Sand T Review Journal Vol 1 No 12012

ISSN : 2231-5160
S & T Review
An International Journal of Science & Technology
Vol. 1, No. 1, 2012
S & T Review
Chief Patron
Shri Suresh Jain, Hon'ble Chancellor, Teerthanker Mahaveer University, Moradabad
Patrons
Shri Manish Jain, Group Vice Chairman, Teerthanker Mahaveer University, Moradabad
Prof. R.K. Mittal, Vice Chancellor, Teerthanker Mahaveer University, Moradabad
Editor-in-Chief
Prof. N.K. Joshi, Director, College of Engineering, Teerthanker Mahaveer University, Moradabad
Editorial Advisory Board
Prof. M.M. Pant, Former Pro-Vice Chancellor IGNOU, New Delhi
Prof. Gregory Gelles, Missouri University of Science & Technology, USA
Prof. Prem Bist, IIT, Chennai
Prof. R.K. Mudgal, Registrar, TMU, Moradabad
Mr. Sanjiva Shankar Dubey, Asia Pacific Executive IGA Service Delivery, IBM
Prof. Om Harsh, University of New England, Australia
Prof. Arijit Sikdar, Wollongong University, Dubai
Prof. Sanjiv D Vaidya, IIM, Calcutta
Prof. Paschal Meehan, LIT, Ireland
Mr. Kapil Lakhtakia, Consultant,World Bank
Dr. R. Sreedhar, Director, Commonwealth of Learning
Prof. D.P Goyal, MDI, Gurgaon
Prof. Maria Kyne, LIT, Ireland
Associate Editors
Dr. B.P. Agrawal
Mr. Shirazul Islam
Mr. Gaurav Pandey
Mr. Namit Gupta
Dr. Avnish Chauhan
Disclaimer: The papers contained in the journal carry the opinion and view of the contributors and not necessarily of the
Editorial Board. The Editorial Board as such shall not be responsible for the authenticity and legality concerns regarding
the submission made in this journal. The entire responsibility shall lie with the author of the papers.
From the Desk of Editor-in-Chief

Welcome to the inaugural issue of the S & T Review, an International Journal of Science and Technology.
The S & T Review is the official publication of the College of Engineering, Teerthanker Mahaveer
University. S & T Review is published for engineers, academicians, scientists, and researchers involved in
engineering, science & technology and its applications to publish high quality and refereed papers.
S & T Review is dedicated to publishing research covering a wide spectrum of applications in the developing
field of engineering and technology such as Architecture, Applied Science, Civil Engineering, Chemical
Engineering, Computer Science & Engineering, Electrical and Electronics Engineering, Mechanical
Engineering, Total Quality Management , Industrial Engineering, Engineering Management, Information
and Communication Technology, Biotechnology and Environmental Science.
The present issue represents the culmination of efforts from many people who believe that education
publication deserves proper exposure and recognition. This first issue of S & T Review is the success of the
reviewers, editorial board and the researchers. In this respect, I would like to take this opportunity to thank
all authors, researchers, reviewers and the editorial board. I also thank the advisory board for their
continuous support towards this endeavor.
S & T Review invites authors to submit papers that discuss and analyze concerns and experiences pertaining
to various branches of engineering and technology. Papers reporting original research and innovative
applications from all parts of the world are welcome. We hope many more will do so and that S & T Review
becomes a prime outlet for research in the area of Engineering and Technology.
Editor-in-Chief
Prof. N. K. Joshi,
Director, College of Engineering
Teerthanker Mahaveer University, Moradabad, UP, India
C
O
N
T
E
N
T
S
1.
A Study Towards Use of Cloud Computing

Nasir Ali and Mohammad Ishrat
01
2.
Modular Neural Network for Detection of Diabetic Retinopathy in Retinal Images

Amol P. Bhatkar and Dr. G. U. Kharat
05
3.
Classification of Customer Behaviour using Artificial Neural Network (ANN) and

Support Vector Mechanism (SVM)
Dr. Mohammad Ahmar Khan and Mr. Mohammad Jameel Hashmi
09
4.
Eigen Spectrum Analysis for Order Reduction of Linear System using Time Moment Algorithm
Vinod Kumar and J.P.Tiwari
16
5.
A Multilevel Cascaded Converter Based Solid State Distribution Transformer

Rakesh Thapliyal, Y.R Sood, R.K Jarial and Navin Chandra Joshi
21
6.
Detection of Hand Movement Using EMG Signal

Antra Saxena, Varsha Gupta, Abid Ali Khan, Omar Farooq and Yusuf U Khan
26
7.
Matlab/Simulink Based Mathematical Modeling and Simulation of an Air Breathing Proton Exchange
Membrane Fuel Cell (ABPEMFC)
A.R.Vijay Babu and P.Manoj Kumar
30
8.
Nonlinear Vibration and Instability Response of Embedded Pipes Conveying Viscose Fluid Using DQM 38
H. Rahimi Pour, M. Ghaytani and R. Kolahchi
9.
Packed Bed Solar Energy Storage: A State-of-Art Review

T. Alam, R.P.Saini and J.S.Saini
43
10.
Friction Stir Welding of Aluminium

Subham Kumar, Vipin Kumar, Gajendra Singh and Mohd. Zaheer Khan Yusufzai
49
11.
Performance and Emission Analysis of a Multi-Cylinder C.I. Engine Fuelled with Diesel-CNG Dual Fuel
Rajesh Kumar, Manish Kumar Chauhan, Naveen Sharma and Amit Pal
55
12.
Distresses on Flexible Pavement and Their Measurement

Vishal Bhatt, S. S. Jain and M. Parida
63
13.
Analysing the Suitability of Waste Materials for Road Construction

Aditya Kumar Anupam, Praveen Kumar G.D. and Ransinchung R.N.
69
14.
Prophylactic Efficacy of Septilin on Skin of Rats Exposed to Gamma Radiation

Enas A. Khalil
76
15.
An Efficient Method for the Calculation of Eigenvector Design Sensitivities

Shibinlal M.K and Aminudin. Hj.Abu
82
16.
Correlation of Bulk Density and Abrasion Resistance of Some Igneous and Metamorpic Rocks
Ademeso Odunyemi Anthony, Adekoya Adeyinka John and Olaleye Boluwaji Muraina
89
A Study Towards Use of Cloud Computing

*
Nasir Ali , Mohammad Ishrat
n the current scenario in India we are keen to implement the e-Governance model. The urban areas are in good position to
avail the services of e-Governance as they have all required infrastructure but in rural areas the biggest problem is the non
availability of proper infrastructure as well as lack of computer aware citizen. Cloud computing can be a future solution to
fulfil that needs. The critical problem discussed for the rural areas is non availability of the necessary infrastructure to
implement the e-Governance services. In this paper we also discussed how to overcome that problems using the applications
of cloud computing .The e-Governance services can be designed in a manner that can be offered through fully customized
service-oriented (SaaS) clouds. In this paper we are presenting the similarities between generic government process and
services and the use of service oriented cloud computing model. Also, we will discuss the major issues in establishing services
oriented grids for governmental organization. We propose a model based framework to implement cloud computing for rural
area e-Governance model.
Keywords: E-Governance, Cloud Computing, SaaS, IaaS, PaaS
1. Introduction
Our Indian government is putting in a lot of effort towards
using applications of e-Governance using ICT and making
the government services more useful and simple for citizens
equally in urban and rural areas. However, there are many
key issues in making the citizen-centric services accessible
in rural areas. For service providers which is government is
most of the cases, it is difficult to reach them and scale up the
functionality. For the rural area population community in
India, accessibility and affordability of ICT infrastructure
are major issues [1]. There is a clear opportunity divide that
exists. To bridge this, we need to leverage on common
elements. We need to build the backend for the last mile
connector and facilitate building a community of
independent rural citizens.
Also, there is a need to build the front-end for the service
provider. We need to provide more ICT application with
good infrastructure to the service providers to use the
platform and to provide last mile access in rural areas.
There is a clear need to build a cloud based e-Governance
system and bring the entire rural community under one roof.
This is possible through cloud computing. To start with
cloud-based services, one need to define a cloud model that
you would like to go with such as public or private cloud,
community cloud or hybrid cloud [2].
* Musanaa College of Technology, Muscat, Sultanate of Oman
E-mail: nasirworld@yahoo.com
S & T Review Volume 1, Number 1, 2012
Public sector should take advantage of those improved

conditions for development and deployment of egovernment solutions. This can be achieved through the
Adoption of new architecture such as cloud computing and
service-oriented architecture, in the public sector. Cloud
computing permits to uniformly cover the whole country
with e-government solutions, independently of divergence
of local administrative units that may be better or worse
prepared to provide e-services. Service-oriented architecture
facilitates provision of compound services covering whole
customer processes, where a customer may be a citizen or an
enterprise. One should understand the security risks
involved. One can take infrastructure-as-a-service (IaaS) as
foundation. Platform-as-a-service (PaaS) builds on IaaS,
Software-as-a-service (SaaS) builds on PaaS, and SaaS is the
most integrated functionality. PaaS helps developers to build
their services on top of a platform. IaaS has limited
application-like features but enormous extensibility [3].
2. Why e-Governance with Cloud Computing
Cloud computing might be one technique for the
implementation of the national e-governance plan (NeGP).
Cloud computing will play important role in the national egovernance plan. There are certainly some advantages that
cloud computing could accrue like start of using eGovernance services by SaaS. Such software as service with
data centres and all other means, which are all bundled
together using cloud computing can be a fairly attractive
model for Indian rural areas . First, many people may want to
use 'pay as you go' model and that generally appeals to the
Indian economy, Second, on the context of e-governance,
there are other aspects of possibility through cloud
computing like different kinds of software platform that are
available for making e-services available in one state or one
part of the country or a department can be used in other area
also when it will be web based using clouds. Once they are
put onto a cloud (may be a government cloud), it could help
in the usage of same by other departments and organizations
to speed up the rolling out of different projects.
4. Proposed Cloud based E-Governance Model

The three main features of our cloud computing based model
are instance availability of services, pay per use model and
massive scalability. A simple Cloud architecture is build and
hence is highly flexible and modular and can integrate with
other systems. It offers the following layers of abstraction as
shown in figure 1.
The figure shows that we can offer e-Governance services
using cloud computing layers that are IaaS , PaaS and SaaS.
Now in next section we will try to explore the services that
can be offered through that 3 layers.
3. Early Work Study

Chandwick and May [3] in the year 2003 proposed an eGovernance framework regarding the interaction between
government and its citizens. They had focused on the
managerial, consultative and participative models of eGovernance. In the year 2005, Grant et al. [7] review report
proposed the strong integration between IT and managerial
aspect of e-Governance in a sophisticated manner in
reflecting the citizen-centric perspective of e-Governance.
Coursey et al. [4] in the year 2008, published in their review
report that e-Governance should have web-technology in
order to integrate the different government information and
services viz. e-participation, e-democracy etc. Again the
existing work in web service performance focuses on the
latest trend of technologies and standards. S.Adreozzi et al.
[1] present a model for rigorous representation of service
characteristics. D. Gouscos et al. [6] present a simple
approach to model certain web service management
attributes. J.P Thomas et al. [11] represent distributed web
service by modelling the flow of messages and methods in a
web service transaction. Tu et al. [10] discuss design
strategies to improve the performance of web service. Levy
et al. [8] present architecture and prototype implementation
of performance management system for cluster based web
service. V. Cardellini et al. [5] consider different categories
of web applications, and evaluate how static, dynamic and
secure web service request affects the performance and
quality of service of distributed web sites. But the entire
above discussed web services are either very much server
centric or device centric in nature. A properly distributed
cloud computing based web services that may fit effectively
in e- Governance model are available in some areas and we
can use this new model in India also to offer e-Governance
service in rural areas.
SAAS
qVarious Services (GSG, G2B, G2E, G2C)

qService Level Agreements
PAAS
qMiddle Ware Solutions

qDatabase Services
qBackup Services
IAAS
qHardware (CPU)
qStorage
qNetwork
Figure 1 : e-Governance cloud model
Layer 1 IaaS
Infrastructure as a service virtualizes the hardware/network
and storage aspects of the data centre. A Storage Area
Network must also be in place in order to offer the eGovernance, services to remote areas. E-Governance
requires a 24x7 infrastructure availability minimizing
downtime. E-Governance applications can assume
unlimited supply of CPU, storage and bandwidth when
operating from cloud. Application designers can focus on
features and usability instead of worrying about scalability
etc. Still, applications perform better on cloud compared to
traditional architecture as the Infrastructure will be in some
urban area and the rural area will use that without having
actual availability. They will access then using nodes like a
simple PDA introduced by Indian government recently
namedAakash.
Layer 2 PaaS
Platform as a service of cloud model offers standard
platforms in terms of providing different kinds of operating

Volume 1, Number 1, 2012
S & T Review
systems, software applications, middleware and integration

systems. Some of the standard platforms they provide are:
Web Based Operating system
Web based Query systems
Web based data base systems
Middle ware services
Work flow management services
Government departments or local area citizen requiring

resources can request and get resources instantly using their
PDA having WiFi or other network connections as compared
to traditional methods where they have to wait till they
purchase, deploy etc. Applications requiring middleware
services can be provided instantly
Layer 3 SaaS
Software as a service model of cloud offers software
applications as a service. Imagine a case of new village of
rural area deciding to move to e-Governance solution for
some application for their citizens. The village need not
purchase applications, hardware and software. They can
make a request for a particular service from the cloud
provider. Applications instances can then be created for their
use. Numerous applications can be provided as standard
services, where departments can request and manage. Some
of the applications can be:
Complaint Resolution System
Employee Management Systems
Tax Filing Systems
E-police, E-court
Municipal Maintenance
Water Boards, Billing, Payment Systems
District Management Solutions
Government office service Desk
Cloud fits in right into the requirements and can offer

excellent service in this regard. Instead of each department
hosting hardware, software and applications, they can get all
the applications for a district instantly to be provisioned and
operational. Hence cloud accelerates the implementation of
e-Governance services. This feature can reduce the cost of EAn International Journal of Science & Technology
Governance to a great extent especially in case of rural areas.

5. Benefits of Proposed Model
As compared to traditional ICT based e-Governance
architecture, the services offered are bound to a physical
machine. One has to maintain the redundancy in terms of the
physical services. This model increases the cost of
deployment and becomes expensive to maintain as the
number of services increase. Cloud computing offers:
On-demand self sufficient service
Ubiquitous network access
Location independent resource
Rapid elasticity
Cloud is making rapid inroads because of the following

advantages:
!
Can reduce IT labour cost by 50% in configuration,

operations, management and monitoring.
Can improve capital utilization by 75% significantly

reducing license costs
Reduce provisioning cycle times from weeks to

minutes
Can reduce end user IT support costs up to 40%
In a traditional infrastructure, there will be one instance of

application per physical server and has an average
utilization of 10%.
6. Conclusion and future work
Thus we have proposed a framework of e-Governance based
on cloud computing for rural areas. Here, we have put
forward the different components of cloud computing and
then specifying the role of each component. We have
introduced a new framework of e- Governance based on
cloud computing, where SaaS as a main service and IaaS as
to provide low cost infrastructure. But here, neither we have
discussed the way of accessing the e-Governance web
services nor the practical implementation of the model.
References
[1] A. V. Anttiroiko., Electronic Government: Concepts,
Methodologies, Tools, and Applications. Information
Science Reference, Hershey, PA, 2008.
[2] Stacy Avery Baird, Government role in developing an

interoperability ecosystem, Proceedings of the 1st
international conference on Theory and practice of
electronic governance, December 10-13, 2007, Macao,
China
[3] Rajkumar Buyya , Chee Shin Yeo , Srikumar Venugopal,
Market-Oriented Cloud Computing: Vision, Hype,
and Reality for Delivering IT Services as Computing
Utilities, Proceedings of the 2008 10th IEEE
International Conference on High Performance
Computing and Communications, p.5-13, September
25-27, 2008.
[4] D. Coursey, and D. Norris, Models of e-Government;
Are they correct? An empirical assessment , Public
Administration, Review, volume 68, Number 3, pp
523-536, 2008
[5] V. Cardellini, , E. Casalicchio, and M. Colajanni,, A
performance study of distributed architectures for the
quality of Web services, in Proceeding 34th Annual
Hawaii International Conference on System
Sciences,2001.
[6] D. Gouscos, M. Kalikakis, and P. Georgiadis, An
approach to modelling Web service QoS and provision
price , in Proceeding 3rd International Conference on
Web Information Systems Engineering Workshops,
pages 121-130, 2003.

[7] G. Grant, , and D. Chau, , Developing a Generic
Framework for E-Government, Journal of Global
Information Management, Volume 13, Number 1, pp 130, year 2005
[8]
R. Levy, J. Nagarajaro, G. Pacific, M. Spreitzer, A.

Tantawi, and A. Youssef, Performance management
for cluster based Web services, in Proceeding
IFIP/IEEE 8th International Symposium on Integrated
Network Management, pages 247-261, 2003
[9] L.F.G Sarmenta, Volunteer Computing, Ph.D. Thesis

in Massachusets Institute of Technology, March
2001
[10] S. Tu, M. Flanagin, Y. Wu, M. Abdelguerfi, E.
Normand, V. Mahadevan, Design Strategies to
improve performance of GIS Web services, in
Proceding International Conference on Information
Technology: coding and Computing (ITCC04),pages
444-448, 2004.
[11] J.P. Thomas, , M. Thomas and G. Ghinea, Modeling
of
Web service flow, in Proceeding IEEE
International Conference on E-Commerce (CEC 03) ,
pages 391-398, 2003.

S & T Review
Modular Neural Network for Detection of

Diabetic Retinopathy in Retinal Images
Amol P. Bhatkar1, Dr. G. U. Kharat2
his paper focuses on Modular feed forward network method to detect diabetic retinopathy in retinal images. The authors
present classification method; the modular feed forward neural network (MNN) to classify retinal images as normal and
abnormal. Publically available database DIARETDB0 including high quality normal and abnormal retinal images is taken for
detection of diabetic retinopathy. Modular Feed Forward Neural Network is designed based on the extracted features of retinal
images and the train N times method. The classification accuracy by MNN classifier was 100% for normal retinal images and
86.67% for abnormal retinal images. The authors have explored such a method using MNN classifier which can detect
diabetic retinopathy by classifying retinal images as normal and abnormal.
Keywords: Diabetic retinopathy, Modular feed forward neural network,
Classification accuracy, Retinal images database DIARETDB0.
1. Introduction
Diabetes is a disease that affects blood vessels throughout the
body, particularly in the kidneys and eyes. When blood
vessels in the eye are affected, the condition is referred to as
diabetic retinopathy (DR). Diabetic retinopathy is a major
public health problem and a leading cause of blindness in the
World [1]. It is caused by changes in the blood vessels of the
retina.
Diabetic retinopathy is a micro vascular complication that
may occur in patients with diabetes. The occurrence of
diabetic retinopathy will result in the disturbance of visual
capability and can eventually leads to blindness. The longer a
person has untreated diabetes; there is higher chance of
developing diabetic retinopathy as it may be converted into
vision loss. Along with diabetes, high blood sugar levels in
long periods can affect small vessels in the retina. Diabetic
retinopathy becomes symptomatic in its later stage. In the
first stage, diabetic patients may not be aware of having
infected by the disease [2]. Therefore early detection of
diabetic retinopathy is very important to avoid vision loss.
Diagnosis of diabetic retinopathy is usually conducted by the
ophthalmologist by employing retinal images of patients.
Ophthalmologist can obtain retinal images from patients to
be diagnosed by using a fundus camera. From the image,
symptoms will be identified manually by an
1. Anuradha Engineering College, Chikhli,
Email: apbhatkar@rediffmail.com
2. Sharadchandra Pawar College of Engineering, Otur,
E-mail: gukharat@gmail.com
ophthalmologist. Therefore more time is required to

diagnose more number of patients. A computerized
screening system can be used for fully automated mass
screening [3]. Such systems screen a large number of retinal
images and identify abnormal
images, which are then further examined by an
ophthalmologist. This would save a significant amount of
workload and time for ophthalmologists, allowing them to
concentrate their resources on surgery and treatment.
Normal structures of retina are the optic disk, macula, and
blood vessels. The characteristic features of diabetic
retinopathy are micro aneurysms, haemorrhages and
exudates [4]. From the set of parameters like vessel ratio,
ratio of exudates area to the total area of the images are
distributed into different groups like normal, severe, mild
and abnormal diabetic retinopathy etc. Neural networks can
be used effectively in data classification [5].
2. DatabaseAcquisition
In order to conduct the experiment for detection of retinal
images, publically available database DIARETDB0 was
used. The current database consists of 130 colour fundus
images out of which 20 are normal and 110 contain signs of
the diabetic retinopathy. The images were taken in the
Kuopio university hospital. The images were dedicatedly
selected, but their distribution does not correspond to any
typical population. Images were captured with few 50 degree
field of view digital fundus cameras with unknown camera
settings. The images contain an unknown amount of imaging
noise and optical aberrations (dispersion, transverse and

lateral chromatic, spherical, field curvature, coma,
astigmatism, distortion). Variance over the visual
appearance of different retinopathy findings can thus be
considered as maximal. However, the data correspond to
practical situations, and can be used to evaluate the general
performance of diagnosis methods. The general
performance corresponds to the situation where no
calibration is performed (no correspondence to the real
world measurements), but where images correspond to
commonly used imaging conditions, i.e., the conditions
typically encountered in hospitals. This data set is referred to
as calibration level 0 fundus images. The database was
divided into two parts for training & testing.
2.1 Feature extraction
Detection of retinopathy involves clinical recognition of
dilation of blood vessels, presence of exudates, lesions or
any other such abnormality in the retinal images. Fundus
image not only provides anatomical information of the retina
in eye, but also the information of these pathological
features. In a healthy retinal image, the features observed are
network of blood vessels, macula and the optic disc. Any
change due to diabetic retinopathy or other disease will cause
variation in these salient features which makes it easy to
diagnose the disease. Fig. 1 shows a healthy gray level
fundus image while Fig. 2 shows the retinal image affected
with diabetic retinopathy.
In this experiment, different features of retinal images were

extracted. The extracted features were: DCT, Entropy, mean,
standard deviation, average, Euler number, contrast,
correlation, and energy. Author's goal was to find out
possible features that can truly differentiate retinal image as
normal or abnormal.
3. Modular Feed Forward Neural Network
As the name indicates, the modular feed forward neural
networks are special cases of Multilayer Perceptron (MLPs)
[7], such that layers are segmented into modules. This tends
to create some structure within the topology, which will
advance specialization of function in each sub-module. In
biology, modular neural networks are very common. In
contrast to the MLP, modular feed forward networks do not
have full interconnectivity between the layers. Therefore, a
smaller number of weights are required for the same size
network (the same number of PEs). This tends to speed the
training and reduce the number of exemplars needed to train
the network to the same degree of accuracy. Fig. 3 shows
Modular feed forward neural network.
Fig. 3 Modular feed forward neural network
4. Overview of Classifier
Fig. 1 Fundus Image of a healthy eye
Fig. 2 Fundus image affected with diabetic retinopathy
It is proposed to develop a decision support system for the

detection of diabetic retinopathy using the artificial
intelligence technique. Authors aim to employ neural
network for the detection of diabetic retinopathy using
features extracted from the retinal images. The fundus
photographs will be collected from data repositories. These
photographs were taken with a fundus camera during mass
screening and then scanned by a flat-bed scanner and saved
as image files. The image files were then analyzed using the
algorithms described in the following section: The block

S & T Review
diagram of the proposed system is shown in figure 4. It

consists of eight modules: Retinal Fundus Image input,
Graphical User Interface (GUI), Pre-processing, Feature
Extraction, Parameter Acquisition, Modular feed forward
Neural Network Classifier, Graphical User Interface (GUI)
and Results (Detection). Here input is the retinal images
from DIARETDB0 Database [6] and after extracting the
image features, Modular neural network was trained to
detect images as normal or abnormal outputs.
Table 2: Modular feed forward networks recognition

results of training dataset based on the performance
Performance
ON
OA
MSE
0.013698564
0.012000019
Min Abs Error
0.0013881
0.000860905
Percent Accuracy
100
100
MSE: The average of the square of the difference between

the desired response and the actual system output (the error).
Fig. 4 Block diagram of proposed system
5. Results
The Modular feed forward neural network was used to test
the proposed features of retinal images. The neural network
was exhaustively designed using one hidden layer with
single neuron. Then progressively numbers of neurons were
increased. Same process was repeated with two hidden
layers. Various parameters were changed progressively to
set optimal neural network with best results and least
complexity.
The experimental results are depicted in following tables.
Table 1 : Modular feed forward networks recognition
results of train dataset
Output /
Desired
Normal
(ON)
Abnormal
(OA)
ON
19
OA
95
Fig. 5 Training MSE graph with

Standard deviation of training dataset

results of test dataset
Output /
Desired
Normal
(ON)
Abnormal
(OA)
ON
OA
13

results of testing dataset based on the performance
Performance
ON
OA
MSE
0.155588824
0.160918599
Min Abs Error
0.019526124
0.003081384
Percent Accuracy
100
86.66666667
Shankar Chatterjee, Jeff Boyd, Edward Hunter, and

Ramesh Jain, Automated diagnosis and image
understanding with object extraction, object
classification, and differencing in retinal images. Br. J
Ophthalmol, vol 83, august 1999.
6. Conclusion
This paper proposes a system for detection of diabetic
retinopathy using neural network. Authors have designed
modular feed forward neural network for detection of
diabetic retinopathy by classifying retinal images as normal
or abnormal. Features of retinal images were based on the
DCT, entropy, mean, standard deviation, average, etc and
were given as the input to the neural network. The train N
Times training method was used to train the network. It was
observed that with 02 hidden layers and 04 neurons the
classification accuracy was 100% for normal retinal images
and 86.67% for abnormal retinal images.
[4]
C.Sinthanayothin, J.Boyce,H. Cook, and T.

Williamson, Automated localisation of optic disc,
fovea, and retinal blood vessels from digital colour
fundus images, Br. J Ophthalmol, vol. 83, august
1999.
[5]
Anil K. Jain Michigan State University, Jianchang

Mao IBM Almaden Research Centre., Artificial
neural networks:Atutorial, 1996.
[6]
S.JeraldJeba Kumar, Madheswaran, Extraction of

Blood Vascular Network for Development of an
Automated
Diabetic Retinopathy Screening
System,International Conference on Computer
Te c h n o l o g y a n d
Development IEEE
10.1109/Icctd.2009.212, 2009.
[7]
R. Rojas: Neural Networks, Springer-Verlag, Berlin,

1996.
References
[1]
[2]
[3]
Meindert Niemeijer,Retinopathy Online Challenge:

Automatic Detection of Microaneurysms in Digital
Color Fundus Photographs,IEEE Transactions on
Medical Imaging vol.29,no.1, January 2010.
D. Klein, B. E Klein, S. E Moss et al The Wisconsin
epidemiologic study of diabetic retinopathy
VII.Diabetic non proliferative retinal lesions, Br. J
Ophthalmol, vol. 94, 1986.
Michael Goldbaum,Saied Moezzi, Adam Taylor,

S & T Review
Classification of Customer Behaviour using Artificial Neural

Network (ANN) and Support Vector Mechanism (SVM)
*
Dr. Mohammad Ahmar Khan , Mr. Mohammad Jameel Hashmi
he paper presents a mechanism, how Artificial Neural Network and Support Vector Machine can be used to find the good
customer. To proactively build lasting relationships with customers, it is thus crucial to predict customer behaviour
because customer behaviour is often non stationary. Training a model based on data extracted from a window of time in the
past yields poor performance on the present. We propose two distinct data mining approaches, using more historical data or
new, unlabeled data, to improve the results for this real-world, large-scale, non stationary problem. Result shows that
percentage of correct prediction of customer behaviour using SVM is better than that of ANN in all cases viz. Never pay month
2 and 4. Also the result in case of Never pay 4 month is more impressive as in this case the rate of correct prediction rate is high
as compared to the Never pay 2 months. This may be because the maturity of data as customer in 2 month may not be stable
with service provider as compared to Never pay month 4.
Keywords : ANN, SVM, Customer Behaviour, Prediction
1. Introduction
The customer churn, switching from one service provider to
another, costs heavily to the wireless tele-communications
industry each year all over the world. To proactively build
lasting relationships with customers, it is thus crucial to
predict customer behaviour. Machine learning has been
applied to churn prediction, using historical data such as
usage, billing, customer service, and demographics.
However, because customer behaviour is often nonstationary, training a model based on data extracted from a
window of time in the past yields poor performance on the
present. We propose two distinct data mining approaches,
using more historical data or new, unlabeled data, to improve
the results for this real-world, large-scale, non-stationary
problem.
A new ensemble classification method, with combination
weights learned from both labelled and unlabeled data, is
also proposed, and it outperforms bagging and mixture of
experts. For good business practice any service provider
want to have good customer. In this chapter we will propose
how artificial neural network and support vector machine
can be used to find the good customer.
2. Customer Behaviour
Customer management involves three main areas:
* Department of CS & IST, Mazoon (University) College,
Muscat, Sultanate of Oman, E-mail: ahmer@mazooncollege.edu.om,
E-mail: jameel@mazooncollege.edu.om
acquisition, development and retention. Companies ealize

that the customer acquisition side is more limited than in the
growth years of the 90's, but still use price and service deals
to attract new customers,
usually now at the expense of a competitor. Managing
customer churn presents its own set of challenges. There are
many reasons a customer might think of churning, and the
switching cost for translating that desire into action is
relatively small due to:
(i) Variety of companies,
(ii) The similarity of their offerings and
(iii) The cheap prices of handsets.
The big challenge these days is to increase customer' loyalty
before subscribers decide to leave and to aim efforts at
customers who are at risk of churning. What if carriers were
able to discover which clients are more susceptible to leave
the company before they decide to call? Companies could
then totally change the retention approach from reactivetype to a proactive-type.
For an industry facing a challenge in core revenue, local
telecom providers are doing little to endear themselves to the
enterprise customers that form their revenue base. According
to the recent Walker Loyalty Report for communications
services, only 34% of local telcos' enterprise customers are
truly loyal to a brand; roughly the same percentage will jump
at the chance to switch carriers. For long-distance the
numbers are similar though Verizon and BellSouth both

score well in that category. That there is no loyalty to telecom
providers shouldn't be a surprise. Both local and longdistance voice service have been commodities for some time.
But that's no excuse. Other industries have commoditized,
with customer loyalty intact. Also, considering what large
Telco's company have invested to shape and maintain their
brands, it should offend carrier executives that they haven't
built fidelity among their biggest customers.
In view of the above it is clear that for a telecom company
good customer is most important than a customer. Good
customer means a customer who will use the product he has
for a long time. As company expense more to join new
customer, the customer should be loyal to the service
provider for a long time to recover the expenses wear by the
company in his joining.
Traditionally a service provider company gives the service to
customer in following steps:
a.
Company collects the contact from other sources.
b.
An executive makes call to the customer and tell him

about the various product of the company.
c.
If customer agreed with any product then he submit

relevant document to the company.
d.
Company verified the document.
e.
After document verification service is offered to the

customer.
f.
Customer uses the service.
In above procedure of connecting the customer a heavy

amount is expanded by the company in hope to recover this
expense in future from the customer. If customer left the
company in short time period then company will be in loss
with respect to that particular customer. Based on the
association time of customer with service provider we can
classify customers into two class good or loyal customer and
bad customer.
Good or Loyal Customer: A customer is said to be good
customer if service provider able to recover the expenses
occurred in joining the customer in association time of
customer with the service provider.
Bad Customer: A customer is said to be bad customer if
service provider not able to recover the expenses occurred in
10
joining the customer in association time of customer with the

service provider.
Return through customer loyalty management is caused by
the optimization of the three most valuable means of
customer loyalty. These are the duration of customer
relation, a revenue orientation and the cost efficiency.
Priority objective is to increase the length of customer
relation because the longer the retention of a customer lasts
the more valuable the customer will be for the company.
With assistance of marketing information systems effective
measures for increasing the length e.g. by preventing losses
thanks to individualised marketing before expected critical
incidents come into effect , could be implemented. In a first
step the market size has to be examined. Second, gaining new
customers and committing existing customers to increase
loyalty to a mobile service provider.
Classification can be used to find good customer using
historical data of customer. In next section the concept of
classification and classification techniques has been given.
3. Classification
Classification is a data mining (machine learning) technique
used to predict group membership for data instances. For
example, you may wish to use classification to predict
whether the weather on a particular day will be sunny,
rainy or cloudy. Popular classification techniques
include decision trees and neural networks.
3.1 Artificial neural network (ANN)
An artificial neural network (ANN), usually called "neural
network" (NN), is a mathematical model or computational
model that tries to simulate the structure and/or functional
aspects of biological neural networks. It consists of an
interconnected group of artificial neurons and processes
information using a connectionist approach to computation.
In most cases, an ANN is an adaptive system that changes its
structure based on external or internal information that flows
through the network during the learning phase. Neural
networks are non-linear statistical data modeling tools. They
can be used to model complex relationships between inputs
and outputs or to find patterns in data.
Types of neural networks
(a) Feed forward neural network: The feed forward neural
network was the first and arguably simplest type of

S & T Review
artificial neural network devised. In this network, the

information moves in only one direction, forward, from
the input nodes, through the hidden nodes (if any) and to
the output nodes. There are no cycles or loops in the
network.
(b) Radial basis function (RBF) network: Radial Basis
Functions are powerful techniques for interpolation in
multidimensional space. A RBF is a function which has
built into a distance criterion with respect to a center.
Radial basis functions have been applied in the area of
neural networks where they may be used as a
replacement for the sigmoidal hidden layer transfer
characteristic in Multi-Layer perceptrons. RBF
networks have two layers of processing: In the first,
input is mapped onto each RBF in the 'hidden' layer. The
RBF chosen is usually a Gaussian. In regression
problems the output layer is then a linear combination of
hidden layer values representing mean predicted output.
The interpretation of this output layer value is the same
as a regression model in statistics. In classification
problems the output layer is typically a sigmoid function
of a linear combination of hidden layer values,
representing a posterior probability. Performance in both
cases is often improved by shrinkage techniques, known
as ridge regression in classical statistics and known to
correspond to a prior belief in small parameter values
(and therefore smooth output functions) in a Bayesian
framework.
(c) Kohonen self-organizing network : The selforganizing map (SOM) invented by Teuvo Kohonen
performs a form of unsupervised learning. A set of
artificial neurons learn to map points in an input space to
coordinates in an output space. The input space can have
different dimensions and topology from the output space
and the SOM will attempt to preserve these.
(d) Recurrent network: Contrary to feed forward
networks, recurrent neural networks (RNs) are models
with bi-directional data flow. While a feed forward
network propagates data linearly from input to output,
RNs also propagate data from later processing stages to
earlier stages.
(e) Stochastic neural networks: A stochastic neural
network differs from a typical neural network because it
introduces random variations into the network. In a
probabilistic view of neural networks, such random

variations can be viewed as a form of statistical
sampling, such as Monte Carlo sampling.
(f) Modular neural networks: Biological studies have
shown that the human brain functions not as a single
massive network, but as a collection of small networks.
This realization gave birth to the concept of modular
neural networks, in which several small networks
cooperate or compete to solve problems.
(g) Other types of networks: Following are special type of
neural networks which do not fit in any of the previous
categories.
a.
Holographic associative memory
b.
Instantaneously trained networks
c.
Spiking neural networks
d.
Dynamic neural networks
e.
Cascading neural networks
f.
Neuro-fuzzy networks
g.
Compositional pattern-producing networks
h.
One-shot associative memory
3.2 Support Vector Machine (SVM)

In the two-class case, a support vector classifier attempts to
locate a hyper plane that maximises the distance from the
members of each class to the optimal hyper plane. The
principle of a support vector classifier is briefly described
next.
Assume that the training data with k number of samples is
n
represented by {xi , yi}, i = 1, , k, where xR is an ndimensional vector and y{- 1,+ 1} is the class label. These
training patterns are said to be linearly separable if a vector w
(which determining the orientation of a discriminating
plane) and a scalar b (determine offset of the discriminating
plane from origin) can be defined so that inequalities (1) and
(2) are satisfied.
w. xi + b +1
for all y = +1
(1)
w. xi + b -1
for all y = -1
(2)
11
The aim is to find a hyperplane which divides the data so

that that all the points with the same label lie on the same
side of the hyperplane. This amounts to finding w and b so
that
yi (w. xi + b) > 0
(3)
If a hyperplane exists that satisfies (3), the two classes is said

to be linearly separable. In this case, it is always possible to
rescale w and b so that
That is, the distance from the closest point to the hyperplane
is 1/||w||. Then (3) can be written as
The hyperplane for which the distance to the closest point is

maximal is called the optimal separating hyperplane (OSH)
(Vapnik, 1995). As the distance to the closest point is 1/||w|| ,
the OSH can be found by minimizing ||w||2 under constraint
(4). The minimisation procedure uses Lagrange multipliers
and Quadratic Programming (QP) optimisation methods. If
i , i = 1,.,k are the non-negative Lagrange multipliers
associated with constraint (4), the optimisation problem
becomes one of maximising (Osuna et.al. 1997):
under constraints i 0, i = 1, ..,k.

a
If i 1. k is an optimal solution of the maximisation

problem (5) then the optimal separating hyperplane can be
expressed as:
The support vectors are the points for which ai > 0 when the
equality in (4) holds.
If the data are not linearly separable, a slack variable i , i
=1,,k can be introduced with i 0 (Cortes and Vapnik
1995) such that (4) can be written as
yi (w . xi + b) - 1 + i 0
and the solution to find a generalised OSH, also called a soft
margin hyperplane, can be obtained using the conditions
(10)
12
The first term in (8) is same as in as in the linearly separable

case, and controls the learning capacity, while the second
term controls the number of misclassified points. The
parameter C is chosen by the user. Larger values of C imply
the assignment of a higher penalty to errors.
Where it is not possible to have a hyperplane defined by
linear equations on the training data, the techniques
described above for linearly separable data can be extended
to allow for non-linear decision surfaces. A technique
introduced by Boser et al. (1992) maps input data into a high
dimensional feature space through some nonlinear mapping.
The transformation to a higher dimensional space spreads
the data out in a way that facilitates the finding of linear
hyper planes. After replacing x by its mapping in the feature
space %x,' equation (5) can be written as:
To reduce computational demands in feature space, it is

convenient to introduce the concept of the kernel function K
(Cristianini and Shawe-Taylor, 2000; Cortes and Vapnik
1995) such that:
Then, to solve equation (11) only the kernel function is

computed in place of computing % x' , which could be
computationally expensive. A number of kernel functions
are used for support vector classifier. Details of some kernel
functions and their parameters used with SVM classifiers are
discussed by Vapnik (1995). SVM was initially designed for
binary (two-class) problems. When dealing with several
classes, an appropriate multi-class method is needed. A
number of methods are suggested in literature to create
multi-class classifiers using two-class methods (Hsu and
Lin, 2002). In this study, a "one against one" approach (Knerr
et al., 1990). In this method, all possible two-class classifiers
are evaluated from the training set of n classes, each
classifier being trained on only two out of n classes. There
would be a total of n (n-1)/2 classifiers. Applying each
classifier to the vectors of the test data gives one vote to the
winning class. The pixel is given the label of the class with
most votes. To generate multi-class SVMs and a radial basis
kernel function (defined as
was used.
Compared with ANN classifier, SVM has three
characteristics when applied to classification task. First of
all, SVM can be generalized well even under small training

S & T Review
sample conditions. Secondly, SVM avoid the effect of

overfitting by choosing optimal hyperplane that can separate
the data in the feature space. Thirdly, SVM can model
nonlinear functional relationships that are difficult to model
withANN techniques [7, 8].
4. Experimental Detail
For example, consider a model which predicts for 10,000

insurance claims whether each case is fraudulent. This model
correctly predicts 9,700 non-fraudulent cases, and 100
fraudulent cases. The model also incorrectly predicts 150
cases which are not fraudulent to be fraudulent and 50 cases
which are fraudulent to be non-fraudulent. The resulting
Table of Confusion is shown in figure 2.
The experiments were carried out on the customer data

available from an Australian mobile company xyz (company
can't be disclosed due to some legal aspects). Available data
has many attributes which gives the detail about the
customers which has association with the mobile company.
In the available dataset (will be provided on request) there
are 36 attributes including one class attribute Never pay. This
attribute indicate that if value is 1 then customer is bad
customer means he will not pay the bill in future and if it is
zero then it is a good customer and he will pay his bill.
The aim of any company is to have Good customer so that
they have more profit and the same time company also wants
more connection so that company may also earn from
incoming calls. All the mobile companies try to find the trade
off between number of customers and loss due to bad
customers. To have minimum loss due to bad customer,
company first calculate the minimum number of association
months to recover the initial or connection cost. Value of
Never pay is decided on the basis of number of month a
customer associated with the company.
In this experiment we have run the classification ANN and
SVM classification method to predict the good and bad
customers. Also we run the experiment for the different
number of association month of customers.
To represent the results of customer classification behaviour
we have use confusion matrix.
In Predictive Analytics, a Table of Confusion, also known as
a confusion matrix, is a table with two rows and two columns
that reports the number of True Negatives, False Positives,
False Negatives, and True Positives. A confusion matrix is a
visualization tool typically used in supervised learning. Each
row of the matrix represents the instances in a predicted
class, while each column represents the instances in an actual
class figure 1. One benefit of a confusion matrix is that it is
easy to see if the system is confusing two classes (i.e.
commonly mislabeling one as another).
5. Result and Analysis

SVM and ANN on customer data have been run in crossvalidation of 10 X 10 folds. The model of SVM used is CSVC and the SVM kernel function is Radial Basis Function
(RBF). The parameters of SVM are as follows:
Parameter values for NeverPay Month 2:
C = 0.100000
Gamma = 0.081577
Number of support vectors used by the model = 7048

Parameter values for NeverPay Month 4:
C = 0.100000
Gamma = 0.009032
Number of support vectors used by the model = 4871
13
Table 7: Confusion Matrix For Validation

data in SVM of Never pay 4 Months
Table 8: Confusion Matrix For Training

data in ANN of Neverpay 2 Months
14

S & T Review
6. Conclusion
Above result shows that percentage of correct prediction of
customer behavior using SVM is better than that of ANN in
all cases viz. never pay month 2 and 4. Also the result in case
of never pay 4 month is more impressive as in this case the
rate of correct prediction rate is high as compared to the
Never pay 2 months. This may be because the maturity of
data as customer in 2 month may not be stable with service
provider as compared to never pay month 4.
In the both case correct prediction of loyal customer i.e.
never pay equals to 0 is high and so both can be used to select
loyal customer. But at the time company may have bad
customer as rate bad customer i.e. never pay equals to 1 is
very low. So we need to have some mixed of these models to
increase the overall correct percentage of prediction while in
above experiment this is almost 70%.
In the next chapter, a hybrid method has been proposed using
SVM ANN. The proposed methods will used SVM and ANN
different ways step by step and one method apply is subset of
attribute and another in different subset of attribute.
References
1.
Sam Anahory, Dennis Murray, Data Warehousing in

the Real World: A practical Guide for Building Decision
Support System, 14th ed., Pearson Education, 2005.
2.
W. Au, C. C. Chan, Yao X., A Novel evolutionary data

mining algorithm with applications to churn
prediction, IEEE Transactions on evolutionary
computation, Vol. 7, No. 6, Dec 2003.
3.
Alex Berson and Stephen J. Smith. Data Warehousing,

Data Mining & OLAP, 4th ed., Tata McGraw Hill,
2004.
4.
Buckinx, D. Van den Poel, Customer base analysis:

partial detection of behaviourally loyal clients in a non
contractual FMCG retail setting, European Journal of
Operational Research 164 (2005) 252-268
5.
W. Buckinx, G. Verstraeten, Van den Poel D.

Predicting customer loyalty using the internal
transactional Database, (2005).
6.
N. Chawla, K. Boyer, L. Hall, Kegelmeyer P. Smote,

Synhetic minority oversampling technique, Journal
ofArtificial Research 16 p321-357 (2002).
7.
Chiang I-Jen, Lin T. Y., and Liu Y., Table

Representations of Granulations Revisited, Rough
Sets, Fuzzy Sets, Data Mining, and Granular
Computing, 10th International Conference, RSFDGrC
2005, Regina, Canada, pp. 728-737, August 31 September 3, 2005.
8.
K. J. Cios and L. Kurgan, Hybrid Inductive Machine

Learning Algorithm That Generates Inequality Rules,
Information Sciences, special issue on soft computing
data mining, pp. 37-84, 2004.
9.
Hu, X., T. Y. Lin and J. Han, A new rough set model

based on database systems, Journal of Fundamental
Informatics, vol.59, pp.135-152, 2004.
10. h t t p : / / w w w. v t t . f i / i n f / j u l k a i s u t / m u u t / 2 0 0 6 /
customer_churn_ case_ study.pdf.
11. Z. Zhang, Y. Lu, and B. Zhang, An effective
partitioning combining algorithm for discovering
quantitative association rules, In Proc. of PAKDD, pp.
241251, 1997.
15
Eigen Spectrum Analysis for Order Reduction of Linear

System using Time Moment Algorithm
Vinod Kumar*, J.P.Tiwari*
mixed method is proposed for finding stable reduced order models of single-input- single-output large-scale systems
using Factor division algorithm and the clustering technique. The denominator polynomial of the reduced order model is
determined by forming the clusters of the poles of the original system and the coefficients of numerator polynomial are
obtained by using the factor division algorithm. This method guarantees stability of the reduced order model when the original
high order system is stable. The methodology of the proposed method is illustrated with the help of examples from literature.
Keywords: Order Reduction, Eigen Spectrum Analysis, Stability,
Time Moment Matching, Integral Square Error (ISE).
1. Introduction
Order reduction is a common approach within the
simulation, control and optimization of complex physical
processes. The mathematical models used in these
computations often result in large scale systems. Order
reduction is often crucial to accelerate the simulation of such
large scale systems. In the advent of new and more complex
technology, engineers often encounter large-scale systems
which are numerically demanding, structurally spacious, and
are not very practical. These large scale systems in turn
create a demand for smaller, less spacious and
computationally faster systems. To achieve this goal,
engineers rely on model reduction technique. The problem
must be approached from a realistic point of view in order to
preserve the characteristic of the original system and reserve
case for troubleshooting and maintenance. A large Number
of methods are available in the literature for order-reduction
of linear continuous systems in time domain as well as in
frequency domain [14]. Further, (Chen [5], Genesio et al.
[1], Elrazaz et al. and many others have studied extensively
different The present attempt is towards evolving an order
reduction method in which both the pole centroid and system
stiffness of the original and reduced order systems are kept
exactly same to obtain the reduced order system poles and
then obtain the numerator by exact time matching. Although
this approach was used for only real poles but we have extend
for complex poles. The paper provides a new mixed method
that uses the eigen spectrum analysis approach [7-8] to
obtained the denominator and the time moment matching

approach to derive the numerator of reduced order model.
2. Statement of Problem
Consider an nth order linear time invariant dynamic system
(i input-j output) described in transfer matrix
* Department of Electrical Engineering, Institute of

Technology, Banaras Hindu University, Varanasi, India
E-mail: vinod.kumar.eee09@itbhu.ac.in,
E-mail: jptiwari.eee@itbhu.ac.in
16

S & T Review
17
4.
18

S & T Review
19
References
[1]
R. Genesio, M. Milanese, A note on the derivation

and
use of reduced order models, IEEE Trans.,
Automat. ControlAC-21(1), pp118122 , 1976
[2]
M.S. Mahmoud, M.G. Singh, Large scale systems

modeling, International Series on systems and
Control, first ed., vol. 3, Pergamon Press, 1981.
[3]
A.K. Mittal, R. Prasad, S.P. Sharma, Reduction of

linear dynamic systems using an error minimization
technique, J. Inst. Eng. India,IE (I) J. EL
84,pp201206 , 2004.
[4]
R. Prasad, J. Pal, Stable reduction of linear systems

by continued fractions, J. Inst. Eng. India IE (I) J.
EL 72 ,pp 113116, 1991.
[5]. C F Chen. 'Model Reduction of Multivariable

Control Systems by means of Matrix Continued
Fraction', International Journal of Control, vol 20,
pp 225, 1974.
6. Conclusion
A new mixed method for computing reduced order models
for high order stable linear time invariant scalar dynamical
system is presented. The method produces stable reduced
order models for stable high order system. In the proposed
method the poles are synthesized by eigen spectrum
analysis and zeroes are determined by time matching. The
used time matching (TM) technique allows matching of the
step response is quite good but, in some cases the reduced
order system develops tendency to become non-minimum
phase. Though not investigated mathematically, the reason
seems to be forced equalization of the system stiffness. The
method is simple, rugged and takes little computational
time. The matching of step response is assured reasonably
well. The algorithm preserves more stability and avoids any
error in between the initial or final values of the responses of
original and reduced order models. Step response of this
method is quite well then other methods.
20
[6]
S. Mukherjee, Order reduction of linear system

using eigen spectrum analysis, J. Inst. Eng. India IE
(I) J. EL 77, pp7679, 1996.
[7]
G. Parmar, S. Mukherjee, Order reduction using

zero spectrum analysis, Int. Conf. Comput. Appl.
Electr. Eng., Recent Advances (CERA-01), I I T
Roorkee, India, pp. 450454, 2002.
[8]
TN Lucas and I F Beat, Model reduction by Least

squares moment matching, Electronic Letters,
vol26, No.15, pp 1213-1215, 1990.

S & T Review
A Multilevel Cascaded Converter Based

Solid State Distribution Transformer
*
Rakesh Thapliyal , Y.R Sood , R.K Jarial , Navin Chandra Joshi
his paper covers some aspects of modelling and simulation of solid state distribution transformer by using multilevel
cascaded converter concept with sinusoidal pulse width modulation (SPWM) controller in three stages. The input and
output stages of solid state distribution transformer are controlled with the help of current and voltage source converters.
Presented configuration consists of front and back end H bridge converters with multi-winding high voltage and high
frequency (HV-HF) transformers. In isolation stage dc link voltage is modulated to high frequency square wave ac that is being
reconverted to low dc voltage, which acts as proportional amplifier. Inverter controls the magnitude of output side ac voltage.
Mathematical model of controller are discussed in a-b-c and d-q reference frame. Solid state transformer is identical with
conventional transformer in terms of electrical performance. With proposed control strategy the dynamic characteristics of
solid state based distribution transformer are verified in Matlab/Simulink.
Keywords: Isolation Bridge; Multilevel Cascaded Converter;
Proportional Amplifier; Solid State Distribution Transformer.
1. Introduction
Solid state transformer (SST) is a new intelligent transformer
also known as Power electronics transformer. SST is
employed not only to perform basic function of conventional
power transformer such as voltage transformation, isolation
and power delivery but also realizes
many additional
features like power quality improvement such as voltage sag,
voltage swells, over voltage, under voltage, voltage
fluctuation, voltage unbalance, harmonics when used in
conjunction with modern distribution system. SST provides
a fundamentally different and more complete approach to
transformer design by using power electronics on the
primary and secondary sides of high frequency transformer
configuration.
For realizing the SST, different topologies have been
proposed in literature [3, 6] where in an AC/AC buck
converter was introduced to transform voltage level directly
and without any isolation transformer. This method is
perhaps the most familiar approach to single phase ac
conversion, but it would bear high stress on semiconductor
devices.
The focus on this paper is to realize a HV to LV solid state
transformer through a proposed control strategy for
* Electrical Engineering Department, NIT Hamirpur, H.P., India,
Email: thapliyalrakesh@gmail.com, E-mail: yrsood@gmail.com,
E-mail: jarial0@gmail.com, E-mail: navinjoshi.gbpec@gmail.com
industrial acceptability for crystallizing the use of SST in the

futuristic distribution substations. Sinusoidal pulse width
modulation (SPWM) technique have been used to control
multilevel cascaded H bridge converter with dc link
connected across isolation stage[4].
In this ppaper brief overview of the configuration and
principle of multilevel cascaded converter cell based
modelling of power electronics transformer is presented in
section II. The control techniques of SST have been covered
in section III. Finally the relative characteristics and
simulation aspects have been carried out in section IV.
2. Principle and Configuration of SST
2.1 Principle of SST
The basic building block configuration of power electronics
based distribution transformer is shown in Fig. 1. Block
diagram model consists of isolated DC/DC converter
between primary and secondary stages.
As can be seen from the Fig. 1, this is a multilevel three
stage design that consists of primary input stage, isolated
kernel and secondary output stage. These three stages
convert the sine wave voltage from one level to another level
without changing input and output power-frequency (50 or
60 Hz). The power-frequency sinusoidal voltage is first
modulated to high frequency (H.F) pulses through input
primary side multilevel converter and then coupled
magnetically with H.F transformer.
21
The low level dc voltage is obtained at output dc link of

isolated kernel, which is being connected through a voltage
source inverter (VSI). Module of VSI comes across
secondary stage that would convert dc voltage to three phase
power-frequency sinusoidal ac voltage [5].
Multilevel cell SST has the following supportive features

1. Identical electrical performance corresponding to
conventional transformer, as voltage transformation and
isolation between primary and secondary sides.
2. It has desirable performance for voltage regulation.
3. It prevents current and voltage harmonics propagating
either sides of transformer.
3. Proposed SST System Control Strategy
SST system employed for distribution resources to achieve
voltage transformation and isolation. Isolation system
prevents harmonics from either end of multilevel converter
cell.
Fig.1. Principle topology diagram of SST
2.2 Configuration of SST

Modules structure of SST meet with requirement of single
phase load or line to line load in a distribution network. Fig. 2
shows SST structure compose of H-bridge cells of power
electronics devices connected with single phase input supply
from generation side. The output in secondary stage consists
of three phase inverter. They convert multilevel dc link
voltage into three phase balanced ac sinusoidal voltage.
There should be desirable control strategy for obtaining high

quality power conversion.
3.1 Input stage multilevel converter cells controller
Primary stage of SST is connected with multilevel H-bridge
converter [3, 11]. The modular structure of this stage directly
converts the input ac voltage into high frequency sinusoidal
square wave. Fig.3. represents PWM pulse generation
through the reference dc link voltages. This type of
techniques used as analog controller.
Fig.3 PWM Controller for input stage
Primary stage Isolation kernel
Secondary stage
Fig.2. Circuit diagram of four level converter based SST
22
Fig.4. d-q decoupled input controller of H bridge

S & T Review
Fig.4. represents schematic comparison of synchronized

voltage pulse with reference voltage pulse and pass over a PI
controller for getting desired current waveforms in d-q
network.
The mathematical model for input controller in synchronous
rotating d-q reference is presented as
Where Ls is the source inductance, is the synchronous

angular velocity of the grid voltage, V1d, V1q and i1d, i1q are d-q
axis voltages and currents and ed, eq are reference dc link
voltages in d-q frame. Eq. (1) will influence the dynamic
system performance of controller in d-q axis.
phase should have capability to provide a clean and constant

output voltage platform for end users. A three phase voltage
source inverter is applied at output stage.
L-C filter is connected across output phase for obtaining pure
sinusoidal voltage. So, the mathematical model of a phase
output stage can be represented as
Where, Cf and Lf are filter capacitance and inductance

respectively.
3.2 Controller for isolated kernel

High level dc voltage coming from the input stage would be
modulated into high frequency square wave ac and then
magnetically coupled with high frequency transformer for
obtaining low voltage square wave ac. This low voltage
square wave ac is reconverted to a fixed level dc voltage.
Open loop PWM control is applied on front and back-end Hbridge through high frequency link transformer [16]. This
simplified model of isolation would change the level of dc
voltage, which is acting as a proportional amplifier.
Relation for proportional control is represented as:
Fig.5. Controller for inverter stage
Fig.5 depicts schematic detail of inner current loop and outer

voltage loop of inverter in both voltage and current
regulators used in PI controlling techniques. Where Vp, Ip and
Vn, In represents phase voltage, current and neutral voltage,
current respectively of an intermediate conversion link.
Corresponding K's represents adjustable loop gains in
relation with reference.
3.4 SST input output characteristics
The power transferred being controlled by the phase angle

difference and the magnitude of the dc voltages at the two
ends as given by:
Table 1. Principle Parameters
Where, Vdc is front end referred input dc link voltage, Vdc1 is

input dc voltage, f is switching frequency, L is leakage
inductance and ddc time delay duty ratio between two bridges.
3.3 Output stage inverter controller
This stage is directly connected to utility sides. SST output
23
A model of SST based upon Matlab/Simulink was prepared

and simulated. The simulation investigates the behaviour of
the input output characteristics of multilevel converter based
SST. In the simulation, Primary voltage of SST is 11kV and
output voltage has been chosen as 230V at rated frequency of
50Hz.Parametric values of SST are enumerated in Table 1.
The simulated results of input and output waveforms are
presented in Fig. 6-10 with and without filter.
24

S & T Review
4. Conclusions
In this paper a control strategy for modeling and simulation
of SST with various controller elements has been drawn,
which reflects step less variation to validate the process of
modeling as per standard practice. The proposed design
consists of three stages i.e. rectifier, isolation stage and
output inversion stage. Using rectifier stage, medium dc
voltage is realized. Rectifier controls input power and reduce
high voltage stress level. Isolated DC/DC converters provide
good regulation between input and output sides. The
simulated result shows SST have good steady state
performance with proposed controlling strategy. These
characteristics highlights through proper selection of power
electronics switches smooth steady state outcome can be
achieved across input side of power substation that
facilitated smooth evacuation of transformed power without
causing impairment to IGBT devices in practice ensuring
reliable operation of SST in futuristic substations.
References
[1] J. L. Brooks, Solid state transformer concept
development, in Naval Material Command. Port
Hueneme, CA: Civil Eng. Lab., Naval Construction
Battalion Centre, 1980.
[2] Kheraluwala, M.N.; Gascoigne, R.W.; Divan, D.M.;
Baumann, E.D.,Performance characterization of a
high-power dual active bridge DC to-DC converter,"
Industry Applications, IEEE Transactions on, vol.28,
no.6, pp.1294-1301, Nov/Dec 1992.
[3] M. Kang, P.N. Enjeti, I.J. Pitel, Analysis and design
of electronic transformers for electric power
distribution system, IEEE Trans. Power Electr. 14
(1999) 11331141
[4] E.R. Ronan, S.D. Sudhoff, S.F. Glover, D.L.
Galloway, A power electronic based distribution
transformer, IEEE Trans. Power Delivery 17 (2002)
537543.
[5] Rodriguez, J.; Jih-Sheng Lai; Fang Zheng Peng,
"Multilevel inverters: a survey of topologies, controls,
and applications," Industrial Electronics, IEEE
Transactions on, vol.49, no.4, pp. 724-738,Aug 2002.
[6] M. Glinka, R. Marquardt, A new ac/ac multilevel
converter family, IEEE Trans. Ind. Electron. 52 (3)
(2005) 662669.
[7] D. Czarkowski, M.K. Kazimierczuk, SPICE
compatible averaged models of PWM full-bridge
DCDC converter, in: Proceedings of the 1992 IEEE

PEMC Conference, vol. 1, 1992, pp. 88493.
[8] K. Harada, F.Anan, K. Yamasaki, M. Jinno, Y. Kawata,
T. Nakashima, K. Murata, H. Sakamoto, Intelligent
transformer, in: Proceedings of the 1996 IEEE PESC
Conference, 1996, pp. 13371341
[9] D. Gerry, P. Wheeler, J. Clare, Power flow
considerations in multi-cellular, multilevel
converters, in: International Conference on Power
Electronics Machines and Drives, vol. 487, 2002, pp.
201205.
[10] Cha, H.J.; Enjeti, P.N., "A three-phase AC/AC highfrequency link matrix converter for VSCF
applications," Power Electronics Specialist
Conference, 2003. PESC '03. 2003 IEEE 34th Annual,
vol.4, no., pp. 1971-1976 vol.4, 15-19 June 2003.
[11] Jih-Sheng Lai; Maitra, A.; Mansoor, A.; Goodman, F.,
"Multilevel intelligent universal transformer for
medium voltage applications," Industry Applications
Conference, 2005. Fourtieth IAS Annual Meeting.
Conference Record of the 2005, vol.3, no., pp. 18931899 Vol. 3, 2-6 Oct. 2005.
[12] Jin Aijuan; Li Hangtian; Li Shaolong, "A three-phase
four-wire highfrequency AC link matrix converter for
power electronic transformer," Electrical Machines
and Systems, 2005. ICEMS 2005. Proceedings of the
Eighth International Conference on, vol.2, no.,
pp.1295-1300 Vol. 2, 29-29 Sept. 2005.
[13] Xiaolin Mao, Raja Ayyanar, and Sixifo Falcones, A
Modular, Interleaved AC-AC Flyback Topology for
Solid State Transformer, in Proc. FREEDM Annual
Conference 2009, North Carolina State University,
Raleigh, NC, May 18-19, 2009, pp. 221-224.
[14] Subhashish Bhattacharya, Tiefu Zhao, Gangyao
Wang, Sumit Dutta, Seunghun Baek, Yu Du, Babak
Parkhideh, Xiaohu Zhou, and Alex Q. Huang, Design
and Development of Gen-1 Silicon based Solid State
Transformer, in Proc. FREEDM Annual Conference
2009, North Carolina State University, Raleigh, NC,
May 18-19, 2009, pp. 69-74.
[15] W. McMurray, Power converter circuits having a high
frequency link, US Patent 3,517,300, June 23, 1970.
[16] Harish K. Krishnamurthy, Control Strategies for a
Universal Fully Modular Power Conversion
Architecture, Ph.D. Dissertation, Arizona State
University, 2008.
25
Detection of Hand Movement Using EMG Signal

Antra Saxena*, Varsha Gupta*, Abid Ali Khan1, Omar Farooq*, Yusuf U Khan2
his paper proposes the use of surface Electromyography (EMG) signal for the identification of extension and gripping
hand movement. Two surface electrodes were placed on two different muscles for acquiring the electromyography
signals from 7 normal healthy subjects. In order to distinguish the hand movements, the statistical features based on mean,
variance, skewness and kurtosis have been extracted. These features were given to three different types of classifiers based on
linear, quadratic and Mahalanobis discriminant function. Using t-test criterion features were ranked and top 3 features were
selected for classification. The comparative classification performance shows best recognition accuracy of 90.2% achieved by
the quadratic classifier.
Keywords: Classification, Discriminant function, Electromyography,
Feature extraction, Human computer interface
1. Introduction
In biomedical research, signals are acquired from any organ
to represent a physical variable of interest. Among the
various types of physiological signals, myoelectric signals
are less researched upon as compared to ECG and EEG.
Electromyography (EMG) is an important tool in the field of
biomedical engineering. EMG signal is generated when
there is any muscle activity. Significant numbers of people
suffer from injuries which may cause restrictions in their
hand movements. The quality of life for these individuals
could be improved by developing a human computer
interface which can control prosthetic hand movement using
Electromyography (EMG) signals [1-4]. This approach is
based on the assumption that EMG patterns carry
information about the hand movements.
Features are extracted from the EMG signals and given to a
classifier for identification of different movements. Various
features such as integral absolute value, difference absolute
mean value, variance, and autoregressive model coefficients
have been used [5]. Different types of classifier based on
discriminant function, Markov's model and artificial neural
network has also been used [4, 5].
This paper proposes the use of statistical features for the
extraction of useful information from EMG signals collected
form two muscles extensor and flexor to detect hand
* Department of Electronics Engineering, Aligarh Muslim University, Aligarh,
E-mail: antra3009@gmail.com,
E-mail: guptavarsha1@gmail.com, E-mail:omarfarooq70@gmail.com
1. Department of Mechanical Engineering, Aligarh Muslim University, Aligarh,
E-mail: abidalikhan@zhcet.ac.in
2. Department of Electrical Engineering, Aligarh Muslim University, Aligarh,
E-mail: yusufkhan.ee@amu.ac.in
26
movement. For classification of collected data, discriminant

function based three classifiers were used. In order to reduce
the dimension of extracted features, and to select features
with higher discrimination capability, ranking was carried
out using t-test criterion. This extraction of knowledge about
the acquired EMG signal will help the researchers/ engineers
working to develop prosthetic arms for gripping functions.
Also it will be helpful to understand the effects of gripping
activity in extensors compared to flexor muscle.
2. Method
An important requirement in biomedical engineering is the
detection of physiological signals. EMG signal is normally a
functionof time with varying amplitude, frequency and
phase. It is a signal which is generated due to
contraction/relaxation of musclesand depends on the
properties of muscles(such as anatomical and
physiological). EMG detectorsare placed on the surface of
skin; it collects signals from different motor unit
2.1 Recording techniques
For the purpose of recording EMG signals, two types of
techniques are available: Surface electromyography and
intramuscular electromyography. Surface EMG is common
and simple to use as it requires placement of the electrodes on
the skin surface. Intramuscular EMG requires a neurologist
or a trained professional who inserts a needle electrode
through the skin into the muscle tissue. The data collected for
this experiment was obtained by using surface EMG
technique. In this electrodes (i.e. EMG preamplifier) were
placed on the skin of the subject (well cleaned before
placement of the electrode), for the purpose to detect a
S & T Review
muscle activity or group of muscle activity during different

movements. The electrodes were placed as per SENIAM8.0.
The location was mid of the muscle for electrode placement.
muscles namely extensor and flexor using acquisition

system. Each EMG signal set has been recorded with
movement starting from rest to full movement.
2.2 Experimental setup and signal recording
2.3 Procedure
In this experiment, five females and 2 males of different age

groups participated. None of the subject had a history of
musculoskeletal injury in their dominant hand. The EMG
data was collected at a sampling rate of 1000 Hz using
Surface EMG sensor (Model: SX230 EMG sensor; Make:
Biometrics Ltd. UK). The pre-amplified signal of EMG was
interfaced to personal computer using 8 channelsDATA Link
system (M/s biometrics Ltd., UK). The earthing strap was
used during EMG recording for reduction of noise. The
signal was recorded using the Biometrics data acquisition
software. The signal was conditioned using a DC blocking
filters and a Low pass filter. Here signal was recorded on two
First of all EMG sensor was attached to the selected extensor

and flexor muscle with die cut medical grade double sided
adhesive tape. The participants were then asked to start the
movement from rest to maximum point of movement for one
minute duration. EMG recording for each movement was
carried out in succession without the setup being altered.
3. Analysis Techniques
The problem of identification of hand movement based on
EMG signal is a standard pattern recognition task. Once the
signal is pre-processed and DC is removed, relevant features
are to be identified first which could be used to discriminate
27
between the two movements using a classifier. The block

diagram of the entire scheme is shown in Fig.1. To evaluate
different frequency components present in the EMG
recording corresponding to two hand movements
spectrogram was calculated (using MATLAB). The results
obtained are shown in Fig. 2 and Fig. 3 Further, the time
domain properties of the signal corresponding to the two
hand movements was evaluated by using a histogram plot as
shown in Fig. 4and 5.
skewness and kurtosis were extracted using one second

frame duration for each muscle recordings. The extracted
features were ranked by t-test criterion using data from all the
subjects. The ranking of features helps to identify good
features to achieve high classification accuracy. The result of
ranking of the features is shown in the Table I. In order to
reduce feature dimension, 6 features were selected out of 8
based on ranking.
3.2 Classification
The purpose of a classifierwas used to identify hand
movement based on the extracted features.Three types of
classifiers (linear, quadratic and Mahalanobis) were
employed in this study. Data set corresponding to one
movement was assigned class 0 and the data set
corresponding to other movement was assigned class 1. 70%
of the data was used for training and rest 30 % for testing.
The average value of EMG signal was evaluated and found
to be close to zero. Mathematically, it is given by:
Variance of EMG signal is calculated by:
Skewness is a measure of symmetry or the lack of symmetry.

A distribution is symmetric if it looks the same to the left and
right of the mean point. For univariatedata , , ............, , the
skewness is evaluated as :
3.1 Feature extraction

A good feature vector is a set that provides provides
maximum separation with minimum dimension. By
analysing the dispersion statistics on the data (see Fig. 4), it
was clear that the two distributions were different and
variance measure can be an important feature for
classification. Statistical features such as mean, variance,
28

S & T Review
4. Results and Discussions
References
The classification results obtained using all the eight features

for three different classifiers are shown in Table II. It showed
that a quadratic classifier gave the best recognition
performance among. This was due to the fact that the
decision boundary between the two classes was not linear.
On the basis of feature ranking as discussed in the previous
section, choosing only 6 features did not degrade the
recognition performance for quadratic and mahalanobis
classifier while for linear classifier it reduces as shown in
Table 2. Thus, it was observed that with fewer features we
can achieve the same performance but with reduce
complexity.
[1]
P. Bornato, T. de Alessio and M. Knaflitz, A

statistical method for the measurement of the muscle
activation intervals from surface myoelectric signal
gait, IEEE Trans Biomed Eng; vol. 45, pp. 287299.1998.
[2]
G. Li, A. E. Schultz andT. A. Kuiken, Quantifying

pattern recognition based myoelectric control of
multifunctional transradial prostheses, IEEE
Trans. Neural Syst. Rehabil. Eng., vol. 18, pp. 185192, 2010.
[3]
G. Li, Y. Li, L. Yu, andY. Geng, Conditioning and

sampling issues of EMG signals inmotion
recognition of multifunction myoelectric
prostheses, Ann. Biomed. Eng., vol. 39(6),
pp.1779-1787, 2011.
[4]
L. Hargrove, K. Englehart and B. Hudgins, A

comparison of surface and intramuscular
myoelectric signal classification, IEEE Trans.
Biomed. Eng., vol. 54, pp. 847-853, 2007.
[5]
S. H. Park and S. P. Lee, EMG pattern recognition

based on artificial intelligence techniques, IEEE
Trans. Rehab. Eng., vol. 6, no. 4, pp. 400405, Dec.
1998.
5. Conclusions
The identification of two kinds of hand movements using
EMG data has been successfully achieved by extracting
simple statistical features in time domain. The variance
feature of the EMG data was found to be the best for
classification of hand movements. Over 90% accuracy was
achieved by using a simple discriminant function based
classifier. It was further observed that the quadratic classifier
gave the best performance even using the reduced set if
features.
29
Matlab/Simulink Based Mathematical Modeling and

Simulation of an Air Breathing Proton Exchange
Membrane Fuel Cell (ABPEMFC)
1
A.R.Vijay Babu , P.Manoj Kumar
he proton exchange membrane (PEM) fuel cell is being investigated as an alternate power source for various
applications like transportation and emergency power supplies. In order to investigate the polarization
characteristic of an air breathing proton exchange membrane fuel cell (ABPEMFC) based on the empirical model, a
novel mathematical model of the PEMFC has been developed with MATLAB/Simulink. By using a fuel cell test
system of the PSG Institute of Advanced Studies (PSGIAS) at Coimbatore, the responses of ABPEMFC were
conducted and analyzed under various operating conditions. The paper presents a novel circuit model for an
ABPEM fuel cell that can be used to design and analyze fuel cell power systems. The model includes the phenomena
like activation polarization, ohmic polarization, and mass transport effect present in an ABPEM fuel cell. The
characteristics obtained through simulation are compared with experimental results obtained on a commercial fuel
cell module.Agood match is found between simulation results and experimental data. The comprehensive results of
simulation manifested that the model is effective and operational. This model will be very useful to optimize the
structure design, improve the operation performance, and develop the real-time control system ofABPEMFC.
Keywords : Air Breathing, Proton Exchange, Membrane,
Fuel Cell (ABPEMFC), Mathematical Model, Polarization
1. Introduction
The fuel cell is one of the most promising sources of
renewable energy for the sustainable future. They can be
considered as green power because they are environmentally
clean, has low emission of oxides of nitrogen and sulphur
and at the same time, they can operate with a very low level
of noise. In addition, they can provide energy in a controlled
way with higher efficiency (around 90%) than conventional
power plants. The fuel cell transforms hydrogen into dc
power. There are different technologies of fuel cell. They are
commonly classified according to temperature and/or the
type of electrolyte. Among others, low-temperature fuel cell
includes proton exchange membrane (PEM), can be
considered a good alternative for the use aboard of electric
vehicles in which simplicity, high specific power and rapid
start-up at different temperatures have a significant
importance [3]. The work process of a PEM fuel cell is
1. School of Energy, PSG College of Technology,
Coimbatore, E-mail: 202vijay@gmail.com
2. Department of Mechanical Engineering,
PSG Institute of Advanced Studies, Coimbatore,
E-mail: manojpanthalingal@gmail.com
30
shown in Fig. 1 and it is suitable to be used in portable power

supply, vehicles, residential and distributed power plants [1],
[2].
Under normal operation, a simple PEMFC typically
produces 0.5 V to 0.9 V. For use in energy generation
systems, where a relatively high power is needed, several
cells are connected in series, arranging a stack that can
supply hundreds of kW. It is expected that in near future
commercially available products will be rated at MW range.
Cells that take up oxygen, for the cathode reaction, from
ambient air by passive means are known as air-breathing
fuel cells. In an air-breathing fuel cell, air flow along the
cathode surface take place because of concentration and
temperature gradient between the cell and the ambient. Airbreathing fuel cells have great potential as power source for
low power portable electronic devices as it eliminates the use
of compressor/blower and humidifier on the cathode side,
thus making the system simple, light and compact. Since an
air-breathing fuel cell take up oxygen, by natural convection,
directly from the surrounding air, the ambient conditions
play a very important role on the cell performance [4].
S & T Review
In this paper, the MATLAB/ Simulink tools based on the

obvious mathematical model of ABPEMFC are used to
establish a simple and effective ABPEMFC model through
which the polarization characteristics of the ABPEMFC with
different test conditions are analyzed. Meanwhile, the
dynamic electrical response of the ABPEMFC was tested by
scribed fuel cell test system 850 e, which has high precision.
The comparison of the test results and the modelling
simulation results show that the model can effectively
predict and evaluate the characteristics of the fuel cell, and it
can be helpful for the system analysis, design optimization,
and real-time control of the PEMFCs.
2. Model Development
Fig. 1 Work process and reaction
principle of the PEMFC
A basic diagram showing the structure of the ABPEM fuel

cell is shown in Fig. 2. The main elements inside the cell are:
end/supporting plates, conductor plates, flow field plates and
membrane electrode assembly (MEA). The electrodes are
composed by a gas diffusion layer and a catalyst layer. Both
layers have a porous, partially hydrophobic, structure.
There is a need for a reliable mathematical model. Such
model can allow the evaluation of the PEMFC dynamic
performance for small size electrical energy generation
systems, reducing cost and time along the design stage and
tests. Such need motivated us to conduct mathematical
modeling to determine the polarization curve of PEM fuel
cell for each operating point. In power generation systems,
the dynamic response is extremely important for the planner
of control and management systems; especially when there is
injection of energy into the grid. So, a special attention
should be given to the dynamic response of the FC.
Fig. 2 Structure of the ABPEMFC.

2.1 Theory ofAMPEMFCs

ABPEMFC is a nonlinear, multiple-input and -output,
strongly coupled, and large-delay dynamic system, the
working process of which is accompanied with
liquid/vapor/gas mixed flow transportation, heat
conduction, and electrochemical dynamical reaction.
Hydrogen and oxygen work as fuel and oxidant, and they
need to be humidified before they are fed into the cell/stack.
In an air-breathing fuel cell, air flow along the cathode
surface take place because of concentration and temperature
gradient between the cell and the ambient. The positively
charged protons diffuse from anode through one side of the
membrane and migrate toward the cathode. The electrons
pass from the anode to the cathode through an exterior circuit
and provide electric power along the way; the by-product
water is drained out with the gases. The work process and
reaction principle are shown in Fig.1. The electrochemical
reactions involved are summarized below.
AtAnode:
Eq. (1) describes the chemical reaction at the anode. The

electrons are transferred to the platinum layer and protons to
the central membrane. Eq. (2) shows what happens at the
cathode. The oxygen reacts with the protons coming from the
membrane and with the electrons fed by the catalyst. The
31
result is water. Finally, Eq. (3) shows the overall reaction.

Platinum catalyst is used to speed up the electrochemical
reaction on both anode and cathode. Hydrogen atoms are
stripped of their electrons, or ionized, at the anode. At the
cathode, the electrons, hydrogen protons, and oxygen from
the air combine to form water [5][12].
2.2 Mathematical model ofABPEMFCs
In order to simplify the analysis, several assumptions are
made as follows.
a)
All gases obey the ideal gas law and are equably
distributed.
b)
Pure hydrogen gas is assumed to be supplied to the

anode from a hydrogen tank and the inlet hydrogen
flow and oxygen flow are assumed to have 100%
relative humidity
c)
The temperatures of hydrogen inside the anode and

oxygen inside the cathode are equal to the stack
temperature.
d)
Pressures in the gas flow channels are constant.
e)
The only reaction product is liquid water.
The ideal standard potential of a hydrogen/oxygen fuel cell

o
at standard state conditions (25 C and 1 atm) is 1.229Vwith
liquid water product. However, the actual cell potential
decreases from its equilibrium potential because of
irreversible losses. There are three types of irreversible
losses, namely activation losses, ohmic losses, and
concentration losses. At low current densities, the activation
losses of oxygen reduction reaction are almost entirely
responsible for the potential drop of the cell. However, at
high current densities, the concentration losses become more
significant. The output voltage of the single cell is given by
Eq .(1) according to the PEMFC output characteristics
empirical equation, which was developed by Amphlett et al.
[13] and Kim et al. [14].
32
In order for a reaction to activate, it is necessary to outstrip

the limit of activation energy. The activation losses represent
the difference from the equilibrium that is needed to get the
reaction started, and this is mainly due to sluggish electrode
kinetics; these losses happen at both anode and cathode side.
It happens that some small amounts of hydrogen diffuse from

anode to cathode whereas some electrons may find a
shortcut through the membrane. These losses appear
insignificant in fuel cell operation, but when the fuel cell is in
the open circuit mode, or when there is a very low current
density, these losses may have a dramatic effect on cell
potential. To simulate this phenomenon, the term iloss is added
to the current i in (11) with a fixed value of 3 mA /cm2 [16].So
in a fuel cell the activation losses are:
Different from the Ohmic losses of the empirical equation

that is already presented, in this model, Ohmic losses consist
of the voltage drop that is caused by cathode and anode
electrical resistances, the equivalent membrane impedance,
and the voltage drop that is caused by the contact resistances
both between the membrane and electrodes as well as the
electrodes and the bipolar plates. It is constant once the cell is
fabricated. It can be shown as

S & T Review
2.3 Single cell modeling
Where, is the water content of the membrane, which is an

adjustable parameter and a function of the relative humidity
of the gas in anode, and has a stoichiometric value.
Concentration polarisation occurs when a reactant is rapidly
consumed at the electrode by the electrochemical reaction so
that concentration gradients are established. In this model,
the effect of concentration losses is also considered, which is
different from previous models. Concentration losses are
caused by mass transportation, which, in turn, affects the
concentration of the hydrogen and oxygen at high current
density. This term is ignored in some models, perhaps
because it is not desirable to operate the cell at regions where
concentration losses are high (efficiency is low). However, if
the cell operates at high current density, this term needs to be
included. The concentration losses can be expressed as
The model was built based upon Eq. (15) multiplied by an

empirical factor (called gain factor from now on) that
approximates the behaviour of the high current field of the
polarisation curve. The gain factor K used is 15 and this
makes the curve decrease earlier than how it would decrease
just through the theoretical formulation. The value K = 15
was chosen once for all.
The model built in Simulink is obtained from Eq. (1) as it is

shown in Fig. 3.For each type of overvoltage there is a sub
system in which the input parameters are inserted and the
relative values of voltage are obtained as output according to
Eqs. (6), (11), (12) and (18).The model gives as output a
steady-state polarisation curve, depending on cell
temperature, hydrogen (Reactant/Fuel) pressure and relative
humidity (Anode dew point temperature).
2.4 Validation of the single cell model
The validation process started from experimental data
carried out ad hoc in laboratory. The cell for the validation
has an active area of 5 cm2, with the following configuration:
-
C atalyst coated membrane (CCM) Nafion_117

membrane with both electrode 40% Pt/C 1.0/1.0
mg/cm2.
Carbon cloth gas diffusion layer (GDL) treated with

PTFE 30% wt on cathode and PTFE 20%wt on anode.
The bipolar plates have a single channel serpentine to

feed the membrane with the gases; hydrogen is
humidified by bubbling the gas through water.
In the Ohmic losses region, i.e. between about 0.2 and 0.7
A/cm2, the agreement between experimental and theoretical
data is very good (Fig. 7-10); it should be remarked that this
current density range is representative of common steady
state operating conditions. At high current density (over 0.8
A/cm2) the comparison between model and experimental
data is less satisfactory because of the influence on losses of
porous media microstructure, of cell water uptake and of lack
of precise reactant gases distribution at the cathode side.
These microstructure features are not properly accounted for
in the present modeling.
Table 1: Simulation Parameters
Temperature
o
( C)
Charge
Transfer
Coefficient
Exchange
Current
Density (iO)
30
0.967
1.02 mA/ cm
40
0.725
1.662 mA/ cm
50
0.552
1.6804 mA/ cm2
60
0.234
1.68104 mA/ cm2
33
The values of simulation parameters are taken from the data

sheets of their respective manufacturers and some of them
are taken from literature [5]-[12] are tabulated in Table 1.
2.5 Electro chemical parameters of fuel cell
Charge transfer coefficient and Exchange current density are
together known as the Electro Chemical Parameters of a fuel
cell. The charge transfer coefficient signifies the fraction of
the interfacial potential at an electrode-electrolyte interface
that helps in lowering the free energy barrier for the electro
chemical reaction. The exchange current density is that
current in the absence of net electrolysis and at zero over
potential. The experimentally calculated values of Charge
transfer coefficient and Exchange current density are
tabulated in Table 2.
Fig. 4. Assembled Single Cell.
Table 2. Electro chemical parameters of an ABPEMFC
Parameter
Value
Specific Resistance of GDL
0.0017 -Cm
Thickness of GDL
0.036 Cm
Specific Resistance of Graphite
0.00231 -Cm
Thickness of Graphite Flow Channel
0.1 Cm
Thickness of Membrane
7 Mil = 0.0178 Cm
Water Drag Coefficient
12
Crossover Current
3 Ma/Cm
Contact Resistance
30 -Cm
Scaling Factor
15
Fig. 5. V-I characteristic of a single ABPEM fuel cell.
3. Experimental Test Facility and Results

The MEA of fuel cell is fabricated in PSGIAS. The
assembled single cell is shown in Fig.4. The fuel cell has
been put to the test using a test system facility composed by
the fuel cell, a controlled load and a easurement and control
unit. The test system facility has a block diagram that is
shown in fig.6. The basic idea can be summarized as follows:
the variable load can be adjusted in order to compute the v-i
characteristic. Fig. 5 shows the measured V-I characteristic
of the fuel cell at a cell temperature of 25oC.
Fig.6. Block Diagram of Fuel Cell Test Station
34

S & T Review
Fig. 7 shows the validation of experimentally measured and

Simulink model V-i characteristics of the fuel cell at a cell
o
temperature of 30 C.

o
Fig. 10. V-I characteristic of an ABPEM fuel cell at 60oC.

o
Fig. 7. V-I characteristic of an ABPEM fuel cell at 30 C.
It is observed from the fig .7-10 that the performance of fuel

cell is degrading as the temperature increases indicating a
sharp decrease in cell voltage and the current density. The
efficiency of the fuel cell is directly proportional to cell
voltage, so as cell temperature increases the efficiency of the
fuel cell will go low.
Nomenclature

o
temperature of 40 C
o
35
Acknowledgement
The authors would like to thank the support of the PSGIAS
for their kind cooperation during this work.
References
[1]
P. Manoj Kumar, Ajit Kumar Kolar, " Effect of

cathode channel dimensions on the performance of an
air-breathing PEM fuel cell, International Journal of
Thermal Sciences, December 2010.
[2]
J. Jia, Q. Li, Y. Wang, Y.T. Cham, M. Han, "Modeling

and Dynamic Characteristic Simulation of Proton
Exchange Membrane Fuel Cell" IEEE Trans. Control
Syst. Technol., vol. 24, no. 16, pp. 283291, Mar.
2009.
[3]
J. Jia, Y.T. Cham, W.K. Au "A Review of PEM Fuel

Cells", World Hydrogen Technologies Convention
(WHTC 2005), Singapore, 03 05 October, 2005
[4]
M. Ceraolo, C. Miulli, and A. Pozio, Modelling static

and dynamic behaviour of proton exchange
membrane fuel cells on the basis of electro-chemical
description, Journal of Power Sources, vol. 113, no.
1, pp. 131 144, 2003.
[5]
T. F. Fuller and J. Newman, Water and thermal

management in solid polymer electrolyte fuel cells,
J. Electro chem. Soc., vol. 140, no. 5, pp. 12181225,
May 1993.
[6]
J.C.Amphlett, R. F. Mann, B. A. Peppley,

P.R.Roberge, andA.Rodrigues,A model predicting
transient responses of proton exchange membrane
fuel cells, J. Power Sources, vol. 61, no. 1/2, pp.
183188, Jul./Aug.1996. (2002) The TMU website.
[Online].Available: http://www.TMU.org/
[7]
J. J. Baschuk and X. Li, Modelling of polymer

electrolyte membrane fuel cells with variable degrees
of water flooding, J. Power Sources, vol. 86,no. 1/2,
pp. 181196, Mar. 2000. FLEXChip Signal Processor
(MC68175/D), Motorola, 1996.
[8]
Y. Wang and C. Y. Wang, Dynamics of polymer
4. Conclusions
The main goal of the work was to construct a valid and
modular model of a ABPEM fuel cell that could be useful for
various portable applications. At first, a generalized steadystate fuel cell model was generated, and it was run and
compared with experimental data. The results of the
validation looked satisfactory; the polarization curves
approximated finely the trends of the experimental data,
certainly at the most typical operating conditions. The model
slightly underestimates the experimental data, probably due
to some simplifying assumptions.
No variable parameters are used in the simulation of the stack
in order to fit the experimental data. The only parameter
introduced to fit simulation with experiments was the gain
factor K used in the modelling of the concentration losses of
the single cell. Once chosen the best value at the level of
single cell model validation, the parameter was no longer
varied and any subsequent simulation on the stack was
performed without any variation of parameters.
The Simulink approach here adopted resulted valid to
describe a large range of fuel cell operations under steadystate conditions; the modelling was quite straightforward
and not so much demanding in terms of computing time.
It can also effectively predict and evaluate the performance
of the fuel cell. Further improvement of the model would be
helpful for the performance analysis, design optimization,
and real-time control.
36

S & T Review
[9]
electrolyte fuel cells undergoing load changes,

Electrochim. Acta, vol. 51, no. 19, pp. 39243933,
May 2006.
proton exchange membrane fuel cell performance

with an empirical equation, J. Electro chem. Soc.,
vol. 142, no. 8, pp. 26702674,Aug. 1995.
J. C. Amphlett, M. Baumertr, and F. Mannr,

Performance modeling of the Ballard mark IV solid
polymer electrolyte fuel cell: Empirical model
development, J. Electro chem. Soc., vol. 142, no. 1,
pp. 915, Jan. 1995.
[11] L. James and D. Andrew, Fuel Cell Systems

Explained, Chichester,U.K.: Wiley, 2003, pp.
1424.
[12] Babir F.PEM, Fuel cells theory and practices,
Oxford, UK: ElseveirAcademic press; 2006.
[10] J. Kim, M. Lees, and S. Srinivasan, Modeling of
37
Nonlinear Vibration and Instability Response of

Embedded Pipes Conveying Viscose Fluid Using DQM
1
H. Rahimi pour , M. Ghaytani , R. Kolahchi
n this paper, nonlinear vibration and instability response of an embedded pipe conveying viscose fluid is
investigated. The pipe is considered as a Timoshenko beam embedded on an elastic foundation which is
simulated by spring constant of the Winkler-model and the shear constant of the Pasternak-model. The external flow
force, acting on the beam in the direction of the flexural displacement is described by the well-known Navier-Stokes
equation. The corresponding governing equations are obtained using Hamilton's principle considering nonlinear
strains and first shear deformation theory. In order to obtain the nonlinear frequency and critical fluid velocity for
clamped supported mechanical boundary condition at two ends of the pipe, Differential quadrature method (DQM)
is used in conjunction with a program being written in MATLAB. The effect of dimensionless parameters such as
aspect ratios of length to radius of the pipe, Winkler and Pasternak modules, fluid velocity and viscosity as well as
the material type of the pipe on the frequencies and instability of pipe are investigated. Results indicate that the
internal moving fluid plays an important role in the instability of the pipe. Furthermore, the nonlinear frequency and
instability increases as the values of the elastic medium constants and viscosity of fluid increases.
Keywords: Nonlinear vibration, Instability, Fluid, DQM.
1. Introduction
A vast variety of structures or structural components used in
civil, mechanical, aerospace and defence engineering are
pipes. These pipes are often used to store and transport highpressure gases and liquids for various hydraulic applications.
A good understanding of their mechanical behaviour,
including vibration, bending and impulse response, is a must
for the successful design and application of pipes in
engineering practice. Vibration analysis of pipes has been
studied by many researchers.
Pipes conveying fluid have become one of the important
structures widely used in engineering, such as those
employed in nuclear reactor, ocean mining, heat exchanger
and drug delivery [1-3]. In such applications, one of the most
important issues is to accurately measure the vibration
characteristics, such as natural frequency, stability and
critical flow velocity of the fluid-conveying systems. It is not
surprising, therefore, that the study on this topic is constantly
expanding in the past decades. Indeed, the vibration and
1. Pars Oil and Gas Co, Iran, E-mail: hrahimipour@pogc.ir
2. NIOPDC, Iran, E-mail: mehdi_mn28@yahoo.com
3. University of Kashan, Kashan, Islamic Republic of Iran,
E-mail: r.kolahchi@kashanu.ac.ir
38
stability of pipes conveying fluid have been studied for more

than six decades, both theoretically and experimentally. A
good review of the related literature was provided by
Padoussis and Li [4].
In general, methods of vibration analysis can be classified as
analytical methods and numerical approaches. It is generally
agreed that the procedure for analytical/closed form
solutions of such vibration problems is tedious and in most
cases, no such solutions might exist at all. This promotes the
popularity of the numerical methods in solving this type of
problems. One class of numerical approaches is differential
quadrature methods (DQMs), which was first introduced by
Bellman et al. [5] in 1972 for solving partial differential
equations.
2. Formulation
Fig. 1 shows the pipes modeled as a Timoshenko beam with
length L, inner radius r1, outer radius r2 and equal thickness h
embedded in an elastic medium. The surrounding medium is
described by the Winkler foundation model with spring
constant k and Pasternak foundation model with shear
constant G. Based on the Timoshenko beam theory, the
displacements of an arbitrary point in the beam along the x-

S & T Review
and z-axes, denoted by U and W respectively, take the form

of [6]
Where U(x, t) and W(x, t) are displacement components in

the midplane, is the rotation of beam cross-section and t is
Fig. 1 Geometry of the pipe modeled

as the nonlocal Timoshenko beam.
time. the von Karman type nonlinear straindisplacement

relations are given by
39
4. Numerical Results
The final converged solutions using the numerical procedure
outlined in section B above are illustrated as nolinear
frequency and critical fluid velocity in Figs 2-6 below.
Fig. 2 and 3 illustrate the effects of aspect ratio (L/R) on the
dimensionless frequency versus fluid velocity and nonlinear
frequency ratio against maximum amplitude, respectively. It
is evident that an increase in the aspect ratio increase
dimensionless frequency and critical fluid velocity. Also,
with increasing L/R, nonlinear frequency ratio increases.
This is because increasing L/R leads to softer pipe.
Fig. 4 and 5 illustrate the influence of the normalized
Pasternak shear modulus (Kg) on dimensionless frequency
versus fluid velocity and nonlinear frequency ratio (i.e. the
dimensionless nonlinear to linear frequency (NL/L) versus
maximum amplitude (wmax), respectively. The result indicate
3. Differential Quadrature Method
The differential quadrature (DQ) method is used to solve the
nonlinear Eqs. (13)-(15) and the associated boundary
conditions to determine the nonlinear free vibration
frequencies of the pipes. The main idea of the differential
quadrature (DQ) method is that the derivative of a function at
a sample point can be approximated as a weighted linear
summation of the function value at all of the sample points in
th
the domain. The functions f = {u, w, } and their k
derivatives with respect to x can be approximated as [7]
that NL/L decreases substantially as harder elastic medium

is employed. Hence, with increasing Pasternak shear
modulus, NL/L decreases. Furthermore, as Kg increases,
the critical fluid velocity and nonlinear frequency increase.
Fig. 6 illustrate the effect of fluid viscosity on the
dimensionless frequency versus fluid velocity. The results
indicating that viscous fluid increases natural frequency very
little. However, during the flow of a fluid through a pipe as a
Timoshenko beam, the effect of fluid viscosity on the
vibration and instability of pipes may be ignored.
where N is the total number of nodes distributed along the xaxis and Cij is the weighting coefficients, the recursive
formula for which can be found in [8]. The cosine pattern is
used to generate the DQ point system
Using DQM, Eqs. (13) to (15) can be expressed in matrix

form as
where M is the ''mass matrix, KL is the linear ''stiffness

matrix and KNL is the nonlinear stiffness matrix.
40
Fig. 2 The effect of geometrical parameter on

nonlinear frequency versus Fluid velocity.

S & T Review
Fig. 3 The effect of geometrical parameter on nonlinear

frequency ratio versus maximum amplitude.
Fig. 6 The effect of fluid viscosity on nonlinear

frequency versus fluid velocity.
5. Conclusion
Fig. 4 The effect of Pasternak foundation on nonlinear

frequency versus Fluid velocity.
This paper investigates the nonlinear free vibration and

instability of pipes based on von Karaman geometric
nonlinearity and Timoshenko beam theory. The differential
quadrature (DQ) method and a direct iterative approach are
employed to obtain the nonlinear vibration frequencies and
critical fluid velocity of pipe with clamped supported.
Results indicate that the internal moving fluid plays an
important role in the instability of the pipe. Furthermore, the
nonlinear frequency and instability increases as the values of
the elastic medium constants and viscosity of fluid increases.
Acknowledgment
I should thank Pars Oil and Gas Co. for having sponsored
this project.
References
[1] M.P. Padoussis and G.X., Li, Pipes conveying fluid: a
model dynamical problem, J Fluids Struct., Vol. 7,
pp.137-204, Nov. 1993.
Fig. 5 The effect of Pasternak foundation on nonlinear

frequency ratio versus maximum amplitude.
[2] S. Rinaldi, S. Prabhakar, S. Vengallator and M.P.

Padoussis, Dynamics of microscale pipes containing
internal fluid flow: Damping, frequency shift, and
stability, J Sound Vib., Vol. 329, pp. 1081-1088, Jan.
2010.
[3] M. Whitby and N. Quirke, Fluid flow in carbon
nanotubes and nanopipes, Nat Nanotechnol., Vol. 2,
pp. 87-94, Sept. 2007.
41
[4] L. Wang, Vibration and instability analysis of tubular

nano- and micro-beams conveying fluid using nonlocal
elastic theory, Physica E ., Vol. 41, pp. 1835-1840,
Feb. 2009.
[5] R.E., Bellman, B.G., Kashef and J., Casti, Differential
quadrature: a technique for the rapid solution of nonlinear partial differential equations, Journal of
Computational Physics., Vol. 10, pp. 4052, Nov.
1972.
[6] G.B., Warburton, Vibration of thin cylindrical shells,
Journal of Mechanical Engineering Science., Vol. 7,
pp. 399407, Jan. 1969.
42
[7] C. Shu and B.E. Richard, Application of generalized

differential quadrature to solve two-dimensions
incompressible Navier-stress equations. International, Journal of Numerical Methods in Fluids.,
Vol. 15, pp. 791798, Nov. 1992.
[8] C. Shu, Y.T. Chew, B.E. Richards, Generalized
differential and integral quadrature and their
application to solve boundary layer equations.
International, Journal of Numerical Methods in
Fluids Vol. 21, pp. 723733, Oct. 1995.

S & T Review
Packed Bed Solar Energy Storage :

A State-of-Art Review
T. Alam*, R.P.Saini*, J.S.Saini1
hermal energy storage is one of the most efficient ways to store solar energy due to it intermittent nature. Packed bed is
generally recommended for thermal energy storage in solar air heaters. A packed bed is a volume of porous media
obtained by packing particles of selected material into a container tank through which the heat transport fluid is circulated.
There have been several studies on packed beds for their performance analysis. These studies include the design of packed
beds, materials used for storage, enhancement of the heat transfer and pressure drop in packed beds. The aim of the present
paper is to review the research work done so far on packed beds. It is concluded that materials other than rocks and pebbles
have been studied in a few studies only. The physical properties, shape and size of the materials are important factors that affect
the performance of the system, so they have to be taken into account for any performance study.
Keywords: Renewable Energy, Heat Storage, Packed Beds, Solar Energy
1. Introduction
Solar energy is a time dependent energy resource. The
energy from sun has intermittent nature, often unpredictable
and diffused; this makes the energy storage critically
important. The storage of energy of a solar process is
necessary if the solar energy is to meet substantial portions of
the energy needs and in order to have uninterrupted supply of
energy in absence of availability of solar energy. Thermal
energy storage is of particular interest and significance in
using this technique for solar thermal applications.
Thermal energy can be stored as sensible heat, latent heat or
chemical energy. In sensible heat storage, heat is stored by
increasing the storage medium temperature. In case of latent
heat storage systems, the energy is stored in phase change
materials .The heat is stored when the material changes
phase from solid to a liquid. Thermo chemical storage is a
technique, which involves chemical reactions. Sensible heat
storage is the most simple and cheap way of energy storage
although there are few advantages of phase change energy
storage over sensible heat storage, but the technological and
economical aspects make sensible heat storage superior.
Packed beds represent the most suitable storage units for airbased solar system.
A packed bed storage system consists of loosely packed solid
material through which the heat transport fluid is circulated.
* Alternate Hydro Energy Centre, Indian Institute of
Technology, Roorkee, India, E-mail: tabish.iitr@gmail.com
1. Mechanical & Industrial Engineering Department, Indian
Institute of Technology, Roorkee, India.
Heated fluid (usually air) flows from solar collectors into a

bed of graded particles from top to bottom in which thermal
energy is transferred during the charging phase as shown
(Fig.1).
Fig. 1 Schematic of packed bed

solar energy storage system
Recovery of this stored energy is usually achieved in

discharging mode by circulating air from bottom to top. The
high heat transfer coefficient between the air and solid causes
quick heat transfer from air to the solid. The particles near the
entrance are heated but the temperature near the exit remains
unchanged and the air comes out of the bed at a temperature
very close to the initial bed temperature. As time progresses a
temperature front passes through the bed and the exit air
temperature begins to rise. When the bed is fully charged its
temperature becomes uniform.
A packed bed in a solar heating system does not operate
normally with constant temperature. During the daytime
dependent conditions like solar radiations, ambient
temperature, collector inlet temperature, load requirements
etc. result in a variable collector outlet temperature. The
optimum size of the storage system depends on the particular
43
application concerned and is a function of several parameters

such as storage temperature, material, storage heat losses,
costs of the storage medium container, heat exchanger and
pumps, cost of auxiliary energy if any, weather data such as
insulations, ambient temperature, wind speed, sky
conditions etc, collector area, efficiency, solar fraction of the
total heat load etc.
2.Analytical Studies
The heat transfer to and from a flowing fluid to a packed bed
has been the subject of many theoretical and experimental
investigations since Schumann's original work. Schumann
(1929) presented one dimensional two phase model for
packed bed system by ignoring the thermal capacity of the
fluid, axial conduction in the fluid and axial conduction in
the bed material. Wyman et.al (1979), Dincer et.al. (1997),
Fath (1998), Hasnain (1998) and Atear (2006) and
conducted extensive studies on different energy storage
techniques and materials used in sensible heat storage
systems. They reported that the choice of material depends
largely on the temperature level of the application. Water is
reported to be the best sensible heat storage liquid because it
is inexpensive and has a high specific heat. However, above
0
100 C, the storage tank must be able to contain water at its
vapor pressure. The difficulties of vapor pressure of water
and the limitations of other liquids can be avoided by storing
thermal energy as sensible heat in solids. Storage in rock
beds is most popular, because the rocks are cheap and
available. Other solid materials including most metals can be
used for thermal energy storage when they are formed into
small balls or cylinders. Water at high pressure, molten salts
and liquid metals are used for sensible heat storage at
intermediate and high temperatures. . The use of metal media
may be advantageous where high thermal conductivity is
required and the cost is of secondary importance. Solid
industrial wastes like copper slag, iron slag, aluminum slag
and copper chips could be used as storage materials for
energy storage.
Riaz (1978) presented a simple model for packed bed, which
accounts for the fluid convective motion, the air-rock heat
transfer, axial bed conduction and internal particle
conduction. Based on a single-phase conductivity model of
packed bed, analytical solutions have been given in simple
closed forms for a variety of inlet temperature conditions.
Maaliiou et al. (1985) have presented a model for
optimization of design parameters of a packed bed. They
44
reported that the net income is the difference between the

economic value of stored heat and two cost factors i.e. the
capital cost and the operating costs. The dominating
operating cost is considered to be the pumping cost of
circulating the air, which depends on the pressure drop
through the bed.
Aly et al. (1990) concluded that one of the main sets of
parameters affecting the design of solid packed beds is the
physical properties of the solid phase used as storage
material. The steel packed bed exhibits a higher storage rate
and capacity relative to the rock bed. The aluminum bed, on
the other hand, shows a superior storage performance
compared with that of rock only during the first 6 hours of
charging. Choudhury et.al. (1995) conducted a theoretical
analysis for optimization of design and operational
parameters of a rock bed thermal energy storage device
coupled with a two-pass single cover solar air heater. They
have reported that rich literature is not available on the
design optimization of the rock bed thermal energy storage
that keeps the cost effectiveness aspects in view.
Ismail & Stuginsky (1999) reported that an increase in
particle diameter reduces pressure drop and increases
Nusselt number. The fluids with higher thermal capacities
are capable of transferring large amounts of energy .The
relevant thermo physical properties of storage material are
thermal capacity and the thermal conductivity. A reduction in
the void fraction leads to an increase in the mass of particles
present in the bed, an increase in thermal storage capacity of
the bed and pressure drop. Crandall et.al. (2004) reported
that the packed beds can have high degree of stratification
and this is a major advantage. Stratification in a rock bed
decreases during the later part of the day, with a decrease in
solar insolation and hence with a decrease in the collector
outlet temperature. They have reported that segmentation of
the bed can improve the stratification.
3. Experimental Work
Furnas (1930) conducted experimental study for heat
transfer from a fluid stream to a bed of broken solids and
concluded that the coefficient of heat transfer varies along a
straight-line w.r.t. gas velocity. The temperature of gas is
reported to have a little effect on heat transfer coefficient and
the degree of packing in a bed having a very large influence
on resistance to fluid flow. A considerable variation in heat
transfer for different materials was observed and the value of

S & T Review
heat transfer coefficient decreased with an increase in

particle diameter.
higher value than the fraction actually available to the

flowing air.
Colburn (1931) conducted an experimental study on heat

transfer between air flowing through a filled tube with
granular materials and presented correlations for heat
transfer coefficient and the heat transfer coefficient was
found to be dependent on the mass flow rate of air and
particle to tube diameter ratio.
Waked (1986) concluded that the selection of the storage

material must be based on different criteria, including,
material properties, type of application, available space and
cost etc. The author reported that the rock could be
successfully used as a storage material. It is well stratified
and more than 60% of the energy stored can be recovered at
almost the maximum storage temperature. Sorour (1988)
concluded that the dimensions of the storage bed limit the
storage efficiency in a pebble bed. Higher flow rate and /or
very small particle dimension produce lower efficiency than
with lower flow rate and with intermediate particle diameter
for small size storage capacity.
Lof et.al. (1948) determined the heat transfer between air and
loose solids in an experimental study. They concluded that
the heat transfer coefficient increases along a straight line
with an increase in airflow rate. However, it decreased along
a straight line with an increase in the particle diameter (size
of element). Littman et.al. (1968) conducted gas particle heat
transfer study in packed beds in low range of Reynolds
number. Dynamic thermal conductivities of the solid phase
can be measured and interpreted by the methods given in this
paper.
Standish et.al. (1970) reported that the shape of packing is a
significant variable in flooding. Besides size distribution, the
particle shape is most likely factor to affect the packing
structure and its properties in randomly packed beds.
Chandra et.al. (1981) in experimental studies found the
pressure drop to depend on rock size, bed porosity and
airflow rate. Coefficient of heat transfer was found to depend
on rock size and flow rate only. No influences of inlet air
temperature or initial rock bed temperature on coefficient of
heat transfer was found.
Courtier et.al. (1982) concluded that the rock beds
represented the most suitable storage units for air-based solar
system.. An excellent agreement is reported to exist between
experimental and theoretical curves obtained for expression
of convective heat transfer coefficient. Beasley et.al. (1984)
shown that the spatial variations in the void fraction have a
significant influence on the dynamic response of both fluid
and solid temperature. A numerical model has been
developed that predicts the 2-d transient response of both
fluid and solid phases.
Hollands et.al. (1984) concluded that in the case of the
unwashed rocks, the fines (very small particles including
dust) substantially increase the pressure drop across a
particulate bed. The fines may form on rocks a coating,
which would have apparent density lower than rocks, so a
void fraction determined by mass measurement may give a
Ammar et.al. (1991) reported that the heat storage in Tafla is

slightly more than that for rock bed, a behavior that can be
attributed to the relatively higher value of thermal
conductivity and specific heat of Tafla. The particles with
smaller diameter i.e. with higher values of interphase surface
area per unit volume cause a large degree of stratification in
the bed. On contrary particles with large diameter degrade
stratification. Audi (1992) tested small sized rocks for the
stability analysis and their possible use as storage materials.
Some of tested rocks such as Tarsand and Zeolite
disintegrated under the operational environment. However,
Jordian Basalt and Limestone demonstrated excellent
stability during the tests and their storage properties were
acceptable.
Sagara et.al. (1991) mentioned that in case of large sized
materials temperature gradient inside the solid couldn't be
ignored. It was found that large size materials like brick &
concrete blocks (Fig 2) have poor thermal performance but
require less power supply to run the fans. They reported that
the large size materials have almost the same thermal
performance as small size materials in a solar heating system
with a heat pump. If the bed is longer, the difference of fan
energy between large and small materials becomes greater
and then in that case, large size materials may be more
favorable as storage materials. They mentioned that
economic evaluation might become a decisive factor for the
ability to utilize large size materials.
45
coefficient has been done and correlations have been given

for Nusselt number. Ozturk et.al. (2003) investigated solar
energy storage in a green house experimentally. They
reported that the solid materials are economically more
attractive for high temperature heat storage than fluids and
their volume requirements are nearly comparable. The
charging and discharging process of a thermal energy storage
system must be analyzed in order to optimize the system
efficiency.
Fig. 2 Material elements used by
Sagara & Nakahara (1991)
Kassaby et.al. (1993) analyzed natural soil available at
Mu'tah site (Jordan) for air based system and water for water
based system in the sensible heat storage (water and air based
storage) systems to study the variations in the amount of
energy stored with respect to time. They have reported that
the use of soil as storage material instead of water is possible,
but additional collector area must be provided. The use of
water system is superior from the heat capacity point of view,
but there are problems in maintaining the system.
Nimr et.al. (1996) reported that earlier models proposed on
packed beds had a number of assumptions and the inlet fluid
temperature was not varied with respect to time, but a packed
bed system receives the energy during charging from a heat
source, which provides it with a variable inlet fluid
temperature. They used rocks as storage material for sensible
heat storage with water as heat transfer fluid. The model
presented here, predicts the temperature distribution in the
column for a variable inlet fluid tempts. Experimental results
have been compared to satisfaction with the results of the
model.
Jalalzadeh et.al. (1997) reported that previous research on
packed bed has been generally concerned with the thermal
performance of the storage materials but has largely ignored
the stability aspects, which are particularly important for
high temperature applications. The salt/ceramic composite
pellets showed poor physical stability while the performance
of zirconium materials was found to be excellent. The ZrO2
medium offered a superior performance over the composite
materials on an equal volume basis. This was attributed to the
high thermal capacity per unit volume of zirconium material.
Nsofor et.al. (2001) conducted an experimental study for
high temperature sensible heat storage. An experimental
investigation of forced convection gas particle heat transfer
46
Singh et.al. (2006) reported that the effects of particle shape

and bed porosity are required to be taken into account while
developing correlations. The correlations have been
proposed for Nusselt number and friction factor as functions
of Reynolds number, Void fraction and sphericity for the
shapes shown (Fig 3). They have compared the obtained
results with earlier studies and have reported the validity of
these correlations to satisfaction.
----------------------
Fig. 3 Material elements used by Singh et.al. (2006)

Nallusamy et.al (2007) analyzed combined latent and
sensible heat storage with water as heat transferring fluid. In
case of constant heat transfer fluid temperature the mass flow
had only a small effect on the rate of charging and the rate of
heat transfer increases in direct proportion with the increase
in inlet temperature of the heat transferring fluid
temperature. The mass flow rate was reported to have a
significant effect on the heat extraction rate. It is concluded
that combined storage system gives better performance than
the conventional sensible heat storage system when there is a
direct mixing of the heat transfer fluid with the storage
material.

S & T Review
Mawire et.al. (2011) presented six different parameters

namely temperature distribution along the height of the
storage tank at different time intervals, the charging energy
efficiency, the charging exergy efficiency, the stratification
number, Reynolds number and Richardson number that
characterize thermal stratification in an oil/pebble-bed TES
system. These parameters are evaluated under six different
experimental charging conditions. The charging exergy
efficiency and Reynolds number give important information
about describing thermal stratification qualitatively and
should be used with care. The charging energy efficiency and
Richardson number have no clear relationship with thermal
stratification.
4. Conclusions
Based on the literature survey of packed bed heat storage
devices conducted under present paper it is concluded that
the packed beds have been widely investigated both
analytically and experimentally. There have been various
analytical studies done on the effects of various parameters
on the performance of packed bed. There is not much
literature available on the optimization on the basis of the
cost effectiveness. Different investigators have used
different materials in the experimental work and given
various correlations for heat transfer and friction factor
(Table 1). The large storage materials are reported to reduce
the fan power considerably, but there are only few studies
reported on the large sized storage materials. Overall, it is
concluded that materials other than rocks and pebbles have
been studied in a few studies only. The physical property of
the solid phase used as storage material is particularly
important. The shape and size of the materials are the other
important factors that affect the performance of the system
and have to be taken into account for any performance study.
The void fraction or the volume of the storage material filled
in the packed bed should also be considered for the heat
transfer and pressure drop characteristics of the packed beds.
References
[1] A.S.A.Ammar & A.A. Ghoneim, Optimization of a
sensible heat storage unit packed with spheres of a
local market, Renewable Energy, Vol.1, No.1 (9195), 1991.
[2] Adnan M. Waked, Solar energy storage in rocks,
Solar & Wind technology Vol.3, No.1 (27-31), 1986.
[3] Ali A.jalalzadeh-Azar, W.Glenn Steele & George

A.Adebiyi, Performance comparison of high
temperature packed bed operation with pcm and
sensible heat pellets, International Journal of Energy
Research, Vol.21 (1039-1052), 1997.
[4] Allan P. Colburn, Heat transfer and pressure drop in
empty baffled, and packed tubes, Industrial
Engineering Chemistry J. Vol.23, No.8 (910-913),
1931.
[5] C.Choudhury, P.M.Chauhan & H.P.Garg, Economic
design of a rock bed storage device for storing solar
thermal energy, Solar Energy Vol.55, No.1 (29-37),
1995.
[6] C.C. Furnas, Heat transfer from a gas stream to a bed
of broken solids, Industrial and Engineering
Chemistry J. (721-731), 1930.
[7] Charles Wyman, James Castle & Frank Kreith, A
review of collector and energy storage technology for
intermediate temperature applications, Solar Energy
Vol.24 (517-540), 1979.
[8] D.M.Crandall & E.F.Thacher, Segmented thermal
storage, Solar Energy 77 (435-440), 2004.
[9] Donald E. Beasley & John A. Clark, Transient
response of a packed bed thermal energy storage,
Int.J.Heat Mass transfer Vol.27, No.9 (1659-1669),
1984.
[10] E.C.Nsofer & George A.Adebiyi, Measurements of
the gas particle convective heat transfer coefficient in a
packed bed for high temperature energy storage,
Experimental Thermal and Fluid Science24 (1-9),
2001.
[11] G.O.C.Lof & R.W.Hawley, Unsteady state Heat
transfer between air and loose solids, Industrial
engineering Chemistry J. Vol.40, No.6 (1061-1071),
1948.
[12] H.H.Ozturk & A. Bascetincelik, Energy and exergy
efficiency of a packed bed heat storage unit for
greenhouse heating, Biosystems Engineering, Vol.86,
No.2 (231-245), 2003.
[13] Hassan E.Fath, Technical assessment of solar thermal
energy storage technologies, Renewable Energy,
47
Vol.14, Nos.1-4 (35-40), 1998.

[14] Howard Littman, Ronald G. Barie & Allen H. Pulsifer,
Gas particle heat transfer coefficients in packed beds
at low Reynolds numbers, I&EC Fundamentals
J.Vol.7, No.4 (554-561), 1968.
[15] Ibrahim Dincer, Sadik Dost & Xianguo Li,
Performance analysis of sensible heat storage systems
for thermal applications, International Journal of
Energy Research.Vol.21 (1257-1171), 1997.
[16] J.Pascal Coutier & E.A.Faber, Two applications of a
numerical approach of heat transfer process within rock
beds, Solar Energy, Vol29, No.6 (451-462), 1982.
[17] K.A.R Ismail & R.Stuginsky, A parametric study on
possible fixed bed models for pcm and sensible heat
storage, Applied Thermal Engineering, Vol.19, No.7
(757-788), 1999.
[18] K.G.T.Hollands & H.F.Sullivan, Pressure drop across
rock bed thermal storage systems, Solar Energy,
Vol.33, No.2 (221-225), 1984.
[19] K.Sagara & N.Nakahara, Thermal performance and
pressure drop of rock beds with large storage
materials, Solar Energy, Vol.47, No.3 (157-163),
1991.
[20] M.Riaz, Transient analysis of packed bed thermal
storage systems, Solar Energy Vol.21 (123-128),
1978.
[21] M.A.Al-Nimr, M.K.Abu-Qudais & M.D.Mashaqi,
Dynamic behavior of a packed bed energy storage
system, Energy Convers.Mgmt.Vol.37, No.1 (23-30),
1996.
[22] A. Mawire, and S. H. Taole, A comparison of
experimental thermal stratification parameters for an
oil/pebble-bed thermal energy storage (TES) system
during charging, Applied Energy 2011; Vol 88: p-p
4766-4778.
[23] M.M.El-Kassaby & A.A.Ghoneim, Comparison of
measured and predicted performance of different heat
storage systems, Renewable Energy, Vol.3, No.8 (849856), 1993.
48
[24] M.M.Sorour, Performance of small sensible heat

energy storage unit, Energy Convers.Mgmt.J.Vol.28,
No.3 (211-217), 1988.
[25] Mahmoud S. Audi, Experimental study of a solar
space heating model using jordian rocks for storage,
Energy Convers.Mgmt.J. Vol.33, No.9 (833-842),
1992.
[26] Meier, A., Winkler, C. and Wuillemin, D., Experiment
for modelling high temperature rock bed storage,
Solar Energy Materials 1991;Vol. 24: 255-264.
[27] N.Nallusamy, S.Sampath & R.Velraj, Experimental
investigation on a combined sensible and latent heat
storage system integrated with constant/varying (solar)
heat source, Renewable Energy 32 (1206-1227) 2007.
[28] N.Standish & J.B.Drinkwater, The effect of particle
shape on flooding rates in packed columns, Chemical
Engineering Science J. Vol.25 (1919-1621), 1970.
[29] O.Ercan Atear, Storage of thermal energy,
Encyclopedia of life support systems, 2006.
[30] O.Maaliou & B.J.Mccoy, Optimization of thermal
energy storage in packed columns, Solar Energy,
Vol.34, No.1 (35-41), 1985.
[31] Pitam Chandra & D.H.Willits, Pressure drop and heat
transfer characteristics of air rock bed thermal storage
systems, Solar Energy Vol.26, No.6 (547-553), 1981.
[32] Ranjit Singh, R.P.Saini & J.S.Saini, Nusselt number
and friction factor correlations for packed bed solar
energy storage system having large sized elements of
different shapes, Solar Energy 80 (760-771) 2006.
[33] S.L.Aly & A.I.El-Sharkawy, Effect of storage medium
on thermal properties of packed bed, heat recovery
Systems & CHPVol.10, No.5/6 (509-517) 1990.
[34] S.M.Hasnain, Review on sustainable thermal energy
storage technologies, part 1: heat storage materials and
techniques Energy Convers. Mgmt. J. Vol.39, No.11
(1127-1138), 1998.
[35] T.E.W.Schumann, Heat transfer: a liquid flowing
through a porous prism, Heat Transfer (405-416),
1929.

S & T Review
Friction Stir Welding of Aluminium

Subham Kumar*, Vipin Kumar*,
Gajendra Singh*, Mohd. Zaheer Khan Yusufzai*
riction stir welding is a solid state welding process that was developed and patented by TWI (The Welding Institute, UK).
Tremendous advantages are built in this new welding process; superior mechanical properties, environmental benefits,
absence of the grain growth region of the HAZ (heat affected zone), easy automation are a few of these benefits. Excessive
work is being carried out world over in this area. Friction stir welding of commercial grade of aluminium was tried out and butt
joint could be successfully produced. Two plates of 3mm thickness and 25mm width were welded. The success of friction stir
welding depends on being able to properly hold the plates being welded together and not allowing them to separate during
welding. For proper clamping of plates a fixture was designed and fabricated. A non-consumable tool is required for friction
stir welding. A specially designed tool of HSS (high speed steel) was used for this purpose. Welding experimental runs were
performed on a robust vertical milling machine with spindle motor of 4 kW. The welds produced were having nice appearance
and there were no visible defects. As with most friction stir welding welds, the welded joint did not require any finishing
operation. The cross section of the weld revealed complete microscopic bonding between the plates. There was no harm done
to the welding tool and it remained suitable for many more welding runs.
Keywords : Friction Stir Welding, Welding of Aluminium, Defect Free Welds
1. Introduction
2. Process Parameters in FSW
Friction Stir Welding (FSW) is a solid-state joining process

in which metal is not melted during the process. In FSW, a
cylindrical-shouldered tool, with a profiled
threaded/unthreaded probe (nib or pin) is rotated at a
constant speed and fed at a constant traverse rate into the
joint line between two pieces of sheet or plate material,
which are butted together.
There can be many factors or variables that can be varied in

FSW and can affect the output response. The
Friction heat is generated between the wear-resistant

welding tool shoulder and nib, and the material of the work
pieces. This heat, along with the heat generated by the
mechanical mixing process and the heat within the material,
cause the stirred materials to soften without reaching the
melting point.
FSW can be used to produce butt, lap, spot, fillet and corner
joints as well as to weld hollow objects, such as tanks and
tubes/pipes, stock with different thicknesses, tapered
sections and parts with three dimensional contours. This
welding process is particularly suited for butt and lap joining
of aluminium alloys which are otherwise quite difficult to
join using conventional arc/fusion welding processes.
* Mechanical Engineering Department , Institute of Technology,
Banaras Hindu University, Varanasi, India
E-mail: subham2009@gmail.com, E-mail: vipinmbd@gmail.com,
E-mail: gajendraitbhu10@gmail.com, E-mial: mzkhan.y@gmail.com
primary variables are tool shape and size, tool rotational

velocity, welding speed, tool tilt and plunge, all of these
variables can affect weld properties significantly.
Tool size and shape is most important variable, to get defect
free welds. Tool should have minimum pin diameter but high
frictional surface to limit stirring area. It should have
sufficient size to provide adequate heating of stirred area and
have resistance against failure by shear. Shoulder should
have adequate size to retain the hot worked material and to
compress it on the rear side. Smaller size of shoulder
diameter can cause loss of material in form of flash and
defective welds. On the other side excessively large shoulder
diameter causes heavy loading on equipment.
Tool RPM and welding speed are other important variables
since tool RPM can be directly related to frictional heat
generated and welding speed sets the upper limit of tool
capacity and equipment. Welding speed and tool RPM also
decide whether enough heat input is being supplied to weld
so as to favourably affect metallurgical and mechanical
properties of the resultant weld.
Plunge depth (the amount by which the shoulder is inserted
49
beneath the work-piece surface) also has a significant impact

on weld quality. Shallow penetration or plunging causes
insufficient compressive force by shoulder to forge the hot
worked material. Excessive plunging may cause material
expulsion from joint line and may result in a defective weld.
Tool tilt is an important parameter as a suitable tilt of tool
toward trailing direction ensures that shoulder holds down
the hot worked material, moves it from front to rear and
compresses it. It is known that 1 to 3 of tool tilt can be used
to give defect free sound welds. Generally the value of this
variable is decided from experience.
4.4 Calculations for aluminium (to be welded)

Table 1 Calculation for process forces and power
Material To Be Welded
Aluminium
3.Advantages of the Process
tool shoulder diameter
D mm
12
Friction stir welding being a solid state welding process is

free of most of the defects usually produced in fusion
welding processes (because of considerably lower heat input
and absence of melting). Apart from this, it has many
advantages like improved mechanical properties of
weldment, grain refining, and improved corrosion
resistance, improved productivity and no requirement of
filler material
pin base diameter
d mm
plate thickness
t mm
tool r.p.m.
640
torque needed
T Nm
18.584
shear yield strength
SY MPa
30
tensile yield strength
Y MPa
90
spindle power needed
Ps kW
1.243
force on tool in x-direction
FX kN
4.43
force on tool in z-direction
FZ kN
10.18
coefficient of friction
0.40
feed
F mm/min 13
4. Experimental Set-Up Development

4.1 Experimental set-up
The experimental setup consisted of work material, the
robust FSW machine, fixture, tool holder, tool, clamping
arrangements, backing plate, nut bolts and the pins.
4.2 Selection of work material
Aluminium plates were chosen for the experiments. During
experiment several weld beads were made on the welding
plates. After gaining some confidence by trial runs then we
planned for butt welding of 4 mm thick aluminium plates.
4.3 Process load calculations
Based on the calculations done above it was decided to select

the machine which should have the following capabilities.
(i)
The machine should be capable of weldingAluminium,
(ii) Should be rigid enough to withstand axial thrust..

(iii) The main spindle motor of 1.243 KW (so that plates of
8-10 mm thickness could be welded)
(iv) The longitudinal feed should be through a motor and
that motor should be of 1.2KW.
As per the calculations made it was found that vertical
milling machines may be used for FSW. The vertical milling
machine available at IT-BHU which was satisfying the
criterion selected.
50

S & T Review
4.5 Universal combination milling machine

The Universal Combination Milling Machine available at
C.I.S. & F. lab of IT-BHU was matching closely with the
requirement and was selected for the experimentations.
Specifications
The specifications of the machine are given
it was necessary to clamp the work pieces firmly. The

existing vertical milling machine was not having any
provision for clamping the work piece as required for
carrying out the butt welding process. In order to overcome
this problem, a dedicated fixture was designed and fabricated
to enable appropriate clamping of the work piece on the
vertical milling machine. The Fig 4.3A shows the picture of
the dedicated fixture which has been developed.
Technical data:
Overall size: 1600 mm x 1020 mm
Clamping area: 1260 mm x 1020 mm
No of T-slots: 4
Distance between T-Slots: 34 mm & 44 mm
Movements
Longitudinal: 450 mm
Transverse: 300 mm
Vertical: 375 mm
Speeds
No of speeds: 32
Range: 50 2040 rpm
Fig. 2 Design of work fixture
Feeds
No of feeds: 21
Range 13-395 mm/min
Power
Cutter spindle motor, Feed motor: 2.8 kW
Coolant pump motor (approx): 0.3 KW
Fig. 3 Setup work fixture
Fig. 1 Vertical Milling Machine

[C.I.S. & F. Lab Shop IT-BHU]
4.6 Design and fabrication of work fixture
For performing the friction stir welding process successfully
The work-pieces for conducting the experiments are cut on

planner machine. Therefore, there is a possibility of having
variation in the width of the plates/work-pieces, which can
create problem at the time of clamping on the fixture. To
cater for such situation, one of the inner side face of the
fixture has been made inclined. Further, to prevent the
collision between tool and clamps of the fixture, the clamps
are machined to a special shape.
51
Fig. 4 Tool Holder Arrangements

In FSW process, the tool gets heated as well. This heat should
not affect the vertical milling machine or the chuck holding
it. So to avoid this, a tool holder arrangement has been
designed and developed. The tool holder consists of a central
vertical shaft which is hollow at the bottom side and threaded
on the outer side so that the collet nut is fastened. The hollow
space provides accommodation for the collet. The tool is
inserted in the collet and tightened with the collet nut. The
tool holder acts as a heat sink and does not allow heat to be
transmitted to the vertical milling machine. The FSW tool
has been fit in the tool holder's chuck which in-turn is fixed to
the chuck of the vertical milling machine. The developed
tool holder arrangement is shown in Fig 4.
5. Selection of Tool Material
In friction stir welding the tool plays an important role. The
process cannot be a successful one without a nonconsumable tool. The tool material is determined based on
the base metal. For materials with low softening temperature
like Aluminium, Tool steel can be used, where as metals like
steel, which has high softening temperature, tungsten
carbide and PCBN (poly crystalline boron nitrate) is used.
The tool grade is selected based on some basic requirements
such as, the tool shall withstand high temperature, should
have good wear resistant properties, should have enough
strength, etc.
On the basis of literature and ease of availability, high speed
steel was selected as the tool material. Because It can
0
withstand temperatures of the order of 1538 C (which is
sufficient for FSW of aluminium) and has high hardness and
wear resistance. high speed steel was bought from a
commercial source to be used as tool material.
52
Fig. 5 Complete experimental set-up at

C.I.S. & F. Lab Shop IT-BHU
6. Welding of Aluminium using FSW Process
When the process is in developmental stage, we are merely
interested in finding mechanical properties and the effect of
mechanical properties. These results can also easily obtained
if instead of bead being deposited on two plates butted
together, a weld bead is deposited on a single piece of
material. This saves the setup time that includes clamping the
plates and aligning them. But the results obtained here can be
accurately applied to welding of two separate plates. This
technique has been widely accepted as a standard technique
by various researches worldwide.
Initially trials were taken on 3 mm thick aluminum plates.
Some initial trial runs were taken to fix the range of process
parameter. The levels of process parameters were selected
based on the initial trial runs, because at the starting of the
experiments it is not known to us the exact parameters in
which welding is to be done for mild steel, as very less
literature were available. The welding parameters such as pin
shape and pin size, shoulder diameter, welding speed, tool
rpm, tool tilt those are set one by one for finalizing them.
The plate in which the FSW is to be made is placed over the
worktable of the vertical milling machine with backing plate
of suitable dimension. The work pieces along with the
backing plate should be rigidly fixed. Care has been taken
while clamping the work piece as it will be subjected to very
high torque and forces as mentioned in Table 1. After
clamping the work piece, the initial parameters are set in the
vertical milling machine. The machine is started and the
worktable is raised so that the rotating tool touches the work
piece. It is raised further in such a way that the pin enters the

S & T Review
work piece and the shoulder slides over the work piece to
generate the sufficient frictional heat required for the
process. Then the table is given the transverse feed, as a
result the friction stir weld is obtained. The same procedure is
repeated for various traverse feeds.
Figure 6 shows the Experimental process of friction stir
welding of aluminum at IT-BHU lab.
8. Conclusion
This paper has discussed FSW of aluminium. The process
parameter should be selected carefully. On selection of
proper process parameters, the elemental composition of the
weld region was found to be same as the base metal which
was used for friction stir welding.
Acknowledgement
We have a great pleasure in expressing our deep sense of
gratitude and indebtedness to Assist. Prof. Mohd. Zaheer
Khan Yusufzai, Department of Mechanical Engineering, ITBHU. We would like to thank him for his esteemed guidance,
encouragement and valuable advice during this paper work.
References
[1] R. S. Mishra and Z. Y. Ma, "Friction stir welding and
processing," Materials Science & Engineering RReports, vol. 50, pp. 1-78, 2005.
Fig. 6 friction stir welding of aluminum

7. Results and Discussions
Visual examination of the plate was carried out after the
welding was completed. When the plate was observed from
the top, good ripples were seen throughout the joint which
were evenly spaced. No excessive flash was observed and
there was firm joining of the two metal plates. A keyhole was
observed at the end of the specimen which was unavoidable
and was formed as a result of the exit of the pin from the
specimen.
[2] Z. H. Fu, D. Q. He, and H. Wang, "Friction stir welding

of aluminum alloys," Journal of Wuhan University of
Technology-Materials Science Edition, vol. 19, pp. 6164, 2004.
[3] H. J. Liu, H. Fujii, M. Masakatsu, and K. Nogi,
"Friction stir welding characteristics of two aluminum
alloys," Transactions of Nonferrous Metals Society of
China, vol. 13, pp. 1108-1111, 2003.
[4] A. C. Somasekharan and L. E. Murr, "Microstructures
in friction-stir welded dissimilar magnesium alloys and
magnesium alloys to 6061-T6 aluminum alloy,"
Materials Characterization, vol. 52, pp. 49- 64, 2004.
[5] W. B. Lee, J. W. Kim, Y. M. Yeon, and S. B. Jung, "The
joint characteristics of friction stir welded AZ91D
magnesium alloy," Materials Transactions, vol. 44, pp.
917-923, 2003.
[6] W.-B. Lee and S.-B. Jung, "The joint properties of
copper by friction stir welding," Materials Letters, vol.
58, pp. 1041-1046, 2004.
Figure 7 Weld obtained after welding.

During the visual examination of the bottom or the other side
of the plate, a through welding of the plate was observed.
There were no specific defects observed in the specimen
[7] Y. Hovanski, M. L. Santella, and G. J. Grant, "Friction

stir spot welding of hot-stamped boron steel," Scripta
Materialia, vol. 57, pp. 873-876, 2007.
[8] T. Watanabe, H. Takayama, and A. Yanagisawa,
"Joining of aluminum alloy to steel by friction stir
53
welding," Journal of Materials Processing Technology,

vol. 178, pp. 342-349, 2006.
[9] K. Kimapong and T. Watanabe, "Lap joint of A5083
aluminum alloy and SS400 steel by Friction Stir
Welding," Materials Transactions, vol. 46, pp. 835-841,
2005.
[10] T. Ishikawa, H. Fujii, K. Genchi, S. Iwaki, S. Matsuoka,
and K. Nogi, "High Quality-High Speed Friction Stir
Welding of 304 Austenitic Stainless Steel," Tetsu to
Hagane-Journal of the Iron and Steel Institute of Japan,
vol. 94, pp. 539-544, 2008.
[11] W. M. Thomas, H. G. Pisarski, I. M. Norris, D. J. Marks,
and J. C. Godden, "An investigation of through-hole
impact testing of weld root imperfections in friction stir
54
welded 12 per cent chromium alloy steel," Proceedings

of the Institution of Mechanical Engineers Part BJournal of Engineering Manufacture, vol. 222, pp.
1043-1054, 2008.
[12] T. Saeid, A. Abdollah-zadeh, H. Assadi, and F. M.
Ghaini, "Effect of friction stir welding speed on the
microstructure and mechanical properties of a duplex
stainless steel," Materials Science and Engineering A,
vol. 496, pp. 262-268, 2008.
[13] Y. S. Sato, P.Arkom, H. Kokawa, T. W. Nelson, and R. J.
Steel, "Effect of microstructure on properties of friction
stir welded Inconel Alloy 600," Materials Science and
EngineeringA, vol. 477, pp. 250-258, 2008.

S & T Review
Performance and Emission Analysis of a Multi-Cylinder

C.I. Engine Fuelled with Diesel-CNG Dual Fuel
*
Rajesh Kumar , Manish Kumar Chauhan ,

*
1
Naveen Sharma and Amit Pal
he smoke is the main problem in the diesel engines. One of the main reasons of increased level of air pollution is
transportation vehicles. Diesel engine produces harmful gases like carbon monoxides, nitrogen oxides, unburnt
hydrocarbon, smoke, soot and particulate matter which are harmful to the environment and human kind [1]. The
environmental pollution can be controlled by utilization of alternative fuels such as Biodiesel and CNG (Compressed Natural
Gas). In this paper the performance testing of Biodiesel and diesel-CNG dual fuel is done on CI engine and the results were
compared with performance of pure diesel. The work was carried out on a 44.5 kW four-stroke, four-cylinder, Tata Indica
diesel engine which was converted into dual fuel mode by substituting the CNG in the intake of suction pipe with the help of air
mixer. Engine was operated in dual-fuel mode with substitution of diesel with CNG. The performance and emission
characteristic of the engine were compared with pure diesel, different blends of Biodiesel (B10, B20 and B30) and diesel-CNG
dual fuel. Knock levels were also analysed in both (pure diesel and diesel-CNG dual fuel) cases by recording the sound
produced by the engine in both cases. These multimedia sounds are then converted to graphical form by using MATLAB
software.
Keywords : Diesel Engine, CNG-Air Gas Mixer, CNG Kit, Smoke Meter, Gas Analyzer
1. Introduction
2. Dual-Fuel Engine Technology
Diesel engines breathe only air, blow by gases from the

crankcase (consisting primarily of air and HC) is rather low.
Due to its low volatility, evaporative emissions from the fuel
tank can also be ignored. However, the low concentration of
CO and unburnt HC in the diesel exhaust are compensated
for by high concentration of NOx. There are smoke particles
and oxygenated HC, including aldehydes and odourproducing compounds which have high nuisance value [2].
NOx emissions, while a heightening of the cetane number

lowers the smoke emission. The dual fuel engine is a diesel
engine that operates on gaseous fuels while maintaining
some liquid fuel injection to provide a deliberate source for
ignition. Such a system attempts usually to minimize the use
of the diesel fuel by its replacement with various gaseous
fuels and their mixtures while maintaining satisfactory
engine performance. There are some problems associated
with the conversion of a conventional diesel engine to dual
fuel operation. At light load, the dual fuel engine tends to
exhibit inferior fuel utilization and power production
efficiencies with higher unburned gaseous fuel and carbon
monoxide emissions, relative to the corresponding diesel
performance. Operation at light load is also associated with a
greater degree of cyclic variations in performance
parameters, such as peak cylinder pressure, torque, and
ignition delay, which have narrowed the effective working
range for dual fuel applications in the past. These trends arise
mainly as a result of the poor flame propagation
characteristics within the very lean gaseous fuel-air mixtures
and originating from the various ignition centers of the pilot.
Dual-Fuel is fitted onto a standard diesel engine which
operates unchanged, except power is generated by mostly
clean natural gas. A measured quantity of natural gas is
Smoke from diesel engines comes in three different hueswhite smoke emitted during cold start idling and at low
loads; blue smoke from the burning of lubricating oil and
additives; and black smoke, a result of incomplete
combustion. Black smoke consists of irregular shaped
agglomerated fine soot/particulates, the formation of which
depends on injector nozzle parameter and type of
combustion chamber (direct or indirect injection). It is a
particular problem with engines that are not well tuned [2].
Impact of fuel quality on emissions Pollution control
depends heavily on the quality of the fuel. In diesel vehicles a
higher density causes higher smoke, CO and
* Department of Mechanical and Industrial Engineering, Indian Institute of
Technology, Roorkee, India, E-mail- barman_rk44@yahoo.co.in,
E-mail: manishku.25@gmail.com, E-mail: sharma.naveen28@yahoo.com
1. Department of Mechanical Engineering, Delhi Technological
University, Delhi, India, E-mail: amitpal1@yahoo.com
55
mixed with the air just before it enters the cylinder and
compressed to the same levels as the diesel engine to
maintain efficiency [3].
fuel tank, manometer, fuel measuring unit, transmitters for

air and fuel flow measurements, process indicator and engine
indicator.
3. CNG asAlternative Fuel
Rotameters are provided for cooling water and calorimeter

water flow measurement. The setup enables study of engine
performance for brake power, indicated power, frictional
power, BMEP, IMEP, brake thermal efficiency, indicated
thermal efficiency, Mechanical efficiency, volumetric
efficiency, specific fuel consumption, A/F ratio and heat
balance. Windows based Engine performance analysis
software package Engine soft is provided for on line
performance evaluation. Emissions are measured with the
help of smoke meter and gas analyser.
The octane rating of natural gas is about 130, meaning that

engines could operate at compression ratio of up to 16:1
without knock or detonation [4]. In the CNG application
policy, many of the automotive makers already built
transportation with a natural gas fuelling system and
consumer does not have to pay for the cost of conversion kits
and required accessories.
Most importantly, natural gas significantly reduces CO2
emissions by 20-25% compare to gasoline because simple
chemical structures of natural gas (primarily methane-CH4)
contain one Carbon compare to diesel (C15H32) and gasoline
(C8H18)[5].
Natural gas composition varies considerably over time and
from location to location. Methane content is typically 7090% with the reminder primarily ethane, propane and carbon
dioxide [6]. At atmospheric pressure and temperature,
natural gas exists as a gas and has low density. Since the
volumetric energy density is so low, natural gas is often
stored in a compressed state at high pressure stored in
pressure vessels.
That natural gas has a high octane rating, for pure methane
the RON = 130 and enabling a dedicated engine to use a
higher compression ratio to improve thermal efficiency by
about 10% above that for a petrol engine, although it has been
suggested that optimized CNG engine should be up to 20%
more efficient, although this has yet to be demonstrated.
Compressed natural gas therefore can be easily employed in
spark ignited internal combustion engines [7].
4. Engine Test Setup
Fig. 1 Experimental setup
4.1 Engine modification

Dual-fuel engine refers to diesel engine operating on a
mixture of natural gas and diesel fuel. Natural gas has a low
cetane rating and is not therefore suited to compression
ignition, but if a pilot injection of diesel occurs within the
gas/air mixture, normal ignition can be initiated. Between 50
and 75% of usual diesel consumption can be replaced by gas
when operating in this mode [9]. The engine can also revert
to 100% diesel operation.
The setup consists of four cylinders, four strokes, Diesel

engine connected to eddy current type dynamometer for
loading. It is provided with necessary instruments for
combustion pressure and crank-angle measurements. These
signals are interfaced to computer through engine indicator
for P- and P-V diagrams. Provision is also made for
interfacing airflow, fuel flow, temperatures and load
measurement.
The set up has stand-alone panel box consisting of air box,
56
Fig. 3 Attachment of Air-CNG gas mixer on diesel engine

S & T Review
To use the diesel-CNG in dual mode we have introduced

CNG from a cylinder of 60 litres capacity shown in figure 3
in which CNG is stored at a high temperature about 200 bar.
A pressure regulator CNG kit shown in figure 4 is used to
convert the pressure of CNG from 200 bar to 4 bar.
Diesel is first ignited as primary fuel when the combustion

takes place this help to ignite our mixture of air-CNG. This
method is one of the easy and cheep conversions of the diesel
engine into diesel-CNG engine. The method of necessary
modification made to convert diesel engine into diesel-CNG
dual mode is shown in Figure 3. The AVL smoke meter and
gas analyser is used to measure the emissions.
Fig. 4 Gas-air Mixer
A CNG flowmeter is used to check the flow of CNG. CNG is

then supplied to the CI engine with the help of a CNG-air
mixer shown in figure 6 through the suction pipe. This
mixture of air and CNG is compressed in compression
stroke. Here we have not used spark plug to ignite the
mixture of air and CNG, but diesel is use as a pilot fuel at the
end of the compression stroke.
Fig. 6: CNG kit

5. Results and Discussion
5.1 Performance parameters
The performance parameters are used to judge that how
efficiently the engine used input energy or how efficiently it
provides the useful energy. The performance parameters of
this test are discussed below
5.1.1 Torque with speed
Fig. 7(a) Comparison of Torque Vs Speed for different fuels
5.1.2 Break power with speed

Fig. 5 CNG cylinders
Figure 7(a) shows the variation of Torque with speed for pure
diesel, different blends of Biodiesel and diesel-CNG dual
fuel mode. In all cases initially the Torque rises sharply with
57
increase in engine speed upto 3000 rpm. The variation of

Torque between 3000-4000 rpm remains almost constant.
Further increase in speed causes slight decrease in Torque.
The pattern of curves is same for all blends. The maximum
Torque is obtained in case of diesel-CNG dual fuel mode is
7.8 kg-m at 3400 rpm.
Figure 7(b) shows the Break Power for pure diesel, blends of
Biodiesel and for diesel-CNG dual fuel. The variation of
Break Power for different fuels is almost similar. Initially for
all fuels the Break power increases sharply up 4000 rpm
except pure diesel. In pure diesel it increases sharply upto
3000 rpm than increases slowly. After 4000 rpm the Break
power increases slowly. Between 4000 rpm to 5000 rpm
break power remains almost constant. In starting at 1000 rpm
B30 has the maximum power. Between 2000 rpm to 4000
rpm Break Power is maximum in case of pure diesel. At
higher speed 5000 rpm B10 has the maximum power.
Fig. 7(b) Comparison of BP Vs

Speed for different fuels
5.1.3 Break thermal efficiency with speed
but at 2000 rpm it increases slightly. At higher load as soon as

speed increases the Break Power remains lower than other
fuels. The maximum power obtained in case of B10 and pure
diesel is 32 kW.Figure
5.1.4 Specific fuel consumption with speed
Fig. 7(d) Comparison of SFC Vs

7(c) shows the variation of Break Thermal Efficiency

(BThE) for pure diesel, blends of Biodiesel and for dieselCNG dual fuel. The pattern of variation of Break Thermal
Efficiency for different fuels is almost similar. At initial
speed Break Thermal Efficiency is almost same for all the
fuels. In starting Break Thermal Efficiency increases with
increase in speed. At 2000 rpm BThE is maximum for B30
and minimum for pure diesel. After 2000 rpm BThE in case
of diesel-CNG increases with speed and remains higher than
all the fuels at higher speed. The maximum BThE is obtained
in case of B30 at 4000 rpm. First BThE increases with speed
and after 4000 rpm BThE decreases with speed. At 5000 rpm
diesel-CNG has maximum BThE.
Figure 7(d) Specific Fuel Consumption (SFC) w.r.to speed.
The variation of SFC is almost similar in case of pure diesel,
B10 and B20. At initial speed the SFC decreases with speed
up to 200 rpm. Between 2000 rpm to 4000 rpm it remains
almost constant at higher speed after 4000 rpm SFC
increases with increase in speed up to 5000 rpm. The
minimum SFC is in case of diesel-CNG dual fuel at 3500
rpm. SFC in case of diesel-CNG dual fuel always remains
lower than the other fuels.
5.2 Emissions results
Fig. 7(c) Comparison of BThE Vs

The Break Power in case of diesel-CNG always remains

lower than pure diesel. In starting it is same as in other fuels
58
The emission results obtained with the help of smoke meter

and gas analyser from the tests performed on the engine are
discussed below.

S & T Review
Figure 8(a) shows the variation of smoke opacity w.r.to

speed. The pattern of variation in opacity is same for all the
fuels. The opacity in case of pure diesel always remains
higher than the other fuels this shows the advantage of using
alternative fuels. The opacity of smoke increases with
increase in speed thus at the higher speed opacity is
maximum for all the fuel used. The opacity in case of dieselCNG throughout the test remains lower than the other fuel
but at higher speed 5000 rpm there is little increase in opacity
due to improper A/F mixture. Maximum Opacity is in case of
pure diesel throughout the test.
increase in engine speed. The pattern of NOX emission is

same for all the fuels. It can be observed from the figure that
the level of NOX emission is lower for diesel-CNG
throughout the test. NOX emission is lowest in case of dieselCNG dual fuel at 1000 rpm. At 5000 rpm NOX emission is
also lowest in comparison to other fuels. While for other
fuels increase in NOX emission is almost same for pure diesel
and B10. NOX emission is higher in case of B20.
5.2.1 Smoke opacity

Figure 8(b) shows the variation of CO emission w.r.to speed.
The CO emission increases with increase in engine speed for
all the fuels. The maximum CO emission is obtained at 5000
rpm in case of diesel-CNG dual fuel. CO emission is higher
for diesel-CNG dual fuel throughout the test. Therefore CO
emission is higher in diesel-CNG dual fuel and lower in pure
diesel throughout the test.
Figure 8(c) shows the variation of HC emission w.r.to speed.
The pattern of increase in HC emission with speed is almost
similar for all the fuels. The HC emission is lowest in
Fig. 8(b) Comparison of CO Vs Speed for different fuels
5.2.3 HC emissions
Fig. 8(c) Comparison of HC Vs Speed for different fuel
Fig. 8(a) Comparison of Opacity Vs

5.2.4 Nitrogen oxides
5.2.2 CO emissions
Case of pure diesel at 1000 rpm and maximum HC emission
is in case of diesel-CNG dual fuel at 5000 rpm. At 3000 and
4000 rpm minimum HC emission is in case of B30. HC
emission for diesel-CNG dual fuel is increasing with
increase in speed in comparison to other fuels. It remains
higher than pure diesel throughout the test.
Figure 8(d) shows the variation of NOX emission with
Fig. 8(d) Comparison of NOX Vs Speed for different fuels
59
negligible.
6. Conclusion
The dual fuel technique is a good method to take the
advantage of both fuels, diesel and alternative fuel (CNG). It
is observed that the diesel engine have some disadvantages
over alternative fuels running at the higher speed like higher
knocking tendency, higher specific fuel consumption and
comparatively higher emissions. On the other hand
dedicated CNG engines converted from diesel engines also
have some drawbacks like higher conversion cost, low
engine power etc. Therefore to take the advantages of both
the fuels a bridge technology in between both the fuel is used.
That's why diesel-CNG dual fuel mode is used by combining
them through the arrangement of supplying of CNG in intake
air suction pipe of the engine. The engine is tested with pure
diesel, blends of Biodiesel (B10, B20 and B30) and dieselCNG dual fuel mode.
The following conclusions have been made from the
experiments
1.
2.
3.
Engine running on the diesel-CNG dual mode

produced less knock at the higher speed comparatively
to the pure diesel. In starting at slow speed in dual fuel
mode knocking was higher but with increase of engine
speed knocking was less in dual fuel case.
The performance parameters were also improved in
case of duel fuel mode. Torque produced in duel fuel
mode was higher up to 4000 rpm after that at high
speed it will start decreasing. In starting torque
produced was 30% more than produced diesel fuel.
There was slight decrease in break power of the engine
at higher speed.
The break thermal efficiency of engine has improved in
diesel-CNG dual fuel mode as compared to other fuels.
At higher speed BThE increased by 14% than pure
diesel mode.
4.
The specific fuel consumption has also reduced due to

the use of CNG. The SFC first decrease up to 3000 rpm
than becomes almost constant and after a specific
speed it starts increasing at higher speed.
5.
The smoke opacity is reducing in our alternative fuels.

The main advantage of using alternative fuels is to
reduce the exhaust emissions. The smoke opacity is
decreasing in case of all the alternative fuels. At 5000
rpm there was slight increase in opacity which is
60
6.
There is slight increase in CO and HC emissions in case

of dual fuel mode. This is the disadvantage of dieselCNG dual fuel mode. This problem can be resolved by
using the emission control technologies or by after
exhaust treatments.
7.
The main advantage of using diesel-CNG dual fuel

mode is reduction in oxides of nitrogen in engine
exhaust emissions.
References
[1] Pal Amit, Maji S. and Arora B.B. et al., Use of CNG
and Diesel in CI Engine in dual fuel mode 2008.
[2] Bernard C. and Baranescu R., Diesel Engines and
emission control 2nd Edn. Elsevier, Oxford, UK. pp:
3-307, 2003.
[3] Ehsan Md. and Barman Sourav, Dual Fuel
Performance of a Diesel Generator Using Compressed
Natural Gas A Project Report, 2008.
[4] Patrick Coroller & Gabriel Plassat, Comparative
Study on Exhaust Emissions from Diesel and CngPowered Urban Buses French Agency of Environment
and Energy Management, Air & Transport Division,
2003.
[5]
Ismail Abdul, Semin, Diesel Engine Convert to Port

Injection CNG Engine Using Gaseous Injector Nozzle
Multi Holes Geometries Improvement American
Journal of Engineering and Applied Sciences, pp: 268278, 2009.
[6] Poulton M. L., Alternative Fuels for Road Vehicles.

1st Edn. Comp. Mechanics Publications, UK. pp: 10110, 1994.
[7] Semin, et al., Effect of Diesel Engine Converted to
Sequential Port Injection Compressed Natural Gas
Engine on the Cylinder Pressure vs. Crank Angle in
Variation Engine Speeds American J. of Engineering
and Applied Sciences, pp:154-159, 2009
[8] Ehsan Md. and Barman Sourav, Dual Fuel
Performance of a Diesel Generator Using Compressed
Natural Gas A Project Report, 2008.
[9] Poulton M. L., Alternative Fuels for Road Vehicles.

S & T Review

[10] Semin, et al., Effect of Diesel Engine Converted to
Sequential Port Injection Compressed Natural Gas
Engine on the Cylinder Pressure vs. Crank Angle in
Variation Engine Speeds American J. of Engineering
and Applied Sciences, pp:154-159, 2009
[11] Srinivasan K.K., The Advanced Injection Low Pilot
Ignited Natural Gas Engine: A Combustion Analysis
J. Eng. Gas Turbines Power, pp: 213-218, 2006.
[12] Kumarappa S. and prabhukumar G.P., Improving the
Performance of Two Stroke Spark Ignition Engine by
Direct Electronic CNG Injection Jordan Journal of
Mechanical and Industrial Engineering, Volume 2,
December, pp: 169-174, 2008.
[13] Fleming R.D. and Allsup J.R.,emission characteristics
of natural gas as an alternative fuel SAE 710833,
1971.
[14] Semin, Rahim Abdul and Bakar R.A., Combustion
Temperature Effect of Diesel Engine Convert to
Combustion Temperature Effect of Diesel Engine
Convert to American Journal of Engineering and
Applied Sciences, pp: 212-216, 2009
[15] Ismail Abdul, Semin, Diesel Engine Convert to Port
Injection CNG Engine Using Gaseous Injector Nozzle
Multi Holes Geometries Improvement American
Journal of Engineering and Applied Sciences, pp: 268278, 2009.
[16] Idris Awang, Bakar R.A., Engine Cylinder Fluid
Characteristics of Diesel Engine Converted to CNG
Engine European Journal of Scientific Research,
Vol.26 No.3, pp: 443-452, 2009.
[17] Kaleemuddin Syed, Rao AmbaPrasad, Development
of Dual Fuel Single Cylinder Natural Gas Engine an
Analysis and Experimental Investigation for
Performance and Emission American Journal of
Applied Sciences, pp: 929-936, 2009
[18] LIU Zhen Tao and FEI Shao Mei, Study of
CNG/diesel dual fuel engine's emissions by means of
RBF neural network Journal of Zhejiang University
Science, pp: 960-965, 2003.
[19] Layford E.J., An Emission and Performance

Comparison of the Natural Gas C-Gas plus Engine in
Heavy-Duty Trucks a report submitted to National
Renewable Energy Laboratory, Colorado 2003.
[20] Patrick Coroller & Gabriel Plassat, Comparative
Study on Exhaust Emissions from Diesel and CngPowered Urban Buses French Agency of Environment
and Energy Management, Air & Transport Division,
2003.
[21] Chelani A.B. and Devotta S., Impact of Change in
Fuel Quality on PM10 in Delhi Bulletin of
Environmental Contamination and Toxicology 75(3):
600607, 2005.
[22] Central Pollution Control Board (CPCB), Vehicular
Pollution Control in Delhi-Initiatives & Impacts a
report of Ministry of Environment & Forests, New
Delhi, 2001.
[23] Agarwal A., Slow Murder: The Deadly Story of
Vehicular Pollution in India Report to Centre for
Science and Environment, New Delhi, 1996.
[24] Bose, R.K. and Nesamani, K.S., Urban Transport,
Energy and Environment A Case of Delhi. A
research report prepared for the Institute of
Transportation Studies, University of California,
Davis, USA. UCD-ITS-RR-00- 11, 2000.
[25] Komiyama K. and Heywood J.B., Predicting NOx
Emissions and Effect of Exhaust Gas Recirculation in
Spark-Ignition Engines, SAE Paper No. 730475,
1973.
[26] Timothy T. et al., Alternative Fuels SAE Inc, 1995.
[27] Heck R.M. and Farrauto R.J., Catalytic air pollution
control John Wiley & Sons, 1995.
[28] Grigorios C. et al., Catalytic Automotive Exhaust
Aftertreatment. pp: 13-14, 1997.
[29] Mltianlec W., Thermo chemical Model of Catalytic
Converter in Exhaust System of Two Stroke Engine
I n t e r n a t i o n a l C o n f e re n c e o n I C E n g i n e s
KONES'2000, Naleczow Poland, September 2000.
[30] Ganesan V., Internal Combustion Engines 2nd Edn.
Tata McGraw-Hill, New Delhi, India, pp: 110-255,
1999.
61
[31] Poulton M.L., Alternative Fuels for Road Vehicles

[32] R.G. Papagiannakis, and D.T. Hountalas,
Experimental Investigation Concerning the Effect of
Natural Gas Percentage on Performance and
Emissions of a DI Dual Fuel Diesel Engine, Applied
Thermal Engineering, 23, pp. 353-365, 2003
[33] Bhandari Kirti, performance and emissions of natural
gas fueled internal combustion engine Journal of
Scientific & Industrial Research vol. 64, pp 333-338,
2005.
62
[34] Kumar P.S., Mahanta P. and Saha U.K., Emission

Characteristics of CNG inAutomobiles, (2002).
[35] Wannatong Krisada, Akarapanyavit Nirod and
Siengsanorh Somchai, Combustion and Knock
Characteristics of Natural Gas Diesel Dual Fuel
Engine journal of Society of Automotive Engineers of
Japan 2007.
[36] Hill P.G. and Douville B.' Analysis of combustion in
diesel engines fueled by directly injected natural gas.
J. Eng. Gas Turbines Power, pp: 143-149, 2000.

S & T Review
Distresses on Flexible Pavement and Their Measurement

Vishal Bhatt*, S. S. Jain*, M. Parida*
raffic load in metro cities increases tremendously in developing countries like India. Problems of urban roads are much
more significant than other roads because of factors like Drainage, higher Traffic Volume and Abrasion of traffic,
underground services, intersections, road side activities and spilling of Petrol, oil and lubricants. Pavement deterioration in
urban network causes great loss directly in terms of accidents, riding quality and indirectly as pollution, time of journey, etc.
There should be a special consideration for pavement maintenance for urban roads. In this study an effort has been made to
study the distress evaluation on urban roads. The study aimed at studying the various types of pavement distresses and their
causes and to suggest the techniques suitable for urban roads. Different emerging techniques for distress measurement in
developed countries were reviewed. Important factors influencing pavement condition and overall performance were
analyzed. Distresses are measured on 2 road sections in Noida to understand the variation of different distress parameters and
their measurement. Distresses like rut-depth, pavement deflection, skid resistance, roughness were measured on these
sections. The measurements were done by Loadman, British skid resistance tester, straight-edge. In addition to these, cracks
measurement, pot-holes and raveling area measurement and traffic volume study was also done.
Keywords: Distress Measurement, Flexible Pavement, Urban Roads
1. Introduction
Roads are the fundamental infrastructure requirement for
economic development of a nation. Roads are an enormous
national investment and require maintenance to keep them in
satisfactory condition and ensure safe passage at an
appropriate speed and with low road user costs. Without
timely maintenance, pavements deteriorate considerably,
leading to higher vehicle operating costs, increased number
of accidents and reduced reliability of transport services.
When the maintenance work can no longer be delayed, it will
often involve extensive rehabilitation, and even
reconstruction, costing many times more than timely
maintenance treatment carried out earlier. Late or inadequate
maintenance will increase the ultimate repair costs,
inconvenience to users and reduce safety. Road maintenance
is therefore an essential function and should be carried out on
timely basis.
2. Objectives
In the recent years due to increase in traffic load the
deterioration of pavement has also increased tremendously.
As roads are the arteries of a country, proper functioning of
them is necessary. Therefore, there is a need to adopt the best
suitable pavement evaluation technique for the maintenance

of the roads which commences with the distress
measurement of the pavement. An attempt is made here to
study the various techniques available for the distress
measurement and also to suggest their suitability to be used
for to urban roads.
The study aimed at studying the various types of pavement
distresses that occurs on flexible roads and different
techniques available to quantify these distresses. Distresses
are measured on Noida Urban roads using the respective
tools depending upon suitability and availability. Study on
variation of distresses on the urban roads is also made to
understand the significance of each parameter.
3. Pavement Distress and its Evaluation
In this section a brief overview of distress parameters occur
on urban roads and devices used for calculating these
parameters is given.
3.1 Pavement distresses
Pavement distresses has been grouped into the following
categories:
!
* Department of Civil Engineering, IIT, Roorkee, India.

Email- vishal2006bhatt@gmail.com,
E-mail: mprdafce@iitr.ernet.in
Cracking
1. Fatigue Cracking
2. Block Cracking
63
3. Edge Cracking
4a. Longitudinal CrackingWheel Path
4b. Longitudinal CrackingNon-Wheel Path
5. Reflection Cracking at Joints
6. Transverse Cracking
Patching and Potholes
1. Patch/Patch Deterioration
2. Potholes
Surface Deformation
1. Rutting
2. Shoving
Surface Defects
1. Bleeding
2. Polished Aggregate
3. Raveling
Miscellaneous Distresses
1. Lane-to-Shoulder Dropoff
2. Water Bleeding and Pumping
3.2 Pavement evaluation

There are various methods of pavement evaluation, which
have been broadly classified in the following two groups
1. Functional evaluation
The various parameters to be observed in functional
evaluation are road roughness, rutting, cracks, potholes and
patches. Following are the functional evaluation methods:
A. Roughness measuring devices
The roughness measuring devices commonly are as given
below in table.
Table1 Roughness measuring devices
64

S & T Review
3.3 Surface distress measurement
3.4.1 Destructive methods for pavement evaluation
Measurement techniques are mostly visual. Older

techniques, uses teams of individuals who drove across
every mile of pavement to be measured. Speeds were usually
quite slow (on the order of 16 km/hr and measurement was
done visually. More current methods record pavement
surface video images at highway speed using a specially
equipped van that is outfitted with high resolution cameras.
Evaluation is either done manually by playing the video
back on specially designed workstations while trained crews
rate the recorded road surface(PSR) or automatically by
computer software (PSI). These ways have many advantages
like higher safety, Accurate and complete distress data, more
effective quality control, more efficient data collection and
moreover video and digital images are available for other
users.
3.4.2 Non destructive methods
Table 4. Non Destructive methods for pavement
structural evaluation
Fig. 1 Digital Analysis of Pavement surface using

Integrated analysis units
Integrated analysis units can collect pavement surface
distress data in the previously described manner as well as
collect data on a variety of other characteristics at highway
speeds such as:
3.4 Structural evaluation

The structural evaluation is done to access the structural
capacity of existing pavements. Structural evaluation is also
important to access the remaining life of the pavement.
Various methods of structural evaluation are as under
65
4. Area of Study
Two urban road sections of Noida were selected to study the
variation of distress parameters for the urban roads. Different
distress parameters are studied for these sections.
Table 5 List of roads under study
rectangles, in case of interconnected and alligator cracks. In

case of single longitudinal and transverse cracks, the crack
length was measured and affected width of the pavement
surface across the length of crack was taken as 50 cm. Thus,
crack area was expressed as percentage of total pavement
area. Separate measurements were taken for cracks of width
up to 3mm (narrow cracks) and width more than 3mm (wide
cracks). It was observed that in majority of cases the
percentage of wide crack area was found to be approximately
one-third of the total crack area.
6.3 Measurement of pothole area
5. Types of Data
!
Traffic characteristics like traffic volume, compositions

of different vehicles, etc.
Roadway characteristics like grade, width, no. of lanes,

earth breams, surface type etc.
Distress Parameters like roughness, rutting, skid

resistance, cracks, raveling, deflection measurements.
6. Data Collection Methods

Traffic data can be collected by using anyone of the
following techniques:
!
Direct observation methods
Video graphic methods
For the traffic volume study direct observation method is

used. Methodology adopted for the collection of various
parameters is described below.
6.1 Surface distress measurements
The type and extent of distress developed at the surface i.e.
cracking, raveling, patch work, potholes, rut depth,
depressions etc. were observed based on the visual condition
survey. The extent and type of distress developed in
quantitative terms were also measured, in addition to the
visual recording of the pavement surface condition. The
information on shoulder type, width, and condition, and
drainage should also be recorded.
6.2 Measurement of crack area
A number of representative test sections of length 50m were
chosen for cracking measurements for each pavement
section. The affected area was marked in the form of
66
Potholes are bowl shaped holes of varying sizes in a surface

layer or extending into base/sub base course. The pothole
area is measured in terms of sq. m., and the depth of each
pothole is also measured to convert into volume of potholes.
The volume is then converted into number of standard
pothole units of one litre volume each. The pothole
measurements are finally expressed as number of pothole
units per km length of the pavement section.
6.4 Rut depth measurements
The transverse deformation across the wheel path is defined
as the rut. The rut depth measurements were taken with a 2m
straight edge, at all deflection points under the wheel path i.e.
at a distance of 0.9m/0.6m from the pavement edge in case of
double lane/single lane carriageway, and maximum value of
rut depth was noted down at each observation point.
6.5 Skid resistance measurements
The skid resistance measurements were taken using a
'Portable Skid Resistance Tester' developed at Transport and
Road Research Laboratory (TRRL). This instrument is also
known as 'British Pendulum Tester'. The characteristics of
the instrument are such as to simulate the sliding conditions
between vehicles with patterned tyres, braking with locked
wheels on wet pavement surface at a speed of 50 kmph. The
quantity measured with the portable tester is termed as Skid
Resistance Value (SRV). The SRV is converted to 'Sideways
Force Coefficient' (SFC, measured with Sideways
Coefficient Routine Investigation Machine - SCRIM), using
the relationship given by equation [Kennedy et al 1990].
SFC = 0.01* SRV
The fall-off in skidding resistance with increased speed on
wet roads depends on the roughness of the surface macro-

S & T Review
texture. Therefore, an additional criteria indicating surface

texture is also required. This is accomplished by measuring
texture depth in mm by 'Sand Patch Method'.
6.6 Measurement of surface deflection using LOADMAN
traffic, cycle rickshaw and cycle were the prominent modes.

Pavement distress parameters were found to be related to
each other, if raveling is found more on a section then
cracking was also high and vice-versa.
LOADMAN is a portable version of the 'Falling Weight

Deflectometer', used to measure the surface deflection of
pavement under standard load conditions. In this instrument,
the deflection caused by the load of a falling weight is
measured by an accelerometer. The measuring result will be
given as the deflection, and the modulus of elasticity 'E'.
6.7 Measurement of raveled area
Raveling is the loss of material from wearing surface. This
distress type is associated with thin surfacing, such as,
surface dressing, seal coat and premix carpet. The effected
area is measured by taking into account area enclosed in
regular geometric shapes such as rectangles, and then it was
expressed as percentage of total pavement area.
6.8 Roughness measurements
Fig. 3 Pavement Distress Values

For Different Road Sections
The pavement roughness should be measured on each

pavement section with 'Fifth Wheel Bump Integrator', towed
by the COTE mobile van, as per the standard procedure. The
instrument should be run at a constant speed of 30 kmph. The
observations are taken on the outer wheel paths in both
direction, at a distance of 0.9m and 0.6m from the edges in
case of two lane and single lane pavements respectively.
Table 6 Parameters Collected on Urban Sections in Noida
Fig. 4 Traffic Flow Pattern in a Day

8. Conclusions
7. Data Analysis and Discussion

Traffic volume was found to follow two peak format because
of work-based trips in the urban area. Traffic volume has
more motorized traffic, especially cars. In non motorized
Traffic on urban roads attains two peaks, one in morning and

another at evening which represents the shift- change time
for different firms and institutions. Riding quality on all the
roads was found satisfactory. In cities, there have been large
scale damage of the existing pavement due to cross cutting at
frequent intervals for service lines, longitudinal cuttings for
providing medians and digging of the roads for various other
purposes. Due to high traffic load on urban roads, it is
important to use a distress measurement technique which
blocks the traffic to minimum extent, and is good in
repeatability and accuracy.
67
[6]
ReddyB. and Veeraragavan A. Structural

performance of in service flexible pavements
Journal of Transportation Engineering, Vol. 123,
No.2, 12122, 1997.
[7]
Chua k. and Xu L. Simple procedure for identifying

pavement distresses from video images Journal of
Transportation Engineering, Vol. 120, No. 3, 4704,
1994.
[8]
Sharma N.K. Loadman and Benkelman Beam

Deflection Studies on Identified Road Sections
M.E Thesis, University Of Roorkee, India, 2000.
Taleb A. and Ibrahim A. and Baker A. Roughness

evaluation of jordan highway network 24th ARRB
Conference Building on 50 years of road and
transport research, Melbourne, Australia, 2010.
[9]
Rokade S , Agarwal P K and Shrivastava R Study

on performance of flexible highway pavements
International Journal of Advanced Engineering
Technology IJAET, Vol.I, Issue III, 312-338, 2010.
Sitaramanjaneyulu K., Kanchan P.M., Sharma

B.M., Nagabhushana M.N. Indian tools for
assessment of maintenance needs 4th International
Conference on Managing Pavements, 1998.
[10]
Mascio P., Piccolo and Cera L. Automated distress

evaluation 4th International SIIV Congress Palermo
(Italy), 2007.
[11]
Rawat, M.M.S. Development of Pavement

Management System for Urban Roads M.E.
Thesis, University of Roorkee, Roorkee, India,
1996.
[12]
Prakash, P. Development of Maintenance

anagement System for Urban Roads.
M.Tech.Thesis, IIT Roorkee, Roorkee, India, 2009.
9. Recommendations
Development of fast, accurate, reliable and portable distress
measuring device to reduce labour and time in pavement
survey is required. Division of the whole network of
pavements in the country into homogeneous groups based on
the pavement layers and materials characteristics and traffic
load factor on the roads so as to develop a common pavement
management system for homogeneous networks can also be
helpful in pavement management.
References
[1]
[2]
[3]
Code of practice for Bituminous surfaces of

highways IRC-82 (1992).
[4]
Guidelines for strengthening of flexible road

pavements using Benkelman Beam Deflection
Technique IRC-81 (1981).
[5]
Hudson W.R. and Uddin, W. Future pavement

evaluation technologies: prospects and
opportunities 2nd North American Pavement
Management Conference, 1987.
68

S & T Review
Analysing the Suitability of Waste Materials

for Road Construction
Aditya Kumar Anupam*, Praveen Kumar*, G.D. Ransinchung R.N.*
ndia produces a huge amount of waste materials as byproducts from different sectors like industrial, construction,
agricultural, etc. These waste materials if not deposited safely it may be hazardous. A large quantity of waste material is
dumped at land filling site, which if investigated properly can be utilized in road construction sector. The utilization of these
waste materials can be an economical and eco-friendly alternative in nearby areas for rural road construction. In India,
research is currently underway to examine the potential for use of some locally available wastes in road construction. The
results, to date, indicate that there is a wide scope for the use of such materials. However, in India, only a few materials have
been used and that too on experimental basis. An attempt has been made in the similar directions for investigating the potential
use of some industrial waste materials viz. fly ash, construction & demolition wastes, rice husk ash, bagasse ash, etc. for road
construction.
Keywords: Rice husk ash, Bagasse ash, Fly ash, Construction & demolition wastes
1. Introduction
2. Review of Paste Studies
The safe disposal of industrial and agricultural waste

products demands urgent and cost effective solutions
because of the debilitating effect of these materials on the
environment and to the health hazards that these wastes
constitute. Landfilling is becoming more expensive and, as
sustainable management of resources is becoming a goal for
many communities, the acceptance of landfilling or
incineration is decreasing.
Cokca, [1] studied the effect of fly ash for the stabilization of
an expansive soil and concluded that the expansive soil can
be stabilized successfully by fly ashes.
To deal with the growing disposal problem of these materials

is an issue that requires co-ordination and commitment on
the part of all parties involved such as government agencies,
companies, the public and professionals. One method to
reduce some portion of the waste disposal problem is by
recycling and utilizing these materials in the construction of
roads. However, such a use should not compromise the
quality and performance of the highway infrastructure nor
create environmental problems.
Nominal research has been done in India to determine the
availability of feasible waste materials and the suitability of
these materials for Indian roads. Yet the use of wastes would
benefit the road sector by providing it with a cheap source of
material and, in some cases, road construction companies
may be able to charge a fee for suitably using waste materials
that would otherwise require expensive treatment for
disposal.
* Department of Civil Engineering, Indian Institute of Technology,
Roorkee, India, Email- addiknit03@gmail.com,
The unsoaked unconfined compressive strength of fly ashsoil mixtures decreases with increases of fly ash. The
decrease in strength may be attributed to decrease in
maximum dry density and increase in optimum moisture
content on addition of fly ash to soils. Chakraborty and
Dasgupta, [2] studied the strength characteristics of
reinforced fly ash by carrying out laboratory triaxial shear
tests. The fly ash was collected from the Kolaghat thermal
power station in India.
A haul road near Clinch River Power Plant, South Western
Virginia was constructed using 14% cement stabilized fly
ash. Although the road was subjected to heavy traffic, it
performed well for many years [3].
Kumar et al. [4] utilized fiber-reinforced fly ash as a subbase
material. Fly ash with 0.2% fiber content has CBR value
16.6%. Therefore fly ash with 0.2% fiber content is suitable
for rural road subbases. For rural roads with higher traffic
IRC 37 (IRC 2001) is to be followed, which states the CBR
requirement of 20 and 30% depending upon the traffic. Fiber
reinforcement of 0.3 and 0.4% will make the fly ash suitable
for these conditions. Resilient modulus (Mr), modulus of
subgrade reaction (k), and field CBR value of fly ash increase
due to reinforcement and mixing of soil.
69
The term rice husk refers to the by-product produced in the

milling of paddy and forms 16-25% by weight of the paddy
processed. In the majority of rice producing countries much
of the husk produced from the processing of rice is either
burnt for heat or dumped as a waste. India alone produces
around 120 million tones of rice paddy per year, giving
around 24 million tones of rice husk per year. Farm income
can be increased both directly and indirectly if economically
profitable means of utilizing rice husk generated are utilized
in industry or road sector. There are many reported uses of
rice husk such as a fuel in brick kilns, in furnaces, in rice mills
for parboiling process, in the raw material for the production
ofxylitol, furfural, ethanol, acetic acid, lignosulphonic acids,
as an cleaning or polishing agent in metal and machine
industry, in the manufacturing of building materials, etc. [5].
Rice husk is an agricultural waste obtained from rice milling.
About 108 tons of rice husks are generated annually in the
world [6]. Rice husk ash includes a huge amount of silica
with high specific surface that is very suitable for activating
the reaction of soil and lime. This matter is not suitable for
cattle feeding and is non-biodegradable. Thus, using RHA as
an additive seems to be economical particularly in regions
having high production capacity.
Muntohar [7] shows that the addition of 6% lime in
combination with RHA principally has a significant effect on
reducing swell and swelling pressure and increase the
durability of soils. Noor et al. [8] studied the influences of
different mix proportions of cement and RHA on compaction
characteristics, soaked and unsoaked unconfined
compressive strength, tensile strength, and durability, and
the results show that RHA usage achieved satisfying values
for soil strength.
According to Ahmad and Shaikh [9], the physical and
chemical properties of sugarcane bagasse ash are found to be
satisfactory and conform to the requirements for class N
pozzolanas (ASTM C618-78). The major oxides commonly
associated with bagasse ash are SiO2, Al2O3, Fe2O3, CaO,
MgO, Na2O, K2O, P2O5 and MnO. The loss on ignition, which
is indicative of total carbon content, is often below the
specified value of 10% for class N pozzolanas.
Ganesan et al. [10] studied on the use of bagasse ash (BA) as
partial cement replacement material in respect of cement
mortars. Up to 20% of ordinary portland cement can be
optimally replaced with well-burnt bagasse ash without any
70
adverse effect on the desirable properties of concrete. The

specific advantages of such replacement are development of
high early strength, a reduction in water permeability and
appreciable resistance to chloride permeation and diffusion.
Goyal et al. [11] reported that SCBA with high specific
surface area, high contents of amorphous silica and calcium
oxide fulfilled the principal requirements of a pozzolanic
material. A little amount of crystallization of silica present in
the ash, may have affected the degree of reactivity. Strength
activity index of SCBA in cement mortar fulfilled the
requirements of pozzolanic materials as per ASTM
standards. Mechanical strength tests confirmed the actual
behavior of the SCBA-blended mortar as 10 to 15%
substitution of OPC with SCBA, gave better strength results
than without SCBA specimen. Khobklang et al. [12]
investigated the effects of the proportion of bagasse ashblended Portland cement in mixtures of lateritic soil
interlocking block. The results revealed that an increase in
the amount of bagasse ash significantly increased water
absorption. The interlocking block using bagasse ash in a
ratio of more than 15% to binder provided higher
compressive strength, compared to that using lower bagasse
ash content, at the testing age of 90 days. Moreover it was
found that an increase in the water to binder ratio
significantly decreased water absorption.
3. Waste Materials used for This Study
The waste materials obtained from different sectors like
industrial, construction, agricultural, etc. which
are
analysed in laboratory for their suitability in road
construction are discussed below.
3.1 Construction and demolition waste
Construction and demolition (C&D) debris is waste material
produced in the process of construction, renovation, or
demolition of structures. Structures include buildings of all
types (both residential and non-residential) as well as roads
and bridges. C&D debris is generally a non-hazardous waste
material. Components of C&D debris typically include
concrete, asphalt, wood, metals, gypsum wallboard, and
roofing. Land clearing debris, such as stumps, rocks, and
dirt, are also included in some state definitions of C&D
debris. Every year more and more billions of tons of
construction and demolition building waste are produced in
the world. Their disposal has become a severe social and
environmental problem in all countries. The possibility of
their recycling and reuse is thus of increasing importance.
S & T Review
3.1.1. Fly ash

Pulverized Coal is being used as fuel in thermal power
generation units to meet the growing demands of electricity.
In India, approximately seventy percent of the total
electricity is produced using this method. Huge quantities of
coal ashes, in form of fly ash and bottom ash, are generated in
the process as a waste and these are collected in ponds in
slurry form which is known as pond ash.
The estimated production of pond ash in India at present is
about cross 100 MT by the year 2000. This huge quantity of
ash is creating problems of disposal and creating
environmental pollution. To overcome these problems,
various processes for utilization of fly ash have been
developed by scientists/engineers in India and abroad, but its
consumption is still quite low.
The properties of conventional aggregates were compared

with recycled aggregate as shown in Table 1.
Based on the experimental results obtained, it can be
concluded that the recycled aggregate can be effectively used
as a road material in different layers as the properties
satisfying the MOSRT&H requirements. Recycled
aggregates found to be relatively soft compared with
conventional aggregate and can be used as a sub-base
material but not in base course and wearing course. Water
absorption of waste material is found to be high compared
with conventional aggregate. The results of recycle
aggregate are found to be within the limits as per MOSRT&H
specifications.
Table1 Laboratory results for different dypes of
building waste materials
3.1.2. Rice husk ash

Rice husk is an agricultural waste obtained from rice milling.
India alone produces around 120 million tones of rice paddy
per year, giving around 24 million tones of rice husk per year.
Rice husk ash (RHA) includes a huge amount of silica with
high specific surface that is very suitable for activating the
reaction of soil and lime. This matter is not suitable for cattle
feeding and is non biodegradable. Thus, using RHA as an
additive seems to be economical particularly in regions
having high production capacity.
3.1.3 Bagasse ash
Bagasse ash is an agricultural by-product of sugar
manufacturing. When juice is extracted from the cane sugar,
the solid waste material is known as bagasse. After the
extraction of all economical sugar from sugarcane, about 4045% fibrous residue is obtained, which is reused in the same
industry as fuel in boilers for heat generation leaving behind
8 -10 % ash as waste, known as sugarcane bagasse ash
(SCBA). This ash is considered as waste and disposed in an
environmentally debilitating manner. It is a very valuable
pozzolanic material if carbon free and amorphous ash could
be obtained by further combustion.
4. Laboratory Investigation of Waste Materials
4.1 Construction and demolition waste
In the present work, use of recycled aggregate from building
waste as base course and sub-base course has been studied in
order to reduce the material transport cost and disposal cost.
4.2 Fly ash

The fly ash used in the study was brought from National
Thermal Power Station situated at Ghaziabad which was
available free of cost. Fly ash is classified as silts of low
compressibility (ML).
Fly ash samples were subjected to thermo gravimetric
analysis (TGA) in a dry air atmosphere at heating rates:
10C/min. Samples weighing 80 mg were heated at
preselected heating rates from ambient temperature to 850C
in a Mettler thermal analyzer. The continuous records of
weight loss and temperature were obtained and analyzed
using a Mettler thermal analysis software system to
determine the weight loss with increase in temperature. The
TGA curve in figure 1 shows that the mass loss for fly ash
even after heating at a high temperature of 650C is only
about 1.7% which shows that the selected fly ash is free of
solid (fixed) carbon. Hence no further heating is required for
the fly ash obtained from thermal power plants before its use
in road construction.
The chemical, physical and engineering properties of ash
depends on the type and source of coal used, method and
71
degree of coal preparation, cleaning and pulverization, type

and operation of power generation unit, ash collection,
handling and storage methods etc.
Figure 2 shows the grain size distribution of collected fly ash
without extra heating. The steep 'S' curve represents the fly
ash contains uniformly graded particles with almost same
size. The grain size distribution curve reveals that fly ash is
fine grained which is better for soil interlocking.
So the properties of fly ash vary from plant to plant and even
within the same plant. The physical and chemical properties
of fly ash tested in laboratory are given in Table 2 and Table 3
respectively.
The results of TGA reveal that fly ash is carbon free material
and is siliceous which makes it suitable as soil stabilizing
agent. Soils with low silt content like clay gives better results
due to higher silt content of the fly ash.
Table 2 Physical Properties of Fly Ash

Rice husk ash is a predominantly siliceous material obtained
after burning of rice husk in a boiler or an open fire. Lime
reactivity test conducted on this ash indicate the fully burned
rice husk ash exhibits greater reactivity.
The collected sample of rice husk ash contains more quantity
of fine grained particles as compared to coarse particle
(figure 4). The steep 'S' curve represents the rice husk ash
contains uniformly graded particles with almost same size.
Table 3 Chemical properties of fly ash
Fig 1. TGA Curve for Fly Ash
Properties
Test Value (%)
SiO2
58.78
Fe2O3
9.31
Al2O3
26.92
CaO
1.77
MgO
0.68
N2O
0.28
K2O
1.44
Loss on Ignition (%) by weight
0.72
4.3 Rice husk ash
Fig 2. Grain size distribution for Fly Ash
72
Rice husk ash samples were subjected to TGA in a dry air

atmosphere at heating rates: 10C/min. Samples weighing
80 mg were heated at preselected heating rates from ambient
temperature to 850C in a Mettler thermal analyzer. The
TGA curve in figure 3 indicates about 1.5% of mass loss for
rice husk ash after heating at a temperature of 600C and at
higher temperatures the mass loss is almost constant. Lesser
mass loss makes it carbon free material and suitable for use in
road sector without further heating.

S & T Review
Table 4 Physical & chemical

properties of rice husk ash
Fig 3. TGA curve for rice husk ash

This waste material having pozzolonic properties can be
utilized in the stabilization for road construction. For this
study, rice husk ash was obtained from paddy mill, Roorkee.
It was fine grained siliceous in nature light weight and grey in
color. The physical properties are given in Table 4.
4.4 Bagasse ash
The bagasse ash used in the study was brought from Uttam
Sugar Mill, Deoband near Roorkee. The ash was obtained
from this mill at a boiler temperature of 750-800C. The
bagasse ash produced at the plant was about 2-2.5 % of the
bagasse used in boiler.
Figure 5 presents the TGA curve for the baggase ash. The
bagasse ash samples were subjected to TGA in a dry air
atmosphere at heating rates: 10C/min. Samples weighing
80 mg were heated at preselected heating rates from ambient
temperature to 850C in a Mettler thermal analyzer. The
continuous records of weight loss and temperature were
obtained and analyzed using a Mettler thermal analysis
software system to determine the weight loss with increase in
temperature.
Properties
Values
SiO2 (%)
72.24
CaO (%)
4.12
MgO (%)
1.7
Fe2O3 + Al2O3 (%)
7.2
Specific Gravity
1.87
Lime Reactivity (kg/cm2)
34
Uniformity coefficient (Cu)
2.00
Coefficient of curvature (Cc)
0.964
The TGA curve shows that in the first stage at the

temperature range of 0-250C the mass loss was hardly 0.81%. The major part of the mass loss took place in the second
stage, in the range 250-550C indicating a steep fall in the
TGA curve. The mass loss of 10% in this stage was due to the
removal of volatile matter. In the third stage the mass loss
was only 0.4% at temperature above 550C. The minimum
temperature required for the oxidation of fixed carbon ranges
from 500 to 550C. This implies that the minimum furnace
temperature should be above 550 C for complete
combustion of bagasse ash to obtain silica.
Fig 5 TGA curve for baggase ash

Fig 4 Grain size distribution for Rice Husk Ash
73
Carbon residues measured after burning indicated that

combustion was almost complete at 600oC but crystallization
of minerals occurred at temperatures higher than 500oC. It
has been reported that at burning temperatures up to 550oC
silica was in amorphous form and silica crystals grew with
time of incineration. Considering the rate of burning and
residence time, the suitable burning condition was identified
as 500oC. At this condition brownish white colour indicated
complete burning of bagasse ash.
Fig 6 Grain size distribution for Baggase Ash

Figure 6 shows the grain size distribution of open burning
followed by controlled burning at 500oC in a thermostatically
controlled electronic furnace of bagasse ash. The flat S-curve
represents the bagasse ash contains the particles of different
size in good proportion such as well graded. The grain size
distribution curve reveals that bagasse is mostly fine grained
which is better for soil interlocking.
To obtain SCBAfor this study, burning was carried out in two
stages as open burning followed by controlled burning at
500oC in a thermostatically controlled electronic furnace.
Table 5 and Table 6 show the chemical and physical
properties of the BArespectively.
Table 6 Physical properties of bagasse ash
74
Values (%)
65.27
4.74
3.11
11.16
1.27
0.52
1.89
5.26
2.11
Values
Specific gravity
2.38
Density (g/cm3)
2.64
Surface area (cm2/gm)
4894
Uniformity coefficient (Cu)
4.61
Coefficient of curvature (Cc)
1.476
As shown in Table 6, BA has low density and higher surface

area (Blaine surface area). Particle size analysis of ash
samples indicated that average size of the ash particles was
30m and 85% particles were of size less than 45m. Kraiwood
et al. [13] has reported that large surface area favours the
pozzolanic reactivity of amorphous silica and other
minerals. The combined chemical composition; (SiO2 +
Al2O3 + Fe2O3) > 70% and CaO >10% testified the pozzolanic
and cementitious nature of SCBA as per ASTM C- 618- 03
specifications.
5. Conclusions
The physical, chemical and strength characteristics of
various waste materials like building construction and debris
(Recycled aggregates), Fly ash, Rice husk ash and Bagasse
ash were tested in laboratory and the results were found to be
within the specified limits. Salient outcomes of the study are
enlisted in following section.
1.
Recycled aggregates found to be relatively soft

compared with conventional aggregate and can be used
as a sub-base material but not in base course and
wearing course. Water absorption of waste material is
found to be high compared with conventional
aggregate. The results of recycle aggregate are found to
be within the limits as per MOSRT&H specifications.
2.
The TGA curve shows that the mass loss for fly ash
even after heating at a high temperature of 650C is
only about 1.7% and 1.5% of mass loss for rice husk
ash after heating at a temperature of 600C which
shows that the selected ash is free of solid (fixed)
carbon. Hence no further heating is required. In case of
bagasse ash, it has been reported that at burning
temperatures up to 550oC silica was in amorphous form
and silica crystals grew with time of incineration.
Considering the rate of burning and residence time, the
Table 5 Chemical properties of bagasse ash

Chemical composition
Silicon dioxide (SiO2)
Aluminum oxide (Al2O3)
Ferric oxide (Fe2O3)
Calcium oxide (CaO)
Magnesium oxide (MgO)
Sulfur trioxide (SO3)
Sodium oxide (Na2O)
Potassium oxide (K2O)
Loss on ignition
Physical properties

S & T Review
suitable burning condition was identified as 500oC.

3.
Fly ash tested in the laboratory was found to have

higher silt content which makes it suitable for
stabilizing soil with low silt content like clay. The grain
size distribution curve reveals that fly ash is fine
grained which is better for soil interlocking.
4.
According to the results, RHA includes 72.24% silica.

It mean that the ash is approximately net silica.
Therefore, it can cause a pozzolan action with soil.
5.
Bagasse ash contains the particles of different size in

good proportion such as well graded which is better for
soil interlocking. It has low density and higher surface
area which favours the pozzolanic reactivity of
amorphous silica and other minerals. The combined
chemical composition; (SiO2 + Al2O3 + Fe2O3) > 70%
and CaO >10% testified the pozzolanic and
cementitious nature of SCBA as per ASTM C- 618- 03
specifications.
6.
Based on extensive laboratory investigations, it is

concluded that the selected four waste materials viz. C
& D waste, fly ash, rise husk ash and bagasse ash are
found suitable for road construction with respect to
their physical/engineering and chemical
characteristics.
References
[1] Cokca, E., Use of Class C fly ashes for the
stabilization of an expansive soil, J. Geotech.
Geoenviron. Eng., vol. 127(7), pp. 568573, 2001.
[2] Chakraborty, D. K., and Dasgupta, S. P., Randomly
reinforced fly ash foundation material, Proc., Indian
Geotechnical Conf., vol. 1, pp. 231235, 1996.
[3] Head WJ, and Seals RK., Design of experimental haul
road and parking lot facilities utilizing power plant
ash, Proceedings of fifth international ash utilization
symposium, pp. 67993, 1979.
[4] Kumar, P. and Singh, S. P., Fiber-reinforced fly ash
subbases in rural roads, Journal of Transportation
Engineering ASCE, vol. 134 (4), pp. 171-180, 2008.
[5] Govindarao, V. M. H. Utilization of rice husk - a

preliminary analysis. J. Science Industrial Research
39, 495-515, 1980.
[6] Alhassan M, Permeability of lateritic soil treated with
lime and rice husk ash, Assumption University
Journal of Thailand, vol. 12(2), pp. 115120, 1999.
[7] Muntohar A S, Uses of RHA to Enhanced LimeStabilized Clay Soil, International Conference of
Geotechnical Engineering, University of Sharjah,
UnitedArab Emirate, 2004.
[8] Noor M, Abdol Aziz A, Suhadi R, Effects of cement
rice husk ash mixtures on compaction, strength and
durability of Melaka Series lateritic soil, Journal of
Institution of Engineers Malaysia, vol. 47, 61 64,
1990.
[9] Ahmad, S.F. and Shaikh, Z., Portland pozzolana
cement from sugar cane bagasse ash, In: Lime and
Other Alternative Cements, Intermediate Technology
Publications, London, pp. 172179, 1992.
[10] Ganesan, K., Rajagopal, K. and Thangavel, K.
Evaluation of bagasse ash as supplementary
cementitious material. Cement & Concrete
Composites, Vol. 29, 515524, 2007.
[11] Goyal, A., Anwar, A.M., Kunio, H. and Hidehiko, O.
Properties of sugarcane bagasse ash and its potential
as cement - Pozzolana binder. Twelfth International
Colloquium on Structural and Geotechnical
Engineering, 10-12 Dec. Cairo Egypt, 2007.
[12] Khobklang, P., Nokkaew, K. and Greepala, V. Effect
of bagasse ash on water absorption and compressive
strength of lateritic soil interlocking block.
Proceedings of the conference on Excellence in
Concrete Construction through Innovation, Kingston
University, United Kingdom, 9 - 10 Sept,181-185,
2008.
[13] Kraiwood, K., Chai, J., Smith, S. and Sekusan, C. A
study of ground coarse fly ashes with different
finenesses from various sources as pozzolanic
materials, Cem. Conc. composites, 23, 335-343,
2001.
75
Prophylactic Efficacy of Septilin on Skin of Rats

Exposed to Gamma Radiation
Enas, A. Khalil*
adiotherapy caused skin injury in patients. Radiation leads to liberation of free radicals. Lipid peroxidation and its
consequences; membrane destruction, are resulted from the action of the free radicals. Septilin is a polyherbal
preparation supplied by the Himalayan Drug Co. (Bombay, India). It has been reported to possess antioxidant property, antiinflammatory, anticutaneous infections, wound healing properties and also prescribed as a health supplement. This
investigation aims to study the protective effect of Septilin on skin of rats exposed to gamma radiation. Fourty male Swiss rats
were used. The rats were divided into four groups: control, whole body gamma irradiated (4GY), Septilin (100mg/kg bw. p.o.)
treated group for a week and Septilin (100mg/kg bw. p.o.) treated group for a week before (4GY) irradiation. Samples of
dorsal skin of rats were processed for haematoxylin and eosin, Mallory's trichrome for determination of collagen fibers and
with bromophenol blue for determination of total protein. The results indicated that many histopathological changes in the
skin of rats exposed to gamma ray (4GY) varied between thin, corrugated and ruptured epidermis. Disturbed hair follicles
diminished sebaceous glands, presence of degenerated areas in the dermal layer and atrophied dermal muscle fibers.
Increased collagen fibers were noted in the epidermal and dermal layers, while degenerated areas appeared negatively stained
of collagen. As well as decreased total protein content in epidermal and dermal layers. Pretreatment of rats with Septilin
(100mg/kg b. w p .o. for 7 days) before - irradiation showing numerous hair follicles in the dermal layer with their sebaceous
glands, but few follicles are still affected, collagen fibers and total protein nearly returned to their normal content. In
conclusion Septilin showed noticeable recovery in the skin tissue.
Keywords: Septilin, Gamma Radiation, Male Rats, Skin
1. Introduction
Radiation interacts with medium through which it passes by
direct and indirect processes to form ion pairs. These ion
pairs rapidly interact with themselves and other surrounding
molecules to produce free radicals. Lipid peroxidation and
its consequences; membrane destruction, are resulted from
the action of free radicals (Walden and Farzaneh, 1990).
Free radicals can be caused by expending high levels of
energy, exposure to radiation, toxic chemicals or
environmental pollutants, and inadequate dietary intake of
antioxidant. Free radicals are normally held in balance by
the body antioxidant defense system, but with an excessive
amount of free radicals, oxidative cell damage can occur. The
generation of free radicals leads to a wide array of skin
diseases (Giampieri et al, 2012). Also the National Cancer
Institute stated that all the external radiation treatments cause
some kind of skin related side effects. Septilin is a plant
mixture extract was supplied by the Himalayan Drug Co.
* Histology Department, National Organization for Drug control
and Research, 12553, Cairo, Egypt,
Email- enasamk@gmail.com
76
(Bombay, India) in a form of tablets or syrup. Its main

ingredients are: Balsamodendron mukul /Commiphora
mukul which is known to increase white blood cell counts,
possess strong disinfecting properties and normalizes lipid
metabolism (Dev,1997); Tinospora cordifolia, inhibiting
growth of bacteria and enhancing the buildup of immune
resistance; Rubia cordifolia purifying blood, immune
regulator and regulate blood pressure(Nair et al,1996);
Emblica officinalis, is an immune function booster,
antioxidant and effective against respiratory complaints;
Moringa pterygosperma functions as antibiotic(Armado et
al,1991) and Glycyrrhiga glabra promotes gastrointestinal
health(Isbruker and Burdock,2006).This herbal preparation
proved to possess immunomodulatory, anti-inflammatory in
vivo(Daswani andYegnanarayan,,2002)), (Khanna and
Sharma,2001)and in vitro(Varma et, al 2011).Also It possess
wound healing properties (Udapa et al.,1989).It is said to be
helpful in treating Gram-positive as well as Gram-negative
infections (Gadekar et al.,1986). Researches displayed that
Septilin is effective in chronic stubborn urinary tract
infection (Bhasin,1990),Tonsillitis (Gadekar et al.,1986
)and infective dermatose (Sharma et al.,1986). It is said to be
S & T Review
useful in the management lower respiratory tract infection (

Bhasin,1990),allergic disorders of upper respiratory tract
(Sanjit,1992),skin and soft tissue infections
(Srivastava,1985 ),infective and inflammatory conditions of
the eye (Sarabhai and Sandeep,1989),Bone and joint
infections ( Ramesh ,et al,1989 ) and has a prophylactic of
diabetic foot ulcer (Singh,2001 ). As well as it protects mice
against whole body lethal dose of gamma radiation(Jagetica
and Baliga,2004) .This research designed to study the
protective effect of administering Septilin on skin injury
caused in rats exposed to whole body gamma radiation.
2. Materials and methods
haematoxlin and eosin (Drury and Wallington, 1980), others

were stained with Mallory's trichrome stain (Dunn, 1974)
for determination of collagen fibers, others were stained with
bromophenol blue (Mazia et al,1953)for determination of
total protein and examined microscopically. The
histochemical interpretation was done using computer image
analyzing system (Leica Model).Estimation of the optical
density of thirty cells in each group was made. The data
obtained were statistically analyzed according to Sendecor
and Coebram (1969).Differences between the group means
were assessed using T-test. P 0.05 was considered
significant.
2 .1Animals
Fourty male Swiss rats weighing 120-140gms were used.
The rats were housed under good hygienic environmental
condition at the National Centre for Radiation Research and
Technology, Cairo, Egypt. The rats were divided into four
groups: control group (referred to as control), whole body irradiated group (4GY) , Septilin (100mg/kg bw. p.o.)
treated group for a week and Septilin (100mg/kg bw. p.o.)
treated group for a week before (4GY) irradiation.
2.2 Plant materials
Septilin is a plant mixture extract was supplied by the
Himalayan Drug Co. (Bombay, India) in a form of tablets. Its
ingredients are (mg): Balsamodendron mukul /Commiphora
mukul,162 ; Tinospora cordifolia,32; Rubia cordifolia,32;
Emblica officinalis,16; Moringa pterygosperma,16; and
Glycyrrhiga glabra,6. Septilin pills were ground, dissolved
in water and orally administered to rats. The dose of Septilin
used in this study was 100 mg/kg b. wt /day for a week before
exposure to gamma radiation.
2.3 Irradiation
Whole body irradiation was achieved through Cesium 137Gamma cell-40 manufactured by the Atomic Energy of
Canada (Ltd.).The dose rate was 1GY/1.42 min at the time of
the experiment.
2.4 Histological and histochemical investigations
Samples of dorsal skin of rats were used for
histopathological studies, fixed in 10% formalin at room
temperature for 24 h, dehydrated through an upgraded series
of ethanol and embedded in paraffin wax. The paraffin
sections were cut at 5m thickness and stained with
3. Results
Skin in control group showed the normal histological
structure of epidermis with continuous keratin, dermis, hair
follicles, blood vessel, sebaceous glands, muscle fibers and
sweet glands at the deep dermis (Fig. 1).Normal distribution
of collagen fibers in the dermal layer with thin layer in the
basement membrane (Fig. 2). As well as normal content of
total protein in epidermal and dermal layers (Fig. 3).The skin
of irradiated rats appeared as highly corrugated, thin and
ruptured epidermis, with highly distributed and atrophied
hair follicles, atrophied muscle fibers in dermal layer and
diminished sebaceous glands with discontinuous
keratin(Fig.
4).Significant increase recorded in collagen
fibers in epidermal and dermal layers meanwhile
degenerated areas appeared negatively stained compared to
control(Fig.
5), Table 1 and histogram(1).Significant
decrease in total protein recorded in epidermal and dermal
layers while degenerated areas appeared negatively stained
compared to control(Figure6) , Table
1 and Histogram
1. Septilin treated group (100mg/kg bw p. o .for 7days)
showing well developed epidermis, dermis and keratin
layers and hair follicles(Figure7a,b), normal content of
collagen fibers in the dermal layer and beneath the epidermal
layer (basement membrane)(Fig 8), normal distribution of
total protein in epidermal and dermal layers compared to
control (Fig 9), Table 1 and histogram(1).Pretreatment of
rats with Septilin (100mg/kg bw p. o .for 7days) before irradiation showing numerous hair follicles in the dermal
layer with their sebaceous glands, but few follicles are still
affected (Figure10). Non-significant change recorded in
collagen fibers and total protein (Fig 11&12), Table 1 and
Histogram 1.
77
4. Discussion
All the external radiation treatments cause some kind of skin
related side effects. Gamma radiation is known to liberate
free radicals in the tissues of experimental rats (Batist etal.,
1986). The generation of free radicals leads to a wide array of
skin diseases (Giampieri et al, 2012). In the present
investigation histological evaluation indicated a significant
change in tissue architecture of skin of rats, which are
exposed to gamma ray (4GY). The skin appeared thin,
corrugated with ruptured epidermis and discontinuous
keratin. Disturbed hair follicles, diminished sebaceous
glands, presence of numerous fatty cells, degenerated areas,
atrophied dermal muscle fibers and increased collagen fibers
in the epidermal and dermal layers, while degenerated areas
appeared negatively stained for collagen and decreased total
protein content in epidermal and dermal layers . These
changes may be due to the liberation of excessive amount of
free radicals and oxidative cell damage (Hopewell, 1990).
Skin of rats administered Septilin (100mg/kg bw) showed
well developed epidermal ,dermal ,keratin layers ,hair
follicles and normal content of collagen fibers in the dermal
layer and beneath the epidermal layer (basement
membrane)and also normal content of total protein .This
result ensures the importance of septiin to skin (Sharma et
al.,1986). Administration of Septilin to rats before exposure
to gamma radiation, showing numerous hair follicles with
their sebaceous glands .Meanwhile few follicles still affected
with absence of fatty cells observed in the radiated group.
Collagen fibers appeared near to their normal content,
together with normal distribution of total protein in
epidermal and dermal layers. This result may be due to its
antioxidant property which could scavenge free radicals,
immunomodulatory, and anti-inflammatory activity and also
it prescribed as a health supplement. As well as it is useful for
skin (Daswani andYegnanarayan,2002, Khanna and
Sharma,2001)
keratin (k),dermis(D),hair follicles(hf),blood

vessel(bv),sebaceous glands(sb) and sweet glands(sg)at the
deep dermis(H&Ex200)
Fig. 1 A photomicrograph of a section in control rat skin

showing normal structure of epidermis (ep) with continuous
78

S & T Review
79
Table 1. The quantitative measurements of the color

density(Pixel) of total protein and collagen fibers in the skin of
control and treated groups of male rats.
10
P> 0.01 Highly Significant, P> 0.05 Significant,

P< 0.05 Non significant
Histogram 1 Quantitative measurements of control

and treated groups of male rats.
5. Conclusions
In conclusion Septilin showed noticeable recovery in the
skin tissue after exposure to gamma radiation.
80

S & T Review
Acknowledgment
The author thank to Professor Mona Ahmed El- Gawish in
Biology Radiation Department at the National Centre for
Radiation Research and Technology, Cairo, Egypt for her
help to achieve this research.
References
[1] Armado, C.; Cabrera,O.; Morales, O and
Menda,P.(1991): Pharmacological properties of
Moringa Oleierra,1Preliminary screening for
a n t i m i c r o b i a l a c t i v i t y. J o u r n a l o f
Ethnopharmacology33, 213-216.
[2] Bhasin, R. (1990): Clinical evaluation of Septilin in
chronic bronchitis .Ind.Practit. 43:83-87
[3] Daswani, B. and Yegnanarayan, R.(2002):
Imunomodulatory activity of Septilin a polyherbal
preparation. Phytother. Res., 16:162-165
[4] Dev, S. (1997):Ethnotherapeutics and modern drug
development: The potential of Ayurveda,Current
Science,73:909-928
[10] Isbruker, R, and Burdock,G(2006): Glycyrrhizin

Toxicity, Regulatory Toxicol Pharmacol,46:167
[11] Mazia, D.;Brewer ,P. and Alert, M.(1953):The
cytochemical staining and measurement of protein
with bromophenol. Biol Bull, 104:57-76
[12] Nair, S.; Nagar, R and R. Gupta,(1996):Dietary antioxidant phenolics and flavonoids in coronary heart
disease . Indian Heart Journal P. 545.
[13] Ramesh, S., Thakur, R.and Bahanu, A. (1989): Septilin
in compound infected injuries. Probe (28),2,138-141
[14] Ross, D. (1984): The anti-infective and anti-bacterial
efficacy of Septilin .Probe, 23:84-87.
[15] Sarabhai,K. and Sandeep,G.(1989):Role of Septilin in
prevention of acute dacryocystitis.Probe(29),1,29-31.
[16] Sanjit, R.(1992):Clinical evaluation of Septilin in
allergic rhinitis. Probe (31)2, 146-156
[17] Sendecor, G.W. and Coebram,W.G.(1969 ): In
''Statistical Methods'' 6th ed . Iowa State Univ. Press.
Anes, Iowa, USA, P.70.
[5] Drury, R .A. and Wallington,E.A.(1980):Carleton's

Histological Techniquies,5th ed, OxfirdUniversity
Press. London, NewYorkToronoto, 1:653-661
[18] Sharma,S.; Agarwal,H.and Dharam,P.(1986): Septilin

in infective dermatoses.Probe,25:156-161.
[6] Dunn,W.L.(1974): Hand book of Histopatholgical and

Histochemical Techniques,3 rd ed., RED WOOD
BURN Limited . Great Britain. 415-416
[19] Singh,K.(2001): Role of combination of multiple

herbal drugs(Septilin) in the prophylaxis of diabetic
foot ulcer double blind trial .Indian Journal of
Dermatology (46),2,83.
[7] G a d e k a r, H ; Vi j a y, A . ; K o m a w q a r, J . a n d
Sonawar,S.(1986): Septilin in acute tonsillitis in
children below 12 years of age Probe,25:164-165
[8] Giampieri F,Alvarez-Suarez JM, Tulipanis, Gonzaitez
Paramas AM,Santos-Buelga C .Bompadre S and
QulesJl(2012):.Photoprotective
Potential of
Strawberry(Fragarria x Ananassa) Extract against UV
Irradiation damage on human fibroblasts . J, Agric
Food chem. Feb.3
[9] Horobin,R,W. and Bancraft ,J.B.(1998):Trouble
Shooting Histology Stains. Churchill Living Stone
Medical Division of professional limited.P.56.
[20] Srivastava, A.1985: Clinical trial of Septilin on ulcers

(specific and non specific and surgical wounds .The
medicine and surgery, 25, 10, 31.
[21] Udupa,A.L .and Kulkami ,D.R. (1989):Wound healing
profile of Septilin Indian .J.Physiol.Pharmacol JanMar,33(1):39-4
[22] Varma, RS; Ashok G, Vidjeishanker S, Patki P. and
Nadakumar, KS (2011):Anti-inflammatory properties
of Septilin in lipopolysaccharide activated monocytes
and macrophage.Immunopharmarcol
immunotoxicol,2011 Mar;33(1):55-63.
81
An Efficient Method for the Calculation of

Eigenvector Design Sensitivities
*
Shibinlal M.K , Aminudin. Hj.Abu
his paper describes a general approach for incorporating the eigenvector normalization condition in the computation of
eigenvector design sensitivities. Design sensitivities for the eigenproblem has received much attention over the past four
decades, however, with little exception, previous work has focused primarily on mass normalized eigenvectors alone. The
approach presented here considers conservative symmetric systems with distinct eigenvalues and addresses first-order
eigenvalue and eigenvector design sensitivities. A rigorous treatment of the eigenvector normalization condition is provided
by using its design derivative to eliminate the singularity issue in the eigenvector design sensitivity system of equations. The
effect of re-scaling eigenvectors is also considered, as is the related computational effort.
Examples illustrate the
computational procedure.
Keywords : Eigenvalue, Sensitivity, Normalization, Eigenvector
1. Introduction
Design sensitivities, often referred to as design derivatives,
have become an integral part of many engineering design
methodologies including design optimization, inverse
analysis, system identification, optimal control, response
surface analysis, damage detection, and approximate reanalysis techniques, among others. Methods for evaluating
design sensitivities have received much attention over the
past four decades, particularly those related to the eigenvalue
and eigenvector problem. While computing eigenvalue
design sensitivities is straight-forward, evaluating
eigenvector derivatives poses several challenges, due in part
to their non-uniqueness, and the possibility of
encountering non-distinct eigenvalues.
Of particular
interest in this paper is the role of eigenvector normalization
on eigenvector design sensitivities.
Design sensitivity analysis is used to quantify the
relationship between parameters that define a design and
computed outputs that are used to measure its performance.
A significant body of research has been devoted to the
computation and application of design sensitivities where
early work focused on linear
Structural systems (see e.g., Adelman and Haftka [1] or
Haug, et al., [2]). This area of research has since expanded to
a broad range of engineering applications as evidenced by
* MJIIT, Universiti Technologi, Malaysia,
Kuala Lumpur, Malaysia,
E-mail: shibinnss@gmail.com
82
the references given in Refs. [3] and [4]. For example, a

detailed discussion of developments related to eigenvalue
design sensitivity analysis appears elsewhere (cf. Adelman
and Haftka [1] or Seyranian, et al., [5]). With little exception
(see e.g.
Adhikari, et al., [6] or Wang [7]), eigenvalue design
sensitivity analyses for vibration problems assume that the
eigenvectors are mass normalized as first described by Fox
and Kapoor [8]. The unity magnitude normalization has also
been employed in the buckling problem as given by Wittrick
[9].
Fox and Kapoor [8] presented two approaches for computing
eigenvector sensitivities for real symmetric systems. Their
first formulation included the design derivative of the mass
normalization condition directly in the computation, which
has since seen little use since this approach destroys the
bandedness of the original eigensystem. The second
approach given in Ref. [8] defines the eigenvector
sensitivities as a linear combination of the system's
eigenvectors. Asimilar approach for eigenvectors with unity
magnitude appeared in Rudisill and Chu [10]. These
methods have been used as a basis for other developments
including eigenvector derivatives for nonsymmetric systems
[11] and non-distinct eigenvalues (see e.g., Frisswell [12] or
Lee and Jung [13]). Unfortunately, evaluations proposed in
Ref. [8] have enjoyed little direct application since they
require the computation of all eigenvectors.
This paper presents a comprehensive approach to include the
eigenvector normalization condition in its design sensitivity
S & T Review
analysis for conservative symmetric systems with distinct

eigenvalues. The approach is based on the premise that the
normalization condition employed to uniquely define an
eigenvector must also be used in the sensitivity calculation.
First- order sensitivities are considered and a procedure is
described for computing eigenvector sensitivities when the
eigenvector is re-scaled. Examples illustrate our evaluation
procedure.
2. The Eigen Problem
To establish a common basis for the eigenvector derivative
evaluations to follow, we first consider the well-known
eigen-problem employed in the analysis of free undamped
vibration and structural stability. The eigen problem derived
from these systems may be written in terms of the eigenvalue
and eigenvector (or mode shape) as
(1)
[K M] = 0
where K is the stiffness matrix and M is the mass matrix in

a vibration analysis or the geometric stiffness matrix in a
buckling analysis, both of which are of dimension N N
where N is the number of free degrees-of-freedom in the
discretized system of equations. In the vibration problem,
the eigenvalue = 2 is the square of the natural frequency
and in a buckling analysis, is the load factor. Here we
consider that both K and M are symmetric and K is positive
definite, which yields real .
Equation (1) has a nontrivial solution only when the
determinant |K M| = 0, providing eigenvalues I, I = 1, 2. . .
N. When each distinct I is substituted into Eq. (1), the matrix
KI M is rank deficient by order 1, rendering each nontrivial
I being determined within an arbitrary scaling factor. A
normalization condition for mode I , defined here by the
scalar function G(I ) = 0, is required to uniquely define each
eigenvector.
Common normalization conditions for
computing eigenvectors are:
Where the p-th component of I is prescribed a value of in

Eq. (3), and the magnitude of I (given as ||I||) is in Eq. (4).
Note that subscripts have been introduced for each function
G to identify the normalization condition of interest in the
discussion to follow, and that other normalizations may be
considered.
Any one of the normalization conditions in Equation (2-4)
may replace an equation in the system Equation (1) to
remove the rank deficiency of [K I M]. Note that the
linearity of the eigenproblem system of equations in
Equation (1) is lost, however, if Equation (2) or (4) is used to
define the eigenvector normalization. Alternatively, when G2
in Equation (3) is employed to specify a component of I, the
linearity of the system in Equation (1) is retained, and the
system matrix is of full rank. Furthermore, we note that its
bandwidth remains unaltered when Equation (3) is
substituted for the p-th row in Equation (1).
3. Eigenvalue and Eigenvector
Design Sensitivity Analysis
Design sensitivities quantify the relationship between design
variables bi , i = 1, 2, . . . , n, which form the n-dimensional
vector b, and the computed eigenpairs (I , I ), I = 1, 2, . . . , N
. To better illustrate the design sensitivity analyses to follow,
we first emphasize the dependence of the eigenproblem in
Eq. (1) on the design b as
[K(b)I(b)M(b)]I(b)=0
(5)
Which is differentiated with respect to the i-th design

variable bi and rearranged to yield
(6)
in a manner that is similar to that presented by Nelson
[14]. Equation (6) may be used to compute eigenvalue and
eigenvector design derivatives as demonstrated below.
The eigenvalue design derivative is derived in a manner
similar to that given elsewhere (see e.g., Nelson [14]) by first
pre-multipying Eq. (6) by the eigenvector Following
mathematical manipulation which employs Eq. (1) and
exploits the symmetry of M and K, we obtain the well know
eigenvalue design sensitivity
(7)
83
Where
84

S & T Review
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
85
5. Numerical Example
86

S & T Review
Table 2 Eigenvector re-scaling factors and their design normalized by

G1 - G3 (from Eqs. 2-4, respectively) for = 1 and = 1.
87
5. Conclusion andAcknowledgment
A comprehensive eigenvalue and eigenvector design
sensitivity evaluation method is presented that directly
incorporates the choice of normalization condition used
when computing an eigenvector. Of particular interest here
are the design sensitivities for conservative, symmetric
systems with distinct eigenvalues. The approach is based on
assuming that the normalization condition used in the
underlying eigenproblem solution be used in the design
derivative calculations. A procedure is also provided for
recomputing eigenvector design derivatives once the
eigenvectors themselves are re-scaled.
Three specific normalization are considered in our
derivation, however, the approach is general enough to
include other eigenvector normalizations of interest.
Examples illustrate the calculation procedure, and the effect
of normalization condition on eigenvector sensitivities.
References
[1] Adelman, H.M and R.T. Haftka. Sensitivity Analysis
of Discrete Structural
System. AIAA Jn. ,
24(5):823832, 1986.
[2] Haug, E. J., K. K. Choi, and V. Komkov. Design
Sensitivity Analysis of Structural Systems. Academic
Press, NewYork, NY, 1986.
[3] Tortorelli, D. A. and P. Michaleris. Design Sensitivity
Analysis: Overview and Review. Inverse Problems in
Engineering,1(1):71105, 1993.
[4] Smith, D.E. and D.A. Tortorelli. Use of Modeling in
Design. In The Encyclopedia of Materials: Science
and Technology, volume 6, pages 57335743. Elsevier
Science, Oxford, 2001.
[5] Seyranian, A.P., E. Lund, and N. Olhoff. Multiple
Eigenvalues in Structural Optimization Problems.
Structural Optimization, 8:207227, 1994.
Eigenderivative
[6] Adhikari, S. and M.I. Friswell.
Analysis of Asymmetric Non-conservative Systems.
International Journal for Numerical Methods in
Engineering, 51:709733, 2001.
[7] Wang, B.P. Eigensolution Derivatives for Arbitrary
Normalized Modes. Shock and Vibration ,
1(4):395399, 1994.
88
[8] Fox, R.L and M.P. Kapoor. Rate of Change of

Eigenvalues and Eigenvectors. AIAA Jn. ,
6(12):24262429, 1968.
[9] Wittrick, W.H. Rates of Change of Eigenvalues, with
Reference to Buckling and Vibration Problems.
Journal of the Royal Aeronautical Society ,
66:590591, September 1962.
[10] Rudisill, C.S and Y. Chu. Numerical Methods for
Evaluating the Derivatives of Eigenvalues and
Eigenvectors. AIAA Jn., 13(6):834837, 1975.
[11] Rogers, L.C. Derivatives of Eigenvalues and
Eigenvectors. AIAA Jn., 8(5):943944, 1970.
The Derivatives of Repeated
[12] Friswell, M.I.
Eigenvalues and Their Associated Eigenvectors.
ASME Jn. of Vibration and Acoustics, 118(7):390397,
1996.
[13] Lee, I.W. and G.H. Jung. An Efficient Algebraic
Method for the Computation of Natural Frequency and
Mode Shape Sensitivities - Part 1. Multiple Natural
F r e q u e n c i e s . C o m p u t e r s a n d S t r u c t u re s ,
62(3):437443, 1997.
[14] Nelson, R.B. Simplified Calculation of Eigenvector
Derivatives. AIAA Jn., 14(9):12011205, 1976.
[15] Friswell, M.I. Calculation of Second and Higher Order
Eigenvector Derivatives. AIAA Jn. of Guidance,
Control and Dynamics, 18(4): 919921, 1995.
[16] Lee, I.W. and G.H. Jung. An Efficient Algebraic
Method for the Computation of Natural Frequency and
Mode Shape Sensitivities - Part 1. Distinct Natural
F r e q u e n c i e s . C o m p u t e r s a n d S t r u c t u re s ,
62(3):429435, 1997.
[17] Murthy, D.V and R.T. Haftka. Derivatives of
Eigenvalues and Eigenvectors of a General Complex
Matrix. InternationalJn. for Numerical Methods in
Engineering, 26:293311, 1988.
[18] Vanderplaats Research and Development, Colorado
Springs, Colorado. GenesisTM , http://www.vrand.com,
2006.
[19] Haftka, R. T. and Z. Gurdal. Elements of Structural
Optimization. Kluwer Academic Publishers, Boston,
MA, third edition, 1992.
S & T Review
Correlation of Bulk Density and Abrasion

Resistance of Some Igneous and Metamorphic Rocks
1
Ademeso, Odunyemi Anthony , Adekoya, Adeyinka John , Olaleye, Boluwaji Muraina
ix igneous and metamorphic rock types were investigated for their petrographic, bulk density and abrasion resistant
characteristics. Petrographic microscope to which a digital camera was attached, balance with necessary equipment set up
for Archimedes' principle measurement and abrasion tester were used to determine the petrographic, bulk density and abrasion
resistant characteristics respectively of the rock types after various sample preparations. Generally, quartz, biotite,
plagioclase, microcline and hornblendeconstitute the major minerals in the thin sections of the rock types and the
microstructures that were exhibited include cleavage, distorted cleavage, mineral inclusions, micro-cracks, bent twin
lamellae, mymerkite and compressed twin plane. Biotite granite has the lowest bulk density of 2.63g/cc while the charnockitic
rock has the highest of 2.81g/cc. The result of abrasion resistance shows that granite gneiss and porphyritic biotite granite have
very low values of 26.50 and 26.65 respectively while lamprophyre and charnockitic rock have very high values of 34.40 and
33.40 respectively. A very high correlation coefficient of 0.9179 was recorded for the relationship between the bulk density
(D)and abrasion resistance (Ha)of the rock types. A model relating the two properties, Ha = 46.79D 97.43, was derived.
Conclusively, bulk density has direct relationship with the abrasion resistance of igneous and metamorphic rocks and it is
therefore used to estimate it.
Keywords: Bulk Density, Abrasion Resistance, Model, Correlation Coefficient
1. Introduction
It has been severally reported that the determination of the
mechanical properties of rocks is tedious, time consuming
and expensive ([1], [2]).Abrasion resistance which measures
the toughness of the rock is one of such mechanical
properties which determination was so described. Bulk
density on the other hand is one of the physical properties
which researchers find a lot easier, faster and cheaper to
determine. The basic uses (particularly as dimension stones
or polished slabs) to
which rocks are put depend on its toughness (hardness),
which abrasion resistance measures. Any means which
makes for easy and quick estimation of this property
particularly during reconnaissance survey is considered
essential. The estimation of abrasion resistance of samples of
rocks will, therefore, help to delineate at reasonable cost and
time, the rocks and outcrops that may be available for
assessment. The bulk density and abrasion resistance of
some selected rock types were therefore studied and
correlated. A relationship was established to help in
estimating the abrasion resistance from the bulk density. The
study area (Fig.1) lies within longitudes 6000'E and 6010'E

and latitudes 7005'N and 7020'N. Six different igneous and
metamorphic rock types were selected from Akure and
Igarra areas of southwestern Nigeria. The rock types include
gneiss (Ak001), granite gneiss (Ak002), Charnockitic rock
(Ak003), biotite granite (Ig001), porphyritic biotite granite
(Ig002) and lamprophyre (Ig003).
Fig. 1 Geological Map of South-western Nigeria Showing

Study Areas (Ademeso, 2011)
1. Geology Department,AdekunleAjasin University, Akungba-Akoko,Ondo State, Nigeria, Email- tonyademeso@gmail.com

2. Applied Geology Department, The Federal University of Technology, Akure, Ondo State, Nigeria, E-mail: yinkadekoya@yahoo.com
3. Mining Engineering Department, The Federal University of Technology,Akure, Ondo State, Nigeria, E-mail: bmolaleye@futa.edu.ng
89
Generally, the areas fall within the basement complex of

southwestern Nigeria which has been variously described by
workers which include ([3], [4], [5], [6]).
2. Methodology
2.1 Field study
The field study was carried out basically to select rock types
for the work and consequently, samples were collected from
the outcrops of the different rock types with the aid of sledge
hammer and Global Positioning System (GPS).
2.2 Petrographic study
Thin sections were prepared from the samples that were
collected from the field and were studied with the
petrographic microscope. Photomicrographs were taken and
analysed with the aid of ImageJas described in [7].
2.3 Determination of specific gravity/bulk density
The specific gravity of the samples was determined using
Archimedes principle and then converted to bulk density as
described in [7].
2.4 Determination of abrasive resistance
The procedure is as stated in [8]. Block specimens of 30mm
by 30mm by 60mm were prepared from the rock samples.
The specimen was weighed to precision and in this case, the
weight was taken twice with the average of the two regarded
as the specimen weight (Ws). The specimen was then placed
in the abrasion testing apparatus and abraded for 225
revolutions. The equipment used for this test was set to
50rpm and this translated to abrading the specimen for 4mins
15sec. The specimen was removed from the machine,
brushed free of dust and weighed to the same precision as for
the original weight [9]. The abrasion resistance was
calculated using the formula below [9].
whereHa = abrasion resistance; G = bulk specific gravity of

the sample; Ws = average weight of the specimen (original
weight plus final weight divided by 2);Wa = loss of weight
during the grinding operation.
3. Result
3.1 Petrographic studies
(i) Gneiss, Ondo Road,Akure

The rock consists of the following minerals (with estimated
modal content): plagioclase (36%), quartz (25%),
biotite(21%), hornblende (7%), microcline (6%),
hypersthene (3%), muscovite (1%), pyroxene (1%)
andmymerkite (0.3%) (Table1).
(ii) Granite Gneiss, Igbatoro Road,Akure
Microscopically, quartz, biotite, plagioclase, microcline and
hornblendeconstitute the major minerals in thin section of
the rock while mymerkiteoccurs sporadically. The modal
percentages of the minerals are 29%, 23%, 21%, 19%, 8%,
4%, and 1% (Table 1). Generally, the platy minerals tend to
be arranged in preferred orientation that defines the foliation
of the rock (Fig. 2B).
(iii) Charnockitic Rocks,Akure
Plagioclase, biotite, quartz, hypersthene, hornblende,
muscovite and orthoclase are the major minerals while
opaque minerals (probably iron oxide) and zircon are the
accessories identified in the thin section of the rock. The
modal analysis of the thin sections of the rock gives 32%,
16%, 16%, 16%, 11%, 4%, 3%, 1% and 1% respectively for
the minerals (Table 1).
(iv) Porphyritic Biotite Granite, Somorika Road, Igarra
Biotite, plagioclase, quartz, microcline and hornblende are
the major minerals in the thin section of the rock while
mymerkite and zircon constitute the accessories. The modal
analysis of the minerals in the thin section is 31%, 30%, 23%,
10%, 4%, 3% and 1%, respectively (Table 1).
(v) Biotite Granite, Ibillo Road, Igarra
Biotite, plagioclase, quartz, microcline and hornblende are
the major minerals in the thin section of the rock while
mymerkite and zircon constitute the accessories. The modal
analysis of the minerals in the thin section is 31%, 30%, 23%,
10%, 4%, 3% and 1%, respectively (Table 1).
(vi) Lamprophyre,Auchi Road, Igarra
Biotite, quartz, plagioclase, hornblende, microcline and
opaque minerals are the major minerals identified in the thin
section of the rock while zircon is the accessory (Fig. 2F).
The result of the modal analysis is 41%, 23%, 21%, 6%, 5%,
5% and 0.5%, respectively, for the minerals (Table 1).
3.1.1 Mineralogy
90

S & T Review
3.1.2
(iii) Charnockitic Rocks
Texture
Generally, three of the rock types (gneiss, granite gneiss and

biotite granite) exhibit medium to coarse grained texture.
The grains display irregular shape with sutured margins
indicating that they are strongly interlocked. The
charnockitic rock is largely coarse grained while the
porphyritic biotite granite displays very large crystals of
feldspar in finer groundmass of quartz, biotite and feldspar.
The charnockitic rocks and porphyritic biotite granite
possess grains with sutured boundaries. The Lamprophyre
possesses coarse biotite grains in a groundmass of very fine
grains of biotite, quartz, feldspar and hornblende, indicating
porphyritic texture.
3.1.3. Microstructures
(i) Gneiss
The charnockitic rocks exhibited the following microstructures: cleavage, distorted cleavage, mineral inclusions,
micro-cracks, bent twin lamellae and compressed twin plane
(Figure. 2C).
(iv) Porphyritic Biotite Granite
The micro-structures that were observed in the thin section of
the rock include: cleavages, mineral inclusions, twinnings
and micro-cracks (Figure. 2D).
3.1.4 Biotite granite
Micro-structures such as bent lamellae, cleavage, bent
cleavage, twinning and mineral inclusions were observed in
the thin sections of the rock (Figure. 2E).
3.1.5 Lamprophyre
The rock exhibited mineral inclusions, micro-cracks,

twinning, partly deformed twinning, cleavages and undulose
extinction as micro-structures (Figure. 2A).
(ii) Granite Gneiss
The rock petrography revealed cleavage, micro-cracks,
mineral inclusions, twinnings and mymerkite as the microstructures exhibited by the rock (Figure. 2B).
This rock type exhibited cleavage and twinning as microstructures in thin section (Figure. 2F).
4. Specific Gravity/Bulk Density
The result of specific gravity test shows that charnockitic
rock has the highest value with an average of 2.81, followed
by lamprophyre with an average of 2.78, porphyritic biotite
granite of 2.69,granite gneiss and gneiss with an average of
2.67 each and by biotite granite with an average of 2.63
91
Figure 2: Photomicrographs of the rock types: A = Gneiss, B = Granite gneiss, C1, C2 = Charnockitic rock,
D = Porphyritic Biotite Granite, E = Biotite Granite and F = Lamprophyre.
(Table 1). The result of the bulk density followed the same
values and trend as for the specific gravity but the values
were expressed in g/cc unit for the different rock types
(Table 1).
Table 1: Result of Specific Gravity/ Bulk density of Rock Samples
92
5. Abrasion Resistance
The results of abrasion resistance test revealed that
lamprophyre (Ig003) has a range of 34.26-34.53 with an
average value of 34.40; charnockitic rock (Ak003), a range
of 32.98-33.97 with an average of 33.40; granite gneiss
(Ak002), a range of 26.27-26.66 with an average of 26.50;
biotite granite (Ig002), a range of 26.5-27.63 with an average
of 27.20; gneiss (Ak001), a range of 26.9-29.36 with an
average 27.80 and porphyritic biotite granite (Ig001), a range
of 25.62-27.78 with an average of 26.70 (Table 2). This
shows that lamprophyre (Ig003) has the highest strength
while granite gneiss (Ak002) has the lowest. This test
measures the degree of toughness of the rock. Generally, all
the rock types that were tested for abrasion strength have
values that are indicative of high mechanical strength.

S & T Review
Table 2: Result ofAbrasion Resistance of Rock Samples
abrasion resistance of all rock types show this positive

correlation.
Fig. 2: Scatter Diagram with Regression Line

Correlating Abrasion Resistance (Ha) with Bulk Density
for the Rock Types.
6.3 Mathematical Model Relating Bulk Density with
Abrasion Resistance
6 Correlation of Bulk Density and

Abrasion Resistance Data
6.1 Scatter Diagram
The results of bulk density and abrasion resistance were
2
plotted in a scatter diagram (Fig. 2). The R was 0.8425 with a
regression equation, Y = 46.796X 97.435.
Since the correlation coefficient is very high, a model

derived from the regression equation relating bulk density
with abrasion resistance, Y = 46.79X 97.43, will most
likely give a good estimation of one parameter from the
other.
The regression equation relating bulk density with abrasion
resistance is:
Y = 46.79X 97.43 ------------------------------
6.2 Line Graph
Which can be re-written as
Figure 3(a) is the line graph with the log2 scale on the vertical
axis showing the variations of abrasion resistance and bulk
density for the rock types. The range of the bulk density of the
rock types is small and therefore the graph presents an almost
straight line but an extended y-axis shows clearly the true
shape of the graph [Fig. 3(b)]. The graphs of abrasion
resistance and bulk density correlate positively as it was
observed that generally the higher the bulk density, the
higher the abrasion resistance and this is corroborated by the
2
values of R and r (0.8425, 0.9179) of the relationships [Table
3(B)]. Essentially, the line graphs of bulk density and
Ha = 46.79D 97.43-----------------------------
(3.1)
(3.2)
Where Ha = abrasion resistance and D = bulk density.

6.4 Test of Reliability of Model
The abrasion resistance of the rock types was estimated with
the model using the bulk density values derived from the
laboratory and the result compared with the abrasion
resistance values determined in the laboratory. The variance,
standard deviation, standard error and Ftest were evaluated
for the two sets of values in accordance with [10], with the aid
of Microsoft Excel (2007). The results of the test gave a mean
93
of 29.30 for the two sets of values; variance of 1.7; standard

deviation of 1.3; Standard Error of 1.47; and an Ftest of 0.86
(Table 4). The standard deviation is very small compared to
the mean indicating that the results are fairly tightly packed
around the mean. Since the range 24.0884 34.5116 [mean
4(standard deviation)] accommodates all estimated values,
none can be regarded as outlier. The model can therefore be
said to be reliable.
[11] and [12] who reported strong correlation between

abrasion resistance and strength parameters. As presented in
Table 3(A), the charnockitic rock and lamprophyre with the
highest bulk density values of 2.81g/cc and 2.78g/cc also
havethe highest abrasion resistance values of 33.38 and 34.4,
respectively.With minor deviations, other rock types, viz,
biotite granite, gneiss, granite gneiss, and porphyritic biotite
granite exhibit a similar relationship between bulk density
and abrasion resistance.
Table 3 Correlation of Abrasion Resistance
and Bulk density of Rock Samples
Fig. 3 (a) Line Graph Relating Bulk Density and

Abrasion Resistance (Ha) for the Rock Types
Note:
1. Gn = gneiss, Ggn = granite gneiss, Chk = charnockitic
rock, Gr = biotite granite, Pgr = porphyritic biotite granite
and Lam = lamprophyre.
2. Density = bulk density of rock types, Ha = abrasion
resistance.
2
Fig. 3 (b) Extended Density Plot

Note: The Scale Of Y-Axis is in log2. Gn = Gneiss,Ggn =
Granite
Gneiss, Chk = Charnockitic Rock, Pgr =
Porphyritic Biotite Granite, Gr = Biotite Granite, Lam =
Lamprophyre.
7. Discussion and Conclusions
The r value of 0.9179 (Table 3B) derived from correlating
bulk density with abrasion resistance indicates a very high
positive correlation. This is consistent with the findings of
94
3.1. R = the square of Pearson product moment correlation

coefficient.
3.2. r = the correlation coefficients between two sets of data.
4.1. a is the table of values while
2
4.1. b is the table of correlation (R and r).

Since abrasion resistance measures the toughness (hardness)
of a rock, it could be said that charnockitic rocks and
lamprophyre are the toughest of the selected rocks that were
studied in this work. Laboratory tests and petrographic
evidences show that these two rocks are the most compact
S & T Review
and have no sign of incipient weathering.
Where:
However, it was noted that the charnockitic rocks exhibited

petrographic evidence of deformation (the micro-structures)
such as micro-cracks, bent lamellae, undulose extinction and
deformed/distorted twinnings. Given the comparatively high
value of abrasion resistance of the charnockitic rock, it
would appear that minor deformation as evidenced in the
microstructures does not significantly affect the abrasion
resistance of the rock and hence its overall hardness. This
conclusion is however contrary to the assertion of [13] that
microstructures affect mechanical strength negatively. The
fine grained texture of the lamprophyre might have been
responsible for the high abrasion resistance it exhibited. This
is consistent with the suggestions of [14] and [15] that fine
grained rocks possess high strength and vice-versa.
Ha = Abrasion resistance determined in the laboratory;
On the other hand, a medium to coarse grained rock

(charnockitic rock) exhibits a similar value which points to
the fact that the effect of texture on abrasion resistance may
not be too significant.
The characteristics that are common to these two rock types
are (i) high bulk density and (ii) lack of signs of incipient
weathering. These characteristics are therefore believed to
have influenced the value of abrasion resistance more than
the micro-structures and texture.
It is therefore concluded that bulk density is a potent factor
that can influence the abrasion resistance of igneous and
metamorphic rocks.
Table 4: Relationship between the abrasion
resistance determined in the laboratory and those
estimated with the derived model using bulk density
values of the rocks.
Ha1 = Abrasion resistance estimated with model;

D = Bulk density of rock types;
Var = variance;
Std = Standard deviation.
Note: Gn = gneiss; Ggn = granite gneiss; Chk =
charnockitic rock; Pgr = porphyritic biotite granite;
Gr = biotite granite and Lam = lamprophyre.
Reference
[1] S. C. Teme The Engineering Geological Significance
of the Point-Load Index Tests in Rock Material
Strength Evaluation. Nigerian Journal of Mining and
Geology, vol. 20 (1&2) 87-93, 1983.
[2] A. Aydin, and A. Basu, The Schmidt hammer in rock
material characterization. Engineering Geology, 66,
p.211-219, 2005.
[3] O.A. Ademeso, Deformation traits in the charnockitic
rocks of Akure area, Southwestern Nigeria. Asian
Journal of Earth Sciences 2(4), pp 113-120, 2009.
[4] O.A. Ademeso, Field and Petrographic relationships
between the charnockitic and Associated Granitic
Rock, Akure Area, SouthwesternNigeria. World
Academy of Science, Engineering and Technology, 71
pp 626-630, 2010.
[5] J. A. Adekoya, O.O. Kehinde-Phillips and A. M.
Odukoya, Geological distribution of mineral
resources in southwest Nigeria. In: Prospects for
investment in mineral resources of Southwestern
Nigeria.A.A. Elueze (Ed.) pp 1-13, 2003.
[6] M. A. Rahaman, Recent advances in the study of the
basement complex of Nigeria. In: Precam. Geol. Nig.,
GSN Publication, pp11-43, 1988.
[7] O. A. Ademeso, Geomechanical Characterization of
some Precambrian Basement Complex rocks in Akure
and Igarra Areas, Nigeria. A PhD thesis submitted to
the Department of Applied Geology of The Federal
University of Technology,Akure, Nigeria. Oct. 2011.
[8] ASTM,Standard test method for abrasion resistance.
95
ASTM Stand. C241 /C241M 09, 2009.

[9] ASTM, Standard test method for determination of
rock hardness by rebound hammer method. ASTM
Stand. 04.09 (D 5873-00), 2001.
[10] D. B. James, Standard error vs standard error of
measurement, shiken: JALT testing and Evaluation
SIG newsletter, vol.3. no.1. Pp. 15-19, 1999.
[11] A. Shakoor, and C.L. Brown Development of a
quantitative relationship between unconfined
compressive strength and los angeles abrasion loss for
carbonate rocks. Bulletin of Engineering Geology
and the Enviroment vol. 53 no. 1 pp 97-103, 1996.
[12] S. Kahraman; S. Sair; O. Gunaydin, and M. Osman,
Empirical methods to predict the abrasion resistance
of rock aggregates. Bull. Of Eng. Geol. and the
96
Env.vol. 66 pp 449-455, 2007.

[13] Liu Hongyuan; Kou Shaoquan; Per-Arne, Lindqvist;
Jan Erik, Lindqvist and Urban kesson, Microscope
Rock Texture Characterization and Simulation of Rock
Aggregate Properties. Report of project 601362/2004, Geological Survey of Sweden.Pp 94,
2005.
[14] F.M. Mendes; L.Aires-Barros; F.P Rodrigues.The use
of modal analysis in the mechanical characterization of
rock masses. In: Proc 1st Int. Cong. Rock Mech.
Lisbon, Vol 1, pp. 217-223, 1966.
[15] T.F.Onodera and K.H.M. Asoka Relationship
between texture and mechanical properties
ofcrystalline rocks.Bull. Int. Assoc. Eng. Geol. 22 pp
173-177, 1980.

S & T Review
SUBSCRIPTION FORM
S & T Review
COLLEGE OF ENGINEERING
TEERTEHNKER MAHAVEER UNIVERSITY
Subscription Charges:
Students/Research Scholars/Alumni (One Year; 2 issues)
100/-
Institutions/Corporate (One Year; 2 issues)
200/-
We are interested in subscribing the Journal S & T Review An International Journal of Science &
Technology. ABank Draft / Cheque bearing number ...............................................................
dated..for Rs..drawn in favour of
Teerthanker Mahaveer University, Moradabad, payable at Moradabad towards subscription is enclosed.
Name................
MailingAddress................
..
City.........PIN/ZIP............. Country...
Phone NoMobile NoFax
E-mail...
Signature with Date
Mail to:
Editor-in-Chief
S & T Review An International Journal of Science & Technology
College Of Engineering
Teerthanker Mahaveer University
NH-24, Moradabad-244001, U.P., India.
Email Id: ijstr@tmu.ac.in
97
STATEMENT OF OWNERSHIP
S & T Review An International Journal of Science & Technology
FORM IV (Rule 8)
1. Place of Publication
Moradabad
2. Periodicity of its Publication
Bi-annual
3. Printer's Name & Address
Mr. Amit Jain, Arihant Graphics,

X/2376, Lane 6A, 2 Raghubar Pura No. 2
Gandhi Nagar, Delhi-110031
4. Publisher's Name & Address
Dr. R. K. Mudgal, Registrar

Moradabad, UP, India
5. Editor-in-chief
Prof. N. K. Joshi
Director, College of Engineering
Teerthanker Mahaveer University,
Moradabad, UP, India
6. Associate Editors
Dr. B. P. Agrawal
Mr. Shirazul Islam
Mr. Gaurav Pandey
Mr. Namit Gupta
Dr. Avnish Chauhan
Moradabad- 244001
7. Nationality
Indian
I, Dr. R. K. Mudgal hereby declare that the particulars given are true to the best of my knowledge
and belief.
Sd/Dr. R. K. Mudgal
Registrar,
98

S & T Review
S & T Review
Bi-Annual Refereed Journal
Guidelines to Authors
S & T Review invites original research and review manuscripts not submitted for publication elsewhere. The Journal considers all
manuscripts on the strict condition that they have been submitted only to the Journal, that they have not been published already. It is
mandatory on the part of the corresponding author to furnish the certificate of copy right at the time of submission of the manuscript.
Certificate of Copyright
Copyright in the text and illustrations of the paper Titled: ..........
..........................................................................................................................................................................................................
by (Author(s) Name).....................
to be published in S & T Review An International Journal of Science & Technology has been transferred to the Teerthanker
Mahaveer University, Moradabad. I confirm that proper citations to the previously reported work have been given and no
data/tables/figures have been quoted verbatim from other publication without giving due acknowledgement and without the
permission of the author(s). The paper is the original work of the author. It has not been published earlier and has not been sent for
publication elsewhere.All copyright requirements in respect of material used directly or indirectly in the paper have been duly met.
Signature
Date:
Place:
Transfer of Copyright : LicensingAgreement is required for all papers. This should be posted or mailed (Scanned copy) at mail Id ijstr@tmu.ac.in.
Signatures and names of all authors : The copyright of the paper will be transferred from the author to the publisher. One original copy and two photostat copies of
the manuscript should be submitted to the editor. The manuscript, after referees' acceptance, will be sent back to the author(s) along with referees' comments. For resubmission, two copies of the revised version of the manuscript, using word processing software such as MS Word (version 2007 onwards) or PDF files (version 9
and onwards), or as an attachment to e-mail should be submitted to the editor.
Preparation of the Manuscript : Manuscripts should be typed in single spaces (11 pt, Times New Roman font) on one side of the bond paper of A4 size. All pages
should be numbered consecutively. Use SI units, and give equivalent SI units in parenthesis, when the use of other units is unavoidable. Symbols should be according
to standard guidelines.
Title Page : The first page should contain a concise and comprehensive title, the names and addresses of the authors and contact details of the corresponding author
(postal address and e-mail). Title must be in 18 pt Regular font.
Short Running Title : The Short Running Title should be a brief phrase describing the contents of the paper not exceeding of 50 characters.
Authors: Names of authors be typed in first letters capital (11 pt)
Addresses ofAuthors : Addresses of the institution (s) where the work was carried out (10 pt.) and e-mail address (9 pt. italic)
Main Headings : Each manuscript should be divided into the following main heading (typed in bold, first letters capital, on the left hand side of the page; 12 pt. with
spacing 12 pt before and after): Abstract, Introduction, Materials and Methods, Results, Discussion, Acknowledgement, References. If the Result and Discussion
sections are combined, a Conclusion section should be added.
Sub-Headings : Typed in flush left, Bold and first letter should be in capital (11 pt with spacing of 9 pt before and after)
Abstract : A self-contained abstract (150-200 words) outlining in a single paragraph the aims, scope and conclusions of the paper must be supplied. It should be
brief, typed in normal 10 pt.
99
Keywords : There should be at least five or six keywords indicating the content of the manuscript. There should be a spacing of 10 pt before and after.
Introduction : The basic aim of the introduction is to provide a clear statement of the problem, literature available on the topic of the given subject, the solution of
problem proposed. It should be comprehensible to the persons in various disciplines of science and technology.
Materials and Methods : Materials and Methods contain the exploration of the experimental work. To describe it, the work done by the author should be explained
in detail. Previously published work or available literature should be cited as reference. Important inferences drawn should be explained in a very comprehensive
manner. Subheadings should be used. Methods in general use need not be described in detail.
Results : For result in a paper, precision and clarity is mandatory. Result should contain only the data found out by the author through his/her experimental work i.e.
the data which are necessary to draw the conclusions. Discussion and detailed interpretation of data should be avoided. Same data should not be presented in both
table and figure form.
Discussion : Discussion is the interpretation of the results obtained in the present work and data available in the literature from the past study. Important inferences
drawn from the result should be clearly highlighted. The way of explanation should be clear and easily understandable. The Results and Discussion sections can
include subheadings, and when appropriate, both sections can be combined.
Conclusion : State the conclusions in a few sentences at the end of the paper.
Tables : Only MS word table format should be used for preparing tables. Do not use tabs and spaces. Do not break contents of a cell across multiple cells. Cell
entries should be clearly belong to a specific row and column. Tables should be prepared using a roman font; italic and bold should not be used for emphasis. Tables
should numbered using Arabic numerals, and they should be cited in numerical order in the text. Titles should identify the table as briefly as possible. Headings
should identify the column as briefly as possible.
Figures Captions : Figures must be numbered using Arabic numerals. Graphics should be prepared using applications capable of generating high resolution GIF,
TIFF, JPEG or Power Point before pasting in the Microsoft Word manuscript file. Figure captions must be in 10 pt Regular font. Captions of a single line must be
centered whereas multi-line captions must be justified. Captions with figure numbers must be placed after their associated figures. Photostat copies and dot matrix
prints will not be accepted.
Mathematical Equations : In case of papers dealing with mathematical analysis, please try to ensure all symbols are defined in the text. All equations should be
numbered in consecutive Arabic numbers written in parentheses that are aligned on the right side of page, with the equation centered on the page. Reference of
equations may be referred as equation (1), equations (2)(4) in the text. Subscripts and superscripts that are labels should not be in italics. Always use the standard
symbols for the representation of different variables.
References : References should be listed at the end of the manuscript in the order of their citation in the body of the text. References should be cited in the text in
Arabic numerals as superscript enclosed within parenthesis. Latest reputed journal articles, books with latest edition must be referred and those references should
only be provided as far as possible.
The following format of references may be consulted:
Research Papers
[1] M. Hasegawa, S. Furutani, S. Doki and S. Okuma,Robust vector control of induction motors using full-order observer in consideration of core loss, IEEE
Transactions on Industrial Electronics, vol.50, no.5, 2003, pp. 912-919.
Books & Proceedings of Conferences
[2] B. K. Bose,Power Electronics andAC Drives, Englewood Cliffs, Prentice Hall, New Jursey, 1986.
[3] T. Niknam, A. M. Ranjbar, and A. R. Shirani, Impact of distributed generation on volt/var control in distribution networks, in Proc. of IEEE Bologna Power
Tech Conf., Bologna, Italy, Jun. 2003, vol. 3, pp. 17.
Thesis & Dissertation
[4] A.Ahmed,Aeration by Plunging Liquid Jet., Ph.D. thesis, Loughborough University of Technology, Leicestershire, UK, 1974.
Patent
[5] K. Miwa, S. Maeda and Y. Murata, "Purification of stevioside by electrolysis", Kokai, Tokkyo, Japan, 79 89,066 (1979).
Peer Review
The manuscripts which require the peer review from author side as per the recommendations of reviewers will be sent back along with reviewers' comments to the
corresponding author. It should be checked carefully and the modified manuscript should be submitted within ten days of receipt. No page proofs will be sent to
author(s).
Multiple Submissions
It is clear policy of the Journal not to consider manuscripts that have been previously submitted to other journals.
100

S & T Review

Sand T Review Journal Vol 1 No 12012

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

Sand T Review Journal Vol 1 No 12012

Încărcat de

Drepturi de autor:

Formate disponibile

ISSN : 2231-5160

Vol. 1, No. 1, 2012

From the Desk of Editor-in-Chief

A Study Towards Use of Cloud Computing

Modular Neural Network for Detection of Diabetic Retinopathy in Retinal Images

Classification of Customer Behaviour using Artificial Neural Network (ANN) and

A Multilevel Cascaded Converter Based Solid State Distribution Transformer

Detection of Hand Movement Using EMG Signal

Packed Bed Solar Energy Storage: A State-of-Art Review

Friction Stir Welding of Aluminium

Distresses on Flexible Pavement and Their Measurement

Analysing the Suitability of Waste Materials for Road Construction

Prophylactic Efficacy of Septilin on Skin of Rats Exposed to Gamma Radiation

An Efficient Method for the Calculation of Eigenvector Design Sensitivities

A Study Towards Use of Cloud Computing

Nasir Ali , Mohammad Ishrat

S & T Review Volume 1, Number 1, 2012

Public sector should take advantage of those improved

4. Proposed Cloud based E-Governance Model

3. Early Work Study

qVarious Services (GSG, G2B, G2E, G2C)

qMiddle Ware Solutions

Figure 1 : e-Governance cloud model

An International Journal of Science & Technology

systems, software applications, middleware and integration

Web Based Operating system

Web based Query systems

Web based data base systems

Middle ware services

Work flow management services

Government departments or local area citizen requiring

Complaint Resolution System

Employee Management Systems

Tax Filing Systems

Water Boards, Billing, Payment Systems

District Management Solutions

Government office service Desk

Cloud fits in right into the requirements and can offer

S & T Review Volume 1, Number 1, 2012

Governance to a great extent especially in case of rural areas.

On-demand self sufficient service

Ubiquitous network access

Location independent resource

Cloud is making rapid inroads because of the following

Can reduce IT labour cost by 50% in configuration,

Can improve capital utilization by 75% significantly

Reduce provisioning cycle times from weeks to

Can reduce end user IT support costs up to 40%

In a traditional infrastructure, there will be one instance of

[2] Stacy Avery Baird, Government role in developing an

pages 121-130, 2003.

R. Levy, J. Nagarajaro, G. Pacific, M. Spreitzer, A.

[9] L.F.G Sarmenta, Volunteer Computing, Ph.D. Thesis

An International Journal of Science & Technology

Modular Neural Network for Detection of

S & T Review Volume 1, Number 1, 2012

ophthalmologist. Therefore more time is required to

noise and optical aberrations (dispersion, transverse and

In this experiment, different features of retinal images were

Fig. 3 Modular feed forward neural network

Fig. 2 Fundus image affected with diabetic retinopathy

It is proposed to develop a decision support system for the

An International Journal of Science & Technology