05 Structural and Vibration Guidelines For Datacom Equipment Centers

This file is licensed to sergio torres morales (retos@correoinfinitum.com).
License Date: 9-26-2012

© 2007, American Society of Heating, refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). For personal use only.
Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAE’s prior written permission.
Structural and Vibration

Guidelines for
Datacom Equipment Centers
This file is licensed to sergio torres morales (retos@correoinfinitum.com). License Date: 9-26-2012
This publication was prepared in cooperation with TC 9.9, Mission Critical Facilities,
Technology Spaces, and Electronic Equipment.
Any updates/errata to this publication will be posted on the

ASHRAE Web site at www.ashrae.org/publicationupdates.
Structural and Vibration

Guidelines for
Datacom Equipment Centers
American Society of Heating, Refrigerating

and Air-Conditioning Engineers, Inc.
ISBN: 978-1-933742-20-5
©2007 American Society of Heating, Refrigerating

and Air-Conditioning Engineers, Inc.
1791 Tullie Circle, NE
Atlanta, GA 30329
www.ashrae.org
All rights reserved.
Printed in the United States of America
Printed on 10% post-consumer waste using soy-based inks.

Cover design by J. Lombardo.
ASHRAE has compiled this publication with care, but ASHRAE has not investigated, and ASHRAE
expressly disclaims any duty to investigate, any product, service, process, procedure, design, or the like
that may be described herein. The appearance of any technical data or editorial material in this publi-
cation does not constitute endorsement, warranty, or guaranty by ASHRAE of any product, service,
process, procedure, design, or the like. ASHRAE does not warrant that the information in the publi-
cation is free of errors, and ASHRAE does not necessarily agree with any statement or opinion in this
publication. The entire risk of the use of any information in this publication is assumed by the user.
No part of this book may be reproduced without permission in writing from ASHRAE, except by a
reviewer who may quote brief passages or reproduce illustrations in a review with appropriate credit;
nor may any part of this book be reproduced, stored in a retrieval system, or transmitted in any way
or by any means—electronic, photocopying, recording, or other—without permission in writing from
ASHRAE.
____________________________________________
Library of Congress Cataloging-in-Publication Data
Structural and vibration guidelines for datacom equipment centers.
p. cm. -- (ASHRAE datacom series)
Summary: "Provides basics for addressing the structural and vibration performance (resistance against wind,
snow, and seismic assault) of datacom equipment centers, focusing on the facility (envelope, beams, columns,
floor slabs, roof slabs), the building infrastructure (power, cooling, flooring, ceiling systems), and the datacom
equipment (servers, storage, tape drives, racks, network equipment)"--Provided by publisher.
Includes bibliographical references and index.
ISBN 978-1-933742-20-5 (softcover)
1. Structural design. 2. Data processing service centers--Design and construction. 3. Buildings--Vibration. 4.
Earthquake resistant design. 5. Data processing service centers--Protection. 6. Electronic digital computers--
Protection. I. American Society of Heating, Refrigerating and Air-Conditioning Engineers.
TH1094.S77 2007
725'.23--dc22
2007046368
ASHRAE STAFF
SPECIAL PUBLICATIONS PUBLISHING SERVICES
Christina Helms David Soltis

Editor Manager
Cindy Sheffield Michaels Tracy Becker
Associate Editor Graphic Applications Specialist
Michshell Phillips Jayne Jackson
Administrative Assistant Publication Traffic Administrator
PUBLISHER
W. Stephen Comstock
Contents
Acknowledgments................................................................................... ix
Part I—Introduction and Best Practices
Chapter 1 Introduction......................................................................... 3
1.1 Overview of this Book............................................................ 3
1.2 Overview of the Datacom Industry ........................................ 5
1.3 Overview of ASHRAE Technical Committee 9.9 ................... 6
1.4 Overview of the ASHRAE Datacom Series ........................... 6
1.5 Document Flow ..................................................................... 7
1.6 Primary Users for This Document ......................................... 7
Chapter 2 Best Practices..................................................................... 9
2.2 Building Structures—New Buildings and Additions ............... 9
2.3 Building Structures—Renovations,
Relocations, and Changes .................................................. 10
2.4 Building Infrastructure ......................................................... 11
2.5 Datacom Equipment............................................................ 11
Part II—Building Structure
Chapter 3 Building Structures Overview ......................................... 15

3.1 Introduction.......................................................................... 15
3.2 Some Basics ....................................................................... 15
3.3 Submissions to Agencies .................................................... 18
3.4 Basic Definitions.................................................................. 19
vi ⏐ Contents
Chapter 4 New Structures .................................................................21

4.1 Initial Investigation ...............................................................21
4.2 Coordination ........................................................................22
4.3 Establish Design Criteria .....................................................22
Chapter 5 Existing Structures and Additions..................................29
5.1 Initial Investigation ...............................................................29
5.2 Coordination of New with Existing Structures......................30
5.3 New Components ................................................................33
5.4 Reinforcement of Existing Structure ....................................34
Chapter 6 Types of Building Structures...........................................37
6.1 Overview .............................................................................37
6.2 Pre-Engineered Metal Buildings. .........................................37
6.3 Braced Frame or Shear Wall-Type Structures. ....................38
6.4 Moment Resisting Frame Systems......................................41
6.5 Combination of Framing Systems........................................41
Part III—Building Infrastructure

Chapter 7 Building Infrastructure Overview ....................................45
7.1 Overview..............................................................................45
7.2 Interior Building Infrastructure .............................................45
7.3 Exterior Building Infrastructure ............................................47
Chapter 8 Structural Considerations for Infrastructure .................55
8.1 Overview..............................................................................55
8.2 Support................................................................................55
8.3 Anchoring ............................................................................57
8.4 Infrastructure Expansion/Contraction ..................................59
Chapter 9 Raised-Access Floor Systems ........................................61
9.1 Overview..............................................................................61
9.2 Raised-Access Floor Components ......................................61
9.3 Raised-Access Floor Structure Design Guidelines..............63
9.4 Seismic Performance of Raised-Access Floors...................66
Chapter 10 Vibration Sources and Control........................................71
10.1 Overview of Vibration Sources ............................................71
10.2 Overview of Vibration Isolation ............................................71
10.3 Selection of Vibration Isolators ............................................73
Structural and Vibration Guidelines for Datacom Equipment Centers ⏐ vii
10.4 Vibration Isolation within the

Datacom Equipment Room ................................................. 73
10.5 Vibration Isolation Adjacent to the
Datacom Equipment Room ................................................. 74
Part IV—Datacom Equipment

Chapter 11 Shock and Vibration Testing
on Datacom Equipment ................................................... 79
11.1 Basic Definitions.................................................................. 79
11.2 Overview of Vibration Sources ............................................ 80
11.3 Datacom Equipment Shock and Vibration Testing .............. 83
11.4 Shock and Vibration Test Guidelines for
Datacom Infrastructure and Cooling Equipment ................. 89
Chapter 12 Seismic Anchorage of Datacom Equipment .................. 95
12.1 Overview ............................................................................ 95
12.2 Nonstructural Seismic Provisions of Building Codes........... 95
12.3 Seismicity in the United States............................................ 96
12.4 Seismic Design Category .................................................... 97
12.5 Properly Applying Anchorage Forces .................................. 97
12.6 Protection of Server Cabinets
on Raised-Access Floors .................................................... 98
Chapter 13 Analysis of Datacom Equipment and
Seismic Anchorage Systems ........................................ 107
13.1 Overview ......................................................................... 107
13.2 Basic Definitions................................................................ 107
13.3 Datacom Equipment Frame............................................... 108
13.4 Finite Element Model Construction and Validation............ 108
13.5 Evaluation of Earthquake Anchorage Systems ................. 110
13.6 Evaluation of Structural Add-On Supports ........................ 113
References and Bibliography.............................................................. 117
Appendix A Codes for Building Structures and
Structural Components.................................................. 121
Appendix B Weights of Materials....................................................... 125
Appendix C Raised-Access Floor Calculations................................ 133
C.1 Raised-Access Floor Panel Loading Definitions................ 133
C.2 Floor Loading .................................................................... 134
viii ⏐ Contents
C.3 Floor Loading Calculations ................................................135

C.4 Structural Guidelines for
Raised-Access Floor Systems...........................................137
Appendix D Data Center Vibration Measurement.............................141
D.1 Overview ...........................................................................141
D.2 Introduction........................................................................141
D.3 Typical Operational Vibration and Shock Testing...............142
D.4 Typical Operational Vibration and Shock
Magnitude Recorded in a Data Center ..............................144
D.5 Monitoring Floor Vibration in Data Centers .......................145
D.6 Best Practices....................................................................146
Appendix E Component Anchorage Forces .....................................149
E.1 Relative Location Factor (1 + 2z/h)....................................149
E.2 Component Importance Factor (Ip)....................................150
E.3 Component Amplification Factor (ap).................................150
E.4 Component Response Modification Factor (Rp) ................151
Glossary of Terms ................................................................................153
Index ......................................................................................................163
Acknowledgments
The information in this guide was produced with the help and support of the
corporations listed below:
AT&T Services, Inc. Intel Corporation

Citigroup JG Pierson
Corgan Associates
Starzer Brady Fagan Associates
Data Aire, Inc.
DLB Associates Tate Access Floors, Inc
IBM Corporation VMC Group
ASHRAE TC9.9 wishes to particularly thank the following people:
• Dr. Roger Schmidt, Dr. Budy Notohardjono, John Quick, and Shawn
Canfield of IBM, Don Beaty and Dan Dyer of DLB Associates Consulting
Engineers, Jeff Trower of Data Aire Inc., Jeffrey Soulages of Intel, Richard
C. Berger of VMC Group, Larry Wong of AT&T Services, Inc., and Bill
Perry of Tate Access Floors, Inc., for their participation as chapter leads,
which included numerous calls, writing, and review.
• Dr. Budy Notohardjono for his leadership in coordinating and leading the
overall development of this book.
• Don Beaty of DLB Associates Consulting Engineers and Dr. Roger Schmidt
of IBM, the past and present chairmen of TC 9.9, for their leadership, vision,
and encouragement for this book.
• Don Beaty of DLB Associates for his leadership on the building structure and
infrastructure portions of the book.
• Steve Felton and Tom Davidson of DLB Associates, Hubert Starzer and Tony
Fagan of Starzer Brady Fagan Associates, Bob Morris of Corgan Associates,
Jerry Estoup of JG Pierson, Elizabeth O’Neill of VMC Group, and Jack Glass
of Citigroup for their participation and improvements to the final document.
ix
Part I
Introduction and
Best Practices
Introduction
1.1 OVERVIEW OF THIS BOOK
Today’s datacom (data and communications) facility managers and operators
understand the importance of protecting their business’ critical data and information
technology (IT) equipment (or datacom equipment), which includes servers as well
as storage, communications, and networking equipment. Data center operators must
therefore implement standards and practices for ensuring the integrity and function-
ality of the equipment within the datacom environment.
High-performance data center facilities accommodate a variety of complex and
sensitive datacom equipment that is vulnerable to internal and external sources of
shock and vibration. Shock and vibration sources are unwanted forces found at some
level in most datacom facilities that over time can degrade both facilities and equip-
ment. The datacom equipment and infrastructure equipment can themselves be
vibration sources within the data center. Datacom equipment manufacturers can
control these vibration sources by reducing the internal vibrations that are transmit-
ted to the surroundings. External sources such as airports, trains, nearby mining
operations (quarry blasting), construction activities, and earthquake and weather
events are other sources of operational shock and vibration. These shock and vibra-
tion sources are transferred through the building structure to the data center and
finally to any operating server and supporting infrastructure equipment. The effect
of these disturbances on the function of the IT and telecommunications equipment
depends on the design or the robustness of the equipment itself. The best practice to
reduce and circumvent the potential disruptive effects of these shock and vibration
sources is to eliminate, minimize, and control the input shock and vibration levels.
The best place to manage shock and vibration is at the originating source, which may
not always be possible.
For users, manufacturers, and installers of servers and storage, telecommuni-
cation, and networking equipment, this publication provides design information and
criteria to allow for continuous data center equipment operation during vibration or
seismic events.
Datacom equipment centers require a focus on the structural and vibration
performance of the facility, the building infrastructure, and the contents (e.g., data-
4⏐ Introduction
com equipment). As the datacom equipment density continues to increase (compac-

tion), the requirements of the datacom facility continue to evolve because:
• the power and cooling infrastructure becomes larger, heavier, and inherently
more structurally challenging
• the datacom equipment itself becomes heavier
• the facility is able to house more datacom equipment and therefore becomes
increasingly important to the owner, which may result in the need for
increased structural resistance against potential threats of high wind, snow,
and seismic and physical assault
In order to maintain high levels of resilience and availability, it is critical to
view these topic in a holistic way. The obvious areas to focus on would be the funda-
mental building envelope and the datacom equipment itself. However, compaction
is causing the scope for a typical data center construction project to be dominated
by the mechanical and electrical portions of the project. Further, keeping the cool-
ing system in operation without interruption has become just as critical as keeping
the power systems operating without interruption. As a result, datacom equipment
centers must consider the structural and vibration performance of:
• the building structure,
• the building infrastructure (power, cooling, flooring, and ceiling systems), and
• the datacom equipment (servers, storage, tape drives, network equipment, etc.).
Fundamentally, the current building codes focus on life safety issues. Therefore,
although they address issues such as weather events, these events are specifically
within the context of reasonably expected conditions and life safety. A datacom facil-
ity must include all the typical life safety issues but also must consider what is
required to keep the facility in operation during and after more extreme natural or
man-made conditions such as a major weather event (e.g., a hurricane or tornado) or
an explosion.
The structural and vibration provisions (design features for withstanding exter-
nal vibration without inducing any failure) are very important but are designed and
integrated into the facility and equipment as an enabler for the successful operation
and performance of the IT equipment, the power systems, and the cooling systems.
For example, some people within the industry are starting to describe the data center
as “the computer” since the IT equipment is so integrated and tightly coupled to the
power and cooling systems.
The typical refresh rate of datacom equipment is often three to five years. This
increases the challenge to provide power, cooling, and structural systems with the
right capacities to support the datacom equipment since these systems typically have
a life span of at least five times that of datacom equipment. The ASHRAE book
Datacom Equipment Power Trends and Cooling Applications (ASHRAE 2005a)
provides a means of predicting power and cooling capacities in the future. A similar
resource does not exist from a structural perspective.
Since structural systems now need to be far more integrated and specialized to
accomplish the needs of today’s data center facilities, it is important for not only
Structural and Vibration Guidelines for Datacom Equipment Centers⏐5
designers but also owners and operators to have a general understanding of the struc-
tural and vibration basics. For example, unintentionally the structural or vibration
system could be seriously compromised by an operator’s not understanding the impor-
tance of the structural and vibration systems’ interactions within the building system.
The intent of this book is to provide some basics for addressing these highly
integrated topics of building structure, infrastructure, and datacom equipment. In
order to effectively address this holistic topic, the contributors to this book include
datacom equipment manufacturers, mechanical/electrical engineers, building archi-
tects, and building structural engineers.
1.2 OVERVIEW OF THE DATACOM INDUSTRY
The data center and telecommunications industry relies on a physical infra-
structure, including datacom equipment centers, for which there is a widely varying
range of size, capacity, level of resilience, configuration, use, staffing strategy, etc.
As a result, it is very important not to overgeneralize or use a cookbook/recipe
approach. Table 1.1 shows typical ranges of applications for various topics but is not
intended to show the absolute extremes.
Table 1.1 Datacom Industry Wide Range of Applications
Topic Range
Space size Small room to the entire building
Building size 500 to over 500,000 ft2 (46.45 to over 46,451.52 m2)
Building configuration Single building to an entire campus
Power and cooling density 5 to 500 W/ft2 (54 to 5382 W/m2)
Equipment weight 30 to 3600 lb (13.6 to 1634.4 kg)
Infrastructure performance/availability Tier 1 to Tier 4
Scope Minor renovation to a new building or campus
Operations staffing Lights out (unmanned) to full 7 × 24 staffing
Use Mixed use to dedicated use
Main function Telecommunications, call center, data center
Backup site No backup site to automatic failover backup site
Lifetime changes Minimal to extreme
Occupancy Owner occupied to tenant occupied
Code Barely meets code to significantly exceeds code
Stories Single story to high-rise
Below 0°F (–18°C), above 100°F (38°C),
Climate
dry/wet, hurricane, tornado prone
Soil type Sand, clay, organic, high water table
6⏐ Introduction
1.3 OVERVIEW OF ASHRAE TECHNICAL COMMITTEE 9.9

Key technical experts of the major IT manufacturers recognized in the late
1980s that power and cooling capacities were going to become increasingly more
challenging for the industry. Further, they saw no vendor-neutral professional soci-
ety holistically addressing the technical aspects of the data center industry. Addi-
tionally, they were seeing increasing need for the collaboration and coordination of
the IT industry and the facilities industry.
Due to ASHRAE’s major international presence and leadership, long history (it
was started in 1894), and major publishing infrastructure (including model codes,
standards, guidelines, courses, etc.), the IT manufacturers saw ASHRAE as the
source to publish unbiased information. As a result, Roger Schmidt (IBM) and Don
Beaty (DLB Associates) started the formal process of providing ASHRAE the justi-
fication for creating a dedicated technical committee for data center facilities.
Since no other vendor-neutral, nonprofit organization existed for data center
facilities, an ASHRAE technical committee (TC) was organized and its members
carefully selected to address the broadest possible scope. For example, even the
committee title, “Data Center Facilities, Technology Spaces, and Electronic Equip-
ment,” reflects a broad perspective (the facility down to the electronics).
TC 9.9 members include experts from IT manufacturers as well as the facility
design, construction, and operation areas. The committee also includes members
from numerous countries around the world to help provide an even broader perspec-
tive. A number of these committee members are neither members of ASHRAE nor
thermal engineers.
The focus of the committee is to identify informational and technical needs of
the data center industry and to meet those needs. Where the committee does not have
a full range of resources or expertise, resources are sought and added to the team.
These needs in some cases are not HVAC based, so the committee and ASHRAE’s
publishing capabilities are employed as a means of meeting the industry’s needs.
TC 9.9 has the following major objectives:
• to produce unbiased technical material on data center HVAC
• to provide unbiased training on data center HVAC
• to provide a forum for publishing unbiased technical material on subjects
other than HVAC for the data center industry
1.4 OVERVIEW OF THE ASHRAE DATACOM SERIES
The ASHRAE Datacom Series is ASHRAE TC 9.9’s primary means to meet the
informational needs of the data center industry. The content is intended to provide
value to both technical and nontechnical readers.
The books vary in that sometimes they are totally independent of previous books
in the series while occasionally they may build on previous books in the series.
At the time of this publication, the following four books have been published,
with another three books having substantial work already completed:
1. Thermal Guidelines for Data Processing Environments (2004)

2. Datacom Equipment Power Trends and Cooling Applications (2005)
3. Design Considerations for Datacom Equipment Centers (2006)

4. Liquid Cooling Guidelines for Datacom Equipment Centers (2006)
1.5 DOCUMENT FLOW

This book is organized into the following main sections:
Part I—Introduction and Best Practices. Part 1 provides an overview of this

book and ASHRAE TC 9.9. It includes the following chapters:
Chapter 1—Introduction
Chapter 2—Best Practices
Part II—Building Structure. Part 2 focuses on the basic building structure,

such as the building envelope, building beams, building columns, floor slabs, and
roof slabs. It includes the following chapters:
Chapter 3—Building Structures Overview

Chapter 4—New Structures
Chapter 5—Existing Structures and Additions
Chapter 6—Types of Building Structures
Part III—Building Infrastructure. Part 3 focuses on power, cooling, flooring,

and ceiling systems. Infrastructure includes exterior equipment yards, such as those
for power and cooling. Part 3 includes the following chapters:
Chapter 7—Building Infrastructure Overview

Chapter 8—Structural Considerations for Infrastructure
Chapter 9—Raised-Access Floor Systems
Chapter 10—Vibration Sources and Control
Part IV—Datacom Equipment. Part 4 focuses on the IT or electronic pack-

aging, including items within a rack, the rack, and a row of racks. Part 4 includes the
following chapters:
Chapter 11—Shock and Vibration Testing on Datacom Equipment

Chapter 12—Seismic Anchorage of Datacom Equipment
Chapter 13—Analysis of Datacom Equipment and Seismic Anchorage Systems
Appendices and Back Matter. These include background material, a bibliog-

raphy, and a glossary of terms.
1.6 PRIMARY USERS FOR THIS DOCUMENT

The intended audience for this publication includes both technical and nontech-
nical readers. Those involved in the design, construction, commissioning, operating,
implementation, and maintenance of datacom equipment centers can all benefit
from this book. In addition, those who develop and/or design electronic, cooling, and
8⏐ Introduction
other infrastructure equipment will benefit from these guidelines. Specific examples
of users of this document include:
• computer equipment manufactures (research and development engineers and

marketing and sales organizations)
• infrastructure equipment manufacturers (cooling and power)
• consultants
• general construction and trade contractors
• equipment operators, IT departments, facilities engineers, and chief informa-
tion officers
Best Practices
2.1 BUILDING STRUCTURES—NEW BUILDINGS AND ADDITIONS
Careful forethought is needed in the early design phases for new buildings and
additions to existing buildings. Datacom equipment centers require a unique set of
design criteria as compared to more traditional spaces. Many services must be
supplied directly to the datacom equipment to meet cooling, power, and communi-
cations needs. These services require support systems that place increased loads on
the building structure.
In addition, the weight of the datacom equipment and its supporting infrastruc-
ture will create the need for special structural consideration, as will any structural
design requirements (such as resistance to hurricanes or explosions) that are above
and beyond code minimums.
Best practices for new buildings and additions to existing buildings are listed in
Table 2.1.
Table 2.1 Best Practices for Building Structures—

New Buildings and Additions
1 Balance the first cost of the building with the long-term adaptability of the structure.
2 Provide above-average clear height to the underside of overhead building structures.
3 Provide a reserve capacity in structural members for collateral loads.
4 Optimize column spacing for maximum adaptability.
Provide special structural inspections as outlined in Chapter 17 of the International
5
Building Code.
6 Consider the building’s structural system versus steel lead-time.
7 Consider the holistic impact of seismic design requirements.
8 Review the benefits of increased soil-bearing capacity.
9 Optimally locate shear walls and braced frames for maximum adaptability.
10 Reduce constraints on penetrations through wall, floor, and roof assemblies.
10⏐ Best Practices
2.2 BUILDING STRUCTURES—RENOVATIONS, RELOCATIONS,

AND CHANGES
Existing structures require an extensive amount of verification to determine

how they will react to their new uses. The best method of such assessment involves
close inspection of the as-built and/or original design drawings. As for new build-
ings, these drawings will include information required by structural engineers to
determine the loading capabilities of roofs, flooring systems, and lateral forces. An
engineer is required to physically survey the structural systems to verify compliance
with the drawings. This survey may require the removal of architectural finishes in
certain areas until the engineer is confident the building was constructed to specifi-
cation. Buildings erected prior to 1975 in areas prone to earthquakes should be
avoided, as building codes prior to 1975 did not include the lateral forces design
criteria found in more recent codes and it is very expensive to upgrade such buildings
to current standards.
Best practices for renovations, relocations, and changes to existing buildings are
listed in Table 2.2.

Renovations, Relocations, and Changes
1 Comprehensively analyze the design and construction of the existing structure.
Expose the existing structure as much as possible (e.g., remove existing ceiling
2
systems).
Establish confidence by comparing construction drawings with visual spot checks of

3
the installed structure.
Evaluate the current code requirements and upgrade the existing structural systems
4
as required.
Evaluate providing independent sub-structures in lieu of upgrading the capacity of

5
the existing structural system.
Provide above-average supervision and inspections during construction of any

6
remedial work.
7 Accurately locate the existing building columns and confirm that they are plumb.
Understand that buildings erected prior to 1975 typically require expensive structural
8
improvements to meet current seismic code requirements.
Recognize that concrete structures are very difficult to verify structurally, with or
9
without drawings.
10 Consider “exit strategies” when spaces are leased.
2.3 BUILDING INFRASTRUCTURE
The building infrastructure supporting datacom equipment can be located in

and around the data center building. Heavy equipment is best located on a ground-
floor level or in an exterior equipment yard. Heavy loading of second-story spaces
or of the roof structure creates unnecessary special lateral loading issues.
The raised-access floor system supporting datacom equipment serves as a
second structural floor that must be able to support the weight of the equipment and
also function in lateral loading conditions.
Mechanical equipment located in the building should be isolated from the struc-
ture to eliminate transfer of vibration to the structure.
Best practices for building infrastructure are listed in Table 2.3.
2.4 DATACOM EQUIPMENT
Datacom equipment is the focus of the data center. The building and infrastruc-
ture designs should be integrated to assure that the reliability and availability of the
datacom equipment is maintained at the desired level. The design and installation of
datacom equipment that can weigh from 300–400 to 3600 lb (1335–1780 to
16,017 N) are key to achieving this goal.
Best practices for datacom equipment installation are listed in Table 2.4, and
best practices for datacom equipment design are listed in Table 2.5.

Building Infrastructure
1 Place heavy equipment and infrastructure systems on the ground level.
2 Locate larger systems outside the building in adjacent equipment yards if possible.
3 Provide a reserve in the capacity of the raised-access floor system.
Consider mechanically fastening raised-access floor pedestals to the structural

4
floor slab.
5 Consider screwing raised-access floor panels to the floor support system.
Install the raised-acess floor system at a temperature very close to the operational
6
temperature.
7 Provide vibration isolation for equipment with rotary components.
Secure all overhead cable trays to a rigid frame that carries the load either to the roof
8
structure or to the floor slab.
9 Organize the structural support of building infrastructure distribution systems.
10 Provide for expansion/contraction of the infrastructure.
12⏐ Best Practices

Datacom Equipment Installation
1 Define the weight and dimensions of the datacom equipment.
Check the building structure—can the building support the datacom equipment
2
loading?
3 Check the raised-access floor—can the it support the datacom equipment loading?
Investigate the installation/relocation route from the loading dock to the final
4
destination.
5 Check for possible high operational shock and vibration.
6 Check the seismicity of the data center.
Consult the equipment manufacturer for unusual relocation routes or high opera-
7
tional shock and vibration.
8 Provide seismic anchoring.
9 Provide vibration isolation for infrastructure equipment with rotary components.
10 Consult the equipment manufacturer when the need arises to relocate the equipment.

Datacom Equipment Design
1 Establish test specifications for the product.
2 Establish directions and magnitude of the static and dynamic loading.
3 Conduct analyses based on the worst-case loading.
4 Define the design goals.
5 Conduct fragility, shippability, operational, and seismic simulation tests.

6 Evaluate test results versus specifications and design goals.
Part II
Building Structure
Building Structures Overview

3.1 INTRODUCTION
Since this book is intended for both technical and nontechnical readers, this
chapter provides some structural basics. These basics are intended neither to be thor-
ough nor to enable a non-engineer to perform structural engineering but rather to
provide some insight into the influences on the design and performance of building
structures.
Structural engineering involves breaking structures down into their most basic
parts—similar to solving algebra, geometry, or trigonometry problems. In simple
terms, structural engineering involves paying attention to the following:
• foundations
• soil or sub-surface bearing elements
• floor slabs
• structural frames
• structural components of wall and roof systems
• structural connections
• other structural elements
The common saying that a building is only as strong as its foundation should not
be dismissed as mere cliché: foundations are critical. One might also say a founda-
tion is only as strong as the soil that supports it, as site selection and facility design
depend greatly on site soil conditions, which may vary over a single site for larger
structures.
Additional elements that are not direct structural elements, such as the building
envelope, skin, surfaces, and cladding, have also have an impact on the structure.
These elements are analogous to ship sails or surfaces that are directly impacted by
exterior forces such as wind, rain, and snow that in turn impact the loading of the
structural system.
3.2 SOME BASICS

As a building’s structure and use change over time, so does the loading of the
structure. Structural design professionals must pull together all the design attributes
16⏐ Building Structures Overview
and determine the load requirements, the impact from natural occurrences, and the
expected design performance of the building now and in the future.
Gathering loading information is often a difficult task, especially when
members of various architectural and engineering disciplines are concurrently
designing the building components, including the mechanical, electrical, plumbing,
and fire protection systems. The owner or tenant and the facility operators also play
a part in the design process.
The required loads as defined by the building code may be sufficient in many
cases, but in every case the design professional must review the structural loading
for the entire project. Building codes are quick to address typical uses, such as office
spaces; however, they seldom, if ever, contain specific requirements for datacom
facilities or address the unique loading conditions imposed by datacom facility
systems and the need to resist threats posed by both man and nature. The designer
must consider the type of loads the facility will bear, as these will influence the
supporting structure, especially during vibratory events.
Loads from datacom equipment, battery or uninterruptible power supply (UPS)
systems, transformers and switchgear, chillers, boilers, large cooling water piping,
cooling towers, roof-mounted mechanical units attached to large ductwork, and fire
protection systems and their associated water distribution systems all greatly impact
the design of elevated floor or roof members. Architectural finishes such as ceilings,
lights, interior lateral wall supports, bearing walls, raised-access floors, large wall
openings, floor slopes, and underground drainage also impact the structural design.
Other elements designers must consider are deflection of members and struc-
ture drift, which vary with the type of structure selected. For financial reasons, pre-
engineered metal buildings are commonly used. Such buildings are typically frame
structures that are relatively flexible and dissipate energy by bending or deforming.
This type of structure requires that nonstructural elements and systems be designed
to accommodate the expected deformations.
Buildings utilizing moment-resisting frames (moment frames) for lateral stabil-
ity will likely have similar considerations. More rigid structures, such as those that
incorporate shear walls or braced frames, will deflect or drift less than a typical pre-
engineered or moment frame structure. If reduced levels of deflection and drift can
be predicted, connecting elements from partitions, exterior walls, and attached
equipment will cost less to construct.
Seismic and wind requirements vary from location to location and are based on
geological data and past damage reports in regions where seismic events have
occurred. A geological investigation called a site-specific seismic study is commonly
completed to determine the type of existing soil and the distance from the building
foundation to the supporting bedrock. These parameters, in combination with code-
prescribed criteria, provide guidance for establishing the percentage of structure
mass to be used in the lateral load analysis, sometimes leading to net savings in over-
all construction costs.
The building design should be as symmetrical as possible to allow seismic
forces to flow more evenly across the structure. Forces induced on an asymmetri-
cally shaped building are considerably greater than those on symmetrically shaped
buildings. Deflections of the asymmetrical structures will also vary and require
greater study in the interconnections of the nonstructural components.
Wind load requirements are based on weather data compiled from sources such
as the National Weather Service. The codes or incorporated standards include maps
and guidance data, but working with the local building department is highly recom-
mended. Although building codes generally allow interpolation between wind speed
contours, some jurisdictions require a set minimum wind speed for the entire juris-
diction based on the highest minimum wind speed in that jurisdiction.
Wind load considerations are critical for not only the performance of a structure
during wind events but also for the performance of the nonstructural components.
Roof uplift and wall anchorage systems as well as equipment tie-downs on the roof
are of great importance and must be considered.
It should be noted that structures are typically designed based on meeting only
the minimum code requirements. Further, the codes are written for life safety and not
for the purpose of keeping data center facilities operational during and after a major
storm.
Facilities can be constructed to withstand wind loads from extremely high wind
forces, such as tornadoes, and wind-borne debris, etc., as shown in Figure 3.1, but
these provisions can be very costly. A cost-benefit analysis should be performed to
confirm the value of storm-resistant design. Such a study should assess structural
upgrades to meet multiple wind-speed conditions (and their associated costs) against
the historic statistical probability of the occurrence of those conditions. The owner
of the structure then can make an informed decision on the level of wind-threat resis-
tance to be incorporated into the project.
Historically, the codes have been prescriptive (cookbook- or recipe-like). The
trend in code content, however, has been toward increasing the complexity and
Figure 3.1 Graphical representation of wind loads on a datacom

equipment center.
including more requirements. This trend is probably driven by events such as Legion-
naire’s disease outbreaks, New Orleans hurricane damage, global warming, etc.
In addition to becoming more complicated, the codes are gradually migrating
toward performance requirements rather than prescriptive requirements.
Appendix A includes details regarding recent code requirements.
Especially in older structures, it is quite common that the plans and specifica-
tions issued to construct the building are no longer available and therefore the design
load capacity of the structure is not known. This can be problematic (and often
costly) for planned datacom-related retrofits and even for seemingly minor building
infrastructure upgrade projects. In such cases, a licensed structural engineer should
review the structure in the field and determine the load carrying capacity of the exist-
ing structural system.
When constructing a datacom equipment center in a leased space, consideration
should be paid to the eventual exit strategy from the facility. Modifications to leased
structures should be kept to a minimum or designed so as to allow them to be
removed at lease termination without impacting the main structural systems.
Pits and trenches can be filled in and free-standing structures can be removed,
but when existing walls or other structural elements are modified such that the struc-
ture would no longer be sound without the added reinforcement, the modifications
must remain in place permanently. Therefore, the terms of a lease should be carefully
checked to determine potential consequences prior to making changes to the build-
ing structure.
3.3 SUBMISSIONS TO AGENCIES

Most planning boards and/or building departments require submission of
geotechnical reports for new buildings or building renovation projects. These reports
will provide information about site preparation and fill placement and compaction
as well as recommended parameters for foundation design.
Structural drawing submission requirements to the state or city building code
departments—the local authority having jurisdiction (AHJ)—vary throughout the
country. At minimum, however, all structural drawings will identify the plans and
elevations of the steel and/or concrete and interactions with the foundations and
upper floor slabs as required. Additionally, the basis of design will need to be iden-
tified by building type, occupancy type, construction classification, and importance
factors related to snow, wind, and earthquakes.
Typically in high wind areas such as southern Florida or earthquake-prone areas
such as California, a full set of calculations for all structural elements signed and
sealed by a licensed professional are required to be submitted with the structural
drawings. In less hazard-prone areas, signed and sealed drawings are often the mini-
mum required documentation.
In cases of lease expiration or termination, the previously submitted and
approved building structure should be on file with the local AHJ. A certificate of
occupancy should be available to the new owner prior to purchase through either the
AHJ or the leasing agent. This will alert the new occupant as to whether the overall
building has been inspected by the local AHJ.
Additions to the building will require further structural submissions, often as

part of a “tenant fit-up” filing. These filings typically exclude the design of the struc-
tural system for the existing building and instead concentrate on how the new struc-
tural elements affect the existing building.
3.4 BASIC DEFINITIONS
dead load: the combined weight of all the permanent construction materials, includ-
ing roofing, framing, and other structural components such as walls, floors, roofs,
ceilings, stairways, built-in partitions, finishes, cladding, and other similarly incor-
porated architectural and structural items and fixed service equipment, including
cranes. The dead load is determined in advance and is assigned a low factor of safety
in the overall building design. All dead loads are considered permanent loads.
collateral load: a type of dead load, the collateral load includes the weight of any
materials other than the permanent construction materials. The collateral load may
include electrical pipes, sprinkler systems, ceilings, lights, and other materials.
live load: the weight of building occupants, furniture, machines, equipment, and the
like. Since the live load is hard to quantify, it carries large factors of safety in the over-
all design of the building. Other live load sources that occur during construction and
maintenance, such as those from tool boxes, temporary generators, etc., must be
taken into consideration as well; these loads may be difficult to predict.
snow load: the maximum weight of snow that will collect on the roof of the building.
The snow load is highly dependant on the ultimate location of the building. The slope
(pitch) of the roof is a factor used to determine snow loads.
In addition to roof snow load, ground snow load must also be taken into consid-
eration. The roof snow load is usually less than the corresponding ground snow load
because snow is often removed from building roofs by melting and wind.
wind load: the force exerted on a structure by strong winds.
seismic load: the force exerted on a structure due to seismic activity.
duration of load: the period of continuous application of a given load, or the aggre-
gate of periods of intermittent applications of the same load.
load factor: a factor that accounts for deviations of the actual load from the nominal
load, for uncertainties in the analysis that transforms the load into a load effect, and
for the probability that more than one extreme load will occur simultaneously.
impact load: the load resulting from moving machinery, such as elevators,
craneways, vehicles, and other similar forces, and kinetic loads, pressure, and possi-
ble surcharge from fixed or moving loads. Note: a craneway is the structure or struc-
tural members upon which a crane travels, which can include columns, support
girders, and rails.
equipment support: those structural members or assemblies of members or manu-

factured elements, including braces, frames, lugs, hangers, or saddles, that transmit
gravity load and operating load between the equipment and the structure.
New Structures
4.1 INITIAL INVESTIGATION
Once a tract of land has been identified as a potential location for a new facility,
preliminary due-diligence activities such as those listed below should commence.
From a structural standpoint, the most important of these activities is the geotech-
nical investigation and its subsequent report.
The geotechnical report will provide information regarding the existing soil
conditions and will provide recommendations for site preparation or soil improve-
ment activities as well as recommended parameters for building foundation design.
As a minimum, it is suggested that the following information be provided in the
geotechnical report:
• Description and geology of the site

• Site preparation recommendations
• Extreme frost penetration depth; scour depth if applicable
• Recommended types of possible foundations
• Seismic soil parameters (in accordance with the local building code)
• Possible total and differential settlements under load
• Water table depth; anticipated seasonal fluctuations if applicable; any perti-
nent local climatological data that might have an impact on the foundation
• Discussion of anticipated construction difficulties, such as rock, expansive
material, etc.
Following a preliminary due-diligence investigation, and once the location of

the new building has been determined, a second, more comprehensive geotechnical
investigation should be initiated. As part of this investigation, soil test borings should
be made within the area of the new building and should extend at least 100 ft
(30.48 m) beyond the perimeter of the facility; typically, these borings are taken on
a 50 × 50 ft (15.24 × 15.24 m) grid.
This comprehensive investigation will provide sufficient information for the
geotechnical engineer to determine the soil improvements necessary to meet the
bearing capacity requirements of the building structure.
22⏐ New Structures
A site-specific seismic study is a form of investigation that is often overlooked

in initial due-diligence activities. Such a study may provide substantial cost and
schedule benefits to later stages of the building construction project by establishing
actual field measurements and data that result in less demanding design factors than
the theoretical design factors stated in the building code.
The outcome of this study is a more accurate determination of the Site Class that
may allow for a reduction in site seismic coefficients (ASCE Standard 7-05, Mini-
mum Design Loads for Buildings and Other Structures, allows a reduction of up to
20% [ASCE 2005]). This in turn could represent a substantial savings on the lateral
force resisting system of the building structure or, even more importantly, could
eliminate the need for seismic restraints on elements of the building infrastructure.
4.2 COORDINATION
All members of the design team must coordinate their documents with the docu-
ments of members from the other disciplines. A free flow of information through a
central management point works best so that a single entity is knowledgeable of
everything that is occurring, including what information has been requested, what
has been furnished, and what is still outstanding. Traditionally, this role is performed
by the architectural firm, but in design/build situations, it can be assigned to the
construction manager or to a contractor.
4.3 ESTABLISH DESIGN CRITERIA

Data centers are somewhat unique in the construction industry because the cost
of the building shell (the structural system, roofing, exterior walls, etc.) is relatively
low when compared to the electrical and mechanical systems and ultimately the
datacom equipment that will be housed within the building shell. Consequently,
traditional thought processes regarding the building structure need to be set aside,
and each aspect of the structure needs to be evaluated in a holistic manner. The goal
of this evaluation is to establish the design criteria for the structural system.
A number of the required structural design criteria are described in the follow-
ing paragraphs and are summarized in Table 4.1.
Table 4.1 Building Structural Design Criteria
1 Establish ceiling clear height
2 Establish column spacing
3 Locate bracing systems
4 Establish all current and future loads
5 Establish frame drift limitations
6 Establish settlement limitations
7 Establish deflection limitations
8 Establish basic target beyond code requirements
4.3.1 Adaptability
When designing any aspect of a datacom equipment center, it is important to
have a clear understanding of how much value the owner places on the adaptability
of the systems being designed.
Datacom equipment has a “refresh rate” of between three and five years; this
means that in a maximum of five years, the cutting-edge datacom equipment of today
will be outdated and will potentially be scheduled for replacement.
The design and construction process for a typical facility is 12 to 18 months; the
anticipated life span of the mechanical and electrical infrastructure is 15 to 20 years;
and the anticipated life span of the building structure is 20 to 50 years. Consequently,
the building structural system may house five to ten vintages of datacom equipment
over its lifetime. It is impossible to predict the structural requirements that will be
imposed by datacom equipment and supporting building infrastructure of the future.
However, it is possible to adopt structural design criteria that will result in a building
structure that can be easily adapted to meet future requirements.
Some specific aspects of adaptability might include:
• Ability to adapt to future heavier floor live loads

• Ability to accept future routings of floor or overhead supported piping, cable
tray, and conduit runs
• Ability to make future structural penetrations
• Ability to bring new large and/or heavy datacom equipment and building
infrastructure equipment into the building
4.3.2 Establish Clear Height

The clear height, or headroom, refers to the distance from the top of the struc-
tural floor slab to the lowest point of the overhead structural system (which may
support a roof system or a building floor above).
In a traditional construction project, there is often a mindset to reduce the clear
height as much as possible since it results in savings on the cost of the building clad-
ding system and, in a multi-story building, may allow for the incorporation of an
additional story. However, datacom equipment centers have an above-average
amount of electrical, mechanical, and IT infrastructure that must be accommodated,
often utilizing raised-access floor (RAF) systems. Consequently, these facilities
demand a clear height more akin to a warehousing building than an office building.
Since the clear height must be established relatively early in the design process,
it is unreasonable to assume that the infrastructure can be coordinated to anything
beyond a conceptual level. A vertical zoning plan can help to establish the required
clear height by assigning mutually exclusive zones for various elements of the infra-
structure, as shown in Figure 4.1.
A clear height that is too low will ultimately result in excessive coordination in
the design and construction of the mechanical, electrical, and plumbing (MEP) infra-
structure; this situation in turn will lead to increased cost and schedule time.
Figure 4.1 Vertical zoning plan example.
Conversely, a clear height that is too high will lead to inefficiencies and an increase
in the cost of the installation of any MEP infrastructure that is supported from the
overhead structure as well as an increase in the cost of the structure itself, which will
require larger column members.
In addition to vertical zones over RAF areas, clear heights are critical for
mechanical and electrical equipment rooms. When designing mechanical and elec-
trical equipment rooms, consider more than the height of the equipment itself. Below
the equipment, account for the additional height of special foundations and house-
keeping pads. Above the equipment, consider radii for conduit and piping bends and
the additional height of pipe flanges and valve bodies. In generator rooms, bear in
mind the height required for exhaust mufflers if located in the room.
4.3.3 Establish Column Spacing

Any columns located within a datacom equipment area can cause a disruption
of the layout of the datacom equipment, resulting in a loss of utilization of this
premium space.
In a structural building system, columns can be spaced widely apart, but as a
result the size, weight, and cost of the structure increase. Consequently, the column
spacing criterion for the building needs to be evaluated and optimized.
One method that can be used to help optimize the column spacing is to perform
the initial programming or layout of the building floor plan. It may be beneficial to have
differing column spacings within different portions of the building. For example,
outside of the datacom equipment area there may be mechanical and electrical equip-
ment rooms where the infrastructure is supported from the structure above and/or is
located on the roof; in these instances, a closer column spacing may be preferable.
Within the datacom equipment area, if the configuration of the datacom equip-
ment is known, it may be possible to configure the column spacing to reduce the
disruption to the equipment.
4.3.4 Establish Load Criterion

Gathering structural loading information is critical. Within a datacom equip-
ment center, it is typical that loads will be suspended from the overhead structural
system and other loads that are placed on the roof or floor above. The magnitude and
location of the loads must ultimately be provided to the structural engineer by the
other design team members (the architect, MEP contractors, fire protection person-
nel, etc.) and often by the building owner.
These loads may take the form of datacom equipment (server racks, networking
equipment, etc.), cooling and power equipment (chillers, generators, fuel tanks, etc.)
and distribution systems (piping, ductwork, cabling, and conduits). Once specifics
are known, equipment cut sheets should be provided to the structural engineer so that
he or she can detail support requirements, seismic bracing, etc.
In the absence of known loads, areas of the building structure can be designed
to support estimated loads. In order to improve the adaptability of the building struc-
ture, it is often beneficial to utilize a fairly high collateral load.
4.3.5 Establish Frame Drift Limitations

Frame drift, or the horizontal displacement of the structure at a floor or roof
level with respect to its original position, is a very important part of the structural
design, and drift limitation requirements can greatly impact the type of structure that
is selected.
Moment resisting frames have a propensity to have greater drift than braced
frames or shear wall systems, so use of such a system can result in greater damage
to items attached to the structure. Base fixity (a number from 0 to 1 used to quantify
how well the base of a frame is attached to the ground, 0 being no connection and
1 being a perfectly rigid connection) can also be an important consideration in help-
ing control drift.
Column bases are often considered to be “pinned,” allowing some rotation of the
column with respect to the footing or even minor rotation of the footing itself, but
this rotation allows for increased drift in the case of a moment resisting frame. There-
fore, to help reduce frame drift, column bases for moment resisting frames are some-
times designed to be “fixed,” or prevented from rotating with respect to the
foundation, by stiffening the column baseplate, strengthening the bolts, and prevent-
ing foundation rotation.
Setting limits on the amount of drift is important to the mechanical and electrical
infrastructure that will ultimately be installed inside the datacom facility. As an
example, if piping is suspended and laterally braced to the structure to resist seismic
forces, the connections from the pipes to the floor-mounted equipment can be
damaged or destroyed if the building structure drifts under seismic or wind loads.
4.3.6 Locate Bracing Systems (Lateral Force Resisting Systems)

An effective way to reduce the amount of drift is through the incorporation of
lateral force resisting systems. Such systems can consist of bracing systems featur-
ing X-bracing, diagonal bracing, a variety of chevron-type bracing, or even rigid
moment frames. Alternately, shear wall systems, which incorporate load-bearing or
non-load-bearing concrete masonry units (CMUs) or concrete walls, can be utilized.
The type of bracing or walls used will depend on the space requirements.
Once internal travel patterns, egress routes, and open space requirements are set,
the bracing system can be located. The final location and type of bracing system
should be coordinated with the needs of the mechanical and electrical distribution
systems to help avoid conflicts. For example, care should be taken to avoid cross
bracing inside exterior or interior walls that conflict with vertical runs of pipe and
conduit or the location of wall-recessed electrical panels.
4.3.7 Establish Settlement Limitations

Buildings settle under load, with 0.5 in. (12.7 mm) generally considered the
preferred upper limit in data center structures. Because subgrade soil conditions are
not homogeneous and in fact can vary widely over very short distances, buildings
often do not settle uniformly. This non-uniform settlement is termed differential
settlement.
The acceptable limit for differential settlement in a datacom building structure
is usually 0.5 in. (12.7 mm) from column to column. This is typically an acceptable
limit for frame distortion, drainage purposes, and suspended material. The problems
caused by this relative elevation difference (such as cracking of concrete floor slabs,
development of stresses in piping and other infrastructure, etc.) are increased by the
deflection of the roof or floor framing members, as discussed in the next section.
4.3.8 Establish Vertical Deflection Limitations

To meet vertical deflection code requirements, the International Building Code
(IBC) (ICC 2006) and similar codes limit vertical deflection effects from live loads
to the supporting member span length, L, in feet or meters divided by 360 if
supported material will be damaged as a result of the deflection. This includes hard
ceilings and other brittle finish materials but does not include lay-in ceilings, cable
trays, wire racks, etc.
The design engineer must use sound judgment concerning the potential for and
possible extent of damage to these materials. Based on past experience, it is
suggested that data center structures be designed for live load vertical deflections of
L/360 or 0.5 in. (12.7 mm) maximum. In the event of very long spans, it is suggested
that the structural steel members, bar joists, or cold roll sections be cambered. This
allows the final position of the beam or joist to be relatively level after total deflection
has taken place.
Because of deflection, possibly combined with the effects of settlement, the
attachments of pipes, conduits, or equipment that are suspended from the structure
must be designed so that they can be adjusted during the installation and testing of
the system. The structure will deflect at the time the empty pipes, racks, and equip-
ment are suspended.
Another incremental deflection will occur when the system is loaded with water
or other liquids or, in the case of rack or cable tray systems, conduits, wires, or
cables. After each loading cycle, the installers must adjust their hangers. Depending
on span lengths, planned loads, installation sequencing, and maintenance issues, it
may be desirable that supporting members be sized so the structure under collateral
load (load from all other loads above the building weight, not including the live load)
deflects no more than L/500.
4.3.9 Establish Horizontal Deflection Limitations

Deflection limitations must also be established for the vertical lateral force
resisting members that experience and resist horizontal loads from seismic and wind
forces. Typically, the limitations will be dictated by the type of wall cladding mate-
rial that is used (e.g., concrete walls should be designed for a maximum horizontal
deflection of L/600). In general, it is important not to allow hard finish material or
equipment to be constructed tight to a member that will experience deflection or drift
of a magnitude that will damage the finish.
4.3.10 Establish Basic Target Beyond Code Requirements

By default, traditional building structure designs are limited to minimum code
requirements. Anything beyond those requirements must be clearly identified and
the additional costs accepted by the owner early in the process. The following exam-
ples indicate some circumstances where the owner may require that structural
elements exceed the code minimum:
• Very high straight-line wind resistance of 140 to over 200 mph (225–322 km/h)
for hurricane and/or tornado resistance
• Very high snow and/or ice loads
• Excessive roof water loads should drains fail
• Setting a building Importance Factor (see Appendix E2 for a detailed explana-
tion of Importance Factor) exceeding the code minimum
• Structural resistance to blast, projectile, vehicle, or other threat force impacts
• Structural resistance to progressive column failure in the aftermath of an ini-
tial column failure due to structural overloading or impact
Existing Structures
and Additions
5.1 INITIAL INVESTIGATION
5.1.1 Geotechnical
When a data center occupies an existing structure, the building owner and the
structure design professionals should obtain the subsurface soils investigation report
and foundation recommendations made for the original building. In addition, the
building owner should perform any additional geotechnical investigation to deter-
mine if subsoil conditions are suitable for a datacom equipment center application.
Specifically, the owner should discover whether the new imposed loads on the struc-
ture can be successfully carried by the foundations and whether new structures can
be properly supported by new foundations. In addition to the previously discussed
items, this investigation should include but not be limited to
• soil borings within the structure to determine water table and moisture in the
subgrade and
• bearing capacity at bottom of slab level and 1–6 ft (0.30–1.83 m) below slab.
If moisture conditions are anticipated or are known to be a problem, a vapor emis-

sions test in several bays should be conducted.
Soil borings should be conducted outside the structure in locations where
mechanical, electrical, and other equipment will be situated. The soil borings are
used to determine if the site subgrade material is adequate to provide support for
foundations and to better understand the potential for ground settlement in the area
planned for miscellaneous structures. If the soil borings and associated calculations
indicate settlement values beyond that which can be tolerated by the new construc-
tion, the design professionals will then determine the type of remediation or deep
foundations applicable for the soils encountered.
The reason for making these soils explorations is that the “in-place” material
outside the structure may have significantly less capacity than what was specified on
the original contract documents. Most structural fill is well compacted within 10 ft
30⏐ Existing Structures and Additions
(3.05 m) of the building, but most of the equipment yard space for a data center
extends far beyond this 10 ft (3.05 m) range.
5.1.2 Existing Structure Plan Review

If the existing base building drawings are available, they should provide infor-
mation on what type of structural system is present, how the structure is braced, and
where the bracing occurs. The construction documents should also include the
design loads that were used as well as the codes under which the structure was
designed, what foundation system was used, the makeup of nonstructural elements,
and the physical layout of the structure. (Older documents may not give portions of
this information.) A cursory review of members using the given loads should be
made to verify the basic building structure. In addition to the structural information,
some MEP and architectural information may be available.
If, as in many cases with older buildings, the existing base building drawings are
not available, considerable research and time, and possibly testing, may be required
to determine the information noted above. For that reason, it is extremely important
to start seeking to obtain the base building drawings and technical specifications as
early in the process as possible.
5.1.3 Site Visit and Review of In-Place Construction

A site visit should be conducted to verify that the structure under consideration
was constructed in general accordance with the construction documents. Visual
checks should be made to determine any missing structural elements such as bridg-
ing, bracing, and bolting. The structural engineer should look for obvious signs of
distress due to possible settlement and for floor slab soundness, wall connections,
roof opening frames and equipment supports, and maintenance of the structure, as
well as signs of movement, vertically as well as horizontally, at any joints.
Following acceptance of the geotechnical findings, the plan review, and the site
visit, the structure design professional will have a good understanding of what will
be required to place the new data center into the existing structure.
5.2 COORDINATION OF NEW WITH EXISTING STRUCTURES
5.2.1 Headroom
Structural clearance is one of the most challenging aspects of converting exist-
ing buildings to data centers. Nowadays, raised-access floor heights are taller, as is
some of the new computer equipment. Thus, overhead clearances for new mechan-
ical and electrical infrastructure areas are often less than desirable. First, floor slabs
are often removed and rebuilt at a depressed level to accommodate the tall genera-
tors, electrical switchgear, and chillers.
When assessing structural clearances, look not only at the typical height to the
underside of typical beams or joists but also to the lower (and more restricting)
height of the deeper beams and joist girders that carry their loads back to the
columns.
Sprinkler system mains can be found or placed within the joist space but may
be located below the joists and girders, thus influencing the headroom available.
Interior roof drains, unit heaters, and other mechanical and electrical equipment can
also be located below the roof joists. All of these elements can be raised or relocated,
but this relocation will increase project cost and will take considerable time and
effort to complete.
5.2.2 Space Planning

If space is critical, it is important that all columns be located and column lines
verified and checked for size and plumbness. Column checks for plumbness will
provide information that will help in space planning. For example, if a 30 ft (9.1 m)
tall column is found to be 2 in. (51 mm) out of plumb and its neighbor is 2 in.
(51 mm) out of plumb in the opposite direction, this will effectively result in 4 in.
(102 mm) of lost space.
In addition, a surveyor should be employed to survey the building’s exterior
gross area and major vertical penetrations, such as stairs, elevator shafts, utility
tunnels, etc., that may restrict use of the available space.
5.2.3 Structural Reinforcing and Bracing Location Possibilities

Changes in building occupancy classification, such as from a warehouse to a
data center, or in loading, such as from an adhered or mechanically fastened roof to
a ballasted roof, can have significant effects on a building’s structural system. Such
changes will require the structural engineer to review the loading on the roof struc-
tural members, columns, and footings as well as their connections, and there is the
potential that the existing structure will need to be reinforced and that additional
bracing systems may be required.
Due to their potential impact on space layout and overall functionality, addi-
tional lateral braces or reinforcing should be considered during the preliminary plan-
ning stages. The effect of new reinforcing depth on headroom, clearances, and
installation of new systems as well as that of required new bracing on the data system
layout should not be overlooked. Possible locations for installation of supplemental
lateral bracing systems should be sought in the initial space planning of the building.
5.2.4 Effect of Loads on the Existing Structure

When constructed, most speculative industrial buildings have a collateral load
capacity of approximately 5 pounds per square foot (psf) (239 N/m2). Collateral
load capacity describes the amount of dead load that the structure can support
beyond its own weight or dead load. With this information and the knowledge of
whether the structure has a fire suppression sprinkler system, the other dead loads
the structure can support can be determined. If the structure is fully sprinklered, the
weight for the sprinkler system will be approximately 2.5–3 psf (120–144 N/m2).
Installed lighting typically adds another 1–2 psf (48–96 N/m2). This will allow
only another 1–1.5 psf (48–72 N/m2) usable load capacity, which is less than that
required for a lay-in ceiling, hangers, and grid. Typical ceiling system materials
weigh approximately 2 psf (96 N/m2) (see weights of materials in Appendix B). In
existing structures where the collateral loads are already 5 psf (239 N/m2), no new
loads may be introduced without the structure being reinforced or new support struc-
tures being added.
Earlier building codes allowed the reduction of live loads on roof members that
carry more than 200 ft2 (19 m2) and on elevated floor members carrying more than
150 ft2 (14 m2). However, newer building codes have revised and even placed restric-
tions on such reductions, and in some cases they have changed required load combi-
nations. Such changes can have substantial impacts on structural capacity.
Live-load reductions have been allowed by building codes that are based on the
presumption that while an area of a column bay may be loaded to the maximum live
load (psf), this uniform load will not actually occur over the entire bay at any single
point in time. This presumption is less true with the uniform loading of large
computer rooms. Structural assessments should look for the use of live-load reduc-
tions that actually lower the actual live-load capacity of the floor due to reductions
taken on some of the major supporting beams and girders.
Design live-load requirements for data centers vary but are often within the
ranges indicated in Table 5.1. Office areas in most buildings are designed for as little
as a 50 psf (2.39 kN/m2) live load plus a 20 psf (0.96 kN/m2) partition live load. The
building code may allow the 50 psf (2.39 kN/m2) live load to be reduced under
certain conditions. This relatively low allowable live load in combination with the
live-load reduction severely limits the possibility of placement of extensive data
centers on the elevated upper floors in most buildings. In existing buildings, care
should be taken to understand the worst-case load conditions of the datacom equip-
ment and the collateral suspended loads of the building infrastructure.
Additionally, the loads imposed by equipment moving into the facility should
be understood. When those loads exceed the structural capacity of the building floor
systems, structural reinforcing is required and/or equipment must be sufficiently
spaced so as not to exceed loads. When spacing is mandated, very clear and well-
documented spacing guidelines should be prepared and given to the owner for use
during equipment move-ins and installations.
Table 5.1 Typical Design Live Loads
Use Design Live Load, psf (kN/m2)
Typical office area 50–100 (2.39–4.79)
Telecom centers 75–150 (3.59–7.18)
Data centers 100–200 (4.79–9.58)
Service corridors 100–150 (4.79–7.18)
Mechanical and electrical rooms 125–250 (5.98–11.97)
High-density filing areas 175–250 (8.39–11.97)
Battery stack areas 600–800 (28.73–38.30)
5.3 NEW COMPONENTS
5.3.1 Introduction
Most new data centers have complex mechanical, electrical, plumbing, and fire
protection system requirements. These systems typically require underground and
above-ground installation of conduits and cable trays carrying conduits. The
mechanical plant requires that large overhead piping systems be supported from
structures overhead. Where overhead plumbing is required, the pipes may be
required to be isolated to prevent leaks from damaging equipment. Fire protection
systems vary greatly by type and layout. The existing system may require partial
replacement or even removal.
5.3.2 Foundation Interference

Existing buildings may be constructed using either shallow or deep foundations.
Shallow foundations can be composed of isolated spread footings, combined foot-
ings, monolithic footings with the slab on grade, or similar specialized mat or ribbed
slab foundations. Foundations can also have deep footings with piers, drilled piers,
driven piles, or augered cast-in-place piles. Driven piles may be acceptable in some
applications, though the vibrations resulting from installation may damage existing
equipment, finishes, or other work that is to remain. Shallow foundations can be
supported on residual material, engineered fill, densified sands (vibro-flotation), etc.
Deep foundations can be supported on suitable subsoil layers or rock, by frictional
resistance of the surrounding soil, or by a combination of these two methods.
When new shallow footings are required to be installed adjacent to existing foot-
ings, it is recommended that the new footings keep clear of the existing footings by
a generous margin. Existing footings were most likely earth formed and may be
larger than those specified on the original contract documents.
Underground piping and conduits will also require the excavation of trenches.
Care must be taken not to penetrate past the foundation influence line, since this
may undermine the existing foundation and cause additional settlement and possi-
bly damage to the existing structure. In a section showing the existing footing and
planned trench, the foundation influence line is a line drawn at a 45° angle outward
from the leading edge of the existing footing to the level of the bottom of trench
excavation. Where it is not possible to locate trenches outside the foundation influ-
ence zone, the geotechnical engineer should be consulted for additional recom-
mendations.
It may be necessary to underpin the existing footing using helical anchors,
driven or rammed minipiles, or similar types of proprietary systems, though such
underpinning is usually very expensive. When saw cutting slabs in structures that are
placed on soil that was stabilized with lime, fly ash, etc. due to the high plasticity of
the soil, care must be taken not to use much water or allow water to penetrate into
the residual and highly plastic material. This material is below the conditioned and
engineered fill material. Soil mitigation of the soils below conduit banks or other
utilities may also be necessary.
5.3.3 As-Built Conditions

Exposed materials can generally be easily reviewed for compliance with the
original contract documents or current needs; however, existing underground condi-
tions are another matter. Existing underground utilities or fill materials can only be
located and determined by a utility locator or by removing a section of slab and
performing a comprehensive geotechnical investigation. Consult with a structural
engineer anytime structural slabs are to be removed or excavations are made near
foundations. Failure to do so and to put in place temporary structural measures can
lead to slab collapse, undermining and failure of foundations, and failure of exposed
piers or columns when their effective length is increased by removal or soils bracing
them in the ground.
5.3.4 Impact Limitation on Existing Structures

Modifications to an existing structure should be limited to nonstructural
elements whenever possible; however, in many cases, this will not be possible. If
elements such as tilt-up walls, precast wall panels, or CMU walls are to be cut or
cored to allow conduits, piping, or ducts to pass through them, such possibly damag-
ing processes must be brought to the attention of the structural engineer and the panel
reinforced if required. If metal wall panels are penetrated, wall panel reinforcement
must be added.
If roof openings are cut into the roof deck, then roof opening frames must be
added. If a roof joist system is used, the joists may require reinforcing to carry the
frames. Existing structures should not be impacted to the extent of removing
columns or bracing systems; however, if such changes to the structure are required,
the impact of these changes on the entire structure must be reviewed by a licensed
structural engineer.
Considerable reinforcing, bracing, and foundation work may be required, often at
great expense, to replace the work proposed for removal. Suspension of material from
the existing structure must be from panel points of the joists using approved hangers
attached through the bottom chords and under-roof clamps from the top chord.
Care must be taken not to exceed the roof collateral load capacity, including
during construction activity. Members must be checked for any concentrated and
uniform loads that are to be applied. All structural assessments and design work
should be performed by a licensed structural engineer with specific load criteria to
design against.
5.4 REINFORCEMENT OF EXISTING STRUCTURE
5.4.1 Introduction
If the existing structure is occupied and the ceilings, ductwork, and sprinklers
are installed, then adding reinforcement in an existing structure is very difficult and
intrusive. The removal of the architectural and mechanical elements will need to be
carried out by a contractor. These removals will impact the schedule and will limit
the amount of other work that can be done in the same areas. If the existing structure
is empty and there is little overhead interference, the possibility of adding reinforce-
ment to joists, beams, joist girders, girders, and columns and reinforcing the foun-
dation, if required to provide sufficient collateral load capacity, is preferred.
5.4.2 Stiffening of Existing Structure

If the drift of the structure is greater than preferred, additional bracing can be
installed if the load path is clear and all elements along that load path are reviewed
for capacity and code compliance. Interior lateral wall supports that are provided by
the roof deck or by the joists must be reviewed. Adequate allowance must also be
made for roof uplift and deflection. Support should be provided to bottom chords of
joists and bottoms of purlins (horizontal structural roof members used to support
roof decking) where walls impart a horizontal load to these members.
5.4.3 Effects on Existing Structure

Major reinforcement additions on the existing structure should be minimized.
If possible, stand-alone support systems should be used in mechanical plant and data
centers to limit the reinforcing on the existing structure.
5.4.4 Reinforcement of Existing Structures to Extend Service Life

Most structures have a service life of 30 to 50 years. If a structure is sound and
is updated to the latest code requirements, it should be possible, with continuing
maintenance, to extend its service life for another 30 to 50 years.
Types of Building Structures

6.1 OVERVIEW
Building structures are to be designed to meet the code requirements for life and
safety as a minimum. Data centers should be designed to meet a higher standard, as
has been discussed in the previous chapters. This chapter focuses on the types of
building structures and their associated issues.
6.2 PRE-ENGINEERED METAL BUILDINGS

Specialty engineers typically design this type of structure. Pre-engineered metal
buildings are designed to support snow, snow drifts, rain, wind, and seismic loads
as required as well as the building’s dead weight and the anticipated occupancy loads
such as those from sprinkler systems, lights, mechanical ducts and systems, ceilings,
wiring, etc. These occupancy loads are called collateral loads and they vary but
should not be less than 20 psf (958 N/m2) for a data center.
Pre-engineered metal buildings typically utilize specially fabricated steel
members produced in the company’s factory or fabrication shop, though typical
framing members are sometimes used. A sample elevation of a pre-engineered metal
building frame is shown in Figure 6.1.
The deflection and drift limits typically used by the pre-engineered metal build-
ing manufacturer may not be sufficient for a datacom equipment center, and more
stringent requirements may need to be imposed.
In addition to the live, wind, and snow loads specified by the applicable building
code, each roof member should be designed to support a collateral load. In the
absence of the specific information needed to define the actual collateral load, it is
recommended that an alternate collateral load be applied in the form of a concen-
trated load of 2000 lb (8898 N) at mid span. In addition to the vertical loads, the struc-
ture must support lateral loads from wind and seismic forces as well as bracing forces
from racks, stanchions, raised-access floors, etc.
The roof insulation, usually supplied and installed by the pre-engineered metal
building supplier, should be carefully reviewed to ensure that it is sufficient to
provide above-code thermal protection to the space.
38⏐ Types of Building Structures
Figure 6.1 Pre-engineered metal building frame.
6.3 BRACED FRAME OR SHEAR WALL-TYPE STRUCTURES
These types of buildings usually utilize typical design and framing methods and
are usually designed by a licensed structural engineering firm, an architectural/engi-
neering firm, or a design/build contractor’s engineer. Similar to pre-engineered
metal buildings, they are designed to support the code-required loads, but rather than
using special or proprietary members, they typically utilize standard steel framing
members.
Buildings utilizing concrete or CMU shear walls typically concentrate such
elements at the building perimeter. Where buildings are large, require expansion
joints, or require additional supports due to load or configuration issues, interior
shear walls may be used. However, because these require considerable space and
may interfere with building layout or equipment routing requirements, braced
frames utilizing X-bracing, diagonal bracing, any type of chevron bracing (“K”,
“V”, etc.), or other bracing may be used. Such bracing can be concentrically located
or eccentrically located depending on design requirements. Sample elevations of
common types of bracing systems are shown in Figures 6.2 through 6.6; a sample
shear wall elevation is shown in Figure 6.7.
The various load paths, or means of transferring the loads through the structure,
should be clearly indicated on the drawings. Of particular importance are bracing
system locations where the horizontal forces from all of the contributing elements
must be collected and transferred to the foundation system. Since such systems must
sometimes be modified due to building usage changes, original design parameters
should always be listed on the project documents.
Figure 6.2 Concentrically braced frame using X-bracing.
Figure 6.3 Concentrically braced frame using diagonal bracing.
Figure 6.4 Concentrically braced frame using K-bracing.
40⏐ Types of Building Structures
Figure 6.5 Concentrically braced frame using chevron bracing.
Figure 6.6 Eccentrically braced frame using diagonal bracing with link
beams.
Figure 6.7 Shear wall.
Figure 6.8 Moment resisting frame.
6.4 MOMENT RESISTING FRAME SYSTEMS

Structures using moment frames to resist lateral loads are similar to the pre-
engineered system in that all of the horizontal loads are gathered and transferred to
frames within the structure. However, they are also similar to shear wall or braced
frame structures in that they typically utilize standard framing members. A sample
elevation of a moment resisting frame is shown in Figure 6.8.
6.5 COMBINATION OF FRAMING SYSTEMS

A variety of combinations of framing systems can be used to best serve the
building’s intended use.
Part III
Overview
7.1 OVERVIEW
Within the context of this book, building infrastructure includes power systems,
cooling systems, and communications systems, including the structural support of
these systems.
There is both interior building infrastructure and exterior building infrastruc-
ture. The interior infrastructure may be located above the datacom equipment (e.g.,
in the ceiling cavity), in the raised-access floor cavity, on the raised-access floor, in
chases, on walls, or on the structural floor slab. The exterior infrastructure may be
located on grade, on elevated platforms, or on the roof of the building itself.
The building infrastructure is one of the more challenging aspects of a build-
ing’s structural design, especially with the significant trend toward 50% to 80% of
the total construction cost being mechanical and electrical. One of the building infra-
structure challenges is the fact that often multiple trades, systems, or subsystems
cohabitate the same area. This complicates load projections and building support
systems. From a structural perspective, the infrastructure can be broken down into
support, anchoring, and expansion and contraction.
Table 7.1 provides an overview of various building infrastructure components
and their associated weights. This table is far from comprehensive, but it neverthe-
less demonstrates the wide range of elements that can exist within the building and
the wide range in structural load capacities that may be required. Tables 7.2 and 7.3
provide some insight into the possible mechanical and electrical equipment loads.
The loads in all three of these tables are general ranges; actual loads should be
obtained for each piece of equipment. In addition, it is important to include the “wet
load” of the equipment (e.g., the load when the equipment is filled with water in the
case of piping and hydronic systems).
7.2 INTERIOR BUILDING INFRASTRUCTURE

The interior building infrastructure may be located within the datacom equip-
ment room or in an ancillary space such as a central cooling plant, a DC power
plant or battery room, a central AC power room, a generator room, or a fuel oil stor-
age room.
46⏐ Building Infrastructure Overview
Table 7.1 Some Infrastructure Elements
Element Weight Range,1 psf (kN/m2)
Mechanical
HVAC piping 10–50 (0.48–2.39)
Storm and sanitary piping 10–15 (0.48–0.72)
Hot and cold water piping 1–5 (0.05–0.24)
Sprinkler piping 2–10 (0.10–0.48)
Ductwork 5–10 (0.24–0.48)
Electrical
Cable and conduit 2–10 (0.10–0.48)
Cable and cable tray 10–50 (0.48–2.39)
Busduct 10–20 (0.48–0.96)
Lighting fixtures 1–2 (0.05–0.10)
General
Suspended ceiling 1–2 (0.05–0.10)
1
Range is a general range and does not include the far extremes.
Table 7.2 Some Mechanical Equipment Loads
Equipment Weight Range,1 psf (kN/m2)
Cooling towers 75–125 (3.59–5.99)
Air-cooled chillers 100–150 (4.79–7.18)
Water-cooled chillers 150–200 (7.18–9.58)
Heat exchangers 200–400 (9.58–19.15)
Pumps 100–225 (4.79–10.77)
CRAC units 75–100 (3.59–4.79)
Rooftop HVAC units 50–75 (2.39–3.5)
Water storage tanks 500–1000 (23.9–47.88)

1
Table 7.3 Some Electrical Equipment Loads
Equipment Weight Range,1 psf (kN/m2)
Substations 100–200 (4.79–9.58)
Transformers 150–225 (7.18–10.77)
Transfer switches 100–200 (4.79–9.58)
Generators 250–400 (11.97–19.15)
Load banks 25–50 (1.20–2.39)
PDUs 150–250 (7.18–11.97)

Switchgear 75–125 (3.59–5.99)
Switchboards 75–125 (3.59–5.99)
Panelboards 25–50 (1.20–2.39)
Motor control centers 50–75 (2.39–3.59)

1
The structural requirements of the building infrastructure within such ancillary

spaces is outside the scope of this document, as there are hundreds of different
components and a myriad of different installation conditions for these spaces.
However, it is recommended that Table 5.1 be referenced at the early stages of
project planning, since it helps identify the range of live loads that may be anticipated
within these various spaces.
The live loads indicated in Table 5.1 represent the weights of equipment that are
supported from the floor slab. It should be noted that in centralized plant areas there
is typically a heavy concentration of piping, power conduits, etc., that is required to be
suspended from the structure above. Figures 7.1 through 7.6 provide some insight into
the interior building infrastructure elements that may need to be structurally supported.
7.3 EXTERIOR BUILDING INFRASTRUCTURE

The exterior building infrastructure consists of equipment and a means of cool-
ing and power distribution into the building. The infrastructure is generally related
to power and cooling systems and includes equipment such as standby generators,
transformers, switchgear, cooling towers, air-cooled chillers, rooftop units, air-
handling units, pumps, and water storage tanks.
As noted previously, the exterior infrastructure may be located on grade (in
equipment yards), on elevated platforms, or on the roof of the building itself.
Figure 7.7 provides an overview of the potential locations of elements of the build-
ing infrastructure exterior to a datacom equipment center.
Attention must be paid to the interface of the building infrastructure with the
building, including an evaluation of the potential for differential settlement between
the grade-mounted equipment and the building. As the distribution from the exterior
Figure 7.1 Piping suspended from an overhead structure.
Figure 7.2 Generator on above-grade floor slab.
Figure 7.3 Piping within a centralized cooling plant.
Figure 7.4 DC plant batteries.
Figure 7.5 Electrical distribution equipment.
Figure 7.6 Water-cooled chiller.
Figure 7.7 Overview of the building infrastructure exterior to a datacom

facility.
infrastructure (piping, conduits, etc.) passes through the exterior wall or roof of the
building, an understanding of the interactions from building drift and deflection is
also important. If designed incorrectly, water intrusion or even physical damage may
occur at these interfaces.
From a structural standpoint, equipment located on grade is most easily handled
and, with the exception of water storage tanks, the load requirements associated with
the building infrastructure are easily achievable through standard soil preparation
techniques. For example, most soils can be improved to allow them to support a bear-
ing pressure of 1,000–1,500 psf (47.88–71.82 kN/m2). However, the load transmit-
ted from vertical water storage tanks may exceed this amount, in which case deep
foundations (piles, piers, etc.) may be required.
Whenever infrastructure is located on the roof of an existing building, an eval-
uation of the existing structure is required to determine if there is sufficient residual
capacity within the existing structural members or if new beams, columns, and foot-
ings need to be engineered and constructed. Figures 7.8 through 7.12) provide some
insight into the exterior building infrastructure components that may need to be
structurally supported.
Like the building itself, all exterior equipment is subject to wind loading of vary-
ing magnitudes and from varying directions (see Figure 3.1). In various regions of the
country there is a higher likelihood of strong wind or storm events such as tornadoes
and hurricanes. In these instances, the structural system and the connections to the
building infrastructure should be evaluated carefully, and consideration should be paid
to enhancing the structural system so that it exceeds the minimum code requirements.
Similarly, the equipment itself should be evaluated to ensure that it is sufficiently
robust; oftentimes, specific equipment components are designed based on maximum
wind speeds (e.g., acoustical enclosures for generators, cooling tower fill, etc.).
Figure 7.8 Air-cooled chillers located on grade.
Figure 7.9 Air-cooled chillers located on structural platform above roof.
Figure 7.10 Drycooler located on roof.
Figure 7.11 Cooling towers on structural platform.
Figure 7.12 Standby generators located on grade.
Structural Considerations
for Infrastructure
8.1 OVERVIEW
Chapter 8 focuses primarily on the building infrastructure that is located within
the datacom equipment room itself. From a structural perspective, the infrastructure
can be broken down into support, anchoring, and expansion and contraction.
8.2 SUPPORT
Experience has shown that within a datacom equipment room, the minimum
clear height requirement under the lowest structural member is typically between
16 and 24 ft (4.88 and 7.32 m). In these rooms, a 2–4 ft (0.61–1.22 m) tall RAF
system is often installed, the datacom equipment itself is typically between 7 and 9 ft
(2.13 and 2.74 m) tall, and distribution systems for power, cooling, and communi-
cations are routed above the equipment.
Within the datacom equipment room, building infrastructure support systems,
which could be in the form of threaded rods, cabling, strut systems, etc., are often
designed and installed in a somewhat haphazard manner. Each subcontractor basi-
cally fastens and supports its infrastructure using whatever is available. This often
leads to unanticipated point loads and complex reactions. The support systems
become even more complicated when the facility is located in a seismic zone.
Figures 8.1 and 8.2 provide some insight into interior building infrastructure compo-
nents that may be installed within a datacom equipment room.
The collateral load capacity is especially important for the installation of build-
ing infrastructure since the different subcontractors are required to coordinate the
suspended loads from the structure and must ensure that the capacity of the structure
is not exceeded.
Traditionally, cable tray systems have caused the most concern in regard to
understanding the allowable load capacity. Cable trays may either be hung from the
overhead structure, supported on the datacom equipment racks, or supported from
a strut system where stanchions are installed through the RAF to transmit the result-
ant loads the structural slab below. Cable trays are predominantly utilized for the
distribution of communications systems (which require copper twisted-pair cabling
or fiber cabling) and DC power cabling (which is typically far heavier than cabling
for AC power systems).
56⏐ Structural Considerations for Infrastructure
Figure 8.1 Building infrastructure suspended from a structure above.
Figure 8.2 Building infrastructure supported from floor slab.
Cable tray systems are utilized because they allow good flexibility for the instal-
lation of cabling both during the initial construction and for future cabling. It is not
uncommon for the quantity of cabling ultimately installed to be 2 to 4 times more
than what was initially conceived and, consequently, the requirements of the struc-
tural support system must be carefully reviewed. It is good practice to identify and
permanently record both the design load and quantity of cables or approximate cable
fill percentages for all cable tray systems.
When considering a new datacom equipment center, a large increase in overall
construction costs is not required to provide for increased structural capacity at the
roof or floor level or in the columns, foundations, and bracing systems. Unfortu-
nately, speculative buildings and often build-to-suit or owner-occupied buildings are
designed precisely to the minimum code and any increased future capacity require-
ments are not accounted for.
Sometimes both cost and schedule benefit from the assignment of specific areas
within the datacom equipment center for the routing of heavy building infrastructure
distribution systems. Within these areas, floor-mounted structural frames may be
utilized to carry the structural loads rather than engineers trying to anticipate or
broadly overdesign the overhead structure.
8.3 ANCHORING
Anchorage of building infrastructure equipment that is not supported on the
RAF but is instead supported directly by a concrete slab (either suspended or on
grade) is quite straightforward. It should be understood that all equipment larger than
400 lb (180 kg) should be securely anchored to the floor per the IBC (ICC 2006).
Examples of equipment that may be located within the datacom equipment room and
that require anchorage are computer room air-conditioning (CRAC) units, power
distribution units (PDUs), and UPSs. In general, such heavy fixed equipment does
not have overturning issues. Of course, this assumption should be confirmed by a
qualified structural engineer.
As lateral loads are the primary concern for anchoring, a direct fastening system
using concrete anchors is the most direct approach. Anchorage to concrete should
use the provisions defined in ACI 318, Building Code Requirementts for Structural
Concrete (ACI 2001). Shot-in anchors, also commonly referred to as powder-driven
fasteners, are hardened steel pins that are driven into masonry or concrete by means
of a powder-actuated tool. Several anchor styles and lengths are produced for differ-
ent fastening requirements. Shot-in anchors should not be used due to their poor
performance in seismic applications. It may be desirable to use anchors that can be
removed and replaced (e.g., concrete screw anchors) if the configuration of the data-
com equipment center will change over time.
There are many standards that include typical details for the anchorage of
mechanical and electrical equipment that can be consulted for additional guidance,
such as ASCE Standard 7-05 (ASCE 2005) and the IBC (ICC 2006). Two special situ-
ations in data centers deserve additional discussion. One is CRAC units that sit on top
of frames to raise them to the level of the RAF, and the other is mechanical or electrical
equipment that requires vibration isolation.
8.3.1 CRAC Units on Frames

Many CRAC units installed in data centers are raised on steel platforms,
commonly referred to as floor stands, so they are flush with the height of the RAF
and to enable the air discharge on the bottom of the unit to be unobstructed. In these
cases, the following recommendations apply:
• The floor stand should be manufactured by the same vendor as the CRAC unit
to ensure that they are engineered to go together.
• The CRAC unit should mate perfectly with the floor stand and be properly
attached to it with bolts and vendor-supplied hardware.
• The floor stand should be rigid enough so that no additional amplification is
transmitted to the unit.
• The floor stand should be rigidly mounted to the concrete floor below with
code-compliant anchors.
• The floor stand should be adequately braced to resist seismic forces.
It should be noted that it is not uncommon for the owner’s engineer to check the
vendor calculations of the vendor-supplied steel pedestal to ensure that it is adequate
for the location where it will be installed. A photo of the floor stand of a CRAC unit
is provided in Figure 8.3.
8.3.2 Vibration Isolation
The building infrastructure includes equipment that is a potential source of trou-
blesome vibration—for example, any equipment that includes a rotary component,
Figure 8.3 CRAC unit floor stand.
such as a compressor or a fan. Consequently, the anchoring methods for such equip-
ment must be configured to ensure that any vibratory energy is not transmitted to the
building structure. In general, anchoring means will utilize vibration isolators.
(Refer to Chapter 10 for a more in-depth discussion on this topic.)
8.4 INFRASTRUCTURE EXPANSION/CONTRACTION

One of the challenges of designing supports for piping systems is that temper-
ature changes in the systems cause the piping system to expand and contract. The
exact expansion is a function of several variables, including the coefficient of ther-
mal expansion of the piping material and the length of a given piping run. The ther-
mal expansion of carbon steel, stainless steel, and copper can be found in
Chapter 41, “Pipes, Tubes and Fittings,” of ASHRAE Handbook—HVAC Systems
and Equipment (ASHRAE 2004). If the system is rigidly supported, such as with
anchors, allowance for pipe expansion and contraction must be made to prevent
either (a) excessive stress on the piping, (b) excessive stress on the anchor supports,
or (c) excessive deflection at take-offs and equipment connections.
From a thermal expansion perspective, piping systems that “float” (i.e., those
that have no anchors) might be considered ideal, as thermal expansion stresses typi-
cally only build up between anchors and other fairly rigid connections (such as
guides or wall penetrations). Almost all piping systems have connections to equip-
ment and other restraints, however, and it is important to know the maximum move-
ment and stresses at these points. As such, anchors are typically added to most
systems to provide a reference point at which no movement is allowed.
8.4.1 Expansion Loops

The most common method of providing for expansion and contraction while
concurrently restraining piping systems is the expansion loop. The expansion loop
takes account of the fact that bends in a piping system provide flexibility and stress
release. As shown in Figure 8.4, a piping expansion loop typically consists of four
elbows and a pipe support called a guide to make sure that the pipe expands in the
intended direction.
Figure 8.4 Typical thermal expansion loop.
In addition to the thermal expansion loop, there are several other types of pipe
shapes that can be used to provide for expansion and contraction, including L-bends
and Z-bends. While computer programs are typically used for detailed thermal
stress analysis, ASHRAE (2004) provides basic equations for loops, L-bends, and
Z-bends in Chapter 41, “Pipes, Tubes and Fittings.” In general, use of bends and
loops to absorb the deflection of the piping system is the preferred method, but in
many cases there is insufficient space to accommodate this expansion, especially
if the design temperature difference is high. In these cases, mechanical expansion
joints can be used.
8.4.2 Mechanical Expansion Joints

There are many types of mechanical expansion joints, and specification of the
proper joint is important to make sure that the joint provides the desired expansion
required without either leakage or catastrophic failure. Types of mechanical expan-
sion joints include packed-slip expansion joints, flexible ball joints, metal bellows
expansion joints, rubber expansion joints, and flexible hoses. As with the expansion
loops described above, mechanical expansion joints typically need guides and
anchors to operate properly, and manufacturers’ instructions must be strictly
adhered to. The Expansion Joint Manufacturers Association has standards that may
be of value in comparing different types of joints (EJMA 2003). A typical type of
expansion joint is a metal bellows joint (shown in Figure 8.5) that has a thin wall
material, is rated for the design piping pressure, and has internal convolutions that
act to absorb the thermal expansion of the rigid portion of the piping system through
flexing and bending.
Additional information on piping expansion can be obtained from both
Chapter 41, “Pipes, Tubes and Fittings,” of ASHRAE Handbook—HVAC Systems
and Equipment (ASHRAE 2004) and Chapter 47, “Sound and Vibration Control,”
of ASHRAE Handbook—HVAC Applications (ASHRAE 2003).
Figure 8.5 Bellows-type mechanical expansion joint.
Raised-Access
Floor Systems
9.1 OVERVIEW
Raised-access floors (RAFs) are often installed within datacom equipment
centers. The RAF directly supports all of the datacom equipment racks and several
elements of the building infrastructure; further, it protects and gives access to all of
the underfloor utilities. Collapse or major failure of the RAF would cause extensive
damage to the datacom equipment and building infrastructure and would certainly
impede egress for any personnel trying to get out. Consequently, the RAF is one of
the most critical elements in any datacom facility.
The RAF systems most prevalent within datacom facilities consist of four main
parts:
1. the removable RAF panel

2. the adjustable pedestal head
3. the pedestal base (sometimes called the riser or stanchion)
4. the stringer
These parts are described in some detail in the following section.
9.2 RAISED-ACCESS FLOOR COMPONENTS

The RAF panel can be solid or, when it is required to allow airflow through it,
it may be perforated or an open grate. The panels can be constructed from aluminum,
steel, or concrete sandwiched by thin steel plate. The most common panel size for
datacom equipment centers is 2 × 2 ft (600 × 600 mm). Thickness varies but is typi-
cally 1–2 in. (25–51 mm).
The floor panels are supported by an “understructure” that consists of the pedes-
tal head and base as well as the stringers (see Figure 9.1). These elements may be
constructed of either steel or aluminum.
The pedestal head consists of a flat section to support the panels and may
include tabs or nubs to ensure proper spacing of the access panels. Often tapped
holes are included in order to add corner lock screws to secure the panels to the head.
62⏐ Raised-Access Floor Systems
Figure 9.1 RAF understructure. Reproduced with permission from Tate

(2003).
On some systems, the four corners of the head have separate tabs for attachment of
the stringers. Attached to the underside of the head is a threaded rod stem that sits
inside the pedestal base. A nut is used to adjust the height of the head when the floor
is leveled. The stem portion of the pedestal head can either be cast aluminum or solid
steel threaded rod.
The pedestal base is a tube with a collar on the top to accept the pedestal head
stem and a base plate on the bottom with holes that are used when the base must be
mechanically fastened to the structural floor. They come in many different heights
from 6 to 48 in. (150 to 1.2 m); currently, the most common finished floor heights
for datacom facilities are 18 and 24 in. (450 and 600 mm). The pedestal head often
rests on the base with no mechanical attachment.
The stringers are attached directly to adjacent pedestal heads either by “snap-
ping” them onto the pedestal heads or by screwing them in. They sometimes are U-
shaped or tube shaped and can be either 24 or 48 in. (600 or 1200 mm) long. Gener-
ally, the stringers are installed to provide lateral stability, but some flooring systems
require the stringers to increase the load bearing capacity at the edges of the RAF
panels.
The RAF panels are placed on top of the stringers and may be gravity-held,
which allows fast access to the underfloor space, or bolted to the stringer system.
Stringerless systems, also referred to as bolt-down or cornerlock systems since
they require that each RAF panel be bolted at every corner to the pedestals directly
below them, are also available, but they are generally restricted to RAFheights of 3 to
24 in. (0.08 to 0.61 m) and typically represent an unsuitable solution for datacom
equipment areas.
9.3 RAISED-ACCESS FLOOR STRUCTURE DESIGN GUIDELINES
9.3.1 Structural Capacity of Raised-Access Floor Panels

One of the main purposes of RAFs within datacom equipment centers is to
support the datacom equipment racks and certain elements of the building infra-
structure.
The structural capacity of RAF panels is classified by several different types of
loading scenarios. These scenarios include the maximum allowable concentrated
load, the maximum allowable uniform load, and the maximum allowable rolling
load (which is usually limited to a specific number of passes). Appendix C provides
more information on the definition of these loads.
RAF panels are most commonly classified by their concentrated load capacity.
The floor panels are available in several concentrated load ratings ranging from
1000 to 3500 lb. (4445 to 15,572 N) and can therefore accommodate a variety of
vertical loading requirements at an optimized cost.
Most installations of datacom equipment will require holes or cut-outs in the
RAF panels; these cut-outs allow for cables to be routed through the floor, for the
datacom equipment to be connected to the slab below, etc. Depending on the size and
locations of these holes or cut-outs, the allowable load limits will be as much as 50%
lower than the limits listed by the RAF panel manufacturer.
Where a panel with a hole or cut-outs may be subjected to a high magnitude of
equipment loads or rolling loads, standard practice is to install two additional
support pedestals at opposite sides of the cut-out. Panels with round grommet holes
5 in. (127 mm) in diameter or less need no additional support.
9.3.2 Lateral Capacity of RAF Systems

The loads placed on the RAF panels are transferred via the understructure to the
structural floor system below. Similarly to the floor panels, the RAF pedestals are
available in several load ratings to allow an economical selection to be made based

on the structural capacity required.
The pedestal systems typically are able to comfortably handle the vertical load
requirements; since a pedestal is placed at each corner of every RAF panel, the
required vertical load capacity of the pedestals matches the concentrated load of the
floor panel (plus the self-weight of the system).
The lateral stability of the floor system is often of more concern than the capac-
ity to support vertical loads. The magnitude of the lateral loading is a function of the
height of the RAF, the weight of the equipment supported by the RAF, and the seis-
mic design requirements within the locality of the datacom equipment center.
There is no uniform guideline regarding the height of a RAF system, but the
general range is from 18 to 36 in. (0.45 to 0.91 m); there is currently a trend toward
deeper systems, e.g., 30 to 48 in. (0.76 to 1.22 m), as a means to support the
increased power and cooling requirements of the latest generation of datacom
equipment.
As this trend continues and as the weight of the equipment required to be
supported by the RAF also increases, there is the likelihood that structural framing
systems (consisting of short columns and structural steel beams, potentially with
moment-resistant connections) will begin to gain popularity over the more tradi-
tional pedestal-and-stringer–based understructure. For the purposes of this discus-
sion, however, we will continue to present the more commonly specified pedestal-
and-stringer–based understructure.
Lateral loads are resisted by the RAF system by one of two main methods:
1. Fixed-Base Pedestals: The lateral load is resisted simply by the base pedestal
cantilevering off of the structural floor. Figure 9.2 illustrates an installation that
utilizes fixed-base pedestals.
2. Underfloor Bracing: Supplemental lateral load resistance is provided by the
addition of diagonal braces to transfer lateral load directly to the structural
floor. Figure 9.3 illustrates an installation of underfloor bracing.
9.3.3 Structural Overview of Fixed-Base Pedestals

The pedestals of RAF systems are typically attached to the structural floor slab
via either epoxy/adhesive connections, mechanical anchoring, or a combination of
both methods. Since the pedestal anchor bases are fixed, the pedestals themselves are
able to act as cantilevers and can resist lateral loads.
In stringerless systems, all of the horizontal forces are transferred from the
equipment on the RAF and from the mass of the floor panels, which are bolted to
the pedestals, directly to the top of the pedestal and then through the pedestal in
bending into the slab on grade. Care must be given to the positioning of the pedes-
tals so that they are at least 12 in. (305 mm) from any control joints in the structural
floor slab.
When stringers are installed, the lateral load can be transferred via the stringers
to a number of pedestals. Consequently, a more even distribution of loading occurs
and the overall floor loading may be used as the basis of the structural evaluation
Figure 9.2 Standard RAF understructure with cantilevered fixed-base

pedestals being installed.
Figure 9.3 Braced RAF understructure.
(rather than a detailed analysis of the individual loads of every component located
on the RAF).
However, the main resistance to lateral loading remains the cantilever aspect of
the pedestals. Although theoretically the lateral loading can be transferred to the
perimeter of the RAF area and picked up by structural elements at those locations,
this is usually not the most practical approach because the floor areas of many data-
com equipment rooms are so large that the cumulative force would be massive.
It is important to note that the stringers themselves typically have a small cross-
sectional area and therefore have a limited resistance to bending. When the RAF
panels are installed, the panels and stringers interlock to form a structural
diaphragm. During the construction and installation of the building infrastructure
below the RAF, there may be a number of RAF panels removed and the diaphrag-
matic action may therefore be compromised.
Although it is rare for structural failure of the RAF system to occur as a result
of too many RAF panels being removed, it is quite common for the understructure
to shift slightly, or “rack.” When this racking occurs, it becomes difficult to properly
replace the RAF panels that were removed. Underfloor bracing systems can help
eliminate the potential for racking.
9.3.4 Structural Overview of Underfloor Bracing

The installation of underfloor bracing systems provides additional resistance to
the lateral loads imposed on a RAF system.
The bracing typically consists of diagonal members that are installed to transfer
the lateral load directly to the structural floor. The bracing can take the form of
“kicker” braces that are connected to four sides of an individual pedestal (refer to
Figure 9.3) or longer diagonal braces, or X-braces, that span selected pedestals up
to 20 ft (6 m) apart.
Bracing is generally recommended for RAFs of 24 in. (609 mm) in height or
greater and is ideally installed in two directions, both parallel and perpendicular to
the rows of datacom equipment. Bracing systems can be intermittent as long as a
stringer system is continuous between braces; this pattern is important as continuous
bracing may conflict with the installation of any underfloor building infrastructure
(power distribution, piping, etc.).
9.4 SEISMIC PERFORMANCE OF RAISED-ACCESS FLOORS
9.4.1 Historic Seismic Performance and Testing of

Raised-Access Floors
Unfortunately, there are not many data available for the performance of RAFs
during large earthquakes. One documented study was conducted by the Earthquake
Engineering Research Institute immediately following the 1989 Loma Prieta Earth-
quake that occurred near San Francisco, California. The data is difficult to normalize
since the floors were all different vintages and different floor heights, had different
manufacturers, utilized different lateral bracing systems, and were located through-
out the Bay Area so the ground acceleration varied tremendously. None of the floors
appear to have been subjected to ground motions that would be characterized as the
“design-level” earthquake. There were, however, some interesting observations. Of
the 25 cases reported, only 7 had “significant” damage and the other 18 had only
“minor” damage. Damage to RAF panels was reported in several cases, but no
damage was reported to pedestal bases or anchorage. Most of the damage was the
result of shifting contents on top of the floor (in some cases up to 36 in. [0.9 m]) fall-
ing or damaging other equipment or the floor. Much of this equipment was on casters
that were unrestrained (EERI 1990).
A number of laboratory tests on RAFs and their attached components were
conducted utilizing a computer-controlled dynamic shake table by the Finance,
Insurance and Monetary Services (FIMS) Committee of the Federal Emergency
Management Agency (FEMA). These tests showed that the RAF can have a brittle
(not ductile) quality and may not have additional reserve capacity beyond yielding
or fracture of the first element (FIMS 1987). Several other tests have shown that
panels can “pop out” of the floor system during strong motion.
9.4.2 Recommendations for Improved Seismic Performance of

Raised-Access Floors
When purchasing a RAF for a data center located in an area of high seismicity,
there are a number of items that should be specified to ensure good performance:
• The pedestal base plate should be anchored to the concrete floor with drilled-
in or cast-in-place anchors; lateral forces should not be transferred by friction,
shot-in anchors, or adhesive mastics.
• Avoid stringerless systems, as they often do not have a verifiable load path.
• Floor stringers should be designed to carry axial seismic loads (for at least
one panel) and should be mechanically fastened to the pedestal heads.
• Bracing (if used) should be designed to avoid a buckling mode of failure and
should utilize structural or piping shapes and not electrical conduit.
• If it is constructed of steel, the pedestal riser should be welded to the pedestal
base plate with regular welds and not brittle resistance welds; the user should
also consider requiring physical tests of the pedestal riser for a minimum lat-
eral load.
• All floor panels with equipment attached directly that transmit shear to the
panel should be corner locked with screws to the pedestal head.
• All floor panels with equipment attached directly that transmit overturning to
the floor system (i.e., equipment that does not have its own system holding it
down to the concrete) should not utilize “slip-on” pedestal heads but rather
should have a mechanism to transfer uplift to the pedestal base.
• Pedestals should have maximum deflection limits.
• Adding corner lock screws to all panels in egress pathways should be con-
sidered.
These recommendations are mostly directed toward ensuring that lateral loads
generated by equipment resting on and attached to the RAF floor gets transmitted
to the concrete floor below using a verifiable and sufficient load path. It should be
noted that the above recommendations are good practice even in areas of moderate
or low seismicity. Many of the above recommendations came out of the FIMS guide-
line published in 1987 (FIMS 1987). This document was important because it was
one of the first to go beyond what had been typical industry practice at the time.
Several of these recommendations have been incorporated into building codes as a
better performing alternative to standard RAFs (see the Special Access Floors
section below).
9.4.3 Recommendations for Determining the Adequate Design of

Raised-Access Floor Anchorage
Calculations for the proper anchorage of RAFs for data centers should be
performed by a registered structural engineer or a professional engineer with expe-
rience in earthquake design. Although the anchorage seems simple, there are a
number of areas where mistakes can occur. Some items to consider when performing
or reviewing these calculations are:
• When calculating the component operating weight Wp (discussed in more

detail in Section C.4 of Appendix C and Appendix E) for the raised-acces
floor system, the code requires 25% of the rated-floor live load plus a 10 psf
(0.5 kN/m2) partition allowance to be added; consideration should be given
to using 100% of any equipment attached directly to the floor and 25% of
any equipment supported by but not attached to the floor.
• In lower seismic zones, sometimes not every pedestal base is anchored to the
floor; in such cases, care should be taken to ensure an adequate load path
through the stringers to the pedestal bases that do have anchors.
• In some cases, only two anchors (instead of four) are used to attach the pedes-
tal base plate to the structural floor; these anchors should always be installed
in opposite corners; care should be taken with the direction of loading—it is
often overlooked that the most critical case is overturning perpendicular to the
line through the two anchors.
• When calculating the load on the floor, care should be taken to include both
horizontal and vertical impacts if the system is expected to resist overturning;
it is not conservative to use the design vertical load when doing this check—
the maximum force that can be used is the same Wp used to calculate seismic
horizontal force Fp (discussed in more detail in Section C.4 of Appendix C
and in Appendix E).
• When calculating the force on the anchors, be sure to combine the effects of
tension and shear; often the manufacturer’s anchor capacity report will pro-
vide an interaction ratio—if not, use the one in the applicable building code.
• When calculating the tension force on the anchors, any increase due to the
effects of prying action should be included.
9.4.4 Special Raised-Access Floors

Special RAFs is a new concept that was added to the IBC (ICC 2003) and ASCE
Standard 7-02, Minimum Design Loads for Buildings and Other Structures (ASCE
2002). The intent of the specifications in these codes is to encourage owners and
engineers to use RAFs with good seismic detailing (design features or reinforce-
ments to overcome seismic load). The benefit of using a special RAF is that the value
for the component response modification factor Rp (discussed in more detail in
Appendix E) has been maintained at 2.5. Conversely, if special access floors are not
used, the Rp value has been lowered to 1.5, which will mean a 40% increase in
anchorage forces. In order to qualify as a “special” access floor, the following five
items must be included:
• Connections transmitting lateral forces must consist of mechanical fasteners,

anchors meeting ACI 318 (strength, design of post-installed anchors) (ACI
2001), welding, or bearing.
• Seismic forces cannot be transmitted by friction, powder-actuated fasteners,
or adhesives.
• The analysis of any bracing system includes the destabilizing effects of braces
bucking in compression.
• Bracing and pedestals are structural or mechanical shapes; electrical tubing
shall not be used.
• Floor stringers that are designed to carry axial seismic loads must be mechan-
ically fastened to the supporting pedestals.
10
Vibration Sources
and Control
10.1 OVERVIEW OF VIBRATION SOURCES
The building infrastructure includes equipment that is a potential source of trou-
blesome vibration. Chapter 10 provides an overview of sources of vibration and the
control of vibration.
Within a datacom equipment center, some representative vibratory sources
include standby generators, cooling towers, air-handling units, chillers, and CRAC
units. In fact, any equipment that includes a rotary component such as a compressor
or a fan is a potential source of vibration.
Although there is the potential for vibration caused by a component of the build-
ing infrastructure to be transmitted to the datacom equipment room, most datacom
equipment is fairly tolerant to vibratory loads and there is generally little risk of fail-
ure being caused in this manner. Of more concern is the impact that uncontrolled
vibrations will have on the equipment that is the vibratory source and any distribu-
tion systems that are directly connected to this equipment. The building infrastruc-
ture is by its very nature critical to the successful operation of the datacom
equipment, and any failure in the building infrastructure, such as the loss of power
or cooling, will directly impact the datacom equipment.
It is therefore important to address the vibration isolation of building infrastruc-
ture equipment within a datacom equipment center. Fortunately, the means of vibra-
tion isolation are fairly well known in the industry and there are several publications,
such as the ASHRAE Handbooks, that provide a more in-depth study of this topic.
For the purposes of this publication, we only consider vibration isolation for building
infrastructure equipment, which is located within or immediately adjacent to the
datacom equipment room.
10.2 OVERVIEW OF VIBRATION ISOLATION

Vibration isolation is a method of limiting the unbalanced forces created by a
vibratory source from entering a structure (positive vibration isolation) or of limiting
the structure from passing vibratory energy into a sensitive component (negative
vibration isolation).
72⏐ Vibration Sources and Control
In building infrastructure installations, the required method of vibration isola-

tion is almost exclusively positive isolation. For example, positive vibration isola-
tion would protect the structure from the vibrating source of a large centrifugal fan,
as shown in Figure 10.1.
Isolators can be broadly divided into two simple categories: limited deflection
isolators, which are mainly used for high, disturbing frequency attenuation, and
higher deflection isolators, which are able to attenuate lower disturbing frequencies.
The term disturbing frequency refers to the lowest operational rpm of the rotat-
ing component that is the source of the vibratory energy. While gravity will have its
greatest negative effect on the heaviest moving part of a component, created either
by weight, manufacturing tolerance, or wear, it is proper that the disturbance is
equated with rpm because that event can only occur once with each revolution.
Limited deflection isolators include conical resilient elastomers molded from
different compounds with deflection characteristics ranging from 0.25–0.5 in.
(6.25–12.5 mm). When used and placed in an above-grade location, these isolators
can effectively attenuate disturbing frequencies from an isolated component of
1200 rpm or higher.
As the rotational rpm moves lower or the speed of the rotating component is
controlled by a variable-frequency drive system, springs isolators are used. Higher
spring deflections, l–6 in. (25–150 mm) restore the proper spatial relationship or
ratio between a component’s disturbing frequency and the isolator’s resonant
frequency. Isolator efficiency equations in use since the 1940s have established this
ratio as 3:1, with the isolator moving at a much lower rate than the disturbance.
Which type of isolator should be used, its deflection, and in what applications it
Figure 10.1 Positive isolation of a vibratory source.
should be used often has much to do with the dynamics of the isolated component
and where it is located within the building.
The component’s location—grade or above, indoor or outdoor—enters into
isolator selection. An isolator’s ability to isolate is a function of the rigidity or stiff-
ness of the mass that it is being supported by (e.g., the structural floor system).
For isolators intended for use in above-grade applications, the intrinsic live-load
deflection of the floor must be added to a selected spring deflection that is properly
sized for the component’s disturbed frequency to overcome the lack of stiffness
inherent by design. This relationship is clearly defined in Chapter 47 of ASHRAE
Handbook—HVAC Applications (ASHRAE 2003).
Vibration isolators must not only isolate and restrain equipment but also must
have the ability to accept high transient dynamic loads. These transient loads can be
induced by seismic events, wind, or man-made sources.
10.3 SELECTION OF VIBRATION ISOLATORS

The two critical considerations previously mentioned help with isolator selec-
tion. Another consideration is that vibratory energy from new equipment will be to
some extent different than the energy forces that will be evident after two years of
24-hour/7-day operation. The vibration produced from a component will be notice-
ably worse as internal bearing and shaft wear and tolerances increase. Regardless of
the best of maintenance programs, restoration of components to like-new condition
is very unlikely.
10.4 VIBRATION ISOLATION WITHIN THE

DATACOM EQUIPMENT ROOM
As discussed previously, a typical datacom equipment room may house many
CRAC units as part of the cooling system. Oftentimes, these CRAC units contain a
number of compressors, and each one is a potential source of troublesome vibration.
Although most CRAC units are internally isolated on limited deflection devices,
utilization of these devices is hindered by both isolator space limitations within the
unit itself and by the flexible connectors that are used to connect internal piping to
the compressors. It is not uncommon, for instance, for a 14 in. (356 mm) long flex
connector to have only 6 in. (152 mm) of live usable length since the end ferrules and
connections subtract from its overall length. The connector is therefore acting more
as a misalignment tool rather than as an isolator. If this condition exists, the connec-
tor is incapable of reducing the compressor’s vibratory energy from exciting the
unit’s casing and ultimately the structure. This vibratory energy, coupled with that
of the un-isolated fans along with the air that they produce, moving through a CRAC
unit’s pressurized casing could easily present a high operational vibration environ-
ment within the datacom equipment room.
10.4.1 Datacom Equipment Room at Grade

Vibration control for CRAC units within data center rooms that have a structural
floor slab on grade is fairly simply addressed through the use of limited deflection
elastomers (often referred to as neoprene pads).
74⏐ Vibration Sources and Control
10.4.2 Datacom Equipment Room Above Grade

The situation is somewhat more complicated in instances where the structural
floor slab is elevated. It is typical for datacom equipment centers to have fairly large
distances between columns (e.g., 40 ft [12.19 m]) and, as a consequence, the struc-
tural floor system will experience a fair amount of deflection. For example, if the
deflection is limited to the span length (L) divided by 360, a 40 ft (12.19 m) span
would yield a deflection of 1.3 in. (33 mm) under load.
To effectively control the transmission of vibrations in this instance, the resil-
iency or deflection of the isolator must overcome the structural floor deflection as
well as the unbalanced forces present in the CRAC unit. The increased deflections
inherent to springs make them the isolator of choice when equipment is placed in an
above-grade installation. In most applications, springs selected from a nominal 2 in.
(50.8 mm) deflection series are suitable, giving proper consideration to both the
influence of the floor and the vibratory energy produced by the component.
10.5 VIBRATION ISOLATION ADJACENT TO THE

DATACOM EQUIPMENT ROOM
The following section describes some practical guidelines for the isolation of
building infrastructure equipment that is located outside but immediately adjacent
to the datacom equipment room.
As stated previously, any equipment that includes a compressor or a fan is a
potential source of vibration. In order to effectively isolate such equipment, vibra-
tion isolators should be placed below the equipment, and any connections to distri-
bution infrastructure (such as ductwork, piping, conduits) should be isolated through
the use of flexible connections.
Whenever a datacom equipment center is located in a seismic designated loca-
tion, a consideration should be made for adopting IBC compliance for all building
infrastructure components, including isolators, subject to seismic loads. It is also
important to institute quality assurance field inspection procedures.
10.5.1 Vibratory Sources on Common Slab on Grade

The first case of vibratory sources is where building infrastructure equipment
is located on the same structural slab on grade as the datacom equipment. In all of
those instances, vibratory energy could be a factor experienced on the center floor.
As such, 1 in. (25 mm) deflection isolators for all vibratory components located in
these areas would be applicable, as we would not have to consider floor deflection
affecting isolator performance.
Base-mounted pumps should also be provided with concrete-filled inertia bases
to limit startup amplitudes.
10.5.2 Vibratory Sources on Floor Above

The second case is where building infrastructure equipment is located directly
above the datacom equipment room. In this case, the inherent spring-like floor
deflection of the building structure is a consideration. In order to prevent vibratory
energy from reaching the floor below, it is recommended to universally select 2 in.
(51 mm) of isolator static deflection for all components except the boiler, if appli-
cable, where 1 in. (25 mm) will suffice.
10.5.3 Vibratory Source on Roof Above

In this example, building infrastructure equipment is located on a roof that is
directly above the datacom equipment room. Because of the wind loads imposed on
roof-mounted components, horizontal restraints are a requirement in addition to
vertical restraints.
The 2006 IBC (ICC 2006) enacted additional requirements for both roof-
mounted components and their installed vibration isolators. Within certain coastal
and some inland locations, components mounted on structures having a finished
height of less than 60 ft (18.3 m) are subjected to wind loads including moment and
shear nearly two times their weight. These dynamic forces have dramatic effects on
the isolator, its mounting surface, and the isolated components. Therefore, vibration
isolators must demonstrate their ability to handle dynamic force loads as much as 2
G’s (2× the component weight), calculated to the full height of the component plus
its mounting and structure, which could result in rather high overturning loads.
Equipment that contains compressors and small-diameter fans, such as air-
cooled condensers and air-cooled chillers, is generally satisfactorily isolated by 2 in.
(51 mm) of static deflection, as this is sufficient to overcome the lack of stiffness of
the roof.
Most cooling towers include relatively large variable-speed fans, and as they are
always exposed to the outer environment, their unbalance greatly increases with age.
As such, higher deflections ranging from 3–5 in. (76–127 mm) are typical.
As curb-mounted HVAC units move large quantities of air through a relatively
confined space, their ability to transmit vibratory energy into the structure is high.
Effective isolation can be provided through a structural curb with 3 in. (76 mm) static
deflection spring isolators.
Part IV
Datacom Equipment
11
Shock and Vibration Testing

on Datacom Equipment
acceleration of gravity (g or G): g is the acceleration produced by gravity at the
surface of the earth and has an international standard value of 9.807 m/s2. Measured
accelerations are often expressed as a ratio of the measured acceleration divided by
gravitational acceleration; this unitless ratio is expressed as a G value.
amplitude or magnitude: the maximum value of a measured vibration. Amplitude
may be measured in values of displacement, velocity, or acceleration.
frequency: the reciprocal of a period in cycles per second, sometimes expressed as
Hertz (Hz).
free-fall drop height: the equivalent velocity change due to free-fall drop from a
certain height (h):
velocity change ΔV = (1 + e) ( 2gh )
where e is the coefficient of restitution and is the ratio of velocities after and before
an impact.
half-sine shock pulse: an ideal shock pulse for which the acceleration/time relation
has the shape of the positive (or negative) section of one cycle of a sine wave.
power spectral density (PSD): a measure of the power content with respect to the
frequency for random vibration. The units are g2/Hz and represent the power present
in a 1-Hz-wide square filter. The square root of the integration of the PSD curve with
respect to frequency gives the overall root mean square level of the vibration.
pulse width: half of the period of a sinusoidal wave expressed in milliseconds (ms).
random vibration: a vibration having randomly varying amplitude and frequency
content within specified limits. It is typical of the vibration to which products are
exposed in the operating environment.
resonance: the point at which the natural frequency (resonance frequency) of an
element is equal to the excitation frequency and the point that produces the peak
80⏐ Shock and Vibration Testing on Datacom Equipment
amplification of input amplitude; a condition caused by excitation with an input

frequency that is at or very near an element’s natural frequency. The resonance
frequency causes a maximum output or element amplitude response.
response (of a system): a quantitative expression of the output of the system as a

function of the input.
response spectrum: a method of predicting the response of mechanical systems

(buildings, machinery, computers, etc.) to vibration or shock inputs. The input wave-
form is mathematically applied to a series of single-degree-of-freedom spring/mass
oscillators with a defined amount of damping. The response of the series of oscil-
lators is plotted versus frequency. The effect of the input waveform on systems
having various natural frequencies can then be determined by inspecting the plot (the
response spectrum).
root mean square (rms): a value of a set of numbers that is the square root of the
average of their squared values; a technique to determine the average magnitude of
a dynamic signal, giving the effective energy or power level of that signal.
sinusoidal vibration: a vibration with a sine waveform. Only one frequency is

present. Typical sources of sinusoidal vibration are rotating machinery and electrical
power equipment.
transient vibration: a vibration that has a high amplitude and is of short duration,
typically only a few cycles, and unlikely to produce resonance.
11.2 OVERVIEW OF VIBRATION SOURCES

Whether operating, renovating, or designing a data center facility, it is important
to consider the effects that shock and vibration loads have on the contents. Today,
most datacom equipment and infrastructure equipment manufacturers design prod-
ucts that are capable of tolerating normal ambient vibrations that may occur within
the data center. In many cases, the equipment can survive more severe conditions if
it has been properly secured. The greatest risk for damage for unsecured equipment
comes from overturning and movement of datacom equipment, especially in data-
com facilities located in seismic event areas.
Lateral loads, tipping loads, and shock and vibration loads are not only a conse-
quence of naturally caused events such as earthquakes or wind loads but can also
result from human-created sources. For example, mechanical devices and construc-
tion activities create additional vibration sources within the building. Shock and
vibration loads reaching datacom equipment are typically transmitted to the equip-
ment by way of the building floor. Floors can be excited from sources outside of the
building several miles away or from sources inside the building located just a few
feet away. Some representative vibratory sources may include generator units, recip-
rocating compressors, large unbalanced fan units, cooling distribution units (CDUs),
CRACs, major roadways, street cars or trains crossing rails, forklifts, and nearby
servers. These types of sources can continuously excite the building floor. The
following intermittent and temporary operational sources may also exist and can
have harmful effects on equipment.
• Intermittent shock and vibration from construction or demolition activities

within the building or from neighboring sites.
• An implosion caused by blasting of an older building.
• Pile drivers working in an adjacent property.
• Jackhammering of concrete materials.
• Large earthmoving vehicles grading earth.
Any one of these activities can generate a great deal of ground motion with
disturbances that can travel and be sensed over great distances. The State of Cali-
fornia Department of Transportation (CalTrans) conducted a study (Eagan et al.
2001) between June 1999 and May 2000 to record and document dynamite blast
ground movements. CalTrans established the safe level for ground movement at a
peak particle velocity of 50 mm/s when blasting in urban area. In the CalTrans study,
two recorded events exceeded the established level, with levels at 70 mm/s peak
particle velocities recorded at distances of 18 and 30 m from the blast site.
The effects of vibration transmitted through the building floor can be even more
severe if the floor is excited at resonant frequencies of the equipment or building. For
example, vibratory pile drivers drive their piles at 1600 vibrations per minute and are
very likely to excite nearby buildings. Pile driving works by liquefying the soil next
to the pile. This allows the pile to move easily down into the ground. The amplitude
of the building’s floor can be multiplied to a level where data center equipment walks
across floors, equipment shakes to levels at which soldered joints crack, components
are caused to separate from holders, or heavier power components fail.
In addition to equipment component damage from severe vibration levels, there
are other risks associated with the equipment. The displacement of racked equip-
ment can separate cables from their connectors or disengage power cords from the
equipment. Datacom equipment on RAFs can move across the surface and result in
a rack foot catching or falling into a cable cutout or even tipping over. Even the very
heaviest of datacom equipment and infrastructure equipment can move under
continuous vibratory conditions.
Although not directly related to shock and vibration, a rack tipping over is an
issue that can impact service within the data center. Equipment racks can also over-
turn from unbalanced weight within the cabinet. It is common for telecommunica-
tions equipment, networking equipment, routers, or UPS equipment to weigh 300–
400 lb (1335–1780 N) per shelf, and a larger server can weigh up to 3600 lb (16,017
N). With the majority of the equipment weight supported by the front set of mounting
rails, the cabinet becomes front weight biased, which increases its chances for
tipping over. Large overhead loads that some equipment racks support also add to
the problem. Cables or cooling equipment placed over the racks moves the center of
gravity of these cabinets to an unstable configuration. An accidental bump or an
inadvertent push tips over these cabinets more easily.
It is common practice to install equipment in cabinets on slide-out drawers. This

feature provides easy maintenance access to drives or boards; however, pulled out
drawers create a very unstable configuration for the cabinet. Some cabinets are
equipped with front anti-tipping feet to prevent tip-over. Racks not incorporating
these features are at higher risk of overturning. Securing racks will eliminate tip-over
dangers.
In reality, there are no regions in the continental United States that will be totally
immune from earthquake risks. The majority of the regions in the United States are
not as active as California or Alaska, but all regions have experienced ground
motions from geologic faults, volcanic activity, or landslides. The area in the Central
United States near New Madrid, Missouri, has been the site of some very powerful
earthquakes. The Chicago, Illinois, area is designated as a zone 0 risk area in the
Uniform Building Code (UBC) (ICBO 1997), but even that unlikely region has expe-
rienced 30 significant earthquakes, with the latest event recorded in 1984. In 1974
in a northern Texas area, another designated zone 0 experienced an earthquake
having a magnitude measuring 4.5 on the Richter scale. The event was strong enough
to crack concrete structures and break windows and was felt over a wide area of the
Texas Panhandle.
Other events that can excite movement of structures can come from forces not
related to ground motions. Strong windstorms, tornados, and hurricane-force winds
are other sources that can create significant building vibration levels that travel
through the floor, which, in turn, can create lateral loads. No structure exposed to
wind is immune from these vibration sources.
Building shapes are typically not very aerodynamic, and wind movements
across bluff bodies produce considerable resistance. This wind resistance creates
oscillations that can be felt as vibrations within the building. In many coastal areas,
building codes require that structures be designed to tolerate wind-excited effects
and oscillations for wind speeds to up 110 mph (177 km/h).
Throughout history and in all cultures there have been periods of political insta-
bility, and in today’s world it is not any different. Unfortunately, today’s data center
managers and operators must consider targeted activities by terrorists and other
saboteurs targeting IT facilities. Attacks on critical infrastructures by bombing or
use of other devices of mass destruction are very real. Data centers by their nature
are important business infrastructures for governments and corporations for main-
taining and carrying on business operations. In recent years, explosive devices have
been the main threat to data centers and their contents. The building and data center
should be designed for the possibility of explosions; the equipment inside the build-
ing must be secured against the effect from a large blast. Backup facilities and recov-
ery plans or similar systems of mission critical government and business functions
must be planned and implemented for continuous, uninterrupted operations. Mission
critical describes a system that is absolutely necessary; an activity, device, service,
or system whose failure or disruption of normal business will result in the failure of
business operation (e.g., the communication system of an online business). Depend-
ing on the nature of the services, there may be facilities that require protection
against possible nuclear device threats, as stipulated by some federal government
agencies. For most commercial enterprises, however, the practices for securing
equipment for natural and man-made risks should be satisfactory.
Considerations for reducing shock and vibration risks within the data center
should focus on securing cabinets as the primary defense against vibratory effects.
Shock and vibration isolation, shock absorption, or other mitigation techniques may
be considered but are generally more costly and more difficult to design. Isolation
design and techniques typically require assumptions on ground motion amplitudes,
frequencies, and providing additional space for the device displacement. The device
becomes less effective working outside of these assumed conditions. In very dense
telecom environments, the space necessary for the movement of an isolator may not
be easily provided. The equipment must be permitted to move freely in all horizontal
directions, which will be summarized in later chapters. In nonisolated installations,
cabinets that have been secured from tipping over or impacting adjacent objects at
least permit the equipment to operate up to its designed ambient vibration levels or
Network Equipment-Building System (NEBS) environmental vibration conditions.
It has been demonstrated using shake table tests that equipment will remain func-
tional up to its design conditions when properly installed.
11.3 DATACOM EQUIPMENT SHOCK AND VIBRATION TESTING

Datacom equipment must be designed and constructed to operate in environ-
ments where ground disturbances and normal shipping and handling conditions are
encountered. Therefore, to demonstrate a product’s capabilities, laboratory tests are
typically conducted to confirm that the product can operate in these situations. It is
difficult to determine the appropriate test levels to encompass all field conditions and
situations. Products should be tested to the most severe conditions expected, but
overdesigning should be avoided to keep products cost effective. Datacom equip-
ment designers should have an understanding of their customers’ operating environ-
ments and design the products accordingly. An extreme environmental example
might be a military application where the equipment could be utilized in extreme
acceleration, shock, and vibration conditions. These environments may require very
robust product construction and special packaging to survive. Test levels for mili-
tary-grade equipment are defined using specifications such as MIL-STD-810F, Envi-
ronmental Engineering Considerations and Laboratory Test Method Standard,
(DOD 2000) and MIL-STD-202G, Test Methods of Electronic and Electrical
Component Parts (DOD 2002). For most commercial datacom product applications,
the test levels are not always clearly defined and are typically left to the equipment
manufacturers to determine. Manufacturers establish their own sets of design criteria
and specify those parameters in their published literature (e.g., IBM [1992b]; Noto-
hardjono et al. [2001]). Equipment designers usually provide adequate shock and
vibration margins to cover most user applications. Different datacom specialty
sectors, such as the telecommunications industry, which has large-volume users and
special operating needs requiring unique design parameters, define their own
specific sets of test levels. Design documents used for the telecommunications indus-
try are the Telcordia Technologies, Inc., NEBS GR-63-CORE (Telcordia 2006) and
ANSI T1.329-2002, Network Equipment—Earthquake Resistance (ANSI 2002).
Datacom products are designed in sizes ranging from a small desktop box to
multiple shelves filling a cabinet. Datacom equipment can weigh from as little as a
few pounds to several thousand pounds per shelf space. Shock and vibration testing
is more difficult to determine and test since these days the norm is to mix and match
products from various manufacturers to form a system that provides specific
services. Not all equipment frames are designed alike, and the installation of data-
com products in an open relay rack versus a closed cabinet will have an effect on
shock and vibration performance. Even among open relay racks, there are designs
with very light gauge aluminum uprights in addition to much stronger welded steel
designs. These differences have a direct effect on how well a product performs when
subjected to shock and vibration tests. Cabinets with side panels and doors will
perform better than open cabinets; lightly loaded datacom equipment frames will
perform better than heavily weighted frames.
Unfortunately, there are no established industry standards that provide product
testing guidance. Therefore, datacom equipment testing parameters must be devel-
oped that address all the configurations of the datacom equipment that may exist.
The question to consider is: Will the end user be content with individual product
testing, or will the end user need overall system testing with all of the components
working together? The datacom equipment designer will need to determine the
users’ requirements for demonstrating the datacom equipment’s capabilities and
establish an agreement on how the tests will be conducted with the pass/fail criteria
clearly defined.
Larger mechanical equipment such as HVAC equipment is much more difficult
to test. There are limited test laboratories with large shock and shaker tables avail-
able to conduct tests on large, heavy equipment. There may also be less need to test
mechanical equipment because of the design of these products. The major compo-
nents of these products are much more robust steel assemblies designed to operate
at high hydraulic or air pressures. One area where tests should be required and
performed on large mechanical equipment is the testing of equipment controls,
which includes circuit panels, monitoring equipment, gauge sets, and other electrical
devices governing the operation of the equipment.
Shock and vibration tests should be conducted to demonstrate the products’
capacity to sustain shock levels associated with transportation from the manufac-
turer’s factory, receiving the product at the user’s facility, and installation and oper-
ation in the user’s equipment room. Therefore, several tests need to be designed for
all datacom products. For transportation purposes, boxed and unboxed product
conditions must be able to tolerate shock and vibration loads related to moving the
product from the storage site onto a truck or train. The product will need testing to
simulate being transported across the country on rough roads with very little cush-
ioning provided by the suspension of a truck or train. At many points during its life-
time, the product may be dropped or mishandled (when arriving at its destination,
for example, or when being transported from one floor of the building to another)—
consequently, the product will need to tolerate all movement expected at the data
center. When the datacom equipment is installed within an equipment framework,
the product will live out its life either in a calm, gentle existence or subjected to low-
level steady ground movement with an occasional larger bump from an earthquake
or accidental impact on the frame.
Shock and vibration tests are usually broken down into four conditions: 1) oper-
ational shock and vibration, 2) seismic simulation, 3) ruggedness or fragility, and 4)
transportation shock and vibration. Each test has specific levels and defined param-
eters unrelated to the others. For example, office vibration levels are usually lower
than transportation vibration levels; office tests are run with equipment powered
while transportation testing does not require powering equipment. The seismic
simulation test, which can be categorized as an operational vibration test, has differ-
ent frequency content requirements and significantly higher amplitudes at a
frequency below 10Hz than do operational office tests.
11.3.1 Operational Shock and Vibration

Operational shock and vibration tests define the product-dependent environ-
ments in which products must satisfactorily operate. Shock and vibration sources are
typically produced by environment conditions. For example, nearby operating
equipment, people walking past equipment or dropping materials, the effect of rock
blasting, and transportation vehicles passing near the datacom building are all exam-
ples of vibration contributed by the environment. Shock and vibration sources can
sometimes be self-induced (vibration from fans, blowers, compressors, hard drive
spindles, etc.). Large cooling units such as compressors, water pumps, and powerful
blowers can also transmit vibration levels high enough to impact datacom equipment
operating function. For operational shock and vibration, criteria and tests are neces-
sary both for continuous ground vibration conditions and for simulated earthquake
events. During the operational shock and vibration test, the product will be powered
and operational. These tests verify that the equipment remains functional and is not
affected by the test table motions. The product to be tested should be configured with
similar loads and cabled as if installed in the users’ datacom facilities. End users’
operational vibration test requirements are described in more detail in Appendix D
or are tested using documents such as MIL-STD-810F (DOD 2000) or NEBS GR-
63-CORE (Telcordia 2006) or equipment manufacturers’ internal specifications.
Figure 11.1 shows a typical level that might be found and can also be used as a test
level. In this figure, level 1 is intended for testing floor-mounted datacom equipment
and level 2 is the appropriate level for small datacom equipment normally used on
desks or tabletops or portable and wall-mounted equipment. Typically, the opera-
tional shock test parameter is five 3.5g half sine pulse with a pulse width equal to
3 ms in all three axes (IBM 1990).
11.3.2 Seismic Simulation Test

Testing for earthquake events imposes much higher ground acceleration values,
greater ground displacement for shorter durations, and lower frequencies. There are
several test waveforms available for conducting seismic simulation. Telcordia’s
NEBS GR-63-CORE (2006) developed a generic waveform used for shaker table
tests that simulate a worst-case event of equipment installed in an upper floor of a
Figure 11.1 Typical power spectral density of an operational vibration

encounter in a data center. Reproduced with permission from
IBM (1990).
structure. The waveform contains a short period of very strong motions with periods
at lower levels of ground motion. Peak ground acceleration approaches 1.6g at the
test table, and when taller equipment is mounted to the test table, acceleration levels
could be amplified to much higher levels. Based on studies of seismic events in Japan
(IBM 1992b), IBM Corp. developed seismic simulation testing parameters for IT
equipment.
Seismic simulation testing is performed similar to the operational shock and
vibration testing. The product to be tested should be configured with similar loads
and cabled as if installed in users’ datacom facilities. For telecommunications equip-
ment, the test parameters are defined in NEBS GR-63-CORE (Telcordia 2006). This
document and the ANSI T1.329 (ANSI 2002) documents also contain specific tests
for earthquake conditions and can be used for reference. Most major test laboratories
are familiar with these test specifications and should be able to conduct the tests on
any product. The earthquake tests in these two documents have optional parameters
for zone intensities that product designers should consult for the earthquake zone
designation of the user’s site. It is always safe to assume worst-case conditions and
test to zone 4 parameters, which envelop all other zones if earthquake tests are
required. The International Code Council Evaluation Service (ICC-ES), a subsidiary
corporation of the International Conference of Building Officials (ICBO), published
AC-156, Acceptance Criteria for Seismic Qualification by Shake Table Testing of
Nonstructural Components and Systems, (ICC-ES 2007) and defined the required
response spectrum for equipment. Plots of seismic test parameters for different test
specifications are shown in Figure 11.2 Chapter 13 discusses the analysis of data-
com equipment and seismic anchorage systems.
Figure 11.2 Typical seismic test parameters. Reproduced with permission

from Notohardjono et al. (2001) and Pekcan (2007).
11.3.3 Ruggedness (Fragility)

Ruggedness relates to a product's ability to withstand shipping, installation,
relocation, and operational environments without damage. Ruggedness testing, also
known as fragility testing, is an essential part of the product development process.
Product ruggedness is ensured through shipping shock and vibration testing.
Ruggedness is a key focus item during the new product design phase. Significant
analysis and testing efforts are typically associated with new product and subassem-
bly designs to ensure adequate levels of ruggedness for frames, fragile components,
and assemblies. There are two types of fragility tests: subassemblies tests and rack
or cabinet level tests. Subassemblies are an integral part of datacom equipment and
will normally be mounted in a rack. The subassemblies are then tested and should
meet certain criteria to minimize manufacturing process stress, shipping, and
handling of shock and vibration from the environment. Typically, the subassemblies
tests include thermal shock from –40°C to +60°C followed by a shock test at 100g
3 ms and two 50g 11 ms half sine pulse shocks in all three axis (Notohardjono 1993).
The next test is a 15 min random vibration test profile at 1.04 grms as shown in
Figure 11.3. Normally, there is no attempt to test the subassemblies to failure if the
required test levels are met. In addition, a half-hour sine sweep test at 0.5g per axis
from 0–500 Hz is conducted to identify high transmissibility (output level/input
level) of any component inside the subassemblies. The next level of subassembly
integration testing is done at the rack or frame level. The types of drop testing that
should be considered for rack-mounted products depend on the transportation and
handling environment of the rack. The primary function of the rack is to prevent
damage to critical components. In other words, the rack and shipping pallet design
Figure 11.3 Typical transportation test parameters. Reproduced with

permission from Notohardjono et al. (2004).
should limit the shock transmitted to the critical component to levels below the
fragility limit of the components when mounted in the rack. Typical drop tests for
racks weighing more than 1000 lb (4450 N) are done by performing two tests at 4 in.
(102 mm) free-fall equivalent drop heights followed by 10 drop tests at 2 in. (51 mm)
free-fall equivalent drop heights. Typical vibration tests are a sine sweep at 0.5g for
30 min followed by a random vibration test for 15 min at 0.8 grms, as shown in
Figure 11.3 (Notohardjono et al. 2004).
11.3.4 Transportation Tests

Transportation tests must include shipping and handling tests of packaged and
unpackaged products. These tests include drop tests and vibration tests performed
as a frequency sweep of the product in its “as-shipped” configuration. The packaged
drop tests are typically conducted from defined heights on defined package surfaces
to confirm that the product will not be damaged when landing on any of its package
surfaces. Unpackaged drop tests can be more severe. These tests simulate dropping
the product after it has been removed from its shipping container while moving it to
where the product will be permanently installed. The drop heights for unpackaged
products are usually less than those of packaged products—approximately 3 in.
(76.2 mm) compared to 12 in. (304.8 mm). The product mass will also determine the
drop height, with heavier equipment dropped from a shorter distance.
Although end users have some interest in transportation tests because the
replacement equipment process could delay in-service dates, these tests are more
important to the equipment manufacturer than they are to the end user. The product
manufacturer must ensure that the datacom equipment arrives at the user’s site with-
out damage and in working condition, because if the product is received damaged
it will have to be replaced at the manufacturer’s expense. Given these conditions,
equipment manufacturers should focus on transportation testing.
Transportation vibration tests consist of frequency sweeps of approximately 2–
200 Hz. Typical transportation test parameters are shown in Figure 11.3. Typical
drop tests for racks weighing more than 1000 lb (4450 N) are done by performing
two tests at 3 in. (76 mm) free-fall equivalent drop heights followed by 10 drop tests
at 1 in. (25 mm) free-fall equivalent drop heights. In documents such as Telcordia’s
NEBS GR-63-CORE (2006), there are two test parameters that cover transportation
using smooth or normal means of transport contrasting to rough means of transport
such as aircraft with reciprocating engines. Product designers should determine the
worst-case situation for their products.
11.4 SHOCK AND VIBRATION TEST GUIDELINES FOR

DATACOM INFRASTRUCTURE AND COOLING EQUIPMENT
11.4.1 International Building Code (IBC)

Until the adoption of the International Building Code (IBC) in 2003, there were
no formal standards for manufacturing or criteria for product testing of infrastructure
components associated with either environmental or power conditioning. Unlike the
Telcordia NEBS specifications developed only for telecommunications equipment
(Telcordia 2006), the IBC seismic (and now wind) load code dictates requirements
for all infrastructure components within a structure necessary to continue the build-
ing’s “intended function.” However, the requirements only occur in geographical
locations that can develop substantial loads because of seismic or wind disturbances.
Prior to the existence of the IBC, manufacturers designed equipment with the
usual collection of precautions from good design practice. Product designs included
shipping or hold-down hardware to resist the forces of deceleration or braking
affects transferred to the internal components of the unit during transportation. Addi-
tional steps were taken to ensure against the dislodging of internally mounted assem-
blies in the event of a seismic disturbance commonly referred to as the “black box
theory”—a distrubance whose outcome is not well understood. If a project requires
military approval, military specs are available but hold few details and little infor-
mation concerning how to accomplish the task of verification testing.
As described in the previous section, datacom equipment is primarily
designed and manufactured using the Telcordia’s NEBS GR-63-CORE (2006)
standard or the equipment manufacturer’s internal standards. When looking at the
cooling and infrastructure components that support 24/7 datacom equipment oper-
ations, it is clear that no standard definition has been routinely utilized; however,
one is now in place: NEBS (Telcordia 2006). Should any major component within
the cooling system fail, an environmentally hostile environment will result within
the datacom equipment room. The datacom equipment that the cooling compo-
nents support will also undoubtedly have reduced reliability and shortened life
expectancy.
As mentioned previously, the IBC requirements are primarily for components
located in geographically charted seismic locations. The maximum considered
earthquake ground motion charts (see Figure 1615 of IBC [ICC 2003]) encompass
nearly 40% of the United States. Additional IBC requirements now include coastal
and inland locations from Maine to Texas that are subjected to hurricane-force
winds; fragility requirements remain optional in all other locations. Considering the
meaning of the term mission critical—a system that is absolutely necessary for
continued operation—this topic needs further consideration. Designing for “mission
critical” equipment should cause a relatively minor cost increase to improve and
move a product line to ”compliant” status.
Unlike the shipping hold-down philosophy but similar to Telcordia’s practices,
the IBC established a series of test procedures. These procedures are based on reso-
nance at various frequency levels and the responses of the component and its
subcomponent assemblies to various input frequencies. Through stiffening the
casing, structural members, and internal components, the unit’s natural frequency
moves high enough to place space between the lower frequencies generated by an
earthquake and the passage of some resulting frequencies to the mounted compo-
nents. The ultimate goal is to establish the “on-line” capability of a component
before and after a seismic event.
11.4.2 Test Procedures

The IBC is the culmination of the United States’ major model code-writing
communities, who decided to combine their efforts under the umbrella of the Inter-
national Code Council (ICC). The ICC established minimum requirements and test
procedures in AC-156 (ICC-ES 2007). This acceptance criteria is applicable for
shake table testing (see Figure 11.4) of nonstructural components and systems
(equipment) that have fundamental frequencies greater than or equal to 1.3 Hz.
Where shake table testing proves impractical because of increased unit size, finite
element analysis is permissible in accord with the same guidelines established in the
AC-156 test procedure.
For nonstructural components and systems (equipment) that are included in this
category, Article 6.7.2 of AC-156 (Ip = component importance factor = 1.5—see
Appendices C and E of this book for further discussion) classifies this equipment as
essential to the continued operation of the facility. The code requires that the equip-
ment evaluated meet the functional and operational requirements similarly both
before and after seismic testing.
When selecting nonstructural components and systems (equipment) for testing,
it should be a representative component of the product line if possible. To be repre-
sentative, the equipment needs to share enough structural similarity with other
models in the product line to merit the testing of one unit rather than all units. The
Figure 11.4 Shake table testing.
representative model determination is generally made by the independent agency

connected with the testing. After the model is chosen, the component is affixed to
a shaker table. The shaker table is a hydraulically actuated platform that ideally
accelerates the test product or specimen in two principal horizontal axes and one
vertical axis simultaneously. The acceleration loads that are usually selected repre-
sent the worst-case conditions within a geographical area or location zone require-
ments that the manufacturer needs to meet for compliance. These loads are generated
at a series of predetermined frequencies designed to search for the possible resonant
range of the component. Often, components are fixture-mounted to the table, simu-
lating field-mounting conditions. However, fixture mounting is not necessary since
computer modeling can simulate installation and tie-down methods.
Preliminary testing should begin with a resonant frequency search, a sweep
through the response spectrum to determine the resonant frequencies and damping
properties of the component. While the resonant frequency search is not a seismic
test, low-level amplitudes occurring from 1.3 to 33 Hz will reveal to the component
manufacturer that there may be possible flaws in the structural design. These find-
ings should be investigated and corrected before actual seismic testing occurs.
Generally, design weaknesses can be corrected using additional or different fasten-
ers or fasteners in combination with minor bracing. Once the tester and manufacturer
are satisfied with these preliminary test results, the actual seismic testing can begin.
The procedure for seismic testing is carefully described in AC-156 published by the
ICC-ES (2007).
Once the seismic tests are completed, Section 6.7 of AC-156 requires that a
post-test functional compliance verification be completed. This requires that the
equipment being qualified be capable of performing its intended functions after a
seismic event.
The unit under test, or UUT, must clearly demonstrate the following:
• that the unit will not pose a safety hazard to personnel due to collapse, anchor-
age failure, or the dislodgement of subassemblies;
• that the unit containing hazardous or flammable materials will not release
those materials that could pose a danger to humans or the environment; and
• that the unit deemed essential either by code or by owner requirements for
“mission critical” status satisfies all functional and operational tests.
The “shake test” option is undoubtedly the most practical test method to prove
the component’s real capabilities. Frequently, components are too large or cumber-
some or the manufacturer’s time constraints make it impractical to carry out this
form of testing, however. In these instances, finite element analysis becomes the next
sensible choice.
Finite element analysis or computer modeling create a replica of the component
and examine both material and fastenings. AC-156 (ICC-ES 2007) is also used as the
performance template for finite element analysis modeling, which has the same
requirements as shake table testing. Again, representative component samples are
utilized for the UUT to accelerate the process and to minimize costs. Finite element
analysis analyses are most effective and mainly used to model the unit’s structure.
Performance testing of subcomponents such as motors, compressors, and panel
boards are more efficient using shaker tests than using generated computer model-
ing. The shortcomings of finite element analysis are the accuracy of the modeling
and whether complex assemblies have been reflected properly.
The final compliance option available to manufacturers of components is
referred to as “The Experience Data Alternative,” which means learning from user-
reported data. If it can be proved that an identical unit survived an identical event in
an identical structure, that product will comply with the building code requirements.
To demonstrate this, it requires a manufacturer to use “nationally recognized proce-
dures” for the design and evaluation requirements to validate the component’s seis-
mic capacities. Capacities must meet or exceed the seismic demands established by
the IBC. The main legal issue linked with this form of compliance is the absence of
“nationally recognized procedures.”
Manufacturers that use and successfully meet or exceed the AC-156 seismic
program requirements will display evidence of testing and compliance. The manu-
facturer will rigidly affix a compliance data label (Figure 11.5) to the shipped
component for use by construction and building officials. The identification label
lists, among other things, the testing agency, function, and performance character-
istics of the component and the representative samples utilized as the test basis.
Keep in mind that “online requirements” for a component include all hardware,
accessories, options, and attachments that enable that component to perform its
“intended function.” The obvious items that fall within this category are pipes, ducts,
and electrical connections. Not-so-obvious items that fall into this category are floor
stands, vibration isolation components, and control panels. Any component that the
manufacturer supplies as an accessory or option must therefore be included in the
analysis to have or withstand the same level of vibration as the component itself.
If an owner is truly eager for a datacom space that will “continue to perform,”
there must be a validation of those desires specifying a series of products that will
do just that. For some datacom facilities such as Emergency 911 call centers, hospi-
tals, or government installations, there is no option, it’s the law! A seismic floor stand
as shown in Figure 11.6 may be installed to satisfy the desire to have datacom cool-
ing equipment and infrastructure “continue to perform”.
Figure 11.5 The identification label showing compliance data of a

nonstructural component.
Figure 11.6 Seismic floor stand.
12
Seismic Anchorage
of Datacom Equipment
12.1 OVERVIEW
There are datacom equipment centers located throughout the world, many in
areas where there can be significant earthquakes. Some areas within the United
States, such as most of California, the Pacific Northwest, and Salt Lake City, are
areas that have well-known high seismic risks, but the fact is that all 50 states have
recorded earthquakes. Damage from earthquakes due to falling objects can induce
bodily injury and can seriously impair occupants’ ability to evacuate a building once
the shaking has stopped. In addition, the financial losses due to heavily damaged
equipment or loss of service can be devastating.
This chapter gives an overview of seismic anchoring and lists several key
aspects of protecting data center equipment from the damaging effects of earth-
quakes. This information is not meant to be a substitute for the building code or to
replace the services of a qualified structural engineer that is licensed to perform the
required calculations. Rather, it is meant to give the owner or engineer a general
background in the seismic requirements for nonstructural components as well as to
explore some particular topics in detail with commentary and guidance where appro-
priate.
12.2 NONSTRUCTURAL SEISMIC PROVISIONS OF

BUILDING CODES
What are nonstructural components, anyway? Nonstructural components are all
those items in a data center that are attached to the structure but are not part of the
building structure itself. Typically, this includes architectural components such as
exterior building cladding, interior partition walls, hung ceilings, RAFs, and storage
cabinets. It also includes mechanical and electrical components such as HVAC units,
pumps and compressors, elevators, motor control units, transformers, datacom
equipment racks, and all associated piping, ductwork, conduit, bus ducts, and cable
trays. And, most importantly, it includes all of the datacom equipment on the data
center floor.
Why must these nonstructural items, specifically the datacom equipment, be
anchored? In any location where a building such as a datacom equipment center is
96⏐ Seismic Anchorage of Datacom Equipment
subjected to earthquake forces over a certain threshold, all nonstructural components

must be anchored to the building structure to ensure two objectives. The first objec-
tive is to protect the occupants from moving or falling hazards that may injure them
or impede their egress out of a damaged building. The second is to ensure certain crit-
ical systems will be less damaged and thus will have the best chance of remaining
in operation immediately after an earthquake.
Provisions for the anchorage design for nonstructural components have
changed a great deal in the last decade. To begin, it is instructive to review several
concepts of seismic design in general and the current code provisions that govern
nonstructural components—in particular to illustrate how these may impact the
design values for components used in datacom centers.
12.3 SEISMICITY IN THE UNITED STATES

Prior to 2000, seismicity in the United States was defined in the Uniform Build-
ing Code (UBC) using the seismic zone factor Z (ICBO 1991, 1994, 1997). Engi-
neers and building owners were accustomed to using these zone factors since there
were only a few of them (0, 1, 2A, 2B, 3, and 4) and since the map in the building
code (Figure 16-2 of the 1997 UBC, for example) was extremely clear on the location
where each factor was to be used.
The original zone map was based on pioneering work done by S.T. Algermissen
in 1948 and 1976 (FEMA 1998). The 1976 map was the basis for all codes after 1976.
It was the first map based on probabilities, namely a 10% probability of the “design
earthquake,” as defined by FEMA (1998), being exceeded in a 50-year period. The
1976 map added a zone 4 (to the original zones 0 through 3) to cover California and
was based on peak ground acceleration, the best measure of ground motion at that time.
However, as the years progressed, the maps were modified and the boundaries
became based more on political divisions than on science (for example, zone 4
stopped exactly at the California/Oregon border). A major study by the United States
Geological Survey (USGS) was undertaken in 1996 to develop a new generation of
seismic hazard maps. These new maps not only were established on the latest ground
motion studies collected by USGS in the past 20 years but also incorporated several
new features including PSD basis at both 0.2 and 1.0 s periods and maps for both the
500-year (10/50 probability) and 2500 year (2/50 probability) earthquakes. The map
that was adopted into the 2000 IBC was a revised version of the 1996 USGS 2500-
year map but added a deterministic cap to the ground motions near active crustal
faults. For example, in areas such as California where the fault mechanisms are well
known and where the ground shaking is limited by the actual strength of the under-
lying soil, the probabilistic peak value was “cut off” to a maximum value using deter-
ministic methods.
It is important for the owner to understand that the hazard maps in the 2000
version of the IBC remove seismic zones. In Southern California, for example, the
mapped short-period spectral values (used for nonstructural components) vary from
a low of 0.42g in the San Fernando Valley to a high of 2.48g along the San Andreas
Fault. In addition, ground motion can vary significantly even for locations just a few
miles away from each other, especially in areas of the country that have significant
contours. For owners, this will be a significant challenge because where before one
could simply use “zone 4” datacom racks, there is no longer one design level that
covers every corner of a geographical area.
12.4 SEISMIC DESIGN CATEGORY

In addition to the change in the seismic hazard definition and maps, the other
major change with the IBC was the addition of the seismic design category (SDC),
which is used to determine seismic detailing requirements. The SDC is a combina-
tion of the seismicity of the location, the soil type, and the building occupancy. By
itself, it is the new sort of “zone” that owners should know in order to understand
what types of equipment require anchorage in their datacom centers.
Generally, equipment in datacom centers classified as SDC A and B require
little if any anchorage or seismic design enhancements on the cabinets such as addi-
tional brackets, etc. The same is true for mechanical and electrical equipment desig-
nated SDC A and B. In SDC C, seismic provisions are included but they are fairly
moderate in scope. For example, suspended ceilings are required to be braced, but
only to the Ceilings & Interior Systems Construction Association (CISCA) provi-
sions for low seismicity areas. SDC D, E, and F are considered high seismic cate-
gories and require full seismic detailing. SDC D is the highest category shown within
the code tables (Tables 11.6-1 and 11.6-2 in ASCE Standard 7-05 [ASCE 2005]),
and SDC E and F are reserved for sites with very high seismicity. For mission critical
equipment, seismic anchorage and provisions should be included to assure high
availability and reliability of the equipment regardless of the seismic design category
of the datacom center.
It will take owners some time to feel comfortable with the seismic design cate-
gory concept, primarily because of the way it combines the location and soil type.
If a data center is located on a site with good stiff soil, then the SDC will mirror the
old “zone” factor and will be assigned A, B, C, D, or E based on the mapped short
or 1.0 s response parameter. If a data center is located on a site with soft soil,
however, the SDC can be much higher than anticipated as a result of the amplifica-
tion caused by the soft soil. For example, a data center in downtown Boston (with
low to moderate seismic hazard) on rock soil could be SDC B but one in nearby
Charlestown on a soft fill site could be SDC D, therefore requiring the same seismic
detailing as a data center in Los Angeles.
12.5 PROPERLY APPLYING ANCHORAGE FORCES

The horizontal seismic design force (Fp) is described in detail in Appendix E.
The following commentary should be used when applying Fp:
• it should be applied at the center of gravity of the component,

• it should be distributed relative to the component’s mass distribution,
• it should be applied independently in both horizontal directions (longitudinal
and transverse), and
• it should be combined with static loads using proper load combinations.
It should also be understood that many codes, including the IBC, require the
addition of vertical seismic forces to the horizontal forces in order to quantify the
entire effect of the earthquake on a component. Equations 12.4-1 and 12.4-2 of
ASCE Standard 7-05 (ASCE 2005) should be used for this purpose.
12.6 PROTECTION OF SERVER CABINETS ON

RAISED-ACCESS FLOORS
Perhaps the most important items in the data center that need to be anchored are
the datacom equipment cabinets located on the data center floor. This is a challeng-
ing task for several reasons. First, they are often very heavy, narrow in width, and
prone to toppling over. Second, the equipment cabinets are generally supported by
the RAF, which means they are more difficult to anchor than if they were simply
sitting on the building’s structural floor. Third, there is often a desire to keep the data-
com equipment in continuous operation immediately after a major earthquake. A
data center owner should pay special attention to the approach to anchor datacom
cabinets in order to protect their data, investment, and operations.
12.6.1 Component Amplification Factor on Raised-Access Floor

In Section 3 of Appendix E, the component amplification factor (ap) is
discussed in the building code equation to account for the dynamic amplification of
equipment supports. This can have a great impact on the design forces used for data-
com equipment racks supported by RAFs. There has long been a debate as to
whether RAFs are considered rigid supports or flexible ones. The answer is unfor-
tunately not straightforward. In terms of dynamics, the answer depends greatly on
the height and horizontal stiffness of the flooring system, how large the floor is, the
boundary conditions at the edge of the floor, and the mass being supported by the
system.
It seems clear that a RAF utilizing diagonal bracing can be considered rigid for
this purpose and can utilize an ap of 1.0. However, other types of RAFs can be
considered flexible. Tests have shown that small portions of 2 ft (600 mm) high
RAFs have a horizontal frequency between 10 and 15 Hz that would define them as
flexible per the building code (period < 0.06 s). However, larger floor systems with
much heavier masses in actual installations have been shown to achieve 35 to 40 Hz
frequencies, which would define them as rigid. Without a detailed analysis and/or
testing program of the actual installation, it would be difficult to assume that the floor
system is rigid with complete certainty. In most cases, it would make sense to be
conservative and assume that the RAF system is a flexible support to any equipment
attached directly to it.
12.6.2 Forces on Datacom Equipment Cabinets

Datacom equipment cabinets are often grouped together in long rows (or aisles)
with space to walk between the aisles and access to the fronts and backs of the units.
Earthquakes produce motion in all three directions and, as described in Appendix E,
all directions must be considered when determining the proper anchorage for a piece
of equipment. The effects on the equipment cabinet from the two horizontal direc-
tions of earthquake motion are the most obvious and are often effectively resisted
using clips at the bottom of the cabinet feet, attached directly to the cabinet frame,
to resist the horizontal shear in each direction. If these clips are attached directly to
the RAF panel, corner lock screws are needed to transfer this shear load to the pedes-
tal head.
When the equipment cabinet is much taller than it is wide, it is prone to over-
turning (forces attempting to tip the cabinet over). Datacom equipment cabinets are
often ganged together using bolts in the longitudinal direction (down-aisle) so that
the cabinets act as one much larger unit. This greatly reduces the impact of over-
turning forces in the down-aisle direction. However, the transverse (or cross-aisle)
direction still needs to be managed because it produces forces coupled to the legs at
the bottom of the cabinet where one is in tension and the other is in compression. The
load path of both the down-aisle and the cross-aisle configurations must be consid-
ered. Tension forces can be handled in a number of different ways and will be
explored in the next section, but the compression force should not be overlooked. A
tall equipment cabinet with a large horizontal force and a narrow width can produce
a large compression force that must be resisted by the RAF panel and understructure
system. It is possible for the small hard castors supporting the equipment cabinet to
punch through or break a floor panel as a result of overturning compression.
12.6.3 Techniques for Anchoring Datacom Equipment to

Prevent Overturning
There are many different techniques for anchoring datacom equipment cabinets
to RAFs to resist the effects of overturning. The methods of attaching the datacom
equipment cabinet to the concrete floor of the data center will vary with the manu-
facturer. Five specific strategies are presented here, but this list is by no means
exhaustive. The choice of which anchoring method to use depends on the owner
deciding on the optimum balance of performance, cost, installation time, and avail-
ability. In some extreme cases, especially when higher-than-normal code perfor-
mance is desired, actual tests are highly recommended to ensure the whole system
behaves as expected. Most of these techniques only deal with overturning in the
transverse direction of cabinets. As mentioned above, in the longitudinal direction
the cabinets are ganged together by rigidly attaching adjacent cabinets to each other
at several locations over the entire cabinet height (see Figure 12.1).
The following are the five methods of anchoring datacom equipment:
1. Vertical Anchor Rods: A common installation method for datacom equipment

racks prone to overturning is to use vertical steel threaded rods that attach the
cabinet directly to the concrete floor below. These anchors are located at each
corner of the cabinet and serve to transfer the tension due to overturning through
the RAF directly to the floor. These systems can be fairly simple and inexpen-
sive. However, creating the opening in the RAF panel is not trivial, and the
attachment at the floor can be hard to locate directly under the cabinet. A better
seismic detail is to use a piece of slotted metal framing attached directly to the
Figure 12.1 Hardware attaching adjacent server cabinets together.
floor underneath the rack at each side to allow better positioning of the rod side
to side. These anchor rods do not resist horizontal shear without a special addi-
tional feature such as an additional clip to prevent horizontal movement; there-
fore, they are often used in conjunction with clips for this purpose. Another
improvement is to use a turnbuckle to achieve easy tensioning of the rod and to
induce a compressive load to resist the horizontal force. Schematic details of
this technique are shown in Figures 12.2 and 12.3 and photographs of this
installation method are shown in Figures 12.4 and 12.5. The improved detail
using slotted metal framing is shown in Figure 12.6.
2. Splay Cables: Another common anchoring option is to use wire rope cables,
called splay cables, to restrain the cabinet in all directions. Splay cables have
the added benefit of resisting loads in horizontal shear as well as in vertical
overturning. The disadvantage of this method is that since the wire rope only
resists tension, four cables are needed, one in each direction at each corner of
the cabinet. Wire rope can be easily threaded through the holes in the corners
of server cabinets and can be secured with wire rope clamps (three minimum
each end). In general, the wire should be located 45° to vertical and horizontal
(hence the term splayed). Wire rope has the benefit that it can be worked
through existing obstacles and does not have to be perfectly located in order to
be effective. Wire rope installs can also utilize turnbuckles to ensure each rope
is tight, but this is often not necessary, as the slack in each rope will tighten as
the cabinet shifts on the RAF during an earthquake. Photos of this installation
method are shown in Figures 12.7 and 12.8
3. Rigid Platforms: Similar to air-conditioning equipment that is placed on RAFs,
some datacom equipment cabinet vendors manufacture a rigid steel base that
Figure 12.2 Vertical anchor rod detail at server cabinet. Reproduced with
permission from Notohardjono (2003).
Figure 12.3 Vertical anchor rod detail at floor. Reproduced with permission
from Notohardjono (2003).
Figure 12.4 Anchor rod.
Figure 12.5 Vertical anchor rod detail at floor.
Figure 12.6 Vertical anchor rod detail at floor using slotted metal framing.
Figure 12.7 Splay cable installation.
Figure 12.8 Splay cable installation.
can support the equipment cabinet directly so that it does not sit directly on the
RAF. Although this makes bracing the datacom equipment cabinet much easier
and ensures excellent performance, there are a few drawbacks. First, if the data-
com equipment is not located exactly at the edges of a RAF panel, the RAF
panels must be cut and the new free edges supported by either edge angles or
additional RAF stanchions. Second, the datacom equipment cannot be easily
moved in the future, as the datacom equipment platform would have to be
moved as well.
4. Top Cabinet Snubber: Sometimes it is not possible to anchor the server cabinets
to the floor below. When this happens, another option is to add “snubbers” to the
top of the equipment cabinets to help restrain them from overturning. There are
several specific considerations with this method. First, the bracing must be stiff
and strong enough to resist the often large overturning forces on the cabinets.
Second, these large forces must be transferred to the structure somewhere, and
often the ceiling is not nearly adequate for this purpose—in some cases, the ceil-
ing must be greatly reinforced or the bracing must continue up through the ceil-
ing to the floor or roof structure above. Third, above-server bracing makes
moving the servers difficult in the future and makes running cables in raceways
above the datacom equipment very difficult. For a heavy datacom equipment
cabinet, the top cabinet snubber can be combined with one of the three anchoring
methods above. A schematic detail of this technique is shown in Figure 12.9, and
a photograph of this installation method is shown in Figure 12.10.
5. Isolation Platforms: A new way to anchor datacom equipment cabinets to the
RAF is to not anchor them at all but rather place them on an isolation platform.
These platforms, using similar techniques as base isolation bearings used in
Figure 12.9 Top cabinet snubber schematic detail. Reproduced with

permission from ASHRAE (2005b).
Figure 12.10Top cabinet snubber installation.
Figure 12.11Isolation platform installation.
buildings though on a much smaller scale, use the same technique to decouple
the motion of the ground from the item being isolated by shifting the funda-
mental resonance frequency. These platforms have the added benefit of being
able to be “tuned” so that the datacom equipment on top of the platform expe-
riences only a specified acceleration level and, thus, the probability that the
datacom equipment will remain operational is improved. There are two impor-
tant details to keep in mind with these systems, however. First, the isolation
bearings in the platform lengthen the period of the cabinet by displacing a large
amount (up to 8 in. [200 mm] in every direction), so an adequate amount of
empty space is needed around the cabinet for the system to work properly and
be unobstructed. Second, the isolation bearings have a discrete amount of
displacement capability, which may not be adequate in larger seismic zones or
for installations many levels above grade. A nonlinear analysis conducted by a
qualified structural engineer should be undertaken to ensure adequate displace-
ment capacity of the bearing being considered. Note that when specifically
detailed, equipment cabinets can still be ganged together side-to-side in a row
using these systems. A photograph of the platforms installed is shown in
Figure 12.11.
13
Analysis of
Datacom Equipment and
Seismic Anchorage Systems
13.1 OVERVIEW
In the preceding chapters, we have discussed the specifications of the building

within which the datacom center is located as well as the design and evaluation of
the datacom center’s RAF system. In this chapter, one method for evaluating the
datacom equipment itself will be examined. In ASCE Standard 7-05 (ASCE 2005),
the Seismic Design Requirements for Non Structural Components section (in
Chapter 13) explains the three evaluation methods used to demonstrate seismic
resistance: analysis, testing, and experience. This chapter demonstrates how the
finite element analysis method can be applied to specific datacom equipment
designs through discussion of two relevant examples. To illustrate the successful
application of this methodology, we examine those systems that anchor datacom
equipment to the data center as well as methods to stiffen the datacom equipment
itself in an effort to resist seismic loads. Complementing this work, Chapter 11
discusses specifications and test parameters for doing actual shock and vibration
testing.
constraint element: a general class of elements that apply kinematic constraints

(rigid link, slider, revolute, point-in-plane, translational) between nodes.
finite element analysis: a numerical technique of solving a wide range of engineer-

ing problems. Finite element analysis is used to determine the stresses and displace-
ments of a mechanical structure. The first step in the analysis is to approximate the
actual geometry by a mesh of small elements bounded by points called nodes. The
next step is to define the boundary conditions, the applied load, and/or the displace-
ment on the elements and nodes.
Guyan reduction solver and Lanczos method solver: numerical procedures used by
finite element analysis as iterative algorithms that are adaptations of power methods
108⏐ Analysis of Datacom Equipment and Seismic Anchorage Systems
to find eigenvalues and the eigenvector of a square matrix. They are particularly
useful for finding decompositions of very large sparse matrices.
parabolic shell element: a type of element used in finite element analysis to repre-
sent a surface.
rigid element: a type of element used in finite element analysis that will not deform
under loading.
normal mode, harmonic frequency, natural frequency, resonance frequency: the

frequency at which a deformable structure will oscillate when disturbed.
13.3 DATACOM EQUIPMENT FRAME

In Chapter 11, a test method was proposed to evaluate the seismic performance
of datacom equipment and its accompanying seismic retention systems. Alterna-
tively, evaluation can be performed using finite element analysis on a simplified
model of the datacom equipment and its retention schemes. First, a method for eval-
uating anchoring methods is illustrated. Second, in the latter part of this chapter, an
example comparing and contrasting the effectiveness of frame stiffening concepts is
shown. The purpose is to show, by way of example, that the finite element analysis
method can be a practical and powerful tool for sizing and predicting failure modes
of datacom equipment under seismic loads.
In this chapter, a datacom equipment frame is used as an example of a generic
piece of datacom equipment, though the methodology can be extended to other types
of equipment in a similar manner.
13.4 FINITE ELEMENT MODEL CONSTRUCTION AND VALIDATION

Before proceeding with any analytic work, we must have confidence that the
finite element model, or the analysis model, correlates well with the actual hardware.
The analysis model used in this chapter was constructed from the simplified
computer-aided design (CAD) geometry of an existing datacom equipment frame
inside of a general purpose finite-element software code, ANSYS Multiphysics
(ANSYS 2006). Simplifying assumptions were made at the interfaces between adja-
cent frame components. A welded joint was assumed to exist along their mating
edges, so coincident nodes from the adjacent component meshes were merged.
To streamline the process of attaching the add-on structural members, groups
of nodes from the finer mesh (frame) were connected to groups of elements from the
coarser mesh (add-on structural members) using constraint equations. Constraint
equations operate like interpolation functions and restrict local strains. This results
in much smaller stress transitions from one mesh to the other. Alternatively, the use
of rigid elements would force some strain values to be zero at the joint. This results
in large errors in local stress values, though the stresses return to correct values away
from the joint. Rigid element connections, however, restrain the element’s face from
deformation as well as restrain the relative motion between the connected nodes.
This increases stiffness and significantly affects stresses in the model, particularly
at the joint. The complete frame model consists of approximately 110,000 parabolic
shell elements (see Figure 13.1).
The finite element model was then solved using a normal-mode dynamics
solver to verify that it indeed accurately can predict the vibration behavior of our
real-world structure. The numerical analysis was conducted using a Lanczos method
solver. The advantage here is that this method is much faster than a traditional Guyan
reduction solver (Golub et al. 1972).
The result from the normal-mode dynamics solver provides a good correlation
with the experimental frame test. Examinations of the resulting mode shapes reveal
that the first and second modes of vibration correspond to the first lateral swaying
mode (6.9 Hz) and first torsional mode (36.35 Hz), shown in Figure 13.2. Corre-
spondingly, the experimental modes were found to occur at 5 and 38 Hz. Therefore,
the analysis results can be said to be a good predictor of the actual structural
response.
Now that correlation has been established between the finite element model and
reality, a series of design scenarios can be evaluated to investigate various earth-
quake anchorage systems as well as in-frame stiffening concepts for this datacom
equipment frame.
Figure 13.1 Finite element model of frame with vertical anchors subjected to
the maximum seismic horizontal design force Fp (defined by
Equation 13.1). Reproduced with permission from
Notohardjono and Canfield (2007).
Figure 13.2 Mode shapes of the frame: (a) lateral sway mode, 6.9 Hz, and
(b) torsional mode, 34.2 Hz. Reproduced with permission from
Canfield and Notohardjono (2004).
13.5 EVALUATION OF EARTHQUAKE ANCHORAGE SYSTEMS
Some of the most important components in the datacom center that need to be
anchored are the datacom equipment frames. There are many different techniques
for anchoring frames to RAF systems to prevent overturning of the equipment. Many
of these anchorage system solutions are discussed and illustrated in Chapter 12. One
common anchoring system used to secure datacom equipment frames is vertical
steel rods that attach to the frame and to the concrete subfloor (depicted as a finite
element model in Figure 13.1). These anchors are located at each corner of the frame
and transfer the large inertial forces of datacom equipment down to the subfloor.
13.5.1 Loading and Setup of Analysis Model
To build an analysis model, first we need to define the model and then the load-
ing and the boundary conditions. For illustration purposes, a generic server with a
total height of 78.7 in. (2 m) is represented as a simple lumped mass suspended in
its frame. We assume that the RAF will support the static weight of the IT equipment
cabinet and that the anchors, with one end connected to the frame and the other ends
fixed to the floor or to the ceiling, will resist the forces that would try to overturn the
frame.
For this load case, let us assume the datacom equipment is installed in a data-
com center in California. The maximum seismic horizontal design force Fp per
Equation 13.3.2 in ASCE Standard 07-05 (ASCE 2005) is given as follows:
Fp = 1.65 SDSIpWp (13.1)
where
SDS = short period spectral acceleration at 1 s period = 2.48g in a high seismic
zone
Ip = component importance factor = 1.5 when the datacom equipment is
required to function after an earthquake
Wp = equipment weight
With the loading and structure defined, a design of experiments is conducted

varying the component’s weight and the height of its center of gravity with respect
to the height of the frame. This is done using two anchoring scenarios, one with four
0.5 in. (12.7 mm) diameter, 12 in. (304.8 mm) length steel rods connected at the four
bottom corners and the other scenario with an additional four steel rods attached at
the top corners of the frame. The end conditions of the anchors were fixed in all trans-
lational and rotational degrees of freedom. The resulting stress analysis is evaluated
in the steel rods and based on the yield strength of the steel (36 kpsi = 250 MPa); a
factor of safety is then calculated. In these two scenarios, the other ends of the rods
are fixed either to the concrete floor or to the concrete floor and the ceiling. For all
of the analyses here, the modulus elasticity of the steel rods and the frame is
29,000 ksi (200 GPa).
13.5.2 Results and Conclusions

The results of the design of experiments for the two anchoring schemes are
shown in Figures 13.3 and 13.4. The data reported for each scenario is the mini-
mum factor of safety of any one of the steel anchors. A factor of safety that is less
than 1.0 signifies that under those specific loads these anchors would not be a reli-
able design. The results for the four anchors at the frame base show that for the
scenario of a heavily loaded frame (5000 lb = 22.25 kN) with a high center of grav-
ity (positioned at 50% or 75% of the cabinet’s height), these 0.5 in. (12.7 mm)
diameter steel rods would be underrated for those applications because their mini-
mum factors of safety fall below 1.0, as shown in Figure 13.3. The results for the
eight anchors–four at the frame base and four at the top of the frame—show that
for all scenarios, these 0.5 in. (12.7 mm) diameter steel rods would be sufficient
for those applications because their minimum factors of safety are above 1.0, as
shown in Figure 13.4. Additionally, by doubling the length of the anchors to 24 in.
(610 mm), the resulting maximum stress in the anchors increased by 23%. This
increased stress result was consistent through different frame weights and various
center-of-gravity positions.
Figure 13.3 Evaluation of four vertical bottom anchor rods for various
equipment configurations. Reproduced with permission from
Notohardjono and Canfield (2007).
Figure 13.4 Evaluation of eight vertical anchor bottom and top rods for
various equipment configurations. Reproduced with
permission from Notohardjono and Canfield (2007).
13.6 EVALUATION OF STRUCTURAL ADD-ON SUPPORTS
13.6.1 Definition of Structural Add-On Supports
Structural add-on supports are another area of investigation and design study for
various types of earthquake retrofit kits that use different classes of structural
elements. Each element is examined and then compared based on its effectiveness
in increasing datacom frame rigidity.
To limit lateral motion during earthquake events, one stiffening concept uses a
triangular support brace in combination with two other support bars (see
Figure 13.5a). This triangular brace is installed with hinged pins on one corner post
of the frame and is restrained on the opposite side by a latch and fastener. In this
configuration, the triangular brace with its hinged pivots provide easy access to the
rack-mounted components when they are installed in the frame.
An alternate option is to use an array of rear support trays (see Figure 13.5b).
In this scenario, seven trays all grouped together are examined. Each support tray is
bolted separately to the frame. The geometry of these support trays allows for instal-
lation without hindering the routing of cables, which are typically spread throughout
the rear of a datacom equipment cabinet.
The section that follows evaluates the variation of these two structural add-on
supports.
Figure 13.5 (a) Triangular brace and support bars and (b) group of support
trays. Reproduced with permission from Canfield and
Notohardjono (2004).
13.6.2 Design of Experiments to Evaluate

Structural Add-On Effectiveness
The effectivenesses of the two structural add-ons shown in Figure 13.5 were
evaluated using six design variations as depicted graphically at the bottom of
Figure 13.6. They are as follows (from left to right in the figure):
1. bare frame
2. single set of tray supports (Figure 13.5b)
3. double set of tray supports (at frame’s front and back)
4. single triangular support (Figure 13.5a)
5. double triangular supports
6. triangular and tray supports
Each variation of the finite element analysis model was solved using the
Lanczos normal mode dynamics solver in ANSYS (ANSYS 2006).
13.6.3 Resulting Shift in First Frame Harmonic

A summary of the resulting shift in the first harmonic frequency is plotted in
Figure 13.6 and provided in Table 13.1. Examining the unpopulated frame without
any design modifications, it was found that the first bending mode occurred at
6.9 Hz. Overall, the analysis of the tray supports proved least effective in shifting the
Figure 13.6 Design study summary: resulting shift in first harmonic

frequency. Reproduced with permission from Canfield and
first harmonic to a higher frequency. On the other hand, all design scenarios using
triangular supports successfully shifted the occurrence of the first mode to 45.0 Hz
and higher. Figure 13.7 depicts the mode shapes of the least effective and most effec-
tive design changes to the frame’s geometry. This evaluation demonstrates that the
cross-bracing provided by the triangular supports is much more successful in
increasing the harmonic frequency or the overall lateral rigidity of the system. In
other words, the harmonic frequency of the frame increases as a result of the
increased frame rigidity. The increased harmonic frequency will lead to a reduced
lateral sway of top of the datacom equipment.
Table 13.1 Summary of Harmonic Frequencies
Single Double Single Triangular Double

Bare
Tray Tray Triangular and Tray Triangular
Frame
Supports Supports Support Supports Supports
First
Harmonic 6.93 12.34 14.8 45 45.3 53.27
(Hz)
Second
Harmonic 35.66 36 36.35 56.3 57.08 61.37
(Hz.)
Figure 13.7 (a) Double tray support design, first mode shape (14.8 Hz) and
(b) double triangular support design, 1st mode shape (53.3 Hz).
Reproduced with permission from Canfield and Notohardjono
(2004).
To protect equipment from undesirable environmental effects such as vibration,

shock, and seismic inputs, the direction and magnitude of these effects must be
completely defined. To aid in this process, a thorough analysis must be accomplished
to define the problem. Techniques such as vibration surveys/studies and performing
a finite element analysis will greatly aid in this definition determination process.
Once the vibration problem is fully understood, the establishment of the vibration
control performance requirement can then be defined.
A robust frame design can be achieved that will provide excellent structural
rigidity and less sway during severe seismic events. The basic frame structure can
be used as designed without additional add-ons in low-end applications that do not
require severe earthquake survivability. A unique structural add-on (the triangular
brace) has demonstrated its ability to increase the rigidity of the frame system as a
whole while remaining a cost-effective solution that can be incorporated when
requirements demand it. This addition of a structural add-on enables the basic frame
design to deliver the required structural performance for datacom center applications
with potentially severe seismic loading.
References and Bibliography

ACI. 2001. ACI 318, Building Code Requirements for Structural Concrete. Farm-
ington Hills, MI: American Concrete Institute.
ANSI. 2002. ANSI T1.329-2002, Network Equipment—Earthquake Resistance.
New York, NY: American National Standards Institute.
ANSYS. 2006. ANSYS Multiphysics. ANSYS, Inc., Canonsburg, PA.
ASCE. 2002. ASCE Standard 7-02, Minimum Design Loads for Buildings and
Other Structures. Reston, VI: American Society of Civil Engineers.
ASCE. 2005. ASCE Standard 7-05, Minimum Design Loads for Buildings and
Other Structures. Reston, VI: American Society of Civil Engineers.
ASHRAE. 1996. ASHRAE Guideline 1-1996, The HVAC Commissioning Process.
Atlanta: American Society of Heating, Refrigerating and Air-Conditioning
Engineers, Inc.
ASHRAE. 2003. 2003 ASHRAE Handbook—HVAC Applications, Chapter 47,
“Sound and Vibration Control.” Atlanta: American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 2004. 2004 ASHRAE Handbook—HVAC Systems and Equipment,
Chapter 41, “Pipes, Tubes and Fittings.” Atlanta: American Society of Heat-
ing, Refrigerating and Air-Conditioning Engineers, Inc.
ASHRAE. 2005a. Datacom Equipment Power Trends and Cooling Applications.
Engineers, Inc.
ASHRAE. 2005b. Design Considerations for Datacom Equipment Centers.
Engineers, Inc.
ASTM. 2006. ASTM D 4728, Standard Test Method for Random Vibration Testing
of Shipping Containers. West Conshohocken, PA: ASTM International.
BOCA. 1999. The BOCA National Building Code. Building Officials and Code
Administrators International, Inc. (now International Code Council), Wash-
ington, DC.
BSSC. 2003. FEMA 450, NEHRP Recommended Provisions for Seismic Regula-
tions for Buildings and Other Structures, Part 2: Commentary, Chapter 6,
118⏐ References and Bibliography
“Architectural, Mechanical, and Electrical Component Design Require-

ments.” Washington, DC: Building Seismic Safety Council of the National
Institute of Building Sciences.
BSSC. 2006. FEMA 451, NEHRP Recommended Provisions Design Examples,
Chapter 13, “Design for Nonstructural Components.” Washington, DC: Build-
ing Seismic Safety Council of the National Institute of Building Sciences.
Canfield, S., and B.D. Notohardjono. 2004. Structural analysis of high-end server
computer frames under earthquake loading. American Society of Mechanical
Engineers, Pressure Vessels and Piping Division PVP, Seismic Engineering
486(1):155–62.
DOD. 2000. MIL-STD-810F, Environmental Engineering Considerations and
Laboratory Test Method Standard. Washington, DC: US Department of
Defense.
DOD. 2002. MIL-STD-202G, Test Methods of Electronic and Electrical Compo-
nent Parts. Washington, DC: US Department of Defense.
Drake, R. 1990. Seismic analysis and design of computer access floors. Proceed-
ings ATC-29, Applied Technology Council, pp. 61–73.
Eagan, J., M. Kermode, L. Skyrman, and L. Turner. 2001. Ground vibration moni-
toring for construction blasting in urban areas. State of California Department
of Transportation Report FHWA/CA/OR-2001/03, Final Report, April, Sacra-
mento.
EERI. 1990. Loma Prieta reconnaissance report. Earthquake Spectra, Supplement
to Volume 6. Oakland, CA: Earthquake Engineering Research Institute.
EJMA. 2003. Standards of the Expansion Joint Manufacturers Association, 8th ed.
Expansion Joint Manufacturers Association.
FEMA. 1998. FEMA 313, Promoting the Adoption and Enforcement of Seismic
Building Codes: A Guidebook for State Earthquake and Mitigation Manag-
ers, Appendix A. Washington, DC: Federal Emergency Management Agency.
FIMS. 1987. Data Processing Facilities—Guidelines for Earthquake Hazard Miti-
gation. Washington, DC: Finance, Insurance and Monetary Services Commit-
tee of the Federal Emergency Management Agency.
Frey, R. 1989. Vibration field survey results and IBM standards revisions. IBM
Internal Technical Report TR 01.B136. Armonk, NY: IBM Corp.
Frey, R.A., B.D. Notohardjono, and R. Sullivan. 2000. Earthquake simulation tests
on server computers. ASME PVP 402(2):1–8.
Golub, G.H., R. Underwood, and J.H. Wilkinson. 1972. The Lanczos algorithm
for the aymmetric Ax = Lambda × Bx problem, STAN-CS-72-270. Stanford
University, Computer Science Department, March, Stanford, CA.
Harris, C.M., and A.G. Piersol. 2001. Harris’ Shock and Vibration Handbook, 5th
ed. New York: McGraw-Hill Professional.
IBM. 1990. Corporate Standard 1-9711-002, Vibration Levels for IBM Hardware
Products, Product Environments, Product Classes. Armonk, NY: IBM Corp.
IBM. 1992a. Corporate Standard 1-9711-007, Operational Shock Levels for IBM
Hardware Products Environments, Product Classes. Armonk, NY: IBM Corp.
IBM. 1992b. Corporate Standard C-B 1-9711-009, Earthquake Resistance for

IBM Hardware Products Guidelines for Design and Testing. Armonk, NY:
IBM Corp.
IBM. 1995. Corporate Standard C-H 1-9711-005, Packaged IBM Products, Test-
ing for Shipment Test Levels and Procedures. Armonk, NY: IBM Corp.
IBM. 2001. Corporate Standard C-S 1-3705-001, Machine Mobility, Stability, Size
and Mass Design Requirements. Armonk, NY: IBM Corp.
ICBO. 1991. Uniform Building Code. Whittier, CA: International Conference of
Building Officials.
Building Officials.
Building Officials.
ICC. 2000. International Building Code. Country Club Hills, IL: International
Code Council.
Code Council.
Code Council.
ICC-ES. 2007. AC-156, Acceptance Criteria for Seismic Qualification by Shake
Table Testing of Nonstructural Components and Systems. Whittier, CA: Inter-
national Code Council Evaluation Service, Inc. www.icc-es.org/criteria/
pdf_files/ac156.pdf.
Meyer, J.D., T.T. Soong, and R.H. Hill. 1998. Retrofit seismic mitigation of main-
frame computers and associated equipment: A case study. Proceedings ATC-
29-1, Applied Technology Council, Redwood City, CA.
NFPA. 2003. NFPA 5000, Building Construction and Safety Code. Quincy, MA:
National Fire Protection Association.
Notohardjono, B.D. 1993. Frame and subassemblies shock and vibration compli-
ance test, tiedown installation manual. IBM Internal Report Doc #
03DEC1993. Armonk, NY: IBM Corp.
Notohardjono, B.D. 2003. Tiedown Installation Manual. IBM Manual PN
16R1105, EC J10559. Armonk, NY: IBM Corp.
Notohardjono, B.D. 2006. IBM data center vibration monitoring. IBM Internal
Technical Report 2006. Armonk, NY: IBM Corp.
Notohardjono, B.D., and S. Canfield. 2007. Finite element analysis of datacom
equipment and earthquake anchorage systems. IBM Internal Report. Armonk,
NY: IBM Corp. Also to be published in American Society of Mechanical
Engineers, Pressure Vessels and Piping Division in 2008.
Notohardjono, B.D., J.S. Corbin, S.J. Mazzuca, S.C. McIntosh, and H. Welz.
2001. Modular server frame with robust earthquake retention. IBM Journal of
Research and Development 45(6):771–82.
Notohardjono, B.D., J. Wilcoski, and J.B. Gambill. 2004. Design of earthquake
resistant server computer structures. ASME Journal of Pressure Vessel Tech-
nology 126(1):66–74.
120⏐ References and Bibliography
Olson, R.A. 1992. Earthquake protection for data systems. Disaster Recovery
Journal 5(2):35–42.
Pekcan, G. 2007. Private communication, University of Reno at Nevada, Reno,
Nevada.
SBCCI. 1997. Standard Building Code. Southern Building Code Congress Inter-
national, Inc. (now International Code Council), Washington, DC.
Tate. 2003. Bolted stringer understructure for concore and all steel access floor
panels-24 in. type 6A. Jessup, MD: Tate Access Floors, Inc.
Telcordia. 2006. Network Equipment-Building System (NEBS) Requirements:
Physical Protection, Generic Requirements. GR-63-Core. Piscataway, NJ:
Telcordia Technologies, Inc.
US Army. 1992. Tri-Service Manual TM 5-809-10, Seismic Design for Buildings.
Washington, DC: Department of Army (Corps of Engineers). www.usace.
army.mil/publications/armytm/tm5-809-10.
Appendix A
Codes for
Building Structures and
Structural Components
The primary purpose of the United States building codes has been to protect the
life and safety of the building occupants and those who are immediately outside of
the building. Building code requirements generally apply to new construction and
to existing structures undergoing significant additions or alterations.
The intent of such codes is to ensure that buildings are capable of supporting
specified gravity loads without structural or architectural distress and that buildings
do not experience structural collapse during major earthquake or wind events at or
exceeding the minimum design level specified by the code. However, it is assumed
that such buildings may experience structural distortions or damage and moderate
to even severe cosmetic damage.
As the result of several natural and man-initiated disasters in the early 2000s,
other considerations have become more important, such as keeping exits open to
allow escape from the building and keeping essential facilities, such as hospitals and
communication centers through which emergency calls are routed, functional after
such events. The economic impacts of any occurrences on the structural and nonstruc-
tural elements of the building are also starting to receive more focused attention.
The use of building codes to establish minimum acceptable design standards
has been common practice since the early 1900s. Since that time, several regional
organizations have developed model building codes that varied across the United
States. In 1994, these regional organizations—Building Officials and Code Admin-
istrators International, Inc. (BOCA), International Conference of Building Officials
(ICBO), and Southern Building Code Congress International, Inc. (SBCCI)—
decided to combine their efforts and formed the International Code Council (ICC),
which published its first major code editions in 2000.
At the time of writing this book, most US states, counties, cities, and govern-
ment agencies have adopted one of the model codes, adapted one of the model codes,
or, in some cases, written their own building codes. In addition, some states adopt
a specific building code (or, often, a group or set of codes covering various functions)
for statewide use, sometimes with state amendments but usually leaving it up to the
local governing agencies to adapt this code and choose how compliance is to be
enforced. Local agencies usually have the prerogative to amend the state-ratified
122⏐ Appendix A—Codes for Building Structures and Structural Components
code or insert their own additional requirements. This may result in significantly
different requirements in adjacent or nearby jurisdictions.
To ensure code compliance, local governmental agencies require that builders
bring their projects through the building plan review process. The plan review
process typically considers the site where the structure will be built, the structure
itself, and all of the building functions, such as electrical, mechanical, plumbing, fire
protection, etc. Small municipalities, such as rural towns and counties, may not be
able to provide these services; therefore, code compliance may be done at the
regional or state level. Larger metropolitan areas have local planning boards that
review, approve, and issue building permits and, upon project completion, issue
certificates of occupancy.
Some states delegate their code adoption and enforcement authority to the
subordinate local jurisdictions but may exempt their own buildings from compliance
with local building codes or local amendments in favor of a statewide building code.
Additional nongovernmental reviews may be required in some instances, such as
lending institutions requiring conformance to local codes and insurance companies
sometimes encouraging conformance to specific codes or guidelines by offering
premium incentives for such compliance.
However, despite these various reviews and approvals that may be required, in
all cases, the design professional bears the ultimate responsibility to ensure that his
or her design meets the applicable building codes, that the design intent is under-
stood by the construction team, and that the structure in its final form is constructed
in general accordance with the contract documents.
The major codes used in the United States today are based on the International
Building Code (IBC), the model code published by the International Code Council.
This model code, now in its third edition (with editions dated 2000, 2003, and 2006),
was based largely on three earlier model codes: The BOCA National Building Code
(BOCA), the Standard Building Code (SBC), and the Uniform Building Code (UBC),
all of which were published in numerous editions.
Some jurisdictions still utilize one or more of these earlier codes. In addition,
the National Fire Protection Association (NFPA) produces a number of codes and
model codes, including NFPA 5000, Building Construction and Safety Code (NFPA
2003), which may be utilized by some jurisdictions. Guidelines, methodologies, and
supportive standards are also developed and provided by various technical societies.
Some of the guidelines and publications from such societies are listed below.
ASCE—American Society of Civil Engineers

This organization has published, among many others:
ASCE Standard 5, Building Code Requirements for Masonry Structures
ASCE Standard 6, Specifications for Masonry Structures
ASCE Standard 7, Minimum Design Loads for Buildings and Other Structures
ASCE Standard 8, Specifications for the Design of Cold-Formed Stainless Steel
Structures
ASCE Standard 11, Guidelines for Structural Condition Assessment of Existing
Buildings
ACI—American Concrete Institute

ACI 318, Building Code Requirements for Structural Concrete
ACI 530, Building Code for Masonry Structures
AISC—American Institute of Steel Construction

AISC 303, Code of Standard Practice for Steel Buildings and Bridges
AISC 335, Specifications for Structural Steel Buildings (Allowable Stress and
Plastic Design)
AISC 341, Seismic Provisions for Structural Steel Buildings
AISC 350, Load and Resistance Factor Design Specifications for Structural
Steel Buildings
AISI—American Iron and Steel Institute

Specification for the Design of Cold-Formed Steel Structures
AWS—American Welding Society

AWS D1.1, Structural Steel Welding Code
AWS D1.3, Structural Welding Code for Sheet Steel
AWS D1.4, Structural Welding Code for Reinforcing Bars
US Department of the Army and Air Force

TM 5-805-13, Raised Floor Systems
TM 5-809-10, Seismic Design for Buildings
DOD—Department of Defense
UFC 4-023-03, Design of Buildings to Resist Progressive Collapse
FM—Factory Mutual
FM Global has produced property loss prevention data sheets for many building
issues or components, including:
1-28, Design Wind Loads
1-29, Roof Deck Securement and Above-Deck Roof Components
1-54, Roof Loads for New Construction
SDI—Steel Deck Institute

SDI 30, Design Manual for Composite Deck, Form Deck and Roof Decks
SDI-DDMO2, Diaphragm Design Manual
SJI—Steel Joist Institute

Standard Specifications and Load Tables for Steel Joists and Joist Girders
Appendix B
Weights of Materials
Table B.1 Minimum Dead Loads
Loads shown below are approximate. Many materials have variable weights. When in
doubt, consult the manufacturer or supplier.
Material Weight, psf Weight, kN/m2

Roof Coverings
Built-Up Composition Roof
3-Ply Ready Roofing 1.0 0.05
3-Ply Felt 1.5 0.07
3-Play Felt with Gravel 5.5 0.26
5-Ply Felt 2.5 0.12
5-Ply Felt with Gravel 6.5 0.31
Single-Ply Roofing
Bituminous, Gravel-Covered 5.5 0.26
Bituminous, Smooth Surface 1.5 0.07
EPDM, Adhered 0.7 0.03
EPDM, Ballasted 12 and up 0.57 and up
Liquid Applied 1.0 0.05
Shingles
Asbestos-Cement 4.0 0.19
Asphalt 2.0 0.10
Wood 3.0 0.14
126⏐ Appendix B–Weights of Materials
Table B.1 Minimum Dead Loads (Continued)

Roof Coverings (Continued)
Tile (for Mortared System, Add 10 psf)
Concrete 10 to 16 0.48 to 0.77
Roman Clay Tile 12.0 0.57
Spanish Clay Tile 19.0 0.91
Ludowici Clay Tile 10.0 0.48
Other roofing materials
Copper or Tin 1.0 0.05
Corrugated Metal Roofing 2.0 0.10
Skylight 8.0 0.38
Slate (3/16 in.) 7.0 0.34
Slate (1/4 in.) 10.0 0.48
Standing Seam Metal Roofing 1.5 0.07
Wood (per Inch) 3.0 0.14
Insulation (per Inch of Thickness)
Cellular Glass 0.7 0.03
Expanded Polystyrene 0.2 0.01
Fiberboard 1.5 0.07
Fiberglass, Rigid 1.5 0.07
Perlite 0.8 0.04
Rigid 1.5 0.07
Urethane Foam with Skin 0.5 0.02
Decking
Wood (3/4 in.) 3.0 0.14
Gypsum Board (1 in.) 4.0 0.19
Insulrock (1 in.) 2.7 0.13
Poured Gypsum (1 in.) 6.5 0.31
Tectum (1 in.) 2.0 0.10
Vermiculite concrete (1 in.) 2.7 0.13

Roof Coverings (Continued)
Decking (Continued)
Plywood (1 in.) 3.0 0.14
1 in. Sheathing (Nominal) 2.1 0.10
2 in. Timber Decking 4.2 0.20
1.5 in. 22 ga. Wide Rib Galvanized 1.8 0.09
3 in. 22 ga. Galvanized 2.5 0.12
Ceilings
Acoustical Fiber Tile 1.0 0.05

Channel System, Suspended 1.0 0.05
Channel System, Suspended Steel, Seismic Rated 2.0 0.10
1/2 in. Gypsum Board 2.2 0.11
MEP Allowance 4.0 0.19
Plaster (1 in. Thick) 8.0 0.38
Sprinkler System 3.0 0.14
Suspended Metal Lath and Cement Plaster 15.0 0.72
Suspended Metal Lath and Gypsum Plaster 10.0 0.48
Walls and Partitions
Wood Studs (Douglas Fir and Southern Pine)

2×4 at 12 in. o.c. 1.4 0.07
2×4 at 16 in. o.c. 1.1 0.05
2×4 at 24 in. o.c. 0.7 0.03
2×6 at 12 in. o.c. 2.2 0.11
2×6 at 16 in. o.c. 1.7 0.08
2×6 at 24 in. o.c. 1.1 0.05

Walls and Partitions (Continued)
Stucco (1 in.) 12.0 0.57
Wood Paneling (1 in.) 2.5 0.12
Glass, Plate, 1/4 in. 3.3 0.16
Cement Plaster (1 in.) 8.0 0.38
Gypsum Plaster (1 in.) 5.0 0.24
Window (Glass, Frame, and Sash) 8.0 0.38
Masonry
Clay Brick (per Wythe)

4 in. Brick 40.0 1.92
8 in. Brick 80.0 3.83
12 in. Brick 120.0 5.75
Hollow Concrete Block (CMU, Lightweight, ≤105 pcf)
4 in. Block 22.0 1.05
6 in. Block 24.0 1.15
8 in. Block 31.0 1.48
12 in. Block 43.0 2.06
Hollow Concrete Block (CMU, Heavyweight, ≤135 pcf):
4 in. Block 29.0 1.39
6 in. Block 43.0 2.06
8 in. Block 39.0 1.87
12 in. Block 54.0 2.59
Other (per Wythe)
4 in. Stone 55.0 2.63
4 in. Glass Block 18.0 0.86
Floors
Steel See Manufacturer’s Data

Floors (Continued)
Concrete (1 in. Thickness)
Regular Weight 12.0 0.57
Semi-Light Weight 9.5 0.45
Reinforcement (Typical Distribution) 0.5 0.02
Raised-Access Floor 10 to 15 0.48 to 0.72
Hardwood (1 in. Nominal) 4.0 0.19
Plywood (per 1 in. Thickness) 3.0 0.14
Ceramic or Quarry Tile (3/4 in.) 10.0 0.48
Linoleum (1/4 in.) 1.0 0.05
Terrazzo Finish (1.5 in.) 19.0 0.91
Wood Joists
Size 12 in. spacing 16 in. spacing 24 in. spacing

2×6 2.2 1.7 1.1
2×8 2.9 2.2 1.5
2×10 3.7 2.8 1.9
2×12 4.4 3.3 2.2
Table B.2 Weights of Materials
Substance Weight, lb/ft3 Weight, kN/m3
Concrete (Reinforced)
Normal Weight, Stone Aggregate 150.0 23.57
Light Weight 110 to 120 17.28 to 18.85

Earth
Clay (Dry) 63.0 9.90
Clay (Damp, Plastic) 110.0 17.28

Earth (Dry) Packed 95.0 14.93
Earth (Moist) Packed 96.0 15.08
Earth (Mud) Packed 115.0 18.07
Sand, Gravel Dry Loose 95 to 105 14.93 to 16.5
Sand, Gravel Dry Packed 100 to 120 15.71 to 18.85
Sand, Gravel Wet 118 to 120 18.54 to 18.85

Liquids
Water, Fresh 62.4 9.80
Water, Ice 57.0 8.96
Water, Sea 64.0 10.06

Masonry and Components
Cast-stone (Cement, Sand, Stone) 144.0 22.63
Cement, Portland 90.0 14.14
Concrete, Cinder 111.0 17.44
Concrete, Slag 138.0 21.68
Concrete, Stone, Sand 150.0 23.57
Brick, Soft 100.0 15.71
Brick, Medium 115.0 18.07
Brick, Hard 130.0 20.43
Lime, Hydrated, Compacted 45.0 7.07
Portland Cement 90.0 14.14
Sand, Clean and Dry 100.0 15.71
Table B.2 Weights of Materials (Continued)
Masonry and Components (Continued)
Sand, River, Dry 106.0 16.65
Stone, Granite, Marble (Piled) 96.0 15.08

Metals and Alloys
Aluminum 170.0 26.71
Brass 526.0 82.65

Bronze 552.0 86.73
Copper 556.0 87.36
Iron, Cast 450.0 70.70
Iron, Wrought 480.0 75.42
Lead 710.0 111.56
Steel, Rolled 490.0 76.99
Tin 459.0 72.12
Zinc, Rolled Sheet 449.0 70.55

Minerals
Coal, Anthracite, Piled 55.0 8.64
Coal, Bituminous or Lignite, Piled 49.0 7.70
Granite 175.0 27.50
Gypsum, Loose 70.0 11.00
Limestone, Crystalline 165.0 25.93
Limestone, Oolitic 135.0 21.21
Marble 173.0 27.18
Quartz 165.0 25.93
Sandstone 145.0 22.78
Slate 172.0 27.02

Timber, US Seasoned
Cedar Red 22.0 3.46
Fir, Douglas Spruce 32.0 5.03
Table B.2 Weights of Materials (Continued)
Timber, US Seasoned (Continued)
Elm 45.0 7.07
Hemlock 29.0 4.56
Maple (Hard) 43.0 6.76
Oak (Live) 59.0 9.27
Oak (White) 46.0 7.23

Pine, Yellow, Long Leaf 44.0 6.91
Pine, Yellow, Short Leaf 38.0 5.97
Redwood, California 26.0 4.09
Spruce 27.0 4.24

Other
Glass 160.0 25.14
Gypsum Wallboard 50.0 7.86
Appendix C
Raised-Access Floor
Calculations
C.1 RAISED-ACCESS FLOOR PANEL LOADING DEFINITIONS
C.1.1 Concentrated Load

The concentrated load capacity of a RAF panel is defined as the capability of
the RAF panel to withstand a load placed on one square inch area with a resulting
deflection no more than 0.100 in. (2.54 mm) and a permanent set no more than 0.010
in. (0.254 mm) when the load is removed (see Figure C.1). The concentrated load
capacity of a RAF panel varies from 1000 lb (4445 N) to 3500 lb (15,572 N) depend-
ing on the type and manufacturer of the floor panel. Note: The term concentrated
load is synonymous with the term point load.
When a multiple-equipment machine configuration is to be installed as a group,
one floor panel may be subjected to two point loads. One of the casters from each
adjacent machine can impart a high load on a floor panel. For a machine weighing
M, the nominal caster load is M/3 and the worst case is M/2. At a given time, only
three of the four casters will bear the total weight of the equipment (a plane is defined
by three points). For equipment that is heavy on one side, the concentrated load
imparted by the caster can be adjusted up.
C.1.2 Uniform Load

The uniform load capacity of a floor panel is defined as the panel’s square-foot
load capacity and is 25% of the panel’s concentrated load capacity. Top surface
deflection resulting from uniform loading will be no more than 0.060 in. (1.52 mm)
and permanent set will be no more than 0.010 in. (0.254 mm) after the load is
removed. The uniform load imparted by an equipment weighing M and with the
machine area equal to A is equal to M/A.
C.1.3 Ultimate Load

The ultimate load capacity of a RAF panel is defined as the maximum load the
floor panel can withstand without failing when a load is applied on a 1 in.2
(645 mm2) area of the panel.
134⏐ Appendix C—Raised-Access Floor Calculations
Figure C.1 Concentrated load capacity.
C.1.4 Rolling Load

The rolling load is defined as the capability of a floor panel to withstand a rolling
load of specific wheel diameter and width and imparting a permanent deformation
no greater than 0.040 in. (1mm). For a piece of equipment weighing M, the estimated
nominal rolling load per caster is M/3 and the worst case is M/2. Rolling load capa-
bility is designated for infrequent heavy equipment loads (based on a 10 pass test)
and for frequent non-equipment loads (based on a 10,000 pass test).
C.2 FLOOR LOADING
C.2.1 Definitions
For the purpose of floor loading calculation, the following definitions will be
used in this chapter.
machine area: the area directly beneath the datacom equipment defined by the
length and width dimensions representing the equipment’s perimeter. The machine
area is represented as A in the formulas in the next section.
service clearance area: the area around the datacom equipment. Service clearance
areas of adjacent data equipment may overlap. The size of the service clearance area
depends on the application of the equipment and the expected environment in which
the equipment will be installed. When the equipment is installed against a wall, the
service clearance area should permit front-end servicing or adequate provisions
should be made to facilitate movement of the equipment as required.
weight distribution area: the area around the datacom equipment. Weight distribu-
tion areas may not overlap. Given that service clearance areas can overlap but weight
distribution areas cannot, when two pieces of equipment are installed next to one
another, only half the area between the equipment can be used for weight distribution
for either equipment. If the result is not sufficient for proper weight distribution, the
distance x between the machines has to be increased until the proper distribution is
achieved. The weight distribution area is represented as S in the formulas in the next
section.
C.2.2 Floor Load Rating

A subfloor load rating or maximum building floor allowable distributed load
varies from building to building—some datacom equipment facilities are
constructed with floors rated to support a uniformly distributed load of 70 lb/ft2
(3350 N/m2). In any case, it is important to verify the allowable floor loading capac-
ity. To estimate the uniform distributed load it is necessary to define the components
of the loads imparted by the datacom equipment including the datacom equipment
installation and the system cable cabling loads.
The live load is the weight imposed around the equipment in the weight distri-
bution area by personnel traffic, test equipment, and various carts, user’s guides,
and manuals. The live load is represented as K1 and is estimated to be 15 lb/ft2
(718 N/m2) for the examples described in the next section.
In many cases, datacom equipment is installed in a RAF environment, which is
a structure used to create a floor area above the building structural floor. The area
below the RAF is used to facilitate cabling and chilled air distribution. The RAF and
the system cables impose an estimated distributed load of 10 lb/ft2 (478 N/m2) on
the building structural floor. This factor is represented as K2 in the examples in the
next section.
Both K1 and K2 factors can be adjusted upward as necessary. For example, high-
performance communication equipment with heavy cables may need to have higher
K2 values.
C.3 FLOOR LOADING CALCULATIONS
C.3.1 General Formulas

The weight of the equipment to be installed must be such that the floor loading
(FL) is less than or equal to the building floor load rating (FLR) (IBM 2001).
The floor loading (FL) is
+ ( K1 × S ) + K2 × ( S + A )
FL = M
---------------------------------------------------------------------- , (C.1)
S+A
where
FLR = maximum floor load rating in N/m2
FL = floor loading in N/m2
M = datacom equipment weight in N
K1 = live load in the weight distribution area at 718 N/m2
K2 = RAF/cable load for the area at 478 N/m2
A = machine area in m2
S = weight distribution area in m2
C.3.2 Floor Loading Calculation Examples

Example 1
Given a frame weighing 5000 N with a footprint 750 × 1525 mm and a weight
distribution area as shown in Figure C.2, the machine floor loading is determined as
follows:
A = 0.75 × 1.525 = 1.144 m2
S + A = (0.75 + 0.76 + 0.76) × (1.525 + 0.76 + 0.76) = 6.912 m2
S = 6.912 m2 – 1.144 m2 = 5.768 m2
FL = {5000 + (718 × 5.768) + (478 × 6.912)} / (6.912) = 1800 N/m2
In this example, the floor loading is less than the allowable 3450 N/m2.
Example 2
A multiple-frame machine configuration is to be installed as a group (four
frames total).
Machine dimensions: 0.75 × 1 m

Machine weight: 6000 N
Service clearance:
Front and rear: 1.2 m
Between frames: 0 m
Left and right of group of frames: 1 m
Use half the service clearance dimensions (1.2/2 = 0.6 m and 1/2 = 0.5 m) for
calculating the weight distribution area.
Figure C.2 Weight distribution area. Reproduced with permission from

IBM (2001).
Check whether the weight distribution area based solely on the service clear-
ance will result in an FL that is less than 3450 N/m2. (Half of the service clearance
area distance is used assuming additional groups of frame systems could be placed
with only the service clearance between them and this group.)
First check the four-frame system as a whole:
M = 6000 × 4 = 24,000 N
A = (0.75 × 1) × 4 = 3 m2
S = {(1 m / 2 × 2) + 0.75 m × 4} × (1.2/2 m + 1 m + 1.2/2 m) – A = 5.8 m2
S + A = 5.8 m2 + 3 m2 = 8.8 m2
FL = {24,000 + (718 × 5.8) + (478 × 8.8)} / (8.8) = 3627 N/m2

In this example, assuming the allowable uniformly distributed floor loading is
3450 N/m2, the service clearance area must be increased in order to increase the
weight distribution area and to reduce the floor loading magnitude.
C.4 STRUCTURAL GUIDELINES FOR

RAISED-ACCESS FLOOR SYSTEMS
Most RAFs in seismic regions are installed with “seismic grade” pedestals
fastened to the concrete subfloor with high-strength pedestal adhesive. Adhesives
eliminate the need for mechanical fasteners or four-leg seismic braces. However,
extreme conditions such as unusually high RAFs, heavy floor loading, and proxim-
ity to major faults (or a combination of these conditions) may require that the pedes-
tals be fastened with mechanical fasteners or reinforced with bracing. The
determination for the right pedestal type and fastening method for a planned RAF
is based on seismic analysis using lateral force equations specified by the building
code applicable to the project. Keep in mind that building codes do not establish seis-
mic performance levels for RAFs—they merely provide equations for determining
the seismic horizontal force on RAFs. The actual seismic understructure require-
ments depend on the seismic criteria and loading requirements established for the
project. The prevailing code utilized to calculate the seismic horizontal force is
ASCE Standard 7-05, Minimum Design Loads for Buildings and Other Structures
(ASCE 2005), in combination with the International Building Code (ICC 2003).
C.4.1 Determining Seismic Horizontal Force, Fp

To determine the adequate pedestal type (or types) and the fastening or reinforc-
ing methods to utilize, the criteria listed below is required to determine the seismic
horizontal force, Fp. Fp is calculated using the equation located in Section 13.3-1 of
ASCE Standard 7-05 (ASCE 2005) and is discussed in more detail in Appendix E of
this book.
Fp = [(0.4apSDSWp)/(Rp/Ip)][1 + 2(z/h)] (C.2)
except that Fp shall not be less than Fp = 0.3 SDSIpWp and where
ap = component amplification factor (Table 13.5-1 of ASCE Standard 7-05,
given as 1.0 for access floors
SDS = design spectral response acceleration
Wp = component operating weight
Rp = component response modification factor, which is either 1.5 or 2.5
depending on whether the pedestals are fastened to the subfloor by
adhesive or mechanical fasteners (Table 13.5-1 of ASCE Standard 7-05)
Ip = component importance factor
z = height in the structure where the access floor is attached to the building
h = roof height of the building with respect to the base
C.4.2 Calculating Design Spectral Response Acceleration, SDS

(Section 1615.1.3 from the IBC [ICC 2003])
The design spectral response acceleration (SDS) value, which is the final accel-
eration coefficient that considers the project location and soil type, is calculated
using the following equation (Equation 16.40 of the IBC [ICC 2003]):
SDS = (2/3)SMS (C.3)
where SMS is the maximum considered earthquake value, FaSs (Section 1615.1.2 of
the IBC), where
SS = mapped maximum spectral acceleration at short periods (Section
1615.1.2 of the IBC)
Fa = site coefficient, which is a function of the mapped spectral acceleration
at short periods (Ss) and the site class (Table 1615.1.2 of the IBC). Site
classes are based on soil types, which are listed in Table 1615.1.2 of the
IBC. When a project’s site class is unspecified, Class D (stiff soil) is
typically used for the design.
C.4.3 Determining Mapped Maximum Spectral Acceleration

at Short Periods, Ss
The mapped maximum spectral acceleration (Ss) used for calculating Fp is the
short period 0.2 s mapped value taken from Figures 1615.1 through 1615.10 in
the IBC.
C.4.4 Calculating Component Operating Weight, Wp

The component operating weight includes the uniformly distributed equipment
live load on the RAF (equipment load not attached to the RAF and as a result typi-
cally reducible by 75%) plus the RAF weight plus 100% of the weight of equipment
and/or partitions attached to the RAF (see Chapter 13 of ASCE Standard 7-05
[ASCE 2005]). Remember that the live-load value used for the RAF design is based
on the expected use of the floor rather than the floor’s rated uniform load capacity.
The IBC designates minimum uniformly distributed live loads of 100 psf for data-
com rooms and 50 psf for offices. Other load values may be substituted at the user’s
discretion.
The component operating weight of a typical datacom RAF with its loads is
calculated using the following:
Wp = [(100.0 psf × 25%) + 10.0 psf + 10.0 psf] = 45.0 psf = 2154 N/m2
where
equipment live load = 100.0 psf = 4788 N/m2
access floor dead load = 10.0 psf = 479 N/m2
partition dead load = 10.0 psf = 479 N/m2
C.4.5 Determining Component Importance Factor, Ip

The component importance factor (Ip) is selected as follows:
Seismic Use Group I: Ip = 1.0 (for all other occupancies not noted below)
Seismic Use Group II: Ip = 1.25 (for structures that have a substantial public
hazard due to occupancy or use)
Seismic Use Group III: Ip = 1.5 (for essential facilities that must function after
an earthquake)
C.4.6 Selecting the Appropriate Pedestal

After the applied seismic horizontal force, Fp, is determined at the RAF level,
the seismic moment that will be applied to the floor pedestal is calculated using the
pedestal’s tributary area and height. The proper pedestal type and fastening method
is then chosen based on the known moment capacity limitations of the various pedes-
tal types available. Capacity limitations of pedestals can be verified by structural
calculations or by physical overturning moment tests.
Appendix D
Data Center
Vibration Measurement
D.1 OVERVIEW
This appendix presents an overview of typical operational shock and vibration
test profiles for datacom and cooling equipment. The test profiles can also be used
as the limits, the acceptable vibration levels for a datacom center. The effect of this
shock and vibration on the functional performance of the datacom equipment
depends on how the equipment cabinet transmits this shock and vibration to critical
components. The acceptable level of shock and vibration can normally be obtained
from the datacom equipment manufacturer. However, the actual level of floor vibra-
tion in a datacom room is not readily available. This appendix is intended to provide
information on how to record the data and what kind of data is needed to evaluate
the effect of vibration on the operation or functionality of the datacom equipment.
In addition, this appendix informs the datacom center operator what vibration
magnitudes are considered normal or high.
D.2 INTRODUCTION
Datacom equipment such as high-performance computer servers, storage serv-
ers, networking equipment, rack-mounted equipment, and cooling equipment can
induce and transmit vibration to surrounding datacom equipment. As the datacom
equipment gets more powerful, more cooling is required, which translates to higher
blower or pump speeds. As a result, higher levels of vibration magnitude will be
transmitted to the floor and to the datacom equipment.
In certain cases, the operator of the datacom equipment can feel the vibration
and express concern as to the effect of this vibration on the operation of the datacom
equipment in the building. Another datacom center operator may be concerned that
the datacom equipment room has excessive vibration from forklift vehicles and
trucks operating in the warehouse that is close to the equipment room. In another
location, the datacom equipment room is on twelfth floor of a 26-story building. The
adjacent building physically attached to the main structure has electrical power
generators that produce levels of vibration that can be felt anywhere inside the build-
ing. All of these examples have vibration environments that may effect the datacom
equipment’s operation. To evaluate the effect of these operational vibration levels,
142⏐ Appendix D—Data Center Vibration Measurement
one needs to know what level of vibration the datacom equipment can tolerate and
the vibration magnitude transmitted to the datacom equipment. The first piece of
information can be obtained from the datacom equipment manufacturer, but each
type of datacom equipment may have different specifications, such as a maximum
g-force level at any frequency range or a certain PSD profile. The second piece of
information is the data center vibration, which may not be readily available. In this
appendix, the typical levels to which datacom equipment can be tested are discussed.
Next, the tests on datacom equipment, the criteria for an acceptable datacom equip-
ment room, and the methodology for datacom equipment room measurements are
presented. Then some actual measurements on datacom equipment rooms are
presented.
In this appendix, no attempts are made to evaluate the effect of operational
vibration (which is dominantly in the vertical direction) on the functional perfor-
mance of the datacom equipment. The datacom equipment manufacturer can eval-
uate the effect of the vibration on their datacom equipment based on the testing and
design of the datacom equipment. Typical operational testing is given for guidance
purposes only.
D.3 TYPICAL OPERATIONAL VIBRATION AND SHOCK TESTING

A typical transient test for a period of 30 min is given in Figure D.1. This profile
represents a probability of less than 0.001 that the vibration levels at the mounting
surface of the product will be greater than the levels specified in Figure D.1. In
general, the specified levels apply to the three orthogonal axes except for a rack-
mount product, where one axis is sufficient. The grms of this profile is 0.1 grms. In
Figure D.1 Typical power spectral density of operational vibration

encounter in a data center. Reproduced with permission from
IBM (1990).
addition, sine-on-random test techniques are recommended to simulate the line

frequency of 50 and 60 Hz at 0.06g. If sine-on-random capability is not available,
15 min sine dwells at 50 and 60 Hz shall be run at 0.06g. For a frequency range from
17–500 Hz, the amplitude of Level 1 vibration shown in Figure D.1 is 2.2 × 10–5 g2/
Hz, and for Level 2, the amplitude is 3.0 × 10–4 g2/Hz.
For heavy rack-mounted datacom equipment, Level 1 is the appropriate test
profile, and for light datacom equipment, Level 2 test profile can be used. The test
profile is applied to the base or at the product’s support points, such as at the casters
or leveling feet. These two levels can also be used as the upper limit for the vibration
transmitted to the data center floor by datacom equipment or cooling units.
A continuous vibration, as shown in Figure D.2, is a vibration that can be present
in a machine throughout its product life. A continuous vibration can produce and
sustain resonance.
Telcordia’s NEBS GR-63-Core specifies somewhat different magnitudes and
frequencies. It also specifies a sine sweep at 0.1g from 5–100 Hz (Telcordia 2006).
Occasionally, some components within the datacom equipment cabinet can
have a “hot swap” capability during which a component is removed and a new
Figure D.2 Typical operational vibration encountered in a data center.

Reproduced with permission from IBM (1990).
component is installed while the system is running. During this operation, a shock
magnitude up to 30g with 3 ms pulse width in the vertical direction and about 15 g
with 3 ms pulse width can occur (IBM 1990).
D.4 TYPICAL OPERATIONAL VIBRATION AND SHOCK

MAGNITUDE RECORDED IN A DATA CENTER
The shock and vibration levels in the previous section were determined by
conducting actual vibration measurements in datacom equipment rooms. The
magnitudes given are an envelope of the peak amplitudes. These values can also be
used as typical limits of vibration amplitude produced by datacom or cooling equip-
ment. Therefore, the limits can also be used as the acceptable datacom equipment
room vibration levels since datacom equipment is tested to meet these limits. Actu-
ally, most datacom equipment can withstand higher magnitudes of operational shock
and vibration. However, datacom equipment manufacturers may not test above the
level indicated in the previous section.
The solid line in Figure D.3 shows the level of vibration shown in Figure D.2.
The dotted line shows the threshold where people start to feel the vibration (Harris
and Piersol 2001), and the dashed line represents a vibration level that is unpleas-
ant to people. Generally, as long as people do not feel the unpleasant vibration
level, the level of datacom equipment room vibration is within the allowable or
acceptable level.
Figure D.3 Average peak acceleration threshold of human perception of

vibration. Reproduced with permission from Notohardjono
(2006).
D.5 MONITORING FLOOR VIBRATION IN DATA CENTERS
Figures D.1 and D.2 depict the expected vibration magnitude in a data center.
Datacom manufacturers require not only the magnitude but also the frequency
related to this magnitude as key information for evaluating the effect of vibration on
the datacom equipment. Generally, recording vibration amplitude data alone,
whether it is in g-force or velocity, requires lower-cost measurement instrumentation
than collecting data in order to produce magnitude and frequency content. However,
it is critical that g or PSD versus frequency data be recorded. The frequency range
of the data should be between 5–500 Hz. Accomplishing this task may require two
sets of data collection; one between 2 and 50 Hz to capture the low frequency content
and the second between 50 and 500 Hz to capture frequencies higher than 50 Hz.
One data center layout is shown in Figure D.4. The cooling units are shown as
A/C boxes; the datacom equipment is shown as S boxes. Vibration was monitored
at nine locations, denoted as I-1, I-2, I-3, II-1, ... III-3.
Figure D.5 illustrates typical acceleration in g for a frequency range from
2–50 Hz, and Figure D.6 shows similar information but at a frequency range
from 5–500 Hz. The maximum g is 0.002g at 220 Hz. This value is an order of
magnitude lower than the limit given in Figure D.2 (0.035). The areas near the
cooling units have higher vibration than the other areas. Figures D.7 and D.8
show the data in a different format. Here, the vibration magnitude is given in the
Figure D.4 Data center layout. Reproduced with permission from

Figure D.5 Floor vibration recorded at a data center—location II-1 from

2–50 Hz. Reproduced with permission from Notohardjono
(2006).
Figure D.6 Floor vibration recorded at a data center—location II-1 from

5–500 Hz. Reproduced with permission from Notohardjono
(2006).
vertical axis as g2/Hz vs. Hz. At 220 Hz, the vertical axis is 5.4 × 10E–5 g2/Hz,
which is somewhat higher than Level 1 but is still below Level 2 (see Figure D.1).
D.6 BEST PRACTICES
1. Level 1 (see Figure D.1) is the appropriate vibration input test level for a heavy
datacom equipment cabinet. This level is intended to mimic the datacom equip-
ment room floor vibration and to test that the functionality of the system is not
affected during the test. For a small and relatively light equipment cabinet,
Level 2 can be used as a typical test level.
2. Vibration amplitudes at 0.07g are most likely the upper limit of vibration at the
base of the datacom equipment. This level can be considered normal data center
vibration and most likely will not induce any functional degradation to datacom
equipment.
3. When a concern arises as to whether the data center’s floor vibration is too high,
an actual measurement of data center vibration is warranted. Both PSD vs.
frequency and g vs. frequency plots are the information needed to evaluate this
concern.
Figure D.7 Power spectral density of floor vibration recorded at a data

center—location II-1 from 2–50 Hz. Reproduced with
Figure D.8 Power spectral density of floor vibration recorded at a data

center—location II-1 from 5–500 Hz. Reproduced with
Appendix E
Component
Anchorage Forces
The horizontal seismic design force is calculated in accordance with the follow-
ing three formulas included in ASCE Standard 7-05 (ASCE 2005):
0.4a p S DS W p
F p = ------------------------------- ⎛⎝ 1 + 2 ---⎞⎠
z
(E.1)
Rp h
Fp
max
= 1.6S DS I p W p (E.2)
Fp
min
= 0.3S DS I p W p (E.3)
where
Fp = horizontal design seismic force
ap = component amplification factor
SDS = design spectral response at short periods
Wp = component operating weight
Rp = component response modification factor
Ip = component importance factor
z = component elevation above grade in the building
h = average roof height of the structure above grade
The above formula for the calculation of lateral forces has changed quite a bit
in the last decade. It is instructive to review a number of the variables that go into
this equation in more detail in order to better understand their impact.
E.1 RELATIVE LOCATION FACTOR (1 + 2z/h)

Under seismic loading, buildings behave dynamically. That is, they sway back
and forth due to the energy caused by the input of earthquake ground motion at the
base of the structure. Although a structure has many significant modes of movement,
the first or characteristic mode (the one with the longest period of oscillation or, in
other words, the lowest frequency of oscillation) often induces the most significant
structural deflection. The IBC assumes this type of behavior for most structures in
150⏐ Appendix E—Component Anchorage Forces
its linear elastic force provisions. A triangular distribution of forces with the largest
force at the roof is assumed by the code because the forces at the roof are higher than
those at the lower floors. This story force is also imparted on all the nonstructural
components attached to the building at each floor.
The relative location factor is used to scale the seismic coefficient at grade to
approximate the dynamic amplification felt by the component when it is located
much higher in the building. The factor ranges from 1.0 (component at grade) to
3.0 (component at roof).
In previous versions of the building code, this factor was (1 + 3z/h), and thus the
maximum multiplier went up to 4.0 for roof-mounted locations. However, subse-
quent research showed that this was too conservative, and the coefficient was
lowered to 2.
It should be noted that the best location for a high-performance datacom equip-
ment center would be on the ground floor, as the seismic forces there would be the
lowest of any location in the building.
E.2 COMPONENT IMPORTANCE FACTOR (Ip)

The IBC uses an “importance” factor to increase design loading for wind, seis-
mic loading, and snow for buildings that contain hazardous materials or buildings
that are required to remain in operation immediately following a damaging earth-
quake or other natural disaster. This increase is meant to reduce the amount of
damage sustained by the structure and both limit the release of chemicals that could
endanger the public and increase the potential that the building will remain in oper-
ation immediately afterward.
The code has a similar importance factor (Ip) for particular nonstructural
components. For normal components, the Ip is 1.0. The code requires a higher impor-
tance factor of 1.5 for three cases: components that are part of the building’s fire or
life safety system (like sprinklers), components that contain hazardous materials,
and components that are required for continued operation of immediate occupancy
facilities. The higher Ip factor increases the seismic design force and also causes
additional components to be braced and anchored with the objective of reducing
damage to these systems.
Many owners of high-performance data centers that are strategic assets should
consider using higher importance factors for certain components in order to protect
against large amounts of damage to these critical components. For example, an
owner should consider using a higher importance factor for the following compo-
nents: attachment of RAFs and anchorage of servers, critical chillers, cooling equip-
ment, and back-up power devices.
E.3 COMPONENT AMPLIFICATION FACTOR (ap)

Just as the relative location factor adjusts the design seismic force for the
dynamic amplification due to position of the component in the height of the building,
the component amplification factor (ap) accounts for the dynamic amplification due
to the support of the component. For example, if a component is rigidly attached to
the ground, the amplification due to the support would be negligible and the compo-
nent would be subjected to only the same force as the floor itself. However, if the
component is supported on a flimsy steel platform, the component would experience
additional force due to the “whipping” of the support directly below it—much like
the end of a snapping bullwhip.
Traditionally, the code has defined ap as a step function between fully rigid and
fully flexible even though neither of these cases really exists in real life. The code
definition of a fully “rigid” support is one with a fundamental period of less than or
equal to 0.06 s. With this stiff of a support, it is assumed that no dynamic amplifi-
cation occurs and the ap can be taken as 1.0. If the fundamental period is greater than
0.06 s (the frequency of the oscillation is lower than 16.7 Hz), the support is consid-
ered flexible and the ap is taken as 2.5. Tables 13.5-1 and 13.6-1 of ASCE
Standard 7-05 (ASCE 2005) list ap values for many common architectural, mechan-
ical, and electrical components.
In real life a support is never really “fully” rigid or flexible but somewhere in
between. Also, the much larger bracing forces required by the code for “flexible”
supports often are quite extensive and expensive. In these cases, the code allows a
more accurate determination of the component amplification factor based on
dynamic analysis, if desired. This can be calculated when reasonably accurate values
for both the fundamental period of the component and the structure are known. The
accurate representation of ap is the ratio of these two periods. The NEHRP Commen-
tary (BSSC 2003) has an excellent discussion of this procedure based on the provi-
sions included in Tri-Service Manual TM 5-809-10, Seismic Design for Buildings
(US Army 1992).
Of special discussion for datacom equipment rooms is the determination of the
component amplification factor for equipment such as equipment cabinets anchored
directly to a RAF.
E.4 COMPONENT RESPONSE MODIFICATION FACTOR (Rp)

The building code uses a response modification factor or R factor to account
for the ductility and energy dissipation capability of different structural lateral
force resisting component systems in buildings. The component systems that can
take more damage and still dissipate the energy input of an earthquake have much
higher R values than component systems that are brittle or break early in an earth-
quake event.
The IBC has a similar factor for nonstructural components: the component
response modification factor (Rp). This factor is intended to represent the amount of
energy absorption capability of a particular component and its attachment to the
larger structure. For brittle elements or components that are expected to buckle
before yielding, the Rp ranges from 1.0 to 2.0. For components that have some mini-
mal amount of energy dissipation capacity, the value of Rp increases to a common
value of 2.5. For highly ductile elements, the value of Rp ranges widely from 3.0 to
12.0. Tables 13.5-1 and 13.6-1 of ASCE Standard 7-05 (ASCE 2005) list Rp values
for many common architectural, mechanical, and electrical components.
152⏐ Appendix E—Component Anchorage Forces
It should be noted that the values of Rp in Table 13.6-1 of ASCE Standard 7-05
have increased dramatically for many components in the latest edition of the code.
These increases acknowledge recent analysis work to develop less conservative
values as well as the good past seismic performance of many of these systems. Also,
many systems are listed with a relatively high Rp factor but, in order to achieve this
good performance, specific seismic detailing is required. Owners should note that
some of the component systems used in high performance datacom equipment
rooms are not specifically listed in the table (except a generic mechanical or elec-
trical component). The owner should decide in advance on the values to be used for
specific datacom equipment room components.
Glossary of Terms
acceleration of gravity (g or G): g is the acceleration produced by gravity at the
surface of the earth and has an international standard value of 9.807 m/s2. Measured
accelerations are often expressed as a ratio of the measured acceleration divided by
gravitational acceleration; this unitless ratio is expressed as a G value.
ACI: American Concrete Institute.
AHJ: the local authority having jurisdiction.
amplitude or magnitude: the maximum value of a measured vibration. Amplitude
ANSI: American National Standards Institute.
ap : component amplification factor. See Section E.3 of Appendix E for further
discussion.
ASCE: American Society of Civil Engineers.
ASHRAE: American Society of Heating, Refrigerating and Air-Conditioning Engi-
neers, Inc.
ASTM International: formerly the American Society for Testing and Materials
(ASTM).
augered cast-in-place pile: see pile.
availability: a percentage value representing the degree to which a system or compo-
nent is operational and accessible when required for use.
ballasted roof: a roof system where the waterproof membrane or the multiple layers
of roof felts are held in place by the weight of ballast stone or concrete pavers.
base fixity: a number from 0 to 1 used to quantify how well the base of a frame is
attached to the ground (1 being a perfectly rigid connection, 0 being no connection).
BOCA: Building Officials and Code Administrators International, Inc.; now Inter-
national Code Council.
154⏐ Glossary of Terms
building loads:
collateral load: a type of dead load, the collateral load includes the weight of
any materials other than the permanent construction materials. The collateral
load may include electrical pipes, sprinkler systems, ceilings, lights, and other
materials.
dead load: the combined weight of all the permanent construction materials,
including roofing, framing, and other structural components such as walls,
floors, roofs, ceilings, stairways, built-in partitions, finishes, cladding, and
other similarly incorporated architectural and structural items and fixed service
equipment, including cranes. The dead load is determined in advance and is
assigned a low factor of safety in the overall building design. All dead loads are
considered permanent loads.
duration of load: the period of continuous application of a given load, or the
aggregate of periods of intermittent applications of the same load.
impact load: the load resulting from moving machinery, such as elevators,
craneways, vehicles, and other similar forces, and kinetic loads, pressure, and
possible surcharge from fixed or moving loads.
live load: the weight of building occupants, furniture, machines, equipment,
and the like. Since the live load is hard to quantify, it carries large factors of
safety in the overall design of the building. Other live load sources that occur
during construction and maintenance, such as those from tool boxes, temporary
generators, etc., must be taken into consideration as well; these loads may be
difficult to predict.
load factor: a factor that accounts for deviations of the actual load from the
nominal load, for uncertainties in the analysis that transforms the load into a
load effect, and for the probability that more than one extreme load will occur
simultaneously.
partition live load: the load carried by a wall that is one story or less in height
used to subdivide the interior space in a building. The loads accounted for are
produced only by the use and occupancy of the structure and not by external
forces such as weather, wind, or seismic loads.
seismic load: the force exerted on a structure due to seismic activity.
snow load: the maximum weight of snow that will collect on the roof of the
building. The snow load is highly dependant on the ultimate location of the
building. The slope (pitch) of the roof is a factor used to determine snow loads.
wind load: the force exerted on a structure by strong winds.
built-up roof: a roof constructed of multiple layers of roof felts laminated together
with a hot or cold applied bitumen, tar, etc., and often finished with a coating of rocks
or aggregate. It is called a built-up roof because it is made (or built up) at a job site.
cabinet: frame for housing electronic equipment that is enclosed by doors and is
stand-alone; this is generally found with high-end datacom equipment.
CDU: cooling distribution unit.
chase: building element within which building infrastructure (ductwork, piping,
etc.) is routed; vertical chases may also be referred to as risers or shafts.
chevron bracing: structural bracing members that start out wide at two points on
top of a component and then gradually converge to the bottom of another (see
Figure 6.5).
CISCA: Ceilings & Interior Systems Construction Association.
clear height: also headroom; refers to the distance from the top of the structural floor
slab to the lowest point of the overhead structural system (which may support a roof
system or a building floor above).
CMU: concrete masonry unit.
collateral load: see building loads.
combined footing: see footing.
component amplification factor: ap . See Section E.3 of Appendix E for further
discussion.
component importance factor: Ip . See Section E.2 of Appendix E for further
discussion.
component operating weight: Wp . See Section C.4.1 of Appendix C for further
discussion.
component response modification factor: Rp . See Section E.4 of Appendix E for
further discussion.
constraint element: a general class of elements that apply kinematic constraints
(rigid link, slider, revolute, point-in-plane, translational) between nodes.
CRAC: computer room air-conditioning unit.
craneway: the structure or structural members upon which a crane travels, which can
include columns, support girders, and rails.
data center: a building or portion of a building whose primary function is to house
a computer room and its support areas. Data centers typically contain high-end serv-
ers and storage products with mission-critical functions.
datacom: an abbreviation for the data and communications industry.
datacom equipment: refers to, but is not limited to, servers, storage products, work-
stations, personal computers, and transportable computers; may also be referred to
as electronic equipment or IT equipment.
dead load: see building loads.
design spectral response acceleration: SDS . See Section C.4.2 of Appendix C for
further discussion.
diaphragmatic action: the resistance to in-plane shear forces (such as from wind)
offered by roof decks or floor slabs that are rigid enough to transfer the forces to the
structural framing system.
drilled pier: see pier.
driven pile: see pile.
duration of load: see building loads.
EERI: Earthquake Engineering Research Institute.
EIA: Electronic Industry Alliance.
EJMA: Expansion Joint Manufacturers Association.
electronic equipment: see datacom equipment.
equipment support: those structural members or assemblies of members or manu-
factured elements, including braces, frames, lugs, hangers, or saddles, that transmit
gravity load and operating load between the equipment and the structure.
FEMA: Federal Emergency Management Agency.
finite element analysis: a numerical technique of solving a wide range of engineer-
ing problems. Finite element analysis is used to determine the stresses and displace-
ments of a mechanical structure. The first step in the analysis is to approximate the
actual geometry by a mesh of small elements bounded by points called nodes. The
next step is to define the boundary conditions, the applied load, and/or the displace-
ment on the elements and nodes.
footing: essentially a base in the ground that will support the weight (load) of the
building structure imposed upon it. The dimensions of a footing vary according to
the soil conditions under the building, the load placed on the footing, and the
construction style of the structure being supported.
combined footing: a type of footing that supports two or more columns.
monolithic footing: concrete footing poured and cast in one piece without
joints.
spread footing: supports the weight (load) from the exterior or foundation walls
of a building.
foundation: the connecting structure between the footing and the structural system
above.
mat foundation: a raft or floating foundation; a continuous footing that
supports a reinforced concrete slab covering a minimum of 75% of the total area
within the exterior walls of a building. It is normally only used when the subsoil
is poor.
ribbed slab foundation: a concrete slab on grade that is continuously cast with
deepened perimeter and interior beams placed in a grid pattern within the foot-
print of the slab.
Fp : seismic horizontal force. See Section C.4.1 of Appendix C for further discussion.
frame drift: the horizontal displacement of the structure at a floor or roof level with
respect to its original position; is a very important part of the structural design. Drift
limitation requirements can greatly impact the type of structure that is selected.
free-fall drop heights: the equivalent velocity change due to free-fall drop from a
certain height (h):
velocity change ΔV = (1 + e) ( 2gh )
where e is the coefficient of restitution and is the ratio of velocities after and before
an impact.
frequency: the reciprocal of a period in cycles per second, sometimes expressed as
Hertz (Hz).
grms: the root mean square value of the acceleration of gravity, g.
Guyan reduction solver: a numerical procedure used by finite element analysis as
an iterative algorithm that is an adaptation of power methods to find eigenvalues and
the eigenvector of a square matrix. It is particularly useful for finding decomposi-
tions of very large sparse matrices.
half-sine shock pulse: an ideal shock pulse for which the acceleration/time relation
has the shape of the positive (or negative) section of one cycle of a sine wave.
harmonic frequency, normal mode, natural frequency, resonance frequency: the
headroom: see clear height.
HVAC: heating, ventilating, and air conditioning.
IBC: International Building Code.
ICBO: International Conference of Building Officials.
ICC: International Code Council.
impact load: see building loads.
Ip : component importance factor. See Section E.2 of Appendix E for further
discussion.
IT: information technology.
IT equipment: see datacom equipment.
Lanczos method solver: a numerical procedure used by finite element analysis as an
iterative algorithm that is an adaptation of power methods to find eigenvalues and the
eigenvector of a square matrix. It is particularly useful for finding decompositions

of very large sparse matrices.
listed: equipment, materials, or services included in a list published by an organi-
zation that is acceptable to the authority having jurisdiction, that is concerned with
the evaluation of products or services, that maintains periodic inspection of produc-
tion of listed equipment or materials or periodic evaluation of services, and whose
listing states that either the equipment, material, or service meets appropriate desig-
nated standards or has been tested and found suitable for a specified purpose
(National Fire Protection Associaton).
live load: see building loads.
load: see building loads.
load, duration of: see building loads.
load factor: see building loads.
machine area: the area directly beneath the datacom equipment defined by the
length and width dimensions representing the equipment’s perimeter.
magnitude or amplitude: the maximum value of a measured vibration. Amplitude
mat foundation: see foundation.
MEP: mechanical, electrical, and plumbing.
minipiles: a soil stability improvement technique, also known as soil nails. See also
pile.
mission critical: a system that is absolutely necessary; an activity, device, service,
or system whose failure or disruption of normal business will result in the failure of
business operation.
moment-resisting frames: structures that resist applied forces through bending or
deflection of their members and connections.
monolithic footing: see footing.
natural frequency, normal mode, harmonic frequency, resonance frequency: the
NEBS: formerly Network Equipment-Building System; provides a set of physical,
environmental, and electrical requirements for a central office of a local exchange
carrier. NEBS is a trademark of Telcordia Technologies, Inc.
NEHRP: National Earthquake Hazards Reduction Program, a United States Govern-
ment agency. Today, there are four primary NEHRP agencies: 1) Federal Emergency
Management Agency (FEMA) of the Department of Homeland Security, 2) National
Institute of Standards and Technology (NIST) of the Department of Commerce
(NIST is the lead NEHRP agency), 3) National Science Foundation (NSF), and 4)
United States Geological Survey (USGS) of the Department of the Interior.
NFPA: National Fire Protection Association.
NIBS: National Institute of Building Sciences.
nonstructural components: all those items in a data center that are attached to the
structure but are not part of the building structure itself. Typically this includes archi-
tectural components such as the exterior building cladding, interior partition walls,
hung ceiling, raised-access flooring, and storage cabinets. It also includes mechan-
ical and electrical components such as HVAC units, pumps and compressors, eleva-
tors, motor control units, transformers, computers, server racks, and all associated
piping, ductwork, conduit, bus duct, and cable trays. And, most importantly, it
includes all of the servers on the data center floor.
normal mode, harmonic frequency, natural frequency, resonance frequency: the
parabolic shell element: a type of element used in finite element analysis to repre-
sent a surface.
partition live load: see building loads.
PDU: power distribution unit; the junction point between the uninterruptible power
supply and the cabinets containing equipment.
perforated floor tile: a tile as part of a raised-access flooring system that is engi-
neered to provide airflow from the cavity underneath the floor to the space. Tiles may
be with or without volume dampers.
pier: a vertical load-bearing member often used as a foundation. See also pile.
drilled pier: a deep foundation system that is constructed by placing fresh
concrete and reinforcing steel into a drilled shaft. Typical shaft diameters range
from 18 to 144 in.
rammed aggregate pier: a foundation soil improvement technique that is
achieved by boring holes in the soil and filling them with a compacted aggre-
gate, thereby strengthening the soil.
pile: a vertical structural shaft that transfers load through weak layers of soil to those
that are capable of supporting such loads. See also pier.
augered cast-in-place pile: a support that is created by removing a column of
soil to a certain depth through the use of an auger and filling the void with
concrete.
driven pile: a support that is driven into the ground by means of a hydraulic ram,
press, or some other means until the support reaches stable solid soil or bedrock.
power spectral density (PSD): a measure of the power content with respect to the
frequency for random vibration. The units are g2/Hz and represent the power present
in a 1-Hz-wide square filter. The square root of the integration of the PSD curve with
respect to frequency gives the overall root mean square level of the vibration.
PSD: see power spectral density.
pulse width: half of the period of a sinusoidal wave expressed in milliseconds (ms).
purlin: horizontal structural roof members used to support roof decking.
rack: a cabinet or frame for housing electronic equipment.
rack-mounted equipment: equipment mounted in an Electronic Industry Alliance
(EIA) or similar cabinet; these systems are generally specified in EIA units, such as
1U, 2U, 3U, etc., where 1U = 1.75 in. (44 mm).
raised-access floor (RAF): a platform with removable panels where equipment is
installed, with the intervening space between it and the main building floor used to
house the interconnecting cables and at times used as a means for supplying condi-
tioned air to the datacom equipment and the room.
RAF: raised-access floor; a platform with removable panels where equipment is
installed, with the intervening space between it and the main building floor used to
house the interconnecting cables and at times used as a means for supplying condi-
tioned air to the datacom equipment and the room.
rammed aggregate pier: see pier.
random vibration: a vibration having randomly varying amplitude and frequency
content within specified limits. It is typical of the vibration to which products are
exposed in the operating environment.
relative location factor (1 + 2z / h): adjusts the design seismic force for dynamic
amplification due to the position of a component in the height of the building. The
factor ranges from 1.0 (component at grade) to 3.0 (component at roof). See
Section E.1 in Appendix E for further discussion.
reliability: a percentage value representing the probability that a piece of equipment
or system will be operable throughout its mission duration. Values of 99.9% and
higher are common in datacom equipment areas. For individual components, the
reliability is often determined through testing. For assemblies and systems, reliabil-
ity is often the result of a mathematical evaluation based on the reliability of indi-
vidual components and any redundancy or diversity that may be employed.
resilience: ability of a system or component to recover and maintain its function
from both expected and unexpected events.
resonance: the point at which the natural frequency (resonance frequency) of an
element is equal to the excitation frequency and the point that produces the peak
amplification of input amplitude; a condition caused by excitation with an input
frequency that is at or very near an element’s natural frequency. The resonance
frequency causes a maximum output or element amplitude response.
resonance frequency, normal mode, harmonic frequency, natural frequency: the

response (of a system): a quantitative expression of the output of the system as a
function of the input.
response spectrum: a method of predicting the response of mechanical systems
(buildings, machinery, computers, etc.) to vibration or shock inputs. The input wave-
form is mathematically applied to a series of single-degree-of-freedom spring/mass
oscillators with a defined amount of damping. The response of the series of oscil-
lators is plotted versus frequency. The effect of the input waveform on systems
having various natural frequencies can then be determined by inspecting the plot (the
response spectrum).
ribbed slab foundation: see foundation.
rigid element: a type of element used in finite element analysis that will not deform
under loading.
root mean square (rms): a value of a set of numbers that is the square root of the
average of their squared values; a technique to determine the average magnitude of
a dynamic signal, giving the effective energy or power level of that signal.
rms: see root mean square.
Rp : component response modification factor. See Section E.4 of Appendix E for
further discussion.
rpm: rotations per minute.
saddles, support saddles: a type of clamp for supporting cables, large vessels, or pipes.
SDC: seismic design category.
SDS : design spectral response acceleration. See Section C.4.2 of Appendix C for
further discussion.
seismic detailing: design features or reinforcements to overcome seismic load.
seismic horizontal force: Fp . See Section C.4.1 of Appendix C for further discussion.
seismic load: see building loads.
seismicity: earthquake activity.
server: a computer that provides some service for other computers connected to it
via a network. The most common example is a file server that has a local disk and
services requests from remote clients to read and write files on that disk.
service clearance area: the area around the datacom equipment. Service clearance
areas of adjacent data equipment may overlap. The size of the service clearance area
depends on the application of the equipment and the expected environment in which
the equipment will be installed. When the equipment is installed against a wall, the
service clearance area should permit front-end servicing or adequate provisions

should be made to facilitate movement of the equipment as required.
sine-on-random vibration: superimposing sinusoidal-on-random waveform.
Normally the power line frequency-related vibration 50 and 60 Hz sine levels are
superimposed on the random waveform through the use of a digital vibration
controller with the appropriate software.
sinusoidal vibration: a vibration with a sine waveform. Only one frequency is
present. Typical sources of sinusoidal vibration are rotating machinery and electrical
power equipment.
snow load: see building loads.
SBCCI: Southern Building Code Congress International, Inc.; now International
Code Council.
spectral response spectrum: the acceleration response of the series of oscillators;
plotted versus frequency when a waveform is mathematically applied to a series of
single-degree-of-freedom spring/mass oscillators with a defined amount of damping.
splay cables: an anchoring method using wire rope cables to restrain a cabinet in all
directions.
spread footing: see footing.
sprinklered: means that a sprinkler system is installed.
stringerless raised-access floor: a raised-access floor without stringers (see
Figure 9.1) where each raised-access floor panel is bolted at every corner to the
pedestals directly below them. Also known as bolt-down or cornerlock systems.
telecom: abbreviation for telecommunications.
transient vibration: a vibration that has a high amplitude and is of short duration,
typically only a few cycles, and unlikely to produce resonance.
UBC: Uniform Building Code.
UPS: uninterruptible power supply.
USGS: United States Geological Survey.
vibro-flotation: a soil improvement technique that uses a vibrating probe to pene-
trate loose granular soil and remove voids, thus compacting and stabilizing the soil.
weight distribution area: the area around the datacom equipment.
wind load: see building loads.
Wp : component operating weight. See Section C.4.1 of Appendix C for further
discussion.
Z: seismicity in the United States a defined in the Uniform Building Code.
Index
A
acceleration of gravity 79, 153, 157
amplitude 12, 25, 27, 51, 63–64, 74, 79–83, 85, 91, 116, 137, 141–46, 153, 158, 160–62
ap 98, 138, 149–51, 153, 155
augered cast-in-place pile 33, 153, 159
authority having jurisdiction (AHJ) 18, 153, 158
availability 4, 11, 97, 99, 153
B
ballasted roof 153
base fixity 153
braced frame 9, 16, 25
built-up roof 154
C
cabinet 81–82, 84, 87, 99–102, 104–106, 110, 160, 162
CalTrans 81
chase 155
chevron bracing 26, 38, 40, 155
clear height 155
collateral load 19, 154
combined footing 1586
component amplification factor (ap) 98, 138, 149–51, 153, 155
component importance factor (Ip) 90, 111, 138–39, 149–50, 155, 157
component operating weight (Wp) 68, 111, 138–39, 149, 155, 162
component response modification factor (Rp) 69, 138, 149, 151, 155, 161
computer room air-conditioning unit (CRAC) 46, 57–58, 71, 73–74, 80, 155
concentrated load/concentrated load capacity 37, 63–64, 133–34
concrete masonry unit (CMU) 26, 34, 38, 128, 155
constraint element 155
cooling distribution unit (CDU) 155
craneway 155
D
data center 3–6, 11, 17, 22, 26, 29–33, 35, 37, 57–58, 67–68, 73, 80–84, 86, 95, 97, 98–99,
107, 110, 141–47, 150, 152, 155, 159
datacom 3–7, 9, 11–12, 16–17, 22–26, 29, 32, 37, 45, 47, 51, 55, 57, 61, 63–64, 66, 71, 73–
75, 80–81, 83–87, 89–90, 93, 135, 139
164⏐ Index
datacom equipment 9, 11–12, 16–18, 23, 45, 47, 55, 57, 61, 63–64, 66, 71, 73–75, 80–81, 85–
86, 89–90, 95, 98–100, 104, 106–11, 113, 116, 134–35, 141–46, 150–52, 155, 158, 160–62
dead load 19, 154
design spectral response acceleration (SDS) 111, 138, 149, 156, 161
diaphragmatic action 156
drift, frame 157
drift, structure 16, 25
drilled pier 159
driven pile 159
duration of load 19, 154
E
earthquake 3, 10, 18, 66–68, 80, 82, 85–86, 90, 95–96, 98–100, 109–11, 113, 116, 119, 121,
138–39, 149–51, 156, 158, 161
electronic equipment 156
element, constraint 155
element, rigid 161
equipment support 20, 156
F
finite element analysis 90, 92, 107–108, 114, 116, 119, 156–57, 159, 161
floor load/floor loading 66, 134–37
footing, combined 156
footing, monolithic 156
footing, spread 156
foundation, mat 156
foundation, ribbed slab 157
Fp 68, 137, 139, 149, 157, 161
fragility 12, 85, 87–88, 90
frame drift 157
free-fall drop height 79, 157
frequency, harmonic 157, 160
frequency, natural 157, 160
G
grms 157
Guyan reduction solver 107, 109, 157
H
half-sine shock pulse 79, 157
handling 83, 87–88
harmonic frequency 157, 160
headroom 155
horizontal design seismic force 68, 137, 139, 149, 157, 161
horizontal force 38, 64, 68, 98–100, 137, 139, 157, 161
HVAC 6, 46, 75, 84, 95, 157, 159
I
impact load 19, 154
infrastructure 3–6, 8–9, 11–12, 18, 22–26, 30, 32, 45–47, 51, 55, 57, 80, 89
International Building Code (IBC) 9, 26, 57, 69, 74–75, 89–90, 92, 96–98, 119, 122, 137–39,
149–51, 157
International Code Council (ICC) 86, 90, 119, 121, 137, 157
Ip 90, 111, 138–39, 149–50, 155, 157
isolation 58, 71, 83, 92, 104, 106
IT/IT equipment 8, 81–82, 86, 110, 155, 157
L
Lanczos method solver 107, 109, 114, 157
live load 19, 23, 26–27, 32, 47, 68, 135, 138–39, 154
load, collateral 19, 154
load, concentrated 37, 63–64, 133–34
load, dead 19, 154
load, duration of 19, 154
load factor 19, 154
load, impact 19, 154
load, live 19, 23, 26–27, 32, 47, 68, 135, 138–39, 154
load, partition live 154
load, rolling 63, 134
load, seismic 19, 154
load, snow 19, 154
load, ultimate 133
load, uniform 32, 34, 63, 133, 138
M
machine area 158
magnitude 12, 25, 27, 51, 63–64, 74, 79–83, 85, 91, 116, 137, 141–46, 153, 158, 160–62
mat foundation 156
minipiles 158
mission critical 82, 90, 92, 97, 158
moment frame 16, 26, 39, 158
N
Network Equipment-Building System (NEBS) 83, 158
nonstructural components 159
normal mode 157, 160
P
parabolic shell element 108–109, 159
partition live load 154
pedestal 11, 58, 61–69, 99, 137–39, 162
pier, drilled 159
pier, rammed aggregate 159
pile 33, 153, 158–59
pile, augered cast-in-place 33, 153, 159
pile, driven 159
power distribution unit (PDU) 159
power spectral density (PSD) 79, 86, 142, 147, 159
pulse width 79, 160
purlin 160
R
rack/rack-mounted equipment 160
raised-access floor (RAF) 11–12, 16, 23–24, 30, 37, 45, 55, 57–58, 61–69, 81, 95, 98–100,
104, 107, 110, 129, 133, 135, 137–39, 150–51, 159–60, 162
raised-access floor panel 11, 61, 63–64, 66–67, 99, 104, 133, 162
rammed aggregate pier 159
random vibration 79, 160
relative location factor 149, 160
reliability 97, 160
resilience 160
resonance/resonance frequency 79, 157, 160
166⏐ Index
response (of a system) 80, 161

response spectrum 80, 161
ribbed slab foundation 157
rigid element 161
rolling load 63, 134
root mean square (rms) 80, 161
Rp 69, 138, 149, 151, 155, 161
S
saddle 161
SDS 111, 138, 149, 156, 161
seismic 3–4, 9–10, 12, 16, 19, 21–22, 25, 27, 37, 55, 57–58, 64, 67–69, 80, 85–87, 89–92, 95–
99, 106–109, 111, 116, 137, 139, 149–50, 152, 154, 157, 160–61
seismic design category (SDC) 97, 161
seismic detailing 161
seismic horizontal force (Fp) 68, 137, 139, 157, 161
seismic load 19, 154
seismicity 12, 67–68, 96–97, 161–62
server 161
service clearance area 161
shaker/shake table test 84–85, 91–92
shear wall 9, 16, 25–26, 38, 40–41
shipping 83, 87–90
snow/snow load 4, 15, 18–19, 27, 37, 150, 154, 162
soil borings 29
spectral response spectrum 162
splay cable 162
spread footing 156
sprinklered 162
stanchion 37, 55, 61, 104
stringer/stringerless 61–64, 66–69, 162
structure drift 16, 25
T
telecom 162
transient vibration 80, 162
U
ultimate load 133
Uniform Building Code (UBC) 162
uniform load 32, 34, 63, 133, 138
uninterruptible power supply (UPS) 162
V
vibration 3–5, 11–12, 33, 58–59, 71–73, 79–80, 160, 162
vibration, random 79, 160
vibration, sine-on-random 162
vibration, sinusoidal 80, 162
vibration, transient 80, 162
vibro-flotation 162
W
waveform 80, 85–86, 161–62
weight distribution area 162
wind/wind load 4, 15–19, 25, 27, 37, 51, 73, 75, 80, 82, 89–90, 121, 150, 154, 156, 162
Wp 68, 111, 138–39, 149, 155, 162

05 Structural and Vibration Guidelines For Datacom Equipment Centers

Încărcat de

Informații document

Titlu original

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

05 Structural and Vibration Guidelines For Datacom Equipment Centers

Încărcat de

Drepturi de autor:

Formate disponibile

This file is licensed to sergio torres morales (retos@correoinfinitum.com).

License Date: 9-26-2012

Structural and Vibration

Any updates/errata to this publication will be posted on the

Structural and Vibration

American Society of Heating, Refrigerating

©2007 American Society of Heating, Refrigerating

Printed on 10% post-consumer waste using soy-based inks.

Christina Helms David Soltis

Part I—Introduction and Best Practices

Part II—Building Structure

Chapter 3 Building Structures Overview ......................................... 15

Chapter 4 New Structures .................................................................21

Part III—Building Infrastructure

Structural and Vibration Guidelines for Datacom Equipment Centers ⏐ vii

10.4 Vibration Isolation within the

Part IV—Datacom Equipment

C.3 Floor Loading Calculations ................................................135

AT&T Services, Inc. Intel Corporation

ASHRAE TC9.9 wishes to particularly thank the following people:

com equipment). As the datacom equipment density continues to increase (compac-

Structural and Vibration Guidelines for Datacom Equipment Centers⏐5

Table 1.1 Datacom Industry Wide Range of Applications

1.3 OVERVIEW OF ASHRAE TECHNICAL COMMITTEE 9.9

1. Thermal Guidelines for Data Processing Environments (2004)

Structural and Vibration Guidelines for Datacom Equipment Centers⏐7

3. Design Considerations for Datacom Equipment Centers (2006)

1.5 DOCUMENT FLOW

Part I—Introduction and Best Practices. Part 1 provides an overview of this

Part II—Building Structure. Part 2 focuses on the basic building structure,

Chapter 3—Building Structures Overview

Part III—Building Infrastructure. Part 3 focuses on power, cooling, flooring,

Chapter 7—Building Infrastructure Overview

Part IV—Datacom Equipment. Part 4 focuses on the IT or electronic pack-

Chapter 11—Shock and Vibration Testing on Datacom Equipment

Appendices and Back Matter. These include background material, a bibliog-

1.6 PRIMARY USERS FOR THIS DOCUMENT

• computer equipment manufactures (research and development engineers and

Table 2.1 Best Practices for Building Structures—

10⏐ Best Practices

2.2 BUILDING STRUCTURES—RENOVATIONS, RELOCATIONS,

Existing structures require an extensive amount of verification to determine

Table 2.2 Best Practices for Building Structures—

1 Comprehensively analyze the design and construction of the existing structure.

Establish confidence by comparing construction drawings with visual spot checks of

Evaluate providing independent sub-structures in lieu of upgrading the capacity of

Provide above-average supervision and inspections during construction of any

10 Consider “exit strategies” when spaces are leased.

Structural and Vibration Guidelines for Datacom Equipment Centers⏐11

2.3 BUILDING INFRASTRUCTURE

The building infrastructure supporting datacom equipment can be located in

2.4 DATACOM EQUIPMENT

Table 2.3 Best Practices for Building Structures—

1 Place heavy equipment and infrastructure systems on the ground level.

3 Provide a reserve in the capacity of the raised-access floor system.

Consider mechanically fastening raised-access floor pedestals to the structural

5 Consider screwing raised-access floor panels to the floor support system.

7 Provide vibration isolation for equipment with rotary components.

9 Organize the structural support of building infrastructure distribution systems.

10 Provide for expansion/contraction of the infrastructure.

12⏐ Best Practices

Table 2.4 Best Practices for Building Structures—

1 Define the weight and dimensions of the datacom equipment.