastic rotation and tension stiffening effect analysis in beams using photogrammetry Anlise da rotao plstica e da contribuio do beto entre fendas em vigas usando fotogrametria
R. N. F. do Carmo a carmo@isec.pt J. Valena b jonatas@isec.pt D. Dias-da-Costa c dias-da-costa@dec.uc.pt
Abstract
Innovative procedures for monitoring experimental tests using photogrammetry and image processing have been recently proposed. This manuscript aims at providing a practical demonstration of the advantages of using these new techniques in experimental tests up to failure. In this scope, focus is given to the curvature, rotation and to the tension stiffening effect, i.e., the contribution of the concrete between cracks to increase the bending stiffness. These parameters are crucial to describe the structural behavior of reinforced concrete beams both in serviceability and in ultimate limit states. These new techniques allow monitoring a high number of points, enhancing the traditional monitoring methods and providing more information which would be impossible to obtain using the traditional methods. Keywords: curvature, plastic rotation, experimental tests, photogrammetry, image processing.
Resumo
A monitorizao de ensaios experimentais tem registado desenvolvimentos significativos nos ltimos anos mediante o recurso fotogrametria e ao processamento de imagem. Neste artigo pretende-se demostrar a aplicabilidade prtica destas novas tcnicas na monitorizao de ensaios at rotura. Neste trabalho dada especial ateno curvatura, rotao plstica e contribuio do beto entre fendas para o aumento da rigidez flexo, i.e., o tension stiffening effect, os quais so fundamentais para caracterizar o comportamento estrutural de vigas de beto armado, tanto em servio como em estado limite ltimo. Verifica-se que estas novas tcnicas permitem monitorizar um nmero muito elevado de pontos, complementando os mtodos tradicionais de monotorizao, e possibilitando a determinao de informao impossvel de obter atravs dos mtodos tradicionais. Palavras-chave: rotao plstica, ensaios experimentais, fotogrametria, processamento de imagem.
CIEC, Instituto Superior de Engenharia de Coimbra, carmo@isec.pt, Rua Pedro Nunes - Quinta da Nora. 3030-199, Coimbra, Portugal; ICIST, Instituto Superior de Engenharia de Coimbra, jonatas@isec.pt, Rua Pedro Nunes - Quinta da Nora. 3030-199, Coimbra, Portugal; c INESC, University of Coimbra, dias-da-costa@dec.uc.pt, Rua Lus Reis Santos, 3030-788 Coimbra, Portugal. School of Civil Engineering, The University of Sydney, NSW2006, Australia.
a b
Received: 29 Nov 2012 Accepted: 14 Apr 2013 Available Online: 12 Jun 2013
2013 IBRACON
1. Introduo
Durante os ensaios experimentais a monitorizao dos elementos estruturais essencial para quantificar os efeitos resultantes da aplicao de aces. As foras, curvaturas, deslocamentos e extenses medidas em determinadas seces permitem caracterizar o comportamento estrutural do elemento em estudo. A observao do mecanismo de rotura igualmente importante, assim como a avaliao do padro de fendilhao desenvolvido nos elementos de beto armado. As grandezas acima referidas so, usualmente, medidas atravs de instrumentos de monitorizao tradicionais, nomeadamente: clulas de carga, alongmetros, demecs, extensmetros e LVDTs. Recentemente foram desenvolvidas novas ferramentas de monitorizao, que recorrem fotogrametria e ao processamento de imagem, permitindo determinar algumas das grandezas mencionadas [1-8]. No entanto, atendendo ao elevado nmero de pontos monitorizados possvel obter dados que so dificeis, ou mesmo inviveis, de determinar com os mtodos tradicionais. A evoluo detalhada da curvatura ao longo de troos de vigas de beto armado um desses exemplos. Neste estudo, procura-se demonstrar como a fotogrametria e o processamento de imagem podem ser empregues no estudo do comportamento flexo de vigas de beto armado, e como essas tcnicas ultrapassam as limitaes dos mtodos tradicionais. Apresenta-se ainda a anlise do tension stiffening effect combinando a informao obtida por fotogrametria e processamento de imagem.
Figura 1 Relao momento-curvatura numa seco de beto armado sujeita flexo antes da cedncia das armaduras
2. Rotao plstica e o tension stiffening effect em vigas de beto armado

Quando se realiza uma anlise linear com redistribuio de esforos ou uma anlise plstica, importante garantir a capacidade de rotao plstica das zonas crticas para que a rotura ocorra como
Figura 2 Evoluo da curvatura e o efeito tension stiffening (exemplo de uma regio sobre o apoio)
488
previsto. Nas situaes mais duvidosas necessrio efectuar uma verificao explcita dessa capacidade. Nesse contexto, fundamental conhecer a relao momento-curvatura ou a relao entre a capacidade de rotao plstica e o parmetro x/d (EC2 [9]). A capacidade de rotao plstica definida como a diferena entre a rotao correspondente carga ltima e a rotao verificada no incio da cedncia das armaduras. Portanto, a rotao plstica pode ser calculada como a integral da curvatura aps a cedncia das armaduras na zona plastificada (Eq. 1).
Segundo o EC2 [9] e outros regulamentos, no clculo da curvatura mdia deve ser realizado atravs de uma ponderao entre o estado no fendilhado e totalmente fendilhado, aplicando a Equao 2.
(1/r)m = .(1/r)II + (1 - ).(1/r)I
(2)
(1)
qpl - rotao plstica lpl - comprimento da rtula plstica 1/r - curvatura total da seco 1/ry - curvatura da seco no instante da cedncia das armaduras es - extenso total das armaduras esy - extenso de cedncia das armaduras d - altura til da seco x - profundidade do eixo neutro A dificuldade na determinao da rotao nas regies crticas deve-se ao facto da curvatura ter um desenvolvimento descontnuo, devido variao da rigidez flexo nas seces fendilhadas e nas seces entre fendas. Por outro lado, na regio da rtula plstica, o pressuposto de que as seces se mantm planas durante a deformao no vlido numa situao prxima da rotura. Estas particularidades dificultam o clculo da rotao a partir da integrao da curvatura ao longo do eixo do elemento. A contribuio do beto entre fendas na resistncia traco origina variaes significativas da rigidez flexo, o tension stiffening effect. A no considerao deste efeito pode conduzir a previses pouco realistas, i.e., se apenas se contabilizar a curvatura de uma seco totalmente fendilhada, obter-se- um valor da rotao superior ao real (Fig. 1).
(1/r)m - curvatura mdia (1/r)I - curvatura numa seco no fendilhada (1/r)II - curvatura numa seco totalmente fendilhada z - coeficiente de distribuio que tem em conta a contribuio do beto traccionado entre fendas Num elemento de beto armado, a deformao da armadura traccionada varivel ao longo do eixo da viga. Consequentemente, a rotao plstica tem tambm uma variao descontnua, dependendo essencialmente da curvatura das seces fendilhadas, sendo pequena a contribuio da curvatura das seces entre fendas. Na Figura 2, a rea preenchida pela trama diagonal corresponde ao integral da curvatura plstica ao longo do comprimento da rtula plstica, o que equivale a afirmar que essa rea igual ao valor da rotao plstica. A Figura 2 permite ainda observar que a curvatura plstica tende concentrar-se nas seces fendilhadas. Bachmann em 1967 [10] props um mtodo expedito para calcular a rotao em troos de viga baseado nas rotaes que ocorrem nas fendas (rotao entre as duas faces da fenda). Este mtodo tem como principal particularidade no determinar a rotao a partir da curvatura. Para se aplicar este modelo de clculo necessrio conhecer o nmero de fendas no troo em anlise, a sua largura, e a profundidade do eixo neutro em cada fenda [11-12]. A fotogrametria e o processamento de imagem so excelentes tcnicas para obter estes dados. Este modelo baseado numa anlise discreta do elemento de beto armado e, consequentemente, h uma descontinuidade da tangente deformada em cada fenda. A Figura 3 exemplifica este clculo para o troo de uma viga numa zona de momentos nega-
Figura 3 Mtodo de Bachmann para o clculo da rotao (exemplo de uma regio sobre o apoio) [11-12]
489
tivos, sendo a rotao total igual ao somatrio das rotaes nas n fendas existentes no troo considerado.
(3)
onde X i e Yi so as coordenadas reais de cada alvo i, colocadas 2 numa grelha com afastamento de 20 20 mm , xi e yi so as coordenadas da imagem correspondentes, e h1 a h9 so os parmetros da homografia, que permitem obter, para qualquer ponto pertencente superfcie do provete, a seguinte relao [16]:
(4)
wi - largura da fenda i d - altura til da seco xi - profundidade do eixo neutro q - rotao
X h Y = h4 1 h7
h2 h5 h8
h3 x h6 y 1 h9
(6)
3. Fotogrametria e processamento de imagem

3.1 Fotogrametria
A fotogrametria pode ser aplicada para calcular o campo de deslocamentos em superfcies de beto, com auxlio de alvos colocados na superfcie dos provetes [1-4]. Em alternativa podem, igualmente, ser utilizadas tcnicas de correlao entre imagens, dispensando, desta forma, a utilizao de alvos. Em ambos os casos, o campo de deformaes pode ser determinado a partir do campo de deslocamentos, aplicando procedimentos standard do mtodo dos elementos finitos (MEF) [5]. Nesta seco descrito o procedimento aplicado para obter o campo de deformaes aps a aquisio de imagem, nomeadamente: i) deteco dos alvos, no qual a posio de cada alvo identificada na imagem; ii) homografia, que permite escalar e orientar as imagens e, desta forma, obter as coordenadas de cada alvo na superfcie do provete; e iii) gerao de uma malha auxiliar, calculada com uma trianguao de Delaunay [13]. A deteco dos alvos foi realizada mediante a aplicao de um algoritmo que utiliza a transformada de Hough [14], o qual permite identificar o centro e o raio de todos os alvos em qualquer fase do ensaio. Em resumo, o raio mdio, em pixels, permite calcular uma transformada paramtrica, da qual resulta um mapa de picos coincidentes com o centro geomtrico dos alvos [14, 15]. Depois de detetar todos os alvos, necessrio obter a posio do sistema em coordenadas globais. Desde que os deslocamentos ocorram no mesmo plano, pode ser estabelecida uma homografia que faa coincidir as coordenadas da imagem com as coordenadas reais. Para esse efeito necessrio resolver, apenas para a fase de referncia, o seguinte sistema de equaes:
onde o factor de escala. U m a vez que o nmero total de incgnitas oito (apenas a relao dos elementos da matriz significativa), o problema sobredeterminado para mais de quatro pontos. Essa questo pode ser resolvida atravs da minimizao dos resduos Ah da Equao5 para n pontos. Pode demonstrar-se que a soluo dada directamente pelo vector prprio correspondente ao menor valor prprio de A T A . De salientar que a homografia define um mapeamento entre dois planos, independentemente das suas posies e orientaes. Portanto, os dois planos (plano da imagem e plano da superfcie real) no tm obrigatoriedade de ser paralelos. Posteriormente, os parmetros da homografia so utilizados para calcular as coordenadas reais dos alvos em cada fase de anlise. O campo de deslocamento obtido pelo clculo das diferenas de coordenadas em relao fase de referncia. Na etapa seguinte, construda uma malha auxiliar atravs de uma triangulao Delaunay, utilizando os alvos de referncia [13]. Finalmente, o campo de deformaes directamente calculado, atravs da matriz de extenso nodal da malha auxiliar (ver mais detalhes em [3]). De referir que todas as imagens adquiridas foram orientadas e dimensionadas escala 1:5, utilizando os parmetros da homografia, uma vez que a resoluo mdia das imagens originais era cerca de 0.2mm/pixel. Posteriormente, esta informao utilizada para medir a abertura das fendas na Seco 5.3.
3.2 Processamento de imagem

O processamento digital de imagem permite detectar descontinuidades na imagem, i.e., pontos onde ocorrem mudanas bruscas no nvel da intensidade dos pixels. Desta forma, possvel detectar e medir fendas em estruturas de beto. Normalmente, a identificao de descontinuidades numa determinada rea da imagem recorre aplicao de edge detectors [6-8], entre outras tcnicas, de forma a obter imagens binrias cuja manipulao permite caracterizar o padro de fendilhao. Como desvantagem, estes mtodos requerem a preparao cuidada da superfcie do provete e a existncia de condies de iluminao adequadas. Para ultrapassar estas limitaes tm surgido abordagens combinadas, onde o campo de deformaes utilizado para restringir a rea da imagem a analisar, aplicando o processamento exclusivamente a reas criticas [15]. Nesta Seco encontram-se descritas, de forma resumida, as operaes de processamento de imagem requeridas para realce e medio das fendas em superfcies de beto. De salientar que a superficie foi prviamente pintada de branco para adquirir um fundo homogneo, evidenciando assim o aparecimento das fendas
(5)
490
Figura 4 Seco transversal, esquema estrutural e diagrama de esforos
durante o ensaio. O procedimento adoptado foi especificamente desenvolvido para monitorizar fendas, combinando processamento digital de imagem e operaes de morfologia matemtica. Este procedimento encontra-se descrito de forma detalhada, incluindo exemplos de aplicao em ensaios experimentais, em [7]. Resumidamente, aps orientao e escala das imagens, de acordo com o descrito na Seco anterior, estas so processadas
aplicando o mtodo de Otsu, que reala as fendas na superfcie do provete atravs da binarizao das imagens. De seguida, so aplicadas operaes de morfologia matemtica (eliminao, unio e prenchimento) para eliminar descontinuidades que so, erradamente, detectadas como fendas (e.g. imperfeies da superfcie). Por fim, pode medir-se directamente qualquer das fendas na imagem escalada.
Figura 5 Vista geral do ensaio
491
Figura 6 Relao fora-tempo do ensaio
4. Programa experimental
4.1 Descrio geral
A viga de beto armado que foi objecto desta anlise tinha 3,0 m de comprimento, 12cm de largura e 27cm de altura. A viga foi produzida com um beto de agregados leves, com uma densidade de 1900 kg/m3, e uma resistncia mdia compresso de 51 MPa aos 28 dias e 54 MPa no dia do ensaio (166 dias aps a betonagem) [17-18]. O mdulo de elasticidade do beto foi tambm
determinado experimentalmente, 24 GPa. Para as armaduras longitudinais e transversais utilizaram-se vares de ao nervurado, laminados a quente, da classe S500NR-SD. Foram realizados ensaios de trao para caracterizar a tenso de cedncia, fsym, e a tenso de rotura do ao, fsum. 545 e 645MPa, respectivamente. A viga encontrava-se simplesmente apoiada, 2,8 m de vo, tendo sido sujeita a um carregamento simtrico, constitudo por uma fora, P, repartida em dois pontos distanciados um metro dos apoios. Deste modo, a regio central da viga ficou submetida, teoricamente, apenas a flexo pura (Fig. 4). A carga principal foi transmitida atravs de um servo-atuador hidrulico, fixo ao prtico de ensaio, o qual tem uma capacidade mxima de 300kN. O ensaio decorreu com controlo de deslocamento, a uma velocidade de 0,01 mm/s. Durante o ensaio foram medidas grandezas que permitem caracterizar o comportamento das vigas ensaiadas, nomeadamente, as reaces de apoio e os deslocamentos em seces crticas. Alm dos mtodos de imagem referidos na Seco3, os ensaios foram monitorizados recorrendo a instrumentao tradicional, nomeadamente, transdutores de deslocamento (LVDTs), verticais e horizontais, colocados na zona de flexo pura, e clulas de carga sob os apoios. Foram utilizados trs LVDTs para medir os deslocamentos verticais da viga: dois foram colocados a 1,0 m de cada um dos apoios (seces onde as cargas foram aplicadas) e o outro foi colocado exactamente a meio vo. Foram ainda utilizados dois LVDTs na horizontal, para medir o encurtamento e o alongamento horizontais da zona de flexo pura da viga (Fig. 5).
4.2 Preparao e aquisio de imagens

De acordo com a experincia adquirida na utilizao de mtodos pticos que recorrem fotogrametria e ao processamento de ima-
Figura 7 Representao do erro na superfcie monitorizada correspondente deteco dos alvos
Tabela 1 Deslocamentos verticais em mm usando os LVDTs e a fotogrametria

Seco a 1,0 m do apoio esquerdo Fotog. LVDT-1 9,60 8,68 17,14 15,97 33,55 32,57 60,49 67,27 Seco a meio vo Fotog. LVDT-2 10,62 9,69 19,65 18,31 40,09 39,21 78,77 83,13 Seco a 1,0 m do apoio direito Fotog. LVDT-3 8,98 9,62 17,12 18,05 34,89 35,81 74,18 72,10
Etapa 1 Etapa 2 Etapa 3 Etapa 4
492
Figura 8 Fendilhao e extenses (etapa 3)
A Padro de fendilhao
gem, o campo de deslocamentos obtido com uma preciso inferior a 0,05 mm. Assim, para medir extenses com uma preciso de 0,5%, valor aceitvel para caracterizar o processo de formao de fendas, a distncia mnima entre os alvos deve ser de 20 mm = 2 x 0,05 / 0,5% (Dias-da-Costa et al. [3]). Os alvos foram colocados numa grelha de 20 x 20 mm2, pintados na superfcie da viga, na face oposta colocao da instrumentao tradicional. As imagens foram adquiridas com uma cmara digital colocada num trip a cerca de 1,75 m da viga. Foi utilizado um disparador remoto para evitar qualquer vibrao do sistema de aquisio fotogrfico. Todas as fotografias foram adquiridas com o mximo de resoluo, i.e., 46083072 pixel. Antes do incio do ensaio foi registada uma sequncia de 10 imagens com o objectivo de avaliar o erro do mtodo na orientao e atribuio de escala s imagens. Esta fase foi considerada a referncia para o ensaio. Durante o ensaio, foram adquiridas imagens sempre que se justificou. Neste estudo, foram seleccionadas quatro imagens, correspondentes s etapas mais relevantes do ensaio (Fig. 6).
Mapa da primeira extenso principal
incio do ensaio experimental foram usadas para estimar o erro da homografia e a preciso das coordenadas de cada. Neste caso, a homografia introduz um erro mdio de 0,236mm e 0,315 mm para um intervalo de 95% de confiana, enquanto o erro na deteco dos alvos 0,031 mm, em mdia, e 0,046 mm para um intervalo com 95% de confiana (Fig. 7). O erro na determinao dos deslocamentos, obtidos pela diferena de coordenadas entre as etapas seleccionadas, corresponde apenas ao erro na deteco dos alvos. Considerando o espaamento adoptado para os alvos, o erro no campo das extenses foi cerca de 0,3%.
5. Anlise de resultados
5.1 Resultados preliminares
Os deslocamentos verticais obtidos atravs da fotogrametria foram comparados com os valores obtidos com a monitorizao tradicional. Verificou-se que as diferenas so relativamente pequenas entre os dois mtodos, em mdia 6% (Tabela 1). A partir dos deslocamentos medidos em cada alvo possvel calcular o campo de extenses (ver Seco 3.1). Na Figura 8, corres-
4.3 Anlise do erro

Como referido, as 10 imagens registadas imediatamente antes do
Figura 9 Relao momento-curvatura mdia determinada usando os LVDTs horizontais
493
Figura 10 Curvatura na seco a 50 mm da extremidade esquerda na regio de flexo pura
Figura 11 Evoluo da curvatura na regio de flexo pura (800 mm)
A Etapa 1
Etapa 2
C Etapa 3
Etapa 4
494
Figura 12 Rotao plstica usando o mtodo 1
A Etapa 3
pondente etapa 3, constata-se que a evoluo das extenses, determinado com o processamento de imagem, est de acordo com o padro de fendilhao (ver Seco 3.2). Adicionalmente, o processamento de imagem pode ser usado para medir a largura das fendas em cada etapa do ensaio. Esta informao ser utilizada para determinar a rotao na Seco 5.3.
Etapa 4
5.2 Determinao da curvatura

A curvatura mdia na regio de flexo pura pode ser determinada atravs da parbola que melhor se ajusta aos deslocamentos verticais. A curvatura determina-se directamente pela 2 derivada da equao da parbola. Para este efeito, a fotogrametria apresenta uma vantagem significativa, pois permite o registo de deslocamentos num nmero elevado de seces, obtendo-se uma aproximao fidedigna para a equao da parbola. A curvatura mdia pode tambm ser determinada medindo o encurtamento e o alongamento horizontais da regio de flexo pura. Como j foi referido, foram utilizados dois LVDTs horizontais com o objectivo de medir essas deformaes. Conhecendo a distncia vertical entre os eixos dos LVDTs e a distncia horizontal entre os pontos de leituras, pode determinar-se as extenses mdias e, consequentemente,
a curvatura. Na Figura 9 apresenta-se a relao momento-curvatura para regio de flexo pura (com 800 mm), identificando-se o incio da cedncia das armaduras, a etapa 3 e a etapa 4. A curvatura, para uma seco especfica e numa dada etapa, determinada com base nas extenses medidas ao longo da altura da seco. Numa primeira fase, necessrio determinar a recta que melhor se ajusta ao desenvolvimento das extenses. Na Figura 10, apresenta-se a evoluo das extenses e o clculo da curvatura para as 4 etapas consideradas na anlise, numa seco a 50 mm do limite esquerdo da regio em flexo pura. Verifica-se que no existe uma linearidade perfeita entre as extenses, porque as seces durante a deformao no se mantm planas devido fendilhao do beto. Recorda-se que a utilizao de alongmetros e demecs apenas permitem obter a curvatura mdia em troos com 100 ou 200 mm (valores limitados pelo comprimento dos alongmetros) e no a curvatura numa seco especfica (Carmo e Lopes [19]). Uma outra desvantagem do uso do alongmetro comparativamente fotogrametria o tempo exigido para a realizao das leituras. A fotogrametria permite uma anlise detalhada da curvatura, pois o espaamento entre miras pode ser bastante reduzido, neste caso foi 20 mm. A evoluo to pormenorizada da curvatura ao longo do eixo da viga permite analisar a influncia do beto entre
Figura 13 Largura das fendas e profundidade do eixo neutro em mm (etapa 3)
495
Tabela 2 Clculo da rotao em cada fenda

Etapa 3 Fenda 1 2 3 4 5 6 7 8 xi (mm) wi (mm) i (10 rad)
-3
86 1,60 92 1,20 100 1,00 87 1,00 99 1,20 91 0,80 109 0,80 120 1,00 i (10-3 rad) =
1,60 1,20 1,00 1,00 1,20 0,80 0,80 1,00 63,2
a tenso de cedncia, o comportamento no linear originado pela fendilhao do beto. Depois das armaduras atingirem a cedncia, existe tambm a influncia da no linearidade da relao tenso - deformao do ao. Neste instante, considera-se que se forma uma rtula plstica. Um dos critrios mais usados para a quantificar a ductilidade de um elemento a avaliar a capacidade de rotao plstica de certas zonas da estrutura [20]. A capacidade de rotao plstica de uma viga a rotao plstica mxima que a viga suporta antes da rotura. Para o clculo da rotao plstica apresentam-se trs mtodos: i) integrar a curvatura aps a cedncia das armaduras na zona plastificada (Eq. 1 e Fig. 12); ii) multiplicar a diferena das curvaturas mdias pelo comprimento de 800 mm (curvaturas determinada pelos LVDTs horizontais) (Fig. 9); iii) mtodo de Bachmann, baseado no somatrio das rotaes que ocorrem entre as duas faces de uma fenda (Eq.s 3 e 4 e Fig. 13). A fotogrametria e o processamento de imagem permitem aplicar o mtodo Bachmann, pois o clculo de todos os dados necessrios so determinados com relativa facilidade e sem necessidade de parar o ensaio para realizar as leituras (Tabelas 2 e 3). Todavia, este mtodo apresenta uma desvantagem. Na fase
Tabela 3 Rotao total obtida segundo o mtodo de Bachmann

Etapa 1 2 3 4 i (10 rad)
-3
Figura 14 Relao rigidez flexo carga aplicada
19,5 63,2 123,5
Tabela 4 Rotao plstica (x 10-3 rad)

Etapa 3 4 Mtodo 1 31,7 90,1 Mtodo 2 30,8 95 Mtodo 3 43,7 104
fendas e das seces fendilhadas na deformao total da viga. Na Figura 11 apresenta-se a evoluo da curvatura ao longo do eixo da viga para as 4 etapas analisadas, podendo-se observar que a curvatura no constante. Na etapa 1 e 2, apesar das fendas ainda no terem uma largura muito significativa verifica-se, como era de esperar, que a curvatura maior nas seces fendilhadas. Entre a etapa 2 e 3, ocorre um aumento da curvatura nas seces fendilhadas, mantendo-se aproximadamente constante os valores para as seces entre fendas. Na etapa 4, observa-se que as seces com as curvaturas mximas tendem a concentrar-se numa zona restrita, com cerca de 300-350 mm.
Figura 15 Curvatura mdia na regio monitorizada (etapa 1)
5.3 Rotao plstica

Os elementos de beto armado tm um comportamento no linear quando sujeito a cargas elevadas. Antes das armaduras atingirem
496
que antecede a cedncia das armaduras, quando a fendilhao pouco significativa, ainda no possvel atingir a preciso necessria para medir a abertura das fendas. A Tabela 4 apresenta um resumo das rotaes plsticas obtidas pelos trs mtodos referidos na regio em flexo pura. Verifica-se que os valores determinados pelo mtodo 1 e 2 so bastante semelhantes, com diferenas inferiores a 5%, e com valores inferiores aos obtidos pelo mtodo 3, principalmente na etapa 3 onde as diferenas registadas atingem os 40%.
5.4 Avaliao do efeito tension stiffening

A rigidez flexo da viga (EI) diminui medida que a carga aplicada aumenta, inicialmente devido fendilhao do beto e, numa fase posterior, porque as armaduras entram em cedncia. Teoricamente, pode determinar-se a rigidez flexo de uma seco para os estados I e II: no estado I considera-se toda a seco de beto e a rea de ao; e no estado II considera-se apenas o beto comprimido e, obviamente, a rea de ao. A rigidez flexo determinada experimentalmente deve estar compreendida entre estes 2 limites, devido ao efeito tension stiffening. Na Figura 14 apresenta-se a evoluo da rigidez flexo (EI) com a carga, P, sendo esta adimensionalizada atravs da relao P/Py (onde Py carga na cedncia das armaduras). A rigidez determinou-se atravs da relao entre o momento aplicado e a curvatura mdia, neste caso determinada usando os LVDTs horizontais. Inicialmente, a rigidez assume valores elevados porque a curvatura muito baixa e, consequentemente, muito sensvel a qualquer variao nas leituras dos LVDTs. O tension stiffening effect particularmente importante na anlise das deformaes das estruturas de beto armado nas condies de servio, como se constata nos regulamentos de projectos de estruturas de beto. Como se referiu no Seco 2, o coeficiente de distribuio z considera a contribuio do beto entre fendas e, portanto, a forma de quantificar o tension stiffening effect. Abaixo, apresenta-se uma anlise deste efeito para a etapa 1, onde a carga aplicada cerca de 70% da carga mxima, o que corresponde aproximadamente s condies de servio. Considerando o valor obtido experimentalmente para a rigidez flexo, 2679 kN.m2, e aplicando a Equao 2 rigidez, obtm-se um valor para o coeficiente de distribuio z de 0,883. Este coeficiente tambm pode ser determinado usando a curvatura calculada atravs da fotogrametria e processamento de imagem. Nesta etapa regista-se um momento flector de 23,5 kN.m. Conhecendo a rigidez no estado I e II, obtm-se os seguintes valores para a curvatura: (1/r)I = 3,9x10-3 m-1 e (1/r)II = 10,5x10-3 m-1. Sendo a curvatura mdia na regio em anlise 9,85x10-3 m-1, o coeficiente de distribuio z 0,902 (Fig. 15). Os valores determinados experimentalmente podem ser comparados com o valor previsto pelo EC2 [9], z = 1 - b.(Mcr/M)2. Neste caso, o b 1 (um s carregamento de curta durao) e o momento de fendilhao Mcr 6 kN.m, considerando a resistncia do beto traco e a seco homogeneizada. Tendo em conta estes dados, o valor do coeficiente de distribuio segundo o EC2 0,935, valor bastante prximo dos valores obtidos experimentalmente.
imagem. Neste trabalho apresenta-se um exemplo prtico destas tcnicas na caracterizao do comportamento estrutural de vigas de beto armado, dando especial ateno rotao plstica e ao tension stiffening effect. A fotogrametria e o processamento de imagem foram utilizadas para monitorizar um nmero bastante significativo de pontos na superfcie de uma viga ensaiada at rotura. Foram analisadas 4 etapas distintas do ensaio, demostrando-se as potencialidades das ferramentas desenvolvidas, nomeadamente, a determinao do campo de extenses na superfcie da viga e a evoluo e caracterizao do padro de fendilhao. Registaram-se valores de deslocamentos verticais similares entre a fotogrametria e os LVDTs ao longo de todo o ensaio. O registo de informao detalhada do campo de extenses na superfcie monitorizada, permitiu determinar a curvatura em seces especficas e avaliar a sua evoluo ao longo do eixo da viga. Esta ltima informao extremamente difcil de obter atravs dos mtodos tradicionais, porque usando o alongmetro, por exemplo, s possvel determinar a curvatura mdia para troos com 100 ou 200 mm (devido limitao da dimenso do equipamento). Recorrendo fotogrametria pode obter-se a curvatura em troos muito mais pequenos, neste caso 20 mm. A evoluo pormenorizada da curvatura ao longo do eixo da viga permitiu o estudo da rotao plstica e do tension stiffening effect. Nesta anlise foi possvel verificar, experimentalmente, que a curvatura plstica concentrou-se nas seces fendilhadas e numa zona restrita da viga, com cerca de 350 mm. A fotogrametria foi ainda utilizada para determinar rotaes entre as faces das fendas, tendo esses valores servido para calcular a rotao plstica na zona em flexo pura. Em suma, o estudo apresentado serviu para validar e demonstrar que a fotogrametria e o processamento de imagem so tcnicas que permitem complementar os dados obtidos pelos mtodos tradicionais de monitorizao, tanto nas condies de servio como numa situao prxima da rotura. Alm disso, devido quantidade de dados que permitem obter, possibilitam a realizao de anlises que seriam inviveis com os mtodos tradicionais de monitorizao.
7. Agradecimentos
Os autores agradecem ao Departamento de Engenharia Civil do ISEC-IPC as condies facultadas para a realizao do ensaio, aos colegas Hugo Costa, Tiago Simes, Ctia Loureno e Diogo Andrade, a colaborao no trabalho experimental.
[01] [02] [03] Hegger J, Sherif A, Grtz S (2004) Investigation of pre-and postcracking shear behavior of prestressed concrete beams using innovative measuring techniques. ACI Structural Journal 101 (2):183-192. Lange J, Benning W, Siering K Crack detection at concrete construction units from photogrammetric data using image processing procedures. In: ISPRS Commission VII Mid-term Symposium Remote Sensing: From Pixels to Processes, Enschede, Netherlands, 8-11 May 2006. pp 493-496. Dias-da-Costa D, Valena J, Jlio E (2011) Laboratorial test monitoring applying photogrammetric post-processing procedures to surface displacements.
6. Concluses
Recentemente, surgiram mtodos inovadores para monitorizao de estruturas, usando a fotogrametria e processamento de
497
Measurement 44 (3):527-538. doi:10.1016/j.measurement.2010.11.014. [04] Valena J, Jlio E, Arajo H (2012) Application of photogrammetry to structural assessment. Experimental Techniques 36 (5). doi:10.1111/j.1747-1567.2011.00731.x. [05] Hoffman ME, Manevitz LM, Wong EK, Geers MGG, De Borst R, Brekelmans WAM (1996) Computing strain fields from discrete displacement fields in 2D-solids. International Journal of Solids and Structures 33 (29):4293-4307. doi:10.1016/0020-7683(95)00240-5 [06] Abdel-Quarter I, Abudayyeh O, Kelly M (2003) Analysis of edge detection techniques for crack identification in bridges. Journal of Computing in Civil Engineering 17 (3):255-263. doi:10.1061/~ASCE!0887-3801~2003!17:4~255! [07] Valena J, Dias-da-Costa D, Jlio ENBS (2012) Characterisation of concrete cracking during laboratorial tests using image processing. Construction and Building Materials 28 (1):607-615. doi:10.1016/j.conbuildmat.2011.08.082 [08] Hutchinson TC, Chen Z (2006) Improved Image Analysis for Evaluating Concrete Damage. Journal of Computing in Civil Engineering 20 (3):210-216 [09] CEN (2004) EN 1992-1-1: Eurocode 2: Design of Concrete Structures - Part 1-1: General Rules and Rules for Buildings, in: European Committee for Standardization (CEN). [10] Bachmann H (1967) Zur plastizittstheoretischen Berechnung statisch unbestimmter Stahlbetonbalken. Technische Hochschule, Zrich [11] Park R, Paulay T (1975) Reinforced concrete structures. Wiley, New York. [12] Buchaim R (2001) A influncia da no-linearidade fsica do concreto armado na rigidez flexo e na capacidade de rotao plstica. So Paulo University. [13] Barber CB, Dobkin DP, Huhdanpaa H (1996) The quickhull algorithm for convex hulls. ACM Transactions on Mathematical Software 22 (4):469-483. doi:10.1145/235815.235821. [14] Ballard D (1981) Generalizing the Hough Transform to Find Arbitrary Shapes. Pattern Recognition 13:111122 [15] Valena J, Dias-da-Costa D, Jlio E, Arajo H, Costa H (2012) Automatic crack monitoring using photogrammetry and image processing. Measurement (in press). doi:10.1016/j.measurement.2012.07.019. [16] Criminisi A, Reid I, Zisserman A (2000) Single view metrology. Int J Comput Vision 40 (2):123-148. doi:10.1023/a:1026598000963. [17] Carmo RNF, Costa H, Loureno C, Andrade D, Simes T (2012) Influence of both concrete strength and transverse confinement on bending behaviour of reinforced LWAC beams. Engineering Structures (in press). doi:10.1016/j.engstruct.2012.09.030. [18] EN 12390, Testing hardened concrete (2009). European Committee for Standardisation, Brussels, Belgium.
[19] [20]
Carmo RNF, Lopes SM (2005) Influence of the shear force and transverse reinforcement ratio on plastic rotation capacity. Structural Concrete, Journal of the fib (Thomas Telford Journal) 6 (3):107-117 CEB (1998) Ductility of reinforced concrete structures. Bulletin dInformation n 242. Lausanne, Switzerland.
498
Numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models Simulao numrica do comportamento da tenso de aderncia entre barras de ao e concreto autoadensvel e convencional utilizando modelos de vigas
F.M. ALMEIDA FILHO a almeidafilho@ufscar.br M. K. EL DEBS b mkdebs@sc.usp.br A.L.H.C. EL DEBS c analucia@sc.usp.br
Abstract
The present study evaluates the bond behavior between steel bars and concrete by means of a numerical analysis based on Finite Element Method. Results of a previously conducted experimental program on reinforced concrete beams subjected to monotonic loading are also presented. Two concrete types, self-compacting concrete and ordinary concrete, were considered in the study. Non-linear constitutive relations were used to represent concrete and steel in the proposed numerical model, aiming to reproduce the bond behavior observed in the tests. Experimental analysis showed similar results for the bond resistances of self-compacting and ordinary concrete, with self-compacting concrete presenting a better performance in some cases. The results given by the numerical modeling showed a good agreement with the tests for both types of concrete, especially in the pre-peak branch of the load vs. slip and load vs. displacement curves. As a consequence, the proposed numerical model could be used to estimate a reliable development length, allowing a possible reduction of the structure costs. Keywords: bond strength, self-compacting concrete, beam, numerical approach, steel-concrete interface.
Resumo
O presente estudo avalia o comportamento da aderncia entre barras de ao e concreto por meio de uma simulao numrica utilizando o mtodo dos elementos finitos. Os resultados de um estudo experimental anteriomente realizado de vigas em concreto armado submeticas a um carregamento monotnico so tambm apresentados. Foram utilizados nesta pesquisa dois tipos de concreto, sendo eles: o concreto auto-adensvel e o concreto convencional (vibrado). Foram utilizadas relaes constitutivas no-lineares para o concreto e o ao para representar o comportamento do modelo numrico proposto com o objetivo de simular o comportamento da aderncia dos ensaios. A anlise experimental mostrou resultados similares para a resistncia de aderncia no concreto auto-adensvel e no concreto convencional com os modelos em concreto auto-adensvel com um melhor comportamento em alguns casos. Os resultados fornecidos pela simulao numrica mostraram boa concordncia com os ensaios realizados para os dois tipos de concreto, especialmente na fase de pr-pico do diagram fora vs. deslizamento e fora vs. deslocamento.Finalmente, o modelo numrico proposto pode ser utilizado para estimar o comprimento de ancoragem mais adequado, permitindo uma possvel um reduo nos custos da estrutura. Palavras-chave: resistncia de aderncia, concreto auto-adensvel, viga, aproximao numrica, interface ao-concreto.
Department of Civil Engineering, So Carlos Federal University, Rod. Washington Lus, km 235, CEP 13565-905, So Carlos, SP Brazil Department of Structural Engineering, So Carlos Engineering School, So Paulo university Av. Trabalhador Socarlense, 400, CEP 13566-590, So Carlos, SP Brazil
a b
Received: 19 Oct 2012 Accepted: 17 Apr 2013 Available Online: 12 Jun 2013
2013 IBRACON
Numerical approach of the bond stress behavior of steel bars embedded in self-compacting concrete and in ordinary concrete using beam models
1. Introduction
Nowadays, the use of high-performance concrete is more and more frequent, due to its economy and versatility. The need of high reinforcement ratios to guarantee the material an appropriate ductility causes many difficulties in the cast operations, requiring extra care to assure the good quality of the structure. Self-compacting concrete (SCC) is an innovative construction material developed for civil constructions in the 1990s that can be the response for this problem. SCC is defined as a mixture that can be cast in any place of the formwork, just through the accommodation due to its own weight [1-2]. This new material is capable of flowing inside the formwork through the reinforcement, filling it out without any compacting equipment. As a result, increase of productivity, reduction of labor costs and improvement of overall quality of the structure can be obtained. [3]. The application of SCC is also expected to improve the flexural behavior of the elements due to its superior filling capability, since an increase of the bond resistance between reinforcement and concrete could indirectly benefit the confinement effects. For low strength concretes, SCC and OC (ordinary concrete) presented similar bond strength, with some peculiarities [4-7]. Particularly in places with high reinforcement rate the fresh properties of SCC surpass the OC [8], producing higher quality elements. For high strength concretes, similar results are expected, since the modulus of elasticity will increase proportionally for both types of concrete, and those are the main properties of concrete affecting the bond strength.
The bond between steel and concrete has been object of study from the middle of the XX century, since the interaction between steel and concrete is considered the main mechanism characterizing the reinforced concrete behavior. As stated before, the obtained bond strength depends on the steel bar and concrete properties, but steel the behavior is know well known. Therefore, the study of bond goes through the complete knowledge of the materials involved in the concrete production. If from the physical point-of-view steel-concrete bond is still not completely understood, the real behavior of the interface is very difficult to be represented by numerical models, being affected by a large number of variables. According to [9], the bond resistance can be divided in three portions. The first one is the adhesion, which consists of the shear resistance between concrete and steel; the second is the friction between surfaces, which is a decisive factor at the ultimate limit state; and the last one is the bearing action, caused by the deformation of the bars in contact with concrete. There are several types of failure associated to the loss of bond between concrete and the steel bar, and the main ones are pullout failure and splitting failure. These failures are strongly influenced by several factors, such as the type of reinforcement (bar, tendons and strings), surface characteristics (flat or rough), bar diameter, presence of confinement reinforcement, distance among the bars, concrete cover, steel bar stresses, concrete quality and others. In the pull-out test of a steel bar from a concrete prism the failure of the concrete occurs nearly the steel bar surface and the mechanism with pure slip would not be possible [10]. If a steel bar is
Figure 1 Beam geometry
Cross section G D 5 F
Dimensions < 16 mm > 16 mm A (cm) B (cm) C (cm) D (cm) E (cm) F (cm) G (cm) 37.5 65.0 18.0 13.0 32.5 10.0 3.0 56.0 110.0 24.0 19.0 51.0 15.0 4.0
Centroid of compressive force Hydraulic jack E Steel hinge
Spreader beam
D 5
LVDT
Bonded zone
LVDT
A
5
10
B
Longitudinal section
500
F.M. ALMEIDA FILHO | M. K. EL DEBS | A.L.H.C. EL DEBS
placed close to the concrete prism surface, the concrete splitting failure occurs. On the other hand, if no reinforcement is added to the concrete prism, the bond strength depends, almost totally, of the concrete compressive strength. According to the current technical literature, the beam test is more reliable to investigate bond behavior, since it reflects the behavior under pure flexure and considers the effects of tension cracks. Few researches were performed using beam models with selfcompacting concrete, being this absence of data a motivation for this research. According to the literature, SCC appears to improve the bond strength, due to its filling ability to involve the reinforcement. This improvement was not significant, but the experimental data proved the use of self-compacting concrete possess, at least, the same behavior of similar models made with ordinary concrete [7, 11].
of reinforced concrete structures. The evaluation of the bond strength of reinforcing bars in structural concrete elements is important to evaluate the feasibility of placement self-compacting concrete. There is some experimental data about the bond behavior comparing self-compacting concrete and ordinary concrete using beam tests, but there is few data related to numerical simulations about the bond phenomena. The importance of this research lies on these two aspects, where it aims to contribute.
2. Summary of the experimental program

The experimental program was part of a wider research on the bond behavior on self-compacting concrete. The beam geometry was based in the model established by [12]. Figure 1 shows the geometry of the adopted beam. The position and the inclination of the bars during the casting had a significant influence on the bond resistance and the specimens cast in the vertical direction presented larger bond resistance than the models cast horizontally. The monotonic displacement applied varied with the bar diameter; so, for 10 mm steel bar, the displacement rate was 0.01 mm/s and for the 16 mm steel bar, the displacement rate was 0.016 mm/s, until failure. The deformation rates were calculated dividing the steel bar diameter by 1000 achieving similar results as Rilem recommendations [7]. Figure 2 shows the test set-up for the beam specimens. The data from these LVDT were used to determine the slip of the steel bar. Electrical resistance strain gauges were placed in five points of the
1.1 Justification
In this paper, the behavior of beams specimens was studied through experimental tests and numerical analysis using the finite elements method. The main objective was evaluating the load vs. displacement and load vs. slip behavior and the bond stress, regarding the influence of the numerical parameters involved in the used software. The secondary objective was to compare, by numerical models, the bond behavior of self-compacting concrete (SCC) with ordinary concrete (OC), and show the bond stress distribution on the contact surface. This research has special significance in construction practice
Figure 2 Beam test set-up

Piston support
Instron's piston Load direction Load cell Instron's hinge Steel hinge Beam specimen Spreader beam
LVDT
LVDT
Support beam Instron's plate support Reaction slab
501
Table 1 Materials content and fresh results for SCC series

Material Cement (kg) Sand (kg) Gravel (kg) Water (kg) Superplasticizer (%) Filler (kg) Silica fume (kg) OC1 365.3 883.9 942.3 260.8 OC2 488.3 766.6 942.4 227.0 SCC1 338.8 854.8 919.1 273.6 0.4% 101.6 SCC2 365.1 815.3 876.7 146.1 0.75% 146.1 36.5 Tests Slump test Slump flow (cm) T50 (s) L-Box test T60 (s) RB V-Funnel Tv (s) SCC1 67.5 1.0 1.0 0.95 1.5 SCC2 61.0 1.0 1.0 0.9 2.0
steel bar, near the bonded region and in the middle of the bar, as shown in Figure 3. According to the technical literature, the application of strain gages on the steel-concrete interface must be avoided due to its high influence on the bond stresses; however, some researches good estimation of the bond stress distribution with this procedure [13-14]. So, the strain gages where placed at the beginning and at the end of the embedment length to measure the strain variation, and an additional strain gage was placed in the middle of the bar (Figure 3). The used cement was Ciminas CP-V Ari Plus (initial high strength cement). The used siliceous sand had density of 2.63 kg/dm3 and absorption of 4.0% and the used crushed gravel had density of 2.83 kg/dm3 and absorption of 1.71%. The used superplasticizer was based on carboxylate, with density of 1.1 kg/dm3 and 20% of solid content. Table 1 shows the materials contents and the results for fresh SCC. Table 2 shows the hardened properties of SCC series and OC series. Figure 4 shows some of the specimens, a beam cast with OC with 10 mm steel bar and a beam cast with SCC with 16 mm steel bar the beams, during tests at the universal test machine, Instron . According to the experimental results, the failure was different
for each series, according to the concrete compressive strength. So, for SCC1 and OC1 series (fc = 30 MPa), the failure was characterized by high slip of the steel bar with low displacement of the beam and, and for SCC2 and OC2 series (fc = 60 MPa), the failure was characterized by the yield of the steel bar with low slip and the beam presenting high vertical displacement.
3. Numerical approach
The failure in steel-concrete interface could be attained by combining Coulombs frictional hypothesis with a bound for the maximum tensile stress, resulting in two different failure modes that could be called sliding failure and separation failure [10]. The sliding failure is assumed to occur in a section when the shear stress exceeds the sliding resistance and should be determined by two parameters: the cohesion (c) and the friction coefficient ( m). In previous studies [15-18] the variation of the frictional coefficient and the cohesion seemed not to affect the general response of the bond in the contact surface. However, the number of elements in the contact surface, and parameters like FKN (normal contact stiffness factor), FKT (tangent contact stiffness factor) and IT (iteration number), presented in Ansys software, affect directly the load vs. slip behavior, according to the adopted bond model [15-16]. But, it is worth to mention that the differences between the two materials (SCC and OC) are restricted to the materials properties.
Figure 3 Steel bar instrumentation
10 mm steel bar 13,75 10 16,25 16,25 10 SG Unbonded zone SG SG
13,75 SG
3.1 Materials
Compressive strength and elasticity modulus of concrete were ob-
Bonded zone SG - Strain Gage 25,0 16 SG 22,0 SG
Table 2 Hardened results for SCC and OC series

Hardened properties fc (MPa) Ec (MPa) fct (MPa) OC1 32.02 27.24 2.182 OC2 50.20 34.30 3.920 SCC1 30.10 27.87 2.450 SCC2 53.30 36.68 4.990
16 mm steel bar 22,0 16 25,0 SG Unbonded zone SG
Bonded zone
502
Figure 4 Beams during tests
tained by tests in cylindrical specimens (10x20cm). Figure 5 shows the experimental behavior of SCC and OC, for each series, and the steel bar behavior assumed in numerical study. As shown at Figure 5, both concretes behavior were practically the same. However, there was an absence of the descending branch of the post-peak of its behavior, which could be achieved by using Popovics formulation [19], shown below (Eq. 1 to 3).
(3)
This formulation takes into account the variation of the concrete compressive strength in the post-peak branch. According to Popovics theory, the relation between the initial modulus of elasticity (Ec) and the secant modulus of elasticity (Ecs) can vary until 4.0 for normal strength concretes and in 1.3 for high strength concretes.
(1)
3.2 Mesh, load and finite elements
(2)
Figure 6 shows the used mesh for the numerical models; due to the symmetry, only a quarter of the beam model was studied. Experimental investigation of the bond stress response was
Figure 5 Stress vs. strain behavior of steel and concrete

35 30 25
70 60 50
700 600
S tr es s (M P a )
St re s s ( M P a )
20 15 10 5 0 -5 -1 0 1 2
OC1 SCC1 Popovics (1973)
40 30 20 10 0
c c
S tr es s (M P a )
500 400 300 200 100 10 mm steel bar 16 mm steel bar 0 2 4 6 8 10 E s (10 mm) = 207.05 GPa E s (16 mm) = 209.18 GPa
OC2 SCC2 E (CAA) = 32,73 GPa E (CC) = 32,61 GPa
-10
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
Strain ()
Strain ()
Strain ()
503
Figure 6 Numerical mesh adopted
A Beam model with 10 mm steel bar. Concrete elements and full model
Beam model with 16 mm steel bar. Concrete elements and full model
Figure 7 Considerations for the numerical model
Beam Hinge Steel bar

performed using the specimen geometry shown in Figure 1. The specimens consisted of a steel bar with 10 and 16 mm of nominal diameter, anchored in 100 and 160 mm embedded length in the concrete beam, respectively. The roughness (including the steel bar ribs) of the steel bar was not considered and a plain contact surface was adopted in the
numerical study. This assumption was made based on previous researches that shown the influence of the steel bar ribs were not significant, due the use of the FKN (normal contact stiffness factor) and FKT (tangent contact stiffness factor) parameters, according to the adopted bond model [15-16]. The finite elements used on the mesh were: for concrete elements, Solid65; for steel elements, Solid45; for contact surface, Conta174 and Targe170 [20]. Figure 7 shows the constraints for the numerical model. The load was applied according to the maximum displacement measured on the test at a constant rate. The mesh was made to allow the same point of application of the load as was applied on the test. Table 4 shows the tests results and the values used for the numerical approach (FKN, FKT and D). Figure 8 and Figure 9 shows the numerical behavior compared with the test results. Table 5 shows the comparison between the numerical and experimental results. According to Table 5, there was good approach of the numerical and experimental results. The slip of the numerical model was less accurate than the displacement prevision for both steel bar models. The beam model with 10 mm steel bar was better represented by the numerical model than the model with 16 mm steel bar. This could be caused by the adopted mesh, which could reduce the accuracy of the numerical approach.
Table 3 Specimens division for tests

Model B-SCC-C30-B10 B-SCC-C30-B16 B-OC-C30-B10 B-OC-C30-B16 B-SCC-C60-B10 B-SCC-C60-B16 B-OC-C60-B10 B-OC-C60-B16 Concrete type SCC SCC OC OC SCC SCC OC OC Concrete compressive strength 30 MPa Bar diameter 10 mm 16 mm 10 mm 16 mm 10 mm 16 mm 10 mm 16 mm
60 MPa
504
Table 4 Specimens division for tests

Model B-SCC-C30-B10 B-SCC-C30-B16 B-SCC-C60-B10 B-SCC-C60-B16 B-OC-C30-B10 B-OC-C30-B16 B-OC-C60-B10 B-OC-C60-B16 Pu (kN) 32.66 61.99 42.35 92.48 33.49 70.77 41.58 90.84 u (mm) 3.97 6.59 27.08 40.96 3.82 7.32 29.87 42.95 u (MPa) 13.00 11.57 16.86 17.25 13.33 13.20 16.55 16.95 su (mm) 0.398 0.938 0.096 0.215 0.295 0.758 0.068 0.660 FKN 3 40 0.0001 0.0001 3 40 0.0001 0.0001 FKT 1/0.15 1 1 1 1/0.15 1 1 1 D (mm) 12.0 18.0 31.0 50.0 12.0 18.0 31.0 50.0
4. Analysis and discussion

The numerical model had the same parameters of the experimental one, like modulus of elasticity and the applied load, justifying the development of one model for each series. In the contact evaluation, the numerical results were compared with those given by Eq. 4 [12].
Figure 10 shows the measurement points adopted for the bond stress evaluation. The evaluated stresses were obtained by the contact elements on the interface and by the concrete elements under the contact surface.
4.1 SCC1 and OC1 series

Figure 11 shows the variation of the bond stresses at the steelconcrete interface during the substep of the failure load. Figure 12 shows the principal stresses in the normal direction of the cross section for the numerical beam model. Also, the detail
u s 40
and
P s k u As
(4)
Figure 8 Numerical approaches of the experimental tests for SCC1 and OC1
505
Figure 9 Numerical approaches of the experimental tests for SCC2 and OC2
of each beam without the steel bar and the steel hinge are shown. According to the numerical results, the variation of the stresses at the steel bar is noticed, mainly for the beam with the 16 mm steel bar. Also, the stress at the steel-concrete interface shows a gradual variation, beginning at the start of the de-
velopment length until the other edge. The stresses shown in Figure 12 are in kN/cm 2 and the positive sign indicates tension. Figure 13 shows the stress variation from the concrete and contact elements.
Table 5 Comparison between numerical and experimental results
506
4.2 SCC2 and OC2 series Figure 10 Measurement points on the bonded zone and on the steel bar
Figure 15 shows the variation of the bond stress at the steel-concrete interface during the substep of the failure load. Figure 16 shows the principal stresses at the cross section normal direction for beam numerical models. Also, the details of each beam without the steel bar and the steel hinge are shown. According to the numerical results, the variation of the stresses at the steel bar can be observed, for both cases. Also, the stress at the steel-concrete interface shows a gradual variation, from the start of the development length until the other edge. Figure 17 shows the stress variation of the concrete and contact elements. According to Figure 17, for both models with 10 and 16 mm, the general behavior was the same. The concrete elements presented the same behavior for the points 1 and 17, while the point 9 presented significant differences in its behavior. For the contact elements, it appears that only the point 1 resists to the steel bar sliding, because, for the other points, the calculated stresses were around zero. Figure 18 shows the comparison between the stresses from the tests results for the strain gages placed on the steel bar and the numerical results. As expected, due to the high strength of the concrete, the results of the strain gages showed that the strains at the steel bar were non-linear for both models, showing that the numerical results led to a satisfactory approach (Figure 18).
Steel bar points of measurement
P 1 5 P 11 P 17 P 6 P 9
P 2
P3 Bonded zone points of measurement P 1 10 mm steel bar P 1 16 mm steel bar
According to Figure 13, the stress behavior on the contact surface and in the concrete elements was almost the same. The point 1 presented the same behavior for both cases while the points 9 and 17 presented a significant difference, showing that the concrete elements have a gradual transfer of stresses, because the stress at point 9 is higher than at point 17. Also, the stress calculated by the contact elements at point 17 was higher than the stress on point 9. Figure 14 shows the comparison between the stresses from the test results of the strain gages placed on the steel bar and the numerical results. As expected, according to Figure 14, the test results from the strain gages showed that the strains at the steel bar were linear for both models, showing that the numerical result led to a satisfactory approach.
4.3 Analysis of the bond behavior between the series

According to the numerical results, the bond stress measured by contact and concrete elements assumed similar values for normal concrete compressive strength, while for high strength concrete there was high difference, mainly provoked by the contact elements. According to Figure 11, the bond stress shows that there was a decreasing of its magnitude along the development length, but the
Figure 11 Stress distribution for the beam models

0 0
-5
-5
16 mm steel bar
S tr es s (c o nc re te ) (M P a )
-10 -15 -20 -25 -30

Concrete elements Contact elements
-10 -15 -20 -25 -30
St re s s (c o n cr et e) ( M P a )
10 mm steel bar
-10 -15 -20 -25 -30
-10 -15 -20 -25 -30
Concrete elements Contact elements
10
12
Measurement points
10
12
14
16
18
Measurement points
S t re s s ( co nt a c t) ( M P a )
St re s s ( co n ta c t) ( M P a )
-5
-5
507
Figure 12 Principal stresses of the numerical beam models
Figure 13 Stress variation for the beam models
508
Figure 14 Comparison between the test results from the strain gages and the numerical result
measured value showed a minimal stress presented in the steel bar surface. This assumption shows that the beginning of the development length was being used resisting the slip, while the other points on surface help only with a minimal amount of stress, distributed in it. For the high strength concrete, both numerical models (with 10 and 16 mm steel bars) showed similar behavior, also observed in the previous series. However, at failure an increase of bond stress at the end of the development length was observed. This behavior may be explained by the nature of the models failure, occurred by the steel bar yielding.
5. Conclusions
The presented paper describes the numerical and experimental investigation performed to evaluate the bond strength. Beam models based on the Rilem recommendation were used, comparing ordinary concrete and self-compacting concrete of same compressive strength. The numerical approach was based on finite element method, using Ansys software. According to the results, the following conclusions can be made: 1. The beam models with self-compacting concrete and ordinary concrete produced similar results, with a small advantage for the ordinary concrete;
Figure 15 Stress distribution on the steel-concrete interface for the beam models
0 0
S tr es s (c o nt a ct ) (M P a )
S tr es s (c o n cr ete ) (M P a )
-20 -30 -40 -50
-10 -15 -20 -25 -30
S tr es s (c o n cr et e) ( M P a )
-10 -15 -20 -25 -30
-10
10 mm steel bar Concrete elements Contact elements
-60 -70 -80 10 12
16 mm steel bar Concrete elements Contact elements
-20
10
12
14
16
18
-30
Measurement points
Measurement points
S tr es s (c o n ta c t) (M P a )
-5
-10
-5
509
Figure 16 Principal stresses of the numerical beam models
Figure 17 Stress variation for the beam models
510
2. The numerical models presented good approach with the test results, mainly for the failure load and for the displacement values; however, the slip results could not be well represented. The differences between the measured slip compared to the numerical results reached almost 54% (B-OC-C30-B10); 3. The stress vs. strain behavior of the steel bar was well represented by the numerical approach, giving reliable results and ensuring the numerical model can represent the test; 4. According to the stress distribution of the steel-concrete interface, both analyzed elements types (contact elements and concrete elements under the contact surface) showed similar results. However, the analysis of the variation of the stress during the test showed a better behavior when considering concrete elements, for both concrete compressive strength. 5. The strain pattern measured by test specimens showed the main strain values in the middle of the steel bar and in the position right before the embedment length. The numerical models showed the same behavior, but with inferior values that those obtained by test specimens. This may be explained by the yield limit established for the steel bar, which reduced the strain of the steel bar in the numerical models. Finally, the utilization of numerical models to represent the bond behavior in a beam test, using ordinary concrete (OC) and selfcompacting concrete (SCC), presented a good approach, showing that the concrete type did not affect the bond response, since the materials properties were similar. Also, according to the results,
the adopted parameters could be extended for others models with different compressive strength and other bar diameters.
6. Notation
t = Bond stress, MPa; tu = Bond stress at the failure load, MPa; Pu = Failure load, kN; k = Assumes 1.25 for f < 16 mm and 1.50 for f 16 mm; ss = Steel bar stress, MPa; ld = Development length, mm; D = Displacement applied by the piston during test, mm; fs = Steel bar diameter, mm; fc = Concrete compressive strength, MPa; su = Slip at the failure load, mm; du = Maximum beam vertical displacement, mm; l = Experimental vs. numerical ratio; fo = Cylinder concrete compressive strength, MPa; e = Strain caused by the fc concrete stress, ; eo = Strain at cylinder concrete failure, ; FKN = Normal contact stiffness factor; FKT = Tangent contact stiffness factor.
7. Acknowledges
To CAPES, CNPq and FAPESP for the financial support. Also, the technical staff at the Structures Laboratory, and to the com-
Figure 18 Comparison between the test results from the strain gages and the numerical result
511
panies Elkem, Holcim, Grace Brasil and Brasil Minas S/A for the material donation.
8. Reference
[01] Bartos, P.J.M. 2000. Measurement of key properties of fresh self-compacting concrete. Proceedings of Measurement, testing and standardization: future needs in the field of construction materials, Paris, 5-6 June, 6 p. [02] Okamura, H. 1997. Self-compacting high-performance concrete. Concrete International, 19 (7): 50-54. [03] Gomes, P.C.C. 2002. Optimization and characterization of high-strength self-compacting concrete. PhD. PhD Thesis, Universitat Politcnica de Catalunya, Escola Tcnica Superior DEnginyers de Camins, Canals i Ports de Barcelona, 140 p., Barcelona, Spain. [04] Dehn F., Holshemacher K., and Weie, D. 2000. Self-compacting concrete (SCC) time development of the material properties and the bond behavior. Proceedings of LACER, No. 5, 10 p. [05] Holshemacher, K., Dehn, F., and Weie, D. 2002. Bond in high-strength concrete influence of rebar position. Proceedings of the 6th international symposium on utilization of high strength / high-performance concrete, Leipzig, Germany, pp.289-298. [06] Almeida Filho, F.M., De Nardin, S., and El Debs, A.L.H.C. 2005. Evaluation of the bond strength of self-compacting concrete in pull-out tests. Proceedings of the Second North American Conference on the Design and Use of Self-Consolidating Concrete and Fourth International RILEM Symposium on Self-Compacting Concrete, Chicago, Illinois. [07] Almeida Filho, F.M. 2006. Contribution to the study of the bond between steel bars and self-compacting concrete. PhD thesis, So Paulo University, So Carlos, 1-308. (in Portuguese) [08] Chan, Y.-W., Chen, Y.-S., and Liu, Y.-S. 2003. Effect of consolidation on bond of reinforcement in concrete of different workabilities. ACI Structural Journal, 100(4): 294-301. [09] Bangash, M.Y.H. 1989. Concrete and concrete structures: numerical modeling and applications. Barking: Elsevier science publishers Ltd., 687p. [10] Nielsen, M.P. 1998. Limit analysis and concrete plasticity. CRC Press, 908p. [11] Isa, M.M., and Aguado, A. 2003. Comportamento da aderncia entre o concreto auto-adensvel de alta resistncia e a armadura. Proceedings of the V Simpsio EPUSP sobre Estruturas de Concreto, So Paulo-SP, Brazil (in portuguese). [12] Rilem-Fip-Ceb, 1973. Bond test for reinforcing steel: 1-Beam test (7-II-28 D). 2-Pullout test (7-II-128): Tentative recommendations. Materials and Structures, 6(32): 96-105. [13] Hamza, A.M., and Naaman, A.E. 1996. Bond characteristics of deformed reinforcing steel bars embedded in SIFCON. ACI Materials Journal, 93(6):578-588
[14] [15] [16] [17] [18] [19] [20]
Mirza, S.M., and Houde, J. 1979. Study of the bond stress-slip relationship in reinforced concrete. ACI Journal, 76(1):19-45. Almeida Filho, F.M., De Nardin, S., and El Debs, A.L.H.C. 2004. Interface ao-concreto: anlise dos parmetros do contato e sua influncia na simulao via MEF. Proceedings of the XXV CILAMCE: Iberian Latin American Congress on Computational Methods, Recife-PE, Brazil (in portuguese). De Nardin, S., Almeida Filho, F.M., and El Debs, A.L.H.C. 2005a. Non-linear analysis of the bond strength behavior on the steel-concrete interface by numerical models and pull-out tests. Proceedings of the ASCE Conference Structures 2005, Vol. 171, New York, USA. De Nardin, S., Almeida Filho, F.M., El Debs, A.L.H.C., and El Debs, M.K. 2005b. Steel-concrete interface: influence of contact parameters. Proceedings of the FIB International Conference: Keep concrete attractive, Budapest, Hungary, 6p. Haach, V.G., Oliveira Filho, J., Almeida Filho, F.M., De Nardin, S., and El Debs, A.L.H.C. 2004. Simulao numrica comparativa do comportamento da aderncia em modelos tridimensionais e axissimtricos de arrancamento. Proceedings of the XXV CILAMCE: Iberian Latin American Congress on Computational Methods, Recife-PE, Brazil (in portuguese). Popovics, S. 1973. A numerical approach to the complete stress-strain curves for concrete. Cement and concrete research, 3(5): 583-599. Ansys. 2002. Manuals on-line.
512

Riem

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Riem

Încărcat de

Drepturi de autor:

Formate disponibile

REVISTA IBRACON DE ESTRUTURAS E MATERIAIS

IBRACON STRUCTURES AND MATERIALS JOURNAL

Ville San Marino Victor Schwaner

Volume 6, Number 3 June 2013 ISSN: 1983-4195

Aims and Scope

Acesso via Internet

Diretoria Binio 2011/2013

Conselho Diretor Binio 2011/2013

Volume 6, Number 3 June 2013 ISSN: 1983-4195

Volume 6, Number 3 (June 2013) p. 360-374 ISSN 1983-4195

S. J. P. J. MARQUES FILHO a sergiopriorifilho@hotmail.com B. HOROWITZ b horowitz@ufpe.br

S. J. P. J. MARQUES FILHO | B. HOROWITZ

Figure 1 Bending moment diagram of a laterally loaded frame

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

2. Impact of joint flexibility

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

Figure 7 Frame joint subjected to a lateral load, VC

3.1 Flexibility of complete frame joints

Figure 8 (a) Adjusted rigid links model; (b) scissors model

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

Substituting VN in Equation (2) for its value in Equation (1), on gets:

3.4 Parametric study of complete frame joints

3.2 Adjusted rigid links model

Figure 9 Distribution of shear stresses in the interior of the joint

3.3 Scissors model

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

Figure 10 Variation of the parameter for 3 meters story height

Figure 12 Finite element anti-symmetric model for L-type joints

Figure 13 Deformed configuration of the complete cross-type joint

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

4.2 Parametric study of concentric joints

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

5.1.1 Cross-type interior joints

Figure 18 Experimental setup for the cross-type joint [13]

5. Results and discussion

Figure 19 Experimental setup for the T-Lateral exterior joint [14]

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

Figure 20 Experimental setup for L-type joint [15]

5.2 Multi-story frames

Table 3 Comparison of actuator force with numerical results L-type joint

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints

Figure 22 Applied loads for frame model 1

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

S. J. P. J. MARQUES FILHO | B. HOROWITZ

Table 4 Geometry of analyzed frames

Column section (cm) 20x40 20x60

Column section (cm) 100x30 130x30 100x20 100x30

Table 5 Results os second order analyses

IBRACON Structures and Materials Journal 2013 vol. 6 n 3

Flexibility modeling of reinforced concrete concentric frame joints