alhosn university journal of engineering & applied sciences

ISSN 2076-8516
www.alhosnu.ae
ALHOSN UNIVERSITY
JOURNAL OF
ENGINEERING
& APPLIED
SCIENCES
CONTENTS
Power Control-based Channel Assignment
Algorithm for Wireless Mesh networks
Aizaz U. Chaudhry, Roshdy H.M. Hafez,
Osama Aboul-Magd
Characterization of Elastic Buckling of
Cylindrical Liners Using Finite Element
Approach
A.M. Sweedan, K.M. El-Sawy
Optimized 3D Mesh for the Finite Element
Analysis of Laminated Composite Plate
with Central Hole
Khaled M. El-Sawy, Abdul Hameed
Arwani
Affordability, Enablement and Sustainability
of Low-income Housing
Mahbubur Rahman
Drilling process Robust Optimization for
6061 Aluminum Using Desirability
Functions
Mahmoud I. Awad, Elsayed Orady,
Chenmaya Dandekar
Neighbourhood Design and Walkability: A
synthesis from Planning, Design, Transportation
and Environmental Health fields
Mohammed Kashef
Qos Improvements in Mobile WiMAX
Networks
P. Rengaraju, C.H. Lung, A. Srinivasan,
R.H.M. Hafez
The Solar-Wind-EV Paradigm: A Strategic
Policy for a Sustainable Energy Future
Zuwei Yu, Tarik Aouam

ALHOSN UNIVERSITY JOURNAL
OF ENGINEERING AND APPLIED SCIENCES
ADVISORY BOARD
(in alphabetical order)
Prof. Ghassan Aouad
Salford University, UK
Prof. Goodarz Ahmadi
Clarkson University, USA
Prof. Hisham Elkadi
University of Ulster, UK
Prof. Jamal A. Abdalla
American University of Sharjah, UAE
Dr. Khaled El-Sawy
United Arab Emirates University, UAE
Dr. Mohamed Lachemi
Ryerson University, Canada
Prof. Mufid Abdul Wahab Samarai
Sharjah University, UAE
Prof. Nizar Al-Holou
University of Detroit Mercy, USA
Prof. Riadh Al-Mahaidi
Monash University, Austrialia
Prof. Sadik Dost
University of Victoria, Canada
Prof. Ziad Saghir
Ryerson University, Canada
2

ISSN 2076-8516
ALHOSN UNIVERSITY
JOURNAL OF
ENGINEERING
& APPLIED
SCIENCES
A bi annual, refereed journal published by
ALHOSN University - Abu Dhabi - UAE
Volume 3 Number 1 July 2010
Chairman
Dr. Nasser Bin Saif Al Mansoori
Editor
Prof. Abdul Rahim Sabouni
Associate Editor
Dr. Hamdi Sheibani
Members
Dr. Adel Khelifi
Dr. Adnan Husnein
Dr. Abdelaziz Soufyane
Dr. Naima Benkari
Managing Editor
Dr. Al Haj Salim Mustafa
Address:
P.O. Box : 38772
Abu Dhabi - UAE
Tel. : +971 2 4070700
Fax : +971 2 4070799
E-mail : engjournal@alhosnu.ae
Website : www.alhosnu.ae
3

AHU J. of Engineering & Applied Sciences 3 (1) 2010
© 2009 ALHOSN University
CONTENTS
Power Control-based Channel Assignment Algorithm for Wireless Mesh networks 7
Aizaz U. Chaudhry, Roshdy H.M. Hafez, Osama Aboul-Magd
Characterization of Elastic Buckling of Cylindrical Liners Using Finite Element Approach 17
A.M. Sweedan, K.M. El-Sawy
Optimized 3D Mesh for the Finite Element Analysis of Laminated Composite Plate with 31
Central Hole
Khaled M. El-Sawy, Abdul Hameed Arwani
Affordability, Enablement and Sustainability of Low-income Housing 51
Mahbubur Rahman
Drilling process Robust Optimization for 6061 Aluminum Using Desirability Functions 75
Mahmoud I. Awad, Elsayed Orady, Chenmaya Dandekar
Neighbourhood Design and Walkability: A synthesis from Planning, Design, 87
Transportation and Environmental Health fields
Mohammed Kashef
Qos Improvements in Mobile WiMAX Networks 107
P. Rengaraju, C.H. Lung, A. Srinivasan, R.H.M. Hafez
The Solar-Wind-EV Paradigm: A Strategic Policy for a Sustainable Energy Future 119
Zuwei Yu, Tarik Aouam
5

AHU J. of Engineering & Applied Sciences 3 (1) : 7-15 (2010)
© 2009 ALHOSN University
POWER CONTROL-BASED CHANNEL ASSIGNMENT
ALGORITHM FOR WIRELESS MESH NETWORKS
Aizaz U. Chaudhry 1 , Roshdy H. M. Hafez 1* , Osama Aboul-Magd 1
1 Department of Systems and Computer Engineering, Carleton University, Ottawa, Canada
ABSTRACT: The assignment of channels to radios in a multi-radio mesh network is a challenging task. In this
paper, we propose a channel assignment algorithm, TICA (Topology-controlled Interference-aware Channelassignment
Algorithm), which significantly improves network throughput by minimizing interference within the
mesh network using a novel approach of controlling the network topology based on power control before
intelligently assigning the channels to the multi radio mesh routers as well as guaranteeing network connectivity.
Key words: Channel assignment algorithm; multi-radio multi-channel; power control; topology control algorithm.
1. INTRODUCTION
In traditional Wireless Local Area Networks (WLANs), each Access Point (AP) is connected to
the wired network whereas in Wireless Mesh Networks (WMNs), only a subset of APs is
connected to the wired network. An AP that is connected to the wired network is called gateway
(GW). APs that do not have wired connections are called mesh routers (MRs) and they connect to
the gateway through multiple hops. Similar to routers in wired networks, mesh routers in a WMN
forward each other’s traffic in order to establish and maintain their connectivity [1].
The IEEE 802.11a standard provides 12 non-overlapping frequency channels. A frequency reuse
pattern similar to those used in cellular networks can be used to assign channels to MRs.
Additionally, if a MR is equipped with more than one transceiver, we can assign multiple channels
to the same router within the interference range. In a large WMN, the total number of radios
(transceivers) is higher than the number of available channels. So, many links between the MRs in
the WMN operate on the same set of channels and interference among transmissions on these
channels decreases their utilization. Therefore, minimizing the effect of interference is required for
the efficient reuse of the scarce radio spectrum. Therefore, the key issue in a Multi-Radio Multi-
Channel (MRMC) WMN architecture is the Channel Assignment (CA) problem which involves
assigning each radio to a channel in such a way that minimizes interference on any given channel
as well as guarantees network connectivity [2].
This paper addresses the channel assignment problem by proposing a centralized CA algorithm.
The innovative element in the proposed CA is controlling the network topology using a power
control design. The main trade off is as follows: If all radios transmit at maximum power , then
the network will have high connectivity but it will also have high interference and low throughput,
on the other hand, if each radio controls its power, then the network connectivity will be less with
lower interference and better throughout. Therefore, power control can give us a mechanism to
trade off connectivity with interference as illustrated in this paper. Specifically, the contributions
of this work are as follows: “A new Topology-controlled Interference-aware Channel-assignment
Algorithm (TICA) which intelligently assigns the available non-overlapping 802.11a frequency
_________________________
* Corresponding Author.
E-mail : hafez@sce.carleton.ca
7

AIZAZ U. CHAUDHRY, ROSHDY H.M. HAFEZ AND OSAMA ABOUL-MAGD
channels to the mesh nodes with the objective of minimizing interference and, thereby, improving
network throughput”.
The rest of the paper is organized as follows. Section II presents the related work. Network
architecture for the proposed model is presented in section III. Section IV explains the channel
assignment problem and sections V and VI discuss the proposed topology control and channel
assignment algorithms. Performance evaluation and results are given in section VII. The paper
finally concludes in section VIII.
2. BACKGROUND
A neighbor-based topology control scheme in [3] is based on a Local Minimum Spanning Tree
Algorithm. A location-based topology control scheme, proposed in [4], is based on an Enhanced
Local Minimum Shortest-Path Tree algorithm. In centralized CA schemes such as Centralized
Hyacinth (C-HYA) [5] and Traffic and Interference Aware Channel Assignment Scheme
(MesTiC) [2], the traffic load is required to be known before assigning channels whereas our
proposed algorithm requires no such knowledge. The Hybrid Multiple Channel Protocol (HMCP)
proposed in [6] requires radios to switch between channels on a per-packet basis. In such cases,
time synchronization and coordination between mesh nodes is required which is not needed in our
proposed algorithm. The Breadth First Search - Channel Assignment (BFS-CA) scheme proposed
in [7] requires certain number of MRs with certain number of radio interfaces to be placed at
certain hops from gateway whereas our proposed algorithm simply requires all MRs to have four
data radios, does not require any careful router placement strategy and works with any placement
of routers as verified by the performance evaluation.
3. NETWORK ARCHITECTURE
We propose a model where each MR is equipped with five radios (i.e. each MR can
communicate over 5 independent channels using five miniature transceivers). The model assumes
that all radios conform to the IEEE802.11a standard and they work in the 5 GHz band. The MR
dedicates one of the 5 radios for control and uses the remaining four radios for communications.
IEEE902.11a has 12 non-overlapping channels. We dedicate channel # 12 to the control
functions at all nodes. We define the term: Maximum Node Degree (MND) to denote the
maximum number of independent radios dedicated for data communications with neighboring
nodes. In our model the MND is 4. According to the work reported in [5], having a total of 11
channels, the use of MND=4 would result in the highest possible throughput and that higher MND
will not increase the throughput.
4. CHANNEL ASSIGNMENT PROBLEM
Given the connectivity graph, the main challenge for CAA is how to assign a channel to each
radio in a way to minimize interference between MRs and ensure network connectivity. In order to
achieve this, the CAA should satisfy the following constraints:
In order to communicate, two nodes within TR of each other need to have a common channel
assigned to their end-point radios
8

POWER CONTROL-BASED CHANNEL ASSIGNMENT ALGORITHM FOR WIRELESS MESH NETWORKS
Links in direct IR of each other should be assigned non-overlapping channels
The number of distinct channels that can be assigned to a MR is bounded by the number of
radios it has
The total number of non-overlapping channels is fixed
Since traffic in a WMN is directed to and from the gateway, so traffic flows are likely to
aggregate at routers close to gateway. Links that are expected to support heavy traffic should
be given more bandwidth than others. So, these links should use a channel that is shared by a
fewer number of nodes. Hence, priority in channel assignment should be given to links starting
from the gateway based on number of nodes that use a link to reach the gateway.
5. TOPOLOGY CONTROL ALGORITHM (TCA)
Initially, the gateway broadcasts a “Hello” message on the control channel, announcing itself as
the gateway. Each MR that receives this hello message on the control channel over its control
radio broadcasts it again and it is flooded throughout the network. The hello message contains a
hop-count field that is incremented at each hop during its broadcast. So, a MR may receive
multiple copies of this message over its control radio but distance of a mesh router from gateway
is the shortest path length (shortest hop count) of hello message received by MR through its
control radio over different paths. In this way, each MR knows the next hop to reach the gateway
using its control radio.
Our proposed TCA controls the network topology by selecting the nearest neighbors for each
node in the network with the objective of minimizing interference among mesh routers. The
proposed TCA is based on the following assumptions:
A. Assumptions
All nodes start with the maximum transmission power.
Each node has its location information.
Each node uses an omni-directional antenna for both transmission and reception.
Each node is able to adjust its own transmission power.
The maximum transmission power is the same for all nodes.
The maximum TR for any two nodes to communicate directly is also the same.
The initial topology graph created, when every node transmits with maximum power, is
strongly connected.
B. Phases of TCA
1) Exchange of information between nodes
In the first exchange, each node broadcasts a Hello message at maximum power containing its
node id and position.
2) Building the Maximum Power Neighbor Table (MPNT)
From the information in the received Hello messages, each node arranges its neighboring nodes
in ascending order of their distance. The result is MPNT. Then, each node sends its MPNT along
with its position and node id to the gateway using its control radio.
9

AIZAZ U. CHAUDHRY, ROSHDY H.M. HAFEZ AND OSAMA ABOUL-MAGD
3) Building the Direct Neighbor Table (DNT)
For each node in the network, the gateway builds a DNT. Based on information in the
neighbor’s MPNT and the total number of nodes in the MPNT of a node v, gateway eliminates a
node w from MPNT of node v if node w is closer to any other node y in the MPNT of node w than
MRs send MPNT to GW
GW runs TCA & builds connectivity graph
Connected
GW checks resulting
network for connectivity
Not connected
GW builds SPT based on minimum power
GW moves to a
higher TCA
For each link in SPT, GW builds link ranking
GW assigns 11 available
channels to 11 highest ranked
links
For 12 th ranked link and onwards,
GW checks channel assignment
within its IR neighborhood
If any node in the SPT
has MND > 4
Connected
No
Yes
GW keeps four links with the
minimum weight
Yes
If GW finds channels that are not
assigned to any link within IR of a link
No
Not connected
GW checks resulting
network for connectivity
GW assigns channel with
highest channel number
GW selects a LIC
GW moves to a
higher TCA
Similarly, GW assigns
channels to all links
Figure 1. Topology-controlled Interference-aware Channel-assignment Algorithm (TICA)
to node v.
If after removing nodes from MPNT of node v, the remaining number of nodes in MPNT of
node v is equal to “x-1”, then gateway selects “x” nearest nodes as neighbors of node v which
results in the DNT. However, after removing nodes from MPNT of node v, if the remaining
number of nodes is greater than or equal to “x”, the result is DNT. We call the above algorithm as
select x for less than x TCA where x is a positive integer.
4) Converting into bi-directional links
For each node in the network, the gateway converts the uni-directional links in the DNT of a
node into bi-directional links. For each uni-directional link, this is done by adding a reverse link in
the DNT of the neighboring node. This converts the DNT into Bi-directional DNT. This results in
the Final Neighbor Table (FNT).
5) Calculating the minimum power required
For each node in the network, gateway calculates minimum power required to reach each of the
nodes in FNT of a node, using appropriate propagation model formulas. If distance between two
nodes u and v is less than cross over distance, i.e. d( u,v ) < Cross_over_dist, Free Space model is
used whereas if d( u,v ) > Cross_over_dist, Two-ray model is used. Cross over distance is
calculated using following expression [8]:
4
hth r
Cross _ over _ dist
10

POWER CONTROL-BASED CHANNEL ASSIGNMENT ALGORITHM FOR WIRELESS MESH NETWORKS
where h t and h r are the antenna heights of the transmitter and receiver respectively.
The free-space propagation model is given by following expression [8]:
2
RxThresh(4d
)
(2)
P
min
2
G t
G r
The two-ray model is given by the following expression [8]:
4
RxThresh(
d )
(3)
Pmin
2 2
GtGrht
hr
where G t and G r are transmitter and receiver antenna gains respectively. RxThresh is the power
required by radio interface of the receiving node to correctly receive the message.
6. CHANNEL ASSIGNMENT ALGORITHM (CAA)
The problem of assigning channels to links in a mesh network is NP-hard based on its mapping
to graph-coloring problem [5] and standard graph-coloring algorithms can not satisfy all of its
constraints. Also, the CA problem for mesh networks looks similar to the list coloring problem
which is NP-complete [9]. Therefore, we propose an approximate algorithm for channel
assignment. Our proposed algorithm, TICA, as shown in Figure 1, has following phases:
A. Topology Control
In order to create the network connectivity graph with the aim of reducing the interference
between MRs, network topology is controlled using power control at each MR. All nodes send
their MPNTs to gateway using their control radio. Note that in order to send its MPNT to gateway,
each MR knows the next hop to reach gateway using its control radio via gateway advertisement
process. Gateway starts with the select 1 for less than 1 TCA and builds FNTs for all nodes.
B. Connectivity Graph
Based on the information of the FNTs of all nodes, gateway builds the connectivity graph. The
gateway checks the resulting network for connectivity to ensure that it can reach any node in the
SPT directly or through intermediate hops. If resulting network is not connected, gateway moves
to a higher TCA by incrementing x in the select x for less than x TCA.
C. Minimum Power based SPT with a MND of 4
After ensuring that the connectivity graph is connected, the gateway builds the SPT based on
the connectivity graph. The metric for path selection is minimum power. While building the SPT,
gateway ensures that each node can have only 4 TR neighbors and builds a SPT with a MND of 4
per node. If any node in the SPT has more than four links, gateway selects those four links for that
node which have the minimum weight and sets weight of all other links to infinity. The gateway
checks the resulting SPT graph for connectivity. If resulting SPT graph is not connected, gateway
moves to a higher TCA.
D. Link Ranking
To assign channels to links of SPT, gateway calculates the rank of each link. If l is the link and
x is node using link l, then rank of link l, i.e. r l , is given by:
r
l
I x , l
(4) I x,l is 1 if node x is using link l and 0
x
otherwise. In case of links with same rank, link whose power of farthest node to gateway is
11

AIZAZ U. CHAUDHRY, ROSHDY H.M. HAFEZ AND OSAMA ABOUL-MAGD
smallest is given highest rank. If there are still links with same rank, link with smallest node ids is
given highest rank.
E. Channel Assignment
The algorithm then assigns a channel to each link according to its rank and it begins with
assigning the 11 available channels to the 11 highest ranked links such that channel 1 is assigned
to the first ranked link.
For the 12th ranked link and onwards, gateway checks the channel assignment of all the links
within IR of both nodes that constitute that link. If gateway finds one or more channels among the
11 available channels that are not assigned to any link within IR neighborhood of the 12th ranked
link, it assigns that channel from the unassigned channels to the 12th ranked link which has the
highest channel number.
1) Least Interfering Channel (LIC)
If gateway cannot find any channel, among the 11 available channels, that is not assigned to
any link within IR neighborhood of the 12th ranked link, it selects a LIC and assigns it to that link.
2) Interference Level(IL)
To find out LIC, gateway builds IL for all 11 channels. LIC is the channel with minimum IL
which means that assigning this channel to the 12th ranked link results in minimum interference in
the network.
In order to build IL for channel one, gateway finds all links within IR of either of the two nodes
that constitute the 12th ranked link that use channel one and calculates IL of each link based on its
rank and distance from the nodes that constitute the 12th ranked link. It sums up the individual ILs
of all links, that use channel one within the IR of either of the two nodes that constitute the 12th
ranked link, to find out the total IL for channel one. This is done by using the following equation:
( IL)
i
r
1
R d
1
r
1
R d
2
r
1
.......
R d
n
(5) where IL is interference level of channel i, r
is rank of the link using channel i, R is maximum rank assigned to a link in SPT, d is distance of
either of nodes on link using channel i to either of the two nodes that constitute 12th ranked link,
is 2 or 4 depending on cross over distance, i is the channel and has value between 1 and 11, n is
link using channel i which is within IR of either of the two nodes that constitute 12th ranked link.
If a link is emanating from either of the two nodes that constitute the 12th ranked link and a
channel has been assigned to that link, then IL for this channel is set to infinity. LIC is selected
using following equation:
IL LIC
minIL1 , IL2
,........, IL11
(6)
Similarly, the gateway assigns channels to all the links in the SPT. Using its control radio,
gateway then sends each mesh node the CA message. For each channel assigned to a MR, channel
assignment contains the channel number and the neighbor node to communicate with using this
channel. The CA message also contains the next hop to reach gateway. Based on channel assigned
to a MR to communicate with a neighbor and its distance to that neighbor, MR applies power
control and adjusts its transmission power accordingly.
12

POWER CONTROL-BASED CHANNEL ASSIGNMENT ALGORITHM FOR WIRELESS MESH NETWORKS
7. PERFORMANCE EVALUATION
The performance of our proposed algorithm, TICA, for MRMC WMNs is compared against a
“Single-Radio Single-Channel” (SRSC) scheme and a “Common Channel Assignment” (CCA)
scheme for multi-radio mesh nodes. In CCA scheme [2], all MRs have four radio interfaces. The
first radio on each MR is tuned to first non-overlapping channel; the second radio is tuned to
second non-overlapping channel and so on. Note that in CCA and SRSC schemes, all nodes
transmit with same the maximum power and use AODV (Ad hoc On-Demand Distance Vector)
routing protocol [10].
F. Simulation Environment
For the performance evaluation via throughput analysis, NS2 (version 2.30) [11] simulation
tool is used. Multi-interface wireless mesh nodes are created in NS2 by modifying the original
node model in NS2 using the procedure given in [12].
All radios are IEEE 802.11a radios and support 12 channels. As per the IEEE 802.11 standard
[13], RxThresh is set equal to -65 dBm in order to achieve a maximum data rate of 54 Mbps
supported by IEEE 802.11a. In order to achieve a strongly connected topology, the maximum
transmission power for all the radios is set to 27 dBm. RTS/CTS is disabled.
G. Network Topologies
Three topologies are used in the evaluation and each consists of 36 MRs distributed in an area
of 500 m × 500 m. Topology 1 is a grid and topology 2 is randomly generated while in topology 3,
called controlled random, physical terrain is divided into cells and a MR is placed randomly in
each cell.
H. Simulation Parameters in NS2
Based on CA by gateway, IEEE 802.11a channels are assigned to links between MRs and
transmission power for each radio of each MR is set accordingly. The physical (PHY) and MAC
(Medium Access Control) layer settings used for simulation are shown in Table I and Table II,
respectively.
All MRs at the periphery of network send traffic to gateway simultaneously. So, this represents
scenario in which multiple flows within mesh network interfere with each other. Each of these
nodes generates an 8 Mbps Constant Bit Rate UDP traffic stream consisting of 1024 byte packets.
They start at 1 second and do not stop until the end of simulation (600 seconds).
Table I. PHY Layer Configuration in NS2
Physical Layer Parameters
Settings
TX/RX Antenna Height (m) 3
Gain of TX/RX Antenna (dB) 1
Packet Capture Threshold (dB) 10
Packet Reception Threshold (watts)
3.16227e-10
Carrier Sense Threshold (watts)
7.91359e-11
Table II. MAC layer Configuration in NS2
13

AIZAZ U. CHAUDHRY, ROSHDY H.M. HAFEZ AND OSAMA ABOUL-MAGD
MAC Layer Parameters
Settings
Minimum Contention Window 15
Maximum Contention Window 1023
Slot Time (micro seconds) 9
SIFS period (micro seconds) 16
Preamble Length (bits) 96
PLCP Header Length (bits) 24
Basic Rate (Mbps) 6
Data Rate (Mbps) 54
80
70
TICA
CCA
SRSC
60
Average Throughput (Mbps)
50
40
30
20
10
0
0 Random Controlled Random Grid
TOPOLOGY
Figure 2. Average Throughput of all schemes for the three topologies
80
70
TICA
CCA
SRSC
60
Average Throughput (Mbps)
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Topology Number
Figure 3. Average Throughput for ten random topologies
80
70
TICA
CCA
SRSC
60
Average Throughput (Mbps)
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Topology Number
Figure 4. Average Throughput for ten controlled random topologies
I. Simulation Results
Figure 2 shows a graphical comparison of average throughput of all schemes for the three topologies.
As compared to CCA scheme, throughput improvement with TICA is 76% for random topology, 1244%
for controlled random topology and 1341% for grid topology. In comparison to SRSC scheme, throughput
improvement with TICA is 319% for random topology, 10863% for controlled random topology and
13941% for grid topology.
14

POWER CONTROL-BASED CHANNEL ASSIGNMENT ALGORITHM FOR WIRELESS MESH NETWORKS
Figures 3 and 4 show average throughput of all schemes for ten different random and controlled random
topologies, respectively. These results clearly indicate that our proposed algorithm, TICA, significantly
outperforms other schemes for all topologies.
8. CONCLUSION
This paper addresses the trade-off between connectivity and throughput of wireless mesh networks. If
nodes are allowed to transmit at maximum power, the connectivity will be high but the excessive
interference will reduce the overall network throughput. This paper proposes and evaluates algorithms that
allow the designer to choose the appropriate balance between connectivity and throughput. The algorithm
assumes that each node has multiple independent radios, and calls for careful calculation of transmitted
power to achieve the required connectivity with minimum interference. The new algorithm is termed TICA
(Topology-controlled Interference-aware Channel Assignment Algorithm). The algorithm performance
was tested on three types of network topologies, Random topology, Controlled Random and Regularly
spaced nodes. The improvement in the overall network throughput due to the proposed TICA is
significantly higher than the conventional CCA scheme and is much higher than the SRSC scheme in all
three topologies. This is due to the fact that power control results in an efficient frequency reuse which
leads to improvement in the network throughput.
9. REFERENCES
[1.] N. Nandiraju et al., “Wireless Mesh Networks: Current Challenges and Future Directions of Web-In-
The-Sky,” IEEE Wireless Communications 14(4), pp. 79-89, August 2007.
[2.] H. Skalli, S. Ghosh, S. K. Das, L. Lenzini, and M. Conti, “Channel assignment strategies for multiradio
wireless mesh networks: Issues and solutions,” IEEE Comm. Magazine, 45(11), pp. 86-93,
November 2007.
[3.] N. Li, J. Hou, and L. Sha, “Design and analysis of an MST-based topology control algorithm,” IEEE
Transactions on Wireless Communications , 4(3), pp. 1195- 1206, May 2005
[4.] F. O. Aron, T. O. Olwal, A. Kurien and Y. Hamam, “Network preservation through a topology
control algorithm for wireless mesh networks,” in Proc. 2nd IASTED Africa Conference on Modelling
and Simulation, September 2008.
[5.] A. Raniwala, K. Gopalan, and T. Chiueh, “Centralized channel assignment and routing algorithms for
multi-channel wireless mesh networks,” ACM MC2R 8(2), pp. 50-65, April 2004.
[6.] P. Kyasanur and N. Vaidya., “Routing and Interface Assignment in Multi-Channel Multi-Interface
Wireless Networks,” in Proc. IEEE WCNC, 4, pp. 2051-2056, March 2005.
[7.] K. Ramachandran, E. Belding, K. Almeroth, and M. Buddhikot, “Interference-Aware Channel
Assignment in Multi-Radio Wireless Mesh Networks,” in Proc. IEEE INFOCOM, April 2006.
[8.] T. Rappaport, “Wireless Communications: Principles and Practice,” 2nd Ed. Prentice Hall, Upper
Saddle River, NJ, 2002.
[9.] K. Rosen, “Discrete Mathematics and its Applications,” McGraw Hill, 1999.
[10.] Ad Hoc On-Demand Distance Vector (AODV) Routing Ptotocol, RFC3561,
http://www.ietf.org/rfc/rfc3561.txt
[11.] The VINT Project, “Network Simulator – NS2,” http://www.isi.edu/nsnam/ns/
[12.] R. A. Calvo and J. P. Campo, “Adding Multiple Interface Support in NS-2,” University of Cantabria,
January 2007.
[13.] "IEEE Standard for IT-Telecom. and Info. Exchange Between Systems-Local and Metropolitan Area
Networks-Specific Requirements - Part 11: Wireless LAN MAC and PHY Specifications," IEEE Std
802.11-2007 (Revision of IEEE Std 802.11-1999) , June 2007
15

AHU J. of Engineering & Applied Sciences 3 (1) : 17-30 (2010)
© 2009 ALHOSN University
CHARACTERIZATION OF ELASTIC BUCKLING OF
CYLINDRICAL LINERS USING FINITE ELEMENT
APPROACH
A. M. Sweedan 1* K. M. El-Sawy 1
1 Civil and Environmental Eng. Faculty of Eng., UAE University, P.O.Box: 17755, Al Ain, U.A.E.
ABSTRACT: The results of a numerical study to evaluate the elastic buckling pressure, thrusts and bending
moments of slender liners are reported. The adopted finite element model accounts for the non-linearity effects of
large deformations and geometric imperfections on the stability of analyzed cylindrical liners. Results of finite
element analysis are used to identify the critical buckling pressure and the associated maximum thrust and moment.
In addition, several analytical and numerical solutions for the same problem are critically reviewed; namely Lo and
Zhang, Thépot, and Madryas and Szot solutions. An extensive finite element analyses are conducted to evaluate the
elastic buckling pressure and the associated thrust and bending moment at buckling of cylindrical liners with various
dimensions. Obtained numerical results are then compared to relevant predictions provided by other solutions to
investigate the influence of approximations adopted by such solutions on the accuracy of the predicted values.
Comparison reveals the adequacy of some of these methods while others are shown to be on the non-conservative
side and, therefore, may be used with caution for design/check purposes.
Key words: Elastic Buckling, Liners, Finite Element Method, Non-Linear Analysis.
1. INTRODUCTION
Despite the fact that most of the damaged pipelines are partially deteriorated, they require
renovation to prevent potential leakage. One solution is to apply an internal flexible polymeric
liner to the pipeline. These liners are often installed in pipelines that lie below the water table,
and as such they are subjected to external hydrostatic pressure which is transmitted through the
leaking cracked pipeline. In this case, the liner carries only the pressures from the contained fluid
and the external groundwater pressure. In addition, due to some installation considerations, the
liner is temporarily subjected to an additional pressure when grout is pumped around the liner to
fill any annular gaps between the liner and its host pipeline.It is experimentally evident that the
encased liner subject to external pressure may buckle and cannot withstand any additional
pressure at a critical pressure, which is well above the one for the unsupported (free) liner
[1,2,3,4 and 5]. For slender liners commonly used in practice it is assumed, and confirmed
experimentally, that geometrical instability (i.e. elastic buckling) occurs first before any material
failure. The study presented herein is restricted to the case of elastic slender liners only. The
typical fundamental buckling mode shape is shown in Fig. 1(b) along with the initial liner’s
geometry depicted by Fig. 1(a). Real conditions in which liners are installed (i.e. non-uniform
groundwater pressure and any possible debris at the bottom of the host pipe) hint that the liner
most probably will buckle at the invert in a one-lobe buckling mode (Fig. 1(b)). In addition,
theoretical potential energy balance of the liner system indicates that the one-lobe mode
corresponds to the minimum critical pressure.
_______________________________________
* Corresponding Author. Tel.: +971-3-762 3154
E-mail : amr.sweedan@uaeu.ac.ae
17

A.M. SWEEDAN, K.M. EL-SAWY
Pressure
Load
Pressure
Load
Soil
R o
R
Liner
Soil
Single Lobe
Liner
gap = d
Rigid Host Pipe
(a) Initial Liner Geometry
Rigid Host Pipe
(b) Single-Lobe Buckling Mode
Fig. 1. Initial geometry and fundamental buckling mode of cylindrical liners.
In the current study, finite element method is utilized to investigate elastic buckling of slender
liners. The induced critical thrust and bending moment at buckling are also evaluated
numerically.
2. PREDICTION OF CRITICAL BUCKLING LOAD
The buckling of an unsupported circular ring subjected to an external pressure was first studied
by Levy [6]. The buckling pressure, P , for a ring of radius, R
cr
o , cross sectional moment of inertia,
I, and modulus of elasticity, E, is defined by
3 EI
(1)
P
cr
=
R
The elastic solution for a pipe with wall thickness, t, is adopted from the ring case by simply
2
adjusting the elastic Young’s modulus E, to be E ' E /(1 ) (where is the Poisson’s ratio for
the liner material) and the moment of inertia I, to be t 3 /12. Thus, P cr
for the pipe case can be
defined as
3
o
3 EI
Pcr
=
2
(1 -
) R
3
o
E'
4
The encased liner stability problem appeared in the literature in many different forms according
to the engineering application being examined. The most common stability problem types are
encased liner subjected to thermal expansion, or under external pressure loading. The two types
differ significantly in their boundary conditions. For the thermal expansion problem, the detached
part of the liner due to buckling has no longer any contact pressure acting on it. On the other
hand, for the external pressure case, the detached portion of the encased liner is still subjected to
the applied pressure. It is therefore important to distinguish between these two types in the
literature review of the stability problem.
Due to the installation techniques of liners and the nature of the lined element, some
imperfections may develop. These imperfections lead to deviation of the encased liner problem
t
R
o
3
(2)
18

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
from the ideal case of tightly fitted cylindrical liner. Three major imperfection types are
considered in literature and are listed as follows:
gap between the liner and the host element,
ovality in the host element which is consequently reflected to the tightly fitted liner, and
longitudinal wrinkle (or wavy imperfection) in the liner itself.
These imperfections are responsible for the reduction of the liner stability and should be
accounted for in the assessment of the liner critical pressure. The work presented herein deals
with the first imperfection type (i.e. gap imperfection or the case of liner).
As a result of this imperfection and as the external pressure increases, the initially loose (or
unsupported) liner deforms until it contacts its surrounding rigid host. Then the behavior of the
liner changes abruptly from the case of an unsupported loose liner to the case of the constrained
one. The constrained liner continues to deform until it becomes geometrically unstable (i.e.
buckles) or material failure occurs. As mentioned earlier, only elastic buckling is considered
herein and therefore this limits the results of this study to the case of slender liners where no
material failure occurs. This would raise a question about the validity of the thin liner
assumption, which will be answered later at the end of this study.
Researchers and code developers have employed the concept of comparing the buckling
pressures for unsupported and encased circular liners. For the case of circular liner the idea has
been to determine an enhancement factor which depends on the liner material, geometry and the
size of the gap between the liner and its rigid host. Now the critical pressure for circular liner,
P cr , can be defined as
P
cr
= K P cr
(3)
where P and
cr
P are the buckling pressures for encased and unsupported liners respectively, and
cr
K is the enhancement factor, which will be discussed later in section 4.
Different solutions have been developed for estimating the critical pressure of the slender liner.
None of these methods have provided a validation of the thin liner assumption they are based
upon. All such methods assumed smooth contact between the liner and the rigid host element (the
most conservative case). They differed mainly in their basic assumptions for strain (or
deformations) and geometry of the detached portion of the liner. Most of the available solutions
have adopted simplifying assumptions that may or may not provide a good approximation for the
liner buckling pressure.
Equations for modelling single-lobe buckling of a cylindrical encased liner have been derived
by Jacobsen [7] in which the single-lobe buckled portion of the liner is approximated by a cosine
function representing three half-waves. The geometry of the buckled shape is shown in Fig. 2(a).
Basic assumptions used by Jacobsen are listed as follows:
the detached part of the liner follows a cosine function (three half-waves) super-imposed on a
circle of radius ,
the thrust in the detached portion of the liner is constant and equals P, and
the thrust in the other portion of the liner (i.e. portion sliding on the walls of the rigid host) is
constant and equals PR.
The equations developed by Jacobsen’s are given as follows:
19

A.M. SWEEDAN, K.M. EL-SAWY
Attached Segment of
the Liner
Pressure
Load
R
Detached Segment
of the Liner
Rigid Circular
Cavity
(a) Jacobsen (1974)
Attached Segment
of the Liner
Pressure
Load
R
Detached Segment
of the Liner
(b) Lo and Zhang (1994)
Rigid Circular
Cavity
Pressure
Load
Attached Segment of
the Liner
Rigid Circular
Cavity
Deformation w 1
Detached Segment
of the Liner
R
Point O
(c) Thépot (2000)
,
n w
,
s u
Initial
gap, d
Fig. 2. Geometry of buckled liner after Jacobsen (1974), Lo and Zhang (1994) and Thépot (2000)
2
9
2
sin( )
1
2
R 4
sin( )
(4)
3
t
2
sin( )
d
sin( )
tan
12
1
sin( )
2R
sin( )
4
20

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
2
9
3
E ' R
12
t
1
2
P 4
(5)
sin( )
sin( )
y t sin( )
PRsin(
)
4R
sin( )
tan( )
1
1
E'
2R
sin( )
E'
t sin( )
t
sin( )
(6)
where R is radius of the rigid host pipe, R o is initial radius of the liner, t is thickness of the liner, d
is the gap between the liner and the rigid host walls, P is the external liner pressure, E’ is
2
E /(1 ) where E is the Young’s modulus of the liner material, is the Poisson’s ratio of the
liner material, y is the yield stress of the liner material, and and are geometrical quantities
and are shown in Fig. 2(a).
3
The major criticism for Jacobsen solution is that he mixed the solution of the elastic buckling
problem represented, by (4) and (5), with the condition for onset of yielding at the outermost
fibres of the most heavily stressed cross-section represented by (6). Thus the solution provided a
value for the pressure that corresponds to first material yielding and, of course, this happens at a
value for the pressure different from the elastic buckling pressure. It is also clear from Jacobsen
solution, that there is no compatibility between thrusts and bending moments in the attached and
detached portions of the liner at their junction point (i.e. at the ends of the detached portion of the
liner). It may also be noted that the two basic assumptions for the thrust distribution along the
liner and deformed shape at buckling should be assessed.
Lo and Zhang [3] studied the effect of the liner’s lack of fit on the critical pressure for curedin-place
polymer liners. Figure 2(b) shows the initial configuration of the problem. The buckling
pressure was derived using simple linear theory. Their basic assumptions are listed as follows.
the detached part of the liner is circular with radius ,
buckling of the detached portion of the liner is analogous to buckling of fixed-ended circular
arch,
thrust in the whole liner is constant and equals PR o where R o is the initial radius of the liner,
and
shortening in liner is due to thrust only, and first order terms are considered in calculating
deformations.
Their equations were based on satisfying the geometrical boundary conditions as follows:
perimeter of deformed liner equals the original undeformed perimeter less the membrane
deformations which can be expressed as a function of the liner pressure, and
liner buckling pressure is analogous to the buckling pressure of a fixed-ended circular arch of
radius (radius of the detached part of the liner).
Lo and Zhang’s solution has been reformulated earlier by El-Sawy and Moore [8] to permit
direct comparison with other solutions. Accordingly, the solution for the liner critical pressure,
P cr , can be represented by the following equations.
21

A.M. SWEEDAN, K.M. EL-SAWY
P = E'
t
R
3
sin( ) t
sin( )
2R
-3
(7)
2
k 1
12
2
2
-3
2
d
sin( ) t d sin( ) t 1
1
k
sin( )
2R
R 2R
sin( )
2R
12
k tan( ) tan( k
)
(9)
(8)
where all symbols are as defined previously and in Fig. 2(b). The solution strategy starts with the
assumption of a value for in (8) followed by the calculation of the corresponding and k
values from (8) and (9). The pressure, P, can then be calculated from (7). The critical pressure,
P cr , corresponds to the value of which maximizes the value of the pressure, P.
The major criticism for Lo and Zhang’s solution is due to their inappropriate analogy between
buckling of the detached portion of the liner and the buckling of the fixed-ended circular arch.
The boundary conditions of the fixed-ended circular arch are inappropriate as they are associated
with unknown fixed end moments at the arch ends. In the case of the liner problem, these end
moments and bending moments along the attached part of the liner have a constant value defined
by the constant change in curvature in the liner after deformation, where the constant change of
curvature equals [( 1/ R o
) (1/ R)]
. This moment distribution in the liner is not consistent with an
arch model featuring fixed ends. Therefore, Lo and Zhang solution will be considered only in
calculating the critical pressure of the liner and will not be used for any thrust or bending moment
calculations.
Recently, Thépot [9] has developed an analytical solution for non-circular sewer liners which,
as reported by Thépot, is a general solution that can be adopted for the case of circular and noncircular
liners with single-lobe and double-lobe buckling modes as well. The solution calculates
the maximum pressure, P cr (i.e., buckling pressure), and thrust, N cr , and maximum bending
moment, M cr , at the mid-point of the detached segment of the liner (i.e., =0) just before buckling
(Fig. 2(c)). Furthermore, Thépot provided an approximate solution which can be written as:
Pcr
E '
0.455 k
2
5
11
5
t
Ro
N
cr
Pcr
1.26 ,
'
'
E Ro
E
3
cr
t
2
'
E Ro
Ro
2
9
1.0
0.41
0.006
2 5 5
and g Ro
11.65
10. 0 (10)
2 3 6
5 5 5
k t
2
0.1
0.053
0.0012
2
1.0
0.17
0.007
2
M (11)
where g is the annular gap between the liner and the host pipe (i.e., g=d/2 for the single-lobe
case), k is the number of buckled lobes (i.e., k=1 for one-lobe buckling model), and is the
perimeter of the liner (i.e., =2R o for the circular liner). Thépot recommended an approximate
solution based on data fitting of his analytical solution for the cases of liner gaps defined by
10. Values for P cr and (Eq. 10) can be rewritten for the case of circular liner with a singlelobe
as:
Pcr
E '
0.218
2
1.0 0.41
0.006
t
R
o
2.2
and d Ro
1.465
10. 0 (12)
Ro
t
4
5
1.2
22

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
More recently, Madryas and Szot [10] have preformed Finite Element analyses and provided a
formula for the critical pressure, P cr , of liner which is based on regression analysis for their
numerical results. It is worth mentioning that their formula was part of finding another formula
for the combined effect of three liner imperfection types, including the gap, ovality, and wavy
imperfections. Madryas and Szot reported an approximate solution for the critical pressure, P cr ,
of the liner as:
2.2
P cr
1
t
(13)
E '
1
d
0.84176
2R
0.89415
t
R
1.16189
2R
It is important to mention that the above formula satisfies the case of no gap and provides the
well established critical pressure of the tightly fitted circular liner (Glock solution [11]) with onelobe
buckling as
4. FINITE ELEMENT MODELING AND ANALYSIS
2.2
P Glock t
(14)
E ' 2R
The Finite Element Method is employed to solve the liner buckling problem. A strip of unit
length of the liner is only considered because of the long cylindrical geometry of the liner (i.e.
plane strain conditions prevail). Node-to-surface contact elements are used to simulate the
contact problem while 3-noded shell elements are used to model the liner. The liner material is
considered linear and elastic and the liner geometry is approximated by an ellipse with a ratio of
Normalized Thrust N/(E'R)*10 5
for d/R=0.00 and t/R=0.01
2.0
Normalized Thrust N/(E'R)*10 6
for d/R=0.05 and t/R=0.01
2.0
Normalized Thrust N/(E'R)*10 6
for d/R=0.10 and t/R=0.01
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
Normalized Thrust N/(E'R)*10 3
for d/R=0.00 and t/R=0.10
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
Normalized Thrust N/(E'R)*10 3
for d/R=0.05 and t/R=0.10
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
Normalized Thrust N/(E'R)*10 3
for d/R=0.10 and t/R=0.10
2.0
1.0
0.0
-1.0
-2.0
-1.0
0.0
1.0
2.0
(a) Initial Liner Geometry
(b) Single-Lobe Buckling Mode
Fig. 3. Distribution of Thrust in liner with d/R=0.0, 0.05, and 0.1
23

A.M. SWEEDAN, K.M. EL-SAWY
1.0005 between the ellipse axes. This initial elliptical imperfection is used to trigger the onelobe
buckling mode. The Finite Element analysis used accounts for non-linearity effects due to
large change of geometry and non-conservative follower pressure load.
A parametric study is conducted to investigate the variation of the buckling load, P cr , with the
liner’s geometry (t/R) and (d/R) and material properties E’. The adopted finite element model is a
revised and extended version of the one that was used earlier by El-Sawy and Moore [8]. The
extended model predicts the critical buckling pressure and the corresponding thrust, N o , and
bending moment, M o , at the mid-point of the detached liner (i.e., the most heavily stressed point
in the liner). This information is essential for conducting more comprehensive engineering design
of such special structure. It is important to mention that the FE solution is the only consistent
solution among all the previously reviewed ones since it does not assume any approximation
(other than the assumption of elastic and smooth thin liner) in modelling the liner behaviour.
5. DISCUSSION OF RESULTS
Figure 3 shows the distribution of the normalized thrust, N/E’R, based on the FEM solution for
different (d/R) and (t/R) ratios. In this figure, it can be noticed that the thrust is almost constant
for all values of (d/R) and (t/R) except for a very localized rise in the thrust magnitude at the
points where the liner starts to contact, and gets supported on, the rigid host. This increase in the
thrust is noticeable for large values of (d/R) where the liner is freer to deform. The maximum rise
in the thrust magnitude is about 23% and occurs for (d/R)=0.1 and (t/R)=0.1. This would show
that the constant thrust along the whole liner is an approximate one, which in general may not
affect the estimation of the critical pressure, but may affect the estimations of the accompanying
thrusts and bending moments.
Normalized Buckling Pressure (P cr / E') x 10 4
20.0
15.0
10.0
5.0
Tightly Fitted Liner (d / R) = 0.0
FEM
Thepot
Lo and Zhang
Madryas and Szot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
24

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
Normalized Buckling Pressure (P cr / E') x 10 4
15.0
10.0
5.0
Loosely Fitted Liner (d / R) = 0.05
FEM
Thepot
Lo and Zhang
Madryas and Szot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
Normalized Buckling Pressure (P cr / E') x 10 4
15.0
10.0
5.0
Loosely Fitted Liner (d / R) = 0.10
FEM
Thepot
Madryas and Szot
Lo and Zhang
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
Fig. 4. Critical pressure for liner with d/R=0.0, 0.05 and 0.10
Figure 4 shows a comparison between the normalized critical pressure, P cr /E’, based on the
different solutions reviewed earlier (i.e., solutions provided by FEM, Lo and Zhang [3], Thépot
[9] and Madryas and Szot [10]) versus the liner thickness-to-radius ratio (t/R) at different radial
gap ratios (d/R=0.0, 0.05, and 0.10). It is clear from the figure that the buckling pressure
decreases as the gap size increases. This decrease is more pronounced for thin liners where the
flexural stiffness-to-axial stiffness ratio is very small. For this case, flexural deformations
dominate the behavior of the liner and the liner depends mainly on the support of its rigid host
element. For thick liners, the axial deformations dominate the behavior and the liner elastic
stability depends mainly on the liner axial stiffness. The comparison shows also the following:
The solution of Lo and Zhang provides the highest estimate for the buckling pressure
(i.e., non-conservative pressure), which may be due to the inconsistent assumption that the
detached segment of the liner behaves similarly to a fixed-ended circular arch. The
maximum difference between Lo and Zhang solution and the results of the FEM reaches
about 35% for thin liners.
On the contrary, the solutions of Thépot and Madryas and Szot provide the most
conservative estimate for the buckling pressure. However, pressure predictions become
highly over-conservative for liners with bigger gaps (i.e., 10). This is attributed to the
fact that the equations provided by Thépot provide reasonable fitting of the finite element
25

A.M. SWEEDAN, K.M. EL-SAWY
results related to gap liners not exceeding a threshold value of 10. The difference
between Thépot solution and the obtained finite element results ranges from -24.7% to
+2.7%, where the -24.7% corresponds to (t/R)=0.04 and (d/R)=0.1.
12.0
Tightly Fitted Liner (d / R) = 0.0
Normalized Bending Moment
[M cr / (E' R 2 )] x 10 5
10.0
8.0
6.0
4.0
2.0
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
14.0
Loosely Fitted Liner (d / R) = 0.05
Normalized Bending Moment
[M cr / (E' R 2 )] x 10 5
12.0
10.0
8.0
6.0
4.0
2.0
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
16.0
Loosely Fitted Liner (d / R) = 0.10
Normalized Bending Moment
[M cr / (E' R 2 )] x 10 5
14.0
12.0
10.0
8.0
6.0
4.0
2.0
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
Fig. 6. Bending moment at the point of maximum stress in liner with d/R=0.0, 0.05 and 0.1
Figure 5 shows the thrusts associated with the buckling pressure (based on FEM and Thépot
[9]) at the point of the maximum stress in the liner. This point corresponds to the mid-point of
the detached part of the liner. One may conclude the following from Figure 5:
26

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
20.0
Tightly Fitted Liner (d / R) = 0.0
Normalized Thrust [N cr / (E' R)] x 10 4
15.0
10.0
5.0
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
15.0
Loosely Fitted Liner (d / R) = 0.05
Normalized Thrust [N cr / (E' R)] x 10 4
12.5
10.0
7.5
5.0
2.5
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
12.5
Loosely Fitted Liner (d / R) = 0.10
Normalized Thrust [N cr / (E' R)] x 10 4
10.0
7.5
5.0
2.5
FEM
Thepot
0.0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
Fig. 5. Thrust at the point of maximum stress in liner with d/R=0.0, 0.05 and 0.10.
By inspection, it is clear that the ratio No/(PcrR) is greater than unity in all of the
considered solutions, where P cr /E’ is shown in Fig. 4 and N o /(E’R) is shown in Fig. 5.
The figure reveals also a very good agreement bet-ween Teapot’s solution and the finite element
results.
The bending moment associated with the buckling pressure (at the point of the maximum
stresses) is shown in Figure 6. In general, the figure shows very good agreement between
Thépot’s solution and the finite element results. For relatively thicker liners, Thépot’s solution
provides more conservative estimates for the bending moment at the point of maximum stresses.
27

A.M. SWEEDAN, K.M. EL-SAWY
Figure 7 shows the enhancement factor, K, versus the liner (t/R) for different values for (d/R),
where the enhancement factor is defined according to (3) and can be written as
3
Pcr Pcr
Pcr
1 d
R
K
4
3
' 4 '
2
(15)
Pcr
E t R E t R t R
o
It is clear that the enhancement in the liner stability is maximized at zero or small values of
(d/R) and small values of (t/R) (i.e. for the case of tightly installed thin liners). For thick liners
and large gaps (i.e. (d/R) = 0.07 to 0.10), the liner enhancement becomes almost constant.
50
Tightly Fitted Liner (d / R) = 0.0
Enhancement Factor K
40
30
20
10
FEM
Thepot
Lo and Zhang
Madryas and Szot
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
7
Loosely Fitted Liner (d / R) = 0.05
6
Enhancement Factor K
5
4
3
2
1
FEM
Thepot
Lo and Zhang
Madryas and Szot
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
5
Loosely Fitted Liner (d / R) = 0.10
4
Enhancement Factor K
3
2
1
FEM
Thepot
Lo and Zhang
Madryas and Szot
0
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
(t / R)
Fig. 7. Enhancement factor for liner with d/R=0.0, 0.05 and 0.10
28

CHARACTERIZATION OF ELASTIC BUCKLING OF CYLINDRICAL LINERS USING FINITE ELEMENT APPROACH
In view of Fig. 7, it can be shown that for the cases of thin liners with large gaps (i.e., small
(t/R) and (d/R)=0.05 and 0.10):
Solution of Lo and Zhang is the most non-conservative solution and may lead to unsafe
estimation of K.
On the other hand, the solution of Madryas and Szot, although based on Finite Element
Analysis, provides the most conservative K, especially at small (t/R) values,
Thépot’s solution is also in a very close agreement to the finite element analysis results.
5. CONCLUSION
The finite element method is used to study elastic stability of liners subjected to uniform
pressure. The adopted model accounts for the non-linearity effects of large deformations and
geometric imperfections on the stability of analyzed liners. Results of finite element analysis are
used to identify the critical buckling pressure and the associated maximum thrust and moment.
A critical review is also conducted to evaluate the available methods for estimating the elastic
critical pressure of slender liners. Comparison is held between finite element results and
predictions that are based on such methods from which the following conclusions may be drawn:
Liner’s buckling pressure decreases as the gap size increases. This decrease is more
pronounced for thin liners where the flexural stiffness-to-axial stiffness ratio is very small and
the behaviour is controlled by flexural deformations of the liner.
On the contrary, for thick liners, the axial deformations dominate the behavior and the liner
elastic stability depends mainly on its axial stiffness.
The solution proposed by Lo and Zhang provides non-conservative prediction of the critical
pressures especially for small values of (d/R). On the contrary, the solution of Thépot
provides more conservative estimate for the buckling pressure. However, pressure predictions
become highly over-conservative for liners with bigger gaps (i.e., 10).
Thepot’s method is shown to be in very good agreement with obtained finite element results
and provides more conservative estimates for the thrust induced in liners with small gaps (i.e.,
10 ).
For thin liners with large gaps the solution of Lo and Zhang is the most non-conservative
solution and may lead to unsafe estimation of the enhancement factor K. Meanwhile, the
solution of Madryas and Szot, although based on Finite Element Analysis, provides the most
conservative K, especially at small (t/R) values.
6. ACKNOWLEDGMENT
Support of this research was provided through research grants No. 03-7-11/99 and 02-7-11/00
from the Research Council of the United Arab Emirates University.
29

A.M. SWEEDAN, K.M. EL-SAWY
7. REFERENCES
[1.] Aggarwal, S. C., and Cooper, M. J. (1984). “External pressure testing of Insituform lining.”
Internal Rep., Coventry (Lanchester) Polytechnic, Coventry, U.K.
[2.] Kyriakides S., Babcock C. D. (1983). Buckle propagation phenomena in pipelines collapse:
the buckling of structures in theory and practice, Proc. IUTAM Symp. on Collapse, London,
Aug. 1982 (edited by J. M. T. Thomson and G. W. Hunt), Cambridge University Press,
75-91.
[3.] Lo, K. H. and Zhang, J. Q. (1994). “Collapse resistance modelling of encased pipes.” Buried
Plastic Pipe Technology, vol. 2, ASTM STP1222, Dave Eckstein, Ed., American Society for
Testing and Materials, Philadelphia, Pa., 97-110.
[4.] Guice L. K., Straughan T., Norris C. R., Bennet R. D. (1994). Long-term structural behavior
of pipeline rehabilitation systems, Louisiana Technical University, Trenchless Technology
Center, Technical Report #302.
[5.] Bakeer, R. M., Barber, M. E., Pechon, S. E., Taylor, J. E., and Chunduru, S. (1999).
“Buckling of HDPE liners under external uniform pressure.” J. Mat. in Civ. Engrg., ASCE,
11(4), 353-361.
[6.] Levy, M. (1884). “Me’memoire sur un nouveau cas inte’grable du proble’m de l’elastique et
l’une de ses applications.” English translation is “Memoir on a new integrable case of the
problem of elasticity and one of its applications.” J. Math. Pure et Appl., Lioville, France,
Series 3, Vol. 10, 5-42.
[7.] Jacobsen, S. (1974). “Buckling of circular rings and cylindrical tubes under external
pressure.” Water Power, 26 (Dec.), 400-407.
[8.] El-Sawy, K., Moore, I. D. (1998). “Stability of liners used to rehabilitate rigid pipes.” J.
Struct. Engrg., ASCE, 124(11), 1350-1358.
[9.] Thepot, O. (2000). “A new design method for non-circular sewer lining.” Trenchless
Technology Res., Vol. 15, No. 1, 25-41.
[10.] C. Madryas, A. Szot (2003). Structural sensitivity of circular sewer liners to geometrical
imperfections, Tunnelling and Underground Space Technology, 18, 421–434.
[11.] Glock, D., 1977. U¨berkritisches verhalten eines starr ummantelten kreisrohres bei
wasserdruck von aussen und temperaturerho¨hung [English translation: Post-critical
behaviour of a rigidly encased circular pipe subject to external water pressure and
temperature rise]. Der Stahlbau 46 (7), 212–217.
30

AHU J. of Engineering & Applied Sciences 3 (1) : 31-49 (2010)
© 2009 ALHOSN University
OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS
OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
*
Khaled M. El-Sawy, 1 Abdul Hameed Arwani 2
1 Civil & Environmental Engineering Department, Faculty of Engineering, United Arab Emirates University,
P.O. Box: 17755, Al Ain, UAE
2 Principal Engineer, Parsons, Abu Dhabi, UAE
ABSTRACT: The stress field around a hole in a laminated composite plate under uni-axial tension exhibits a
very complex pattern. Usually, the Finite Element Method (FEM) is used to calculate such stress field. The
accuracy of the Finite Element (FE) analysis depends mainly on the appropriateness of the mesh used in the
analysis. Adaptive meshing is one of the good options to search for that optimum mesh. Unfortunately, the
adaptive meshing technique involves iterative solutions of the problem with gradual refining of an initial coarse
mesh until an optimum mesh (that satisfies the requested user tolerances) is reached. These iterative solutions
may add an extra cost to the analysis if the initial mesh is very coarse. In addition to this, some problems still
exist with the error estimation used in the adaptive meshing technique especially when analyzing a problem with
multiple materials (which is the case for laminated composite plates). This initiated a parametric study that tries
to find some guidelines for the choice of the optimum mesh without the need to do costly adaptive meshing.
The sub-modeling method has been used to increase the accuracy of the solution around the hole (zones of high
stress concentrations). Different parameters that control the shape of the sub-model mesh have been studied and
recommendations for their optimum values are reported. These parameters are used later in the analysis of a
laminated plate with real sequence of plies and the problem size and run-time is proved to be affordable with
current PC configurations.
Key words: Finite Element Method, Laminated Composite Plate, Perforated Plate
1. INTRODUCTION
Bolts and rivets are usually necessary components of some structural joints. Despite careful
design, practically every structure contains stress concentrations due to bolt or rivet holes. In the
case of laminated structures, the presence of a hole in the laminate introduces significant stress
contributions in the third dimension, which creates a very complicated three-dimensional (3D)
stress field in the vicinity of the hole. This complex state of stresses depends on the stacking
sequence of the laminate, the fiber orientation of each lamina (or ply) as well as the material
properties of the fiber and the matrix [1, 5]. Ultimately, these stress concentrations form a primary
source of damage initiation especially in the form of delamination.
Delamination along the interface of two adjacent laminas has long been recognized as one
of the most important failure modes in laminated composite structures. The growth of
delamination may result in a substantial reduction of strength and stiffness of the laminate. A
primary requirement for the study of delamination in composite plates is to identify clearly
the stress field in the regions where delamination is most likely to happen.
Over the past century, many investigators have considered the problem of having a circular
hole in a composite plate where delamination happens in the vicinity of the hole. Since the
complex stress field around the hole is of great interest to correctly recognize delamination,
many researchers have considered several methods of determining the stresses in an axially
loaded rectangular plate with a centrally located circular hole [1-9]. Experimental studies
have shown that the interlaminar stresses at a free-edge, such as in the case of a hole, can
produce delamination under cyclic loading [1.]. Solution of such a problem (e.g. a uni-axially
loaded rectangular plate with centrally located circular hole) could be analytical or numerical.
______________________________________
* Corresponding Author. Tel.: +971-3-7133516
E-mail : k.elsawy@uaeu.ac.ae
31

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
Analytical or closed-form solutions are quite few in number [2.]. Alternately, much
advancement has occurred in using numerical solutions especially the use of Finite Element
Method (FEM). The presences of complex three-dimensional (3D) stress field in the vicinity
of the hole and the complex structural geometry have turned researchers to the use of FEM in
the last two decades. Two dimensional (2D) solutions are available in the literature [3.] and
provide some insight into the stress field in the neighborhood of the hole edge. Several
authors have recognized the importance of 3D analysis in dealing with such problems. Since
3D analysis is not cheap and needs more resources (time and computer hardware), several
researchers have tried to use hybrid methods, for example, the combination of the 2D
Classical Laminated Theory (CLT) with 3D FEM [4.] or the combination of 2D and 3D FEM
[5.]. Treating the problem in full 3D analysis had encountered difficulties in the past due to
complexity of the problem and lack of computer memory. For this reason the concept of
finding a FE strategy for the 3D analysis that optimizes the solution and reduces the huge
amount of resources required have been emphasized. Experience tells us that far away from
the hole, stresses are relatively smaller and their rates of changing with distance (stress
gradients) are small. However, in the vicinity of the hole, the stresses are relatively higher and
their gradients are very high. Generally, the FE mesh should be fine at stress concentration
regions and coarse far from these regions.
Recently, various approaches of optimizing 3D FE analysis of composite plates having
circular holes have been considered. Adaptive refinement is one of these approaches [6.].
Although adaptive meshing technique is a good candidate, unfortunately the error estimation
due to multiple materials still needs further investigation and verification. Problems with the
error estimation happen most likely at material interfaces (i.e. at laminas interfaces), which
are the regions of interest. In addition, adaptive meshing is not cheap since it involves
iterative process of error estimation, refinement, and mesh regeneration. The adaptive
meshing process usually terminates based on some tolerances, set initially by the analyst, to
ensure that the solution has converged to the required accuracy.
In this study we introduce a FE strategy that optimizes the modeling and analysis of
laminated composite plates with centrally located circular holes. The strategy is based on
detailed parametric analysis conducted to study the effect of different mesh parameters on the
desired results. The final outcome is some recommendations for the optimum mesh
parameters to be used. The conducted parametric analysis is expected to help in putting
guidelines for FE modeling of composite plates with holes and in reducing the costly
modeling by reducing solution time. Moreover, when used with adaptive procedure it could
help in reaching convergence faster. The following sections discuss the briefly the different
aspects of the FE technique and the detailed parametric study.
2. SUB-MODELING TECHNIQUE
To get accurate stress estimations, it is a common practice to use finer element sizes in
regions of the FE model that exhibits high stress concentration and allow for larger element
sizes away from these regions. For the case of laminated composite plate, this implies using a
very fine mesh around the hole while allowing larger element sizes away from the hole. The
use of a fine mesh around the hole would add more cost to the 3D analysis of the laminated
composite plate. A cheaper method of achieving the same goal (i.e., accurate results around
the hole) is the use of the sub-modeling technique.
The sub-modeling technique provides the solution and stress estimations in two stages. The
32

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
first stage is to use a model with a coarse mesh (termed as global model, Fig. 1a) sufficient to
get accurate stresses away from the hole (i.e. away from the localized zone of high stress
concentration). The second stage uses a model with a very fine mesh (termed as sub-model,
Fig. 1b) for the zone around the hole (i.e., zone of interest) and then get interpolated values
for the displacements and rotations from the previously solved global model and prescribe
them as displacement and rotation loads in the fine sub-model. Figure 1 shows the concept of
sub-modeling.
(1a) Global coarse Model with Superimposed High Stress Regions
Sub-models
(1b) Fine sub-models superimposed in the coarse model
Fig 1. Models Used in the Finite Element Sub-modeling Technique
1 Spacing ratio is defined as the ratio of the size of first element to the size of the last element i.e. if the spacing ratio was 14 in certain
direction, then the ratio of the first element length in that direction to the last element length in the same direction is 14.
33

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
r
2a
2W
2L
Fig 2. Problem geometry
3. PROBLEM DEFINITION AND THE FINITE ELEMENT MODEL
Figure 2 shows a laminated composite plate having a central circular hole and being axially
loaded at both ends. The plate is composed of four plies with symmetrical stacking sequence
(typically 0/30/30/0, 0/45/45/0, 0/60/60/0 or 0/90/90/0). The material of each ply is assumed
to be homogenous, linear elastic, and orthotropic. The direction of orthotropy coincides with
the direction of the fibers (i.e. the material is stiff in the direction of fibers and less stiff in the
other perpendicular directions). Geometry is defined by the plate length L, plate width W,
radius of the central hole, a, and the thickness of each ply, h (i.e. total plate thickness is 4h).
The mesh parameters are defined as follows:
N z is the number of elements through the thickness of each ply,
R z is the spacing ratio 1 of elements through the thickness of each ply,
N is the number of elements around the edge of the hole and no spacing ratio is used in this
direction,
N r is the number of elements extending radially out from the edged of the hole edge to the
end of the sub-model, and
is the spacing ratio of the elements in the radial direction.
R r
The ratio of the radius of the hole to the ply thickness, a/h, has been taken to be 4, 20, and
100 to examine three schemes of composite plates; thick, medium, and thin plates.
L = 12a
Uniform displacement
Ux = 0.0125h
W = 6a
N r , R r
N
N z , R z
Sub-model
2h
Material Properties
Ex=139.0 GPa Gxy=5.52 GPa xy=0.3
a S=6h Ey=10.43 GPa Gxz=5.52 GPa xz=0.3
Ez=10.43 GPa Gyz=3.31 GPa yz=0.55
Fig 3. Finite Element model
34

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
Because of the symmetry of the problem with respect to x, y, and z directions, only oneeighth
of the laminate is analyzed. Figure 3 shows the finite element model, prescribed
displacement boundary conditions, mesh variables, and the orthotropic material properties. It
has been shown by [4.] that the plate dimensions L and W are essentially infinite with respect
to the diameter of the hole and that the far-field stresses are reached within a distance of 6
times the ply thickness (i.e. 6h) away form the hole edge. Thus a region of radial distance of
at least S=6h (measured from the edge of the hole) has been considered as the sub-model
where more refinement has been done to depict the complex stress field.
High order isoparametric 20-noded brick elements were utilized in the analysis. These
elements are more tolerant of curved boundaries than standard 8-node brick elements and
avoid “locking” due to high aspect ratios.
Figure 3 shows the mesh parameters. The default values for these parameters are as follows
unless otherwise noted: N z =6, R z =14, N r =10, R r =40, and N =30.
4. VERIFICATION FOR THE SUB-MODEL METHOD
It is essential to check the accuracy of the FE analysis using the sub-modeling technique
before using it in performing the parametric analysis. This would ensure that stresses are not
affected by the location of the sub-model boundaries.
Two FE models, for the problem geometry shown in Fig. 3 with a/h=20, were used for
comparison purposes (Fig. 4). The first one, MODEL-1, has a very fine mesh over the entire
model while the second, MODEL-2, has used the sub-modeling technique which involves the
use of two meshes for the global model and the sub-model. The mesh for the zone of interest
in MODEL-1 (i.e. in the neighborhood of the hole defined by dimension S in Fig. 3) matches
exactly the mesh of the sub-model of MODEL-2 and has been based on the parameters N z =6,
R z =14, N r =10, R r =40, and N =30. Figure 4 shows the different meshes used for both models.
(4a) Mesh for MODEL-1
(4b) Mesh for MODEL-2
35

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
(4c) Mesh for sub-model of MODEL-2
Fig 4. Finite Element models to test the reliability of sub-modeling technique
Since the interlaminar stresses, which are most likely, the source of delamination, are very
sensitive to the shape of the mesh, a comparison between the stresses resulting from the two
models would indicate how accurate MODEL-2 relative to MODEL-1 is. From comparison,
it is noticed that results from MODEL-2 show excellent agreement with the ones of MODEL-
1 with reduction in solution time up to 80%. Figures 5a through 5c show the comparison
between the results for MODEL-1 and MODEL-2. Interlaminar stresses are shown along
paths extend radially form the edge of the hole at 0, 45, and 90 degrees where the 0-degree
path coincides with the x-axis and the 90-degree path coincides with the y-axis. All results are
normalized against the far field stress, avg , due to applied displacement ( avg = 74.1 MPa).
It is also essential to verify that the chosen boundaries for the sub-model are far enough
from the stress concentrations around the hole (i.e. distance S is sufficient). We can do this by
comparing stresses along the cut boundaries with those along the corresponding locations on
the global model mesh of MODEL-2. If the results are in good agreement, this indicates that
proper boundaries have been chosen. Figure 6 shows excellent agreement between x and y
stresses in the sub-model and the corresponding stresses in the coarse model. Stresses are
compared at the material interface between the two plies.
DESCRIPTION OF THE PARAMETRIC ANALYSIS
A parametric study for the problem (shown in Fig. 3) is performed to identify the optimum
values for the parameters defining the FE mesh of the sub-model (i.e. N z , R z , N r , R r , and N ).
The problem geometry is defined by the plate length L, width W, radius of the central hole, a,
and the ply thickness, h. The default values for the mesh parameters are N z =6, R z =14, N r =10,
R r =40, and N =30 (unless otherwise noted).
To verify that the FE results (for any mesh parameters) have reached the required accuracy,
there should be a measure for that. The rate of change of the maximum interlaminar stress
(i.e. the maximum of rz , zz , or z at the circular boundary of the hole) with any mesh
parameter is selected to estimate the accuracy of the results of any mesh. The maximum
interlaminar stress has be chosen for this purpose since it is, generally, responsible for the
delamination happening in most of the laminates composite plates and since other stresses are
small and change abruptly, which consequently leads to fluctuation in the measurements of
the rate of change of the stress between any two consecutive runs.
Figure 7 shows different interlaminar stress components along the boundary of the hole for
different stacking sequences. The maximum interlaminar stress is found to be the component
36

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
of the stress z along the boundary of the hole and between =70 o and =80 o , where is
measured from the x-axis counterclockwise. The previous observation is valid for all the
examined stacking sequences and hole sizes.
0.06
zz / avg
0.04
0.02
0.00
0-degree (MODEL-1)
45-degree (MODEL-1)
90-degree (MODEL-1)
0-degree (MODEL-2)
45-degree (MODEL-2)
90-degree (MODEL-2)
-0.02
-0.04
0 1 2 3 4 5
(r-a)/h
(5a) Stress zz (Comparison between MODEL-1 and MODEL-2)
0.4
0.3
45-degree (MODEL-1)
45-degree (MODEL-2)
z / avg
0.2
0.1
0.0
0 1 2 3 4 5
(r-a)/h
(5b) Stress z (Comparison between MODEL-1 and MODEL-2)
0.06
0.04
0.02
rz / avg
0.00
-0.02
-0.04
-0.06
0-degree (MODEL-1)
45-degree (MODEL-1)
90-degree (MODEL-1)
0-degree (MODEL-2)
45-degree (MODEL-2)
90-degree (MODEL-2)
-0.08
0 1 2 3 4 5 6
(r-a)/h
(5c) Stress rz (Comparison between MODEL-1 and MODEL-2)
Fig 5. Comparisons between MODEL-1 and MODEL-2
37

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
2.0
1.5
Normalized Stress
1.0
0.5
x - MODEL-1
y - MODEL-1
x - MODEl-2
y - MODEL-2
0.0
-0.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(6a) (Distance Along the Cut-Boundary) / a
Fig 6: Comparisons between Stresses in MODEL-1 and MODEL-2 at the Boundaries
/ avg
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
z [0/90]
z [0/60]
z [0/45]
z [0/30]
zz
[0/90]
zz
[0/60]
zz
[0/45]
zz
[0/30]
rz
[0/90]
rz
[0/60]
rz
[0/45]
rz
[0/30]
a/h=20, N z
=8, R z
=14, N r
=10, R r
=40, and N =30
-0.6
0 10 20 30 40 50 60 70 80 90
degree
Fig 7. Interlaminar stresses along the hole boundary for different stacking sequences
Figure 8 shows z , the maximum stress, for different plate thickness (thick with a/h=4,
medium with a/h=10, and thin with a/h=100) with [0/90/90/0] stacking sequence. It is worth
mentioning that maximum stress appears to be more sensitive to the change in the a/h ratio
than the change in the stacking sequence. Later in this study, we will find that increasing the
number of elements through the thickness N z from 8 to 12 for the case of a/h=4 will improve
the accuracy of the results to reach the same accuracy for the cases of a/h=20 and 100.
The parametric analysis is performed by varying one mesh parameter at a time while fixing
all the others. Comparison between the maximum interlaminar stress z of any two
consecutive solutions (i.e. consecutive meshes based on the considered mesh parameter)
provided a good measure of accuracy of the mesh. The FE solution is assumed to converge to
the accurate value whenever the difference in the maximum interlaminar stress (i.e. z at
angle between =70 o and =80 o ) between any two consecutive computer runs becomes less
than 1.0%, and the value of the parameter at the last run is taken as the optimum value for this
case. At this stage, further refinement of the mesh would enhance the results by less than
1.0%, which is not economical compared to the cost of the solution time of the refined mesh.
38

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
1.4
1.2
[0/90], N z =8, R z =14, N r =10, R r =40, and N =30
1.0
0.8
z
/ avg
0.6
0.4
0.2
0.0
-0.2
-0.4
0 10 20 30 40 50 60 70 80 90
degree
a/h=100
a/h=4
a/h=20
Fig 8. Stress z along the hole boundary for different hole sizes
6. RESULTS OF THE FE PAREMETRIC ANALYSIS
In the parametric analysis each mesh parameter was examined alone while fixing other parameters
at their default values until convergence is obtained. The following sections report the effect of
changing these parameters on the accuracy of the solution. All the forthcoming results (reported
in figures 9-21) show the variation of the %difference between the interlaminar stress z at the
hole boundary of any two successive meshes with the angle (at which the difference in stress is
calculated). The shaded areas shown in Figs. 9-21 correspond to the location of the maximum
interlaminar stress z (i.e. at between 70 o and 80 o ).
Symbols shown in the figure legend refer to the varying mesh parameter and its value for
the two successive meshes. For example, in Fig. 9 the varying mesh parameter is N z and the
figure shows two cases; difference in z between meshes with N z =6 and N z =7, and difference
in z between meshes with N z =7 and N z =8. One may notice that the symbols for the
difference between any two successive runs show a discontinuity at =20º-30º (+ to -
discontinuity). This is because values of the stress z at these locations are very small and
approach zeros (Fig. 8) and therefore any slight stress difference will represent a huge
%difference.
A. Analysis of Parameters N z and R z
Since we are concerned with the interlaminar stresses, N z and R z are expected to be the most
important parameters since they are participating considerably in controlling the element size
at the material interface. The parametric analysis started with initial values of N z and R z as 6
and 14, respectively. The analysis has been performed for different stacking sequences and
different plate thickness (i.e. different a/h ratios). Figures 9-15 show the results of such
parametric analysis for N z and R z .
It is clear from Figs. 9-12 for the case of R z =14, N r =10, R r =40, N =30, S=6h and different
stacking sequences, that N z converges (i.e. produce %difference less than 1%) at number of
elements N z between 8 and 10. We can notice also that for thick plates (i.e. a/h=4) more
elements are needed through the ply thickness while for thin plates (i.e. a/h=100) fewer
elements are needed. This conclusion is general because it is observed that the interlaminar
stresses increase in the neighborhood of the hole (i.e. more stress concentration) as a/h ratio
decreases, which agrees with Choi and Folias 1994. In general we could conclude that for
thick plates (i.e. a/h=4) N z =10 could be enough, for medium ones (i.e. a/h=20) N z =8 is
39

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
enough, and for thin ones (i.e. a/h=100) less number could be reliable (N z =6).
For R z parameter it is noticed that changing this parameter does not affect the results
significantly (Fig. 15). In addition, it does not affect the solution time and cost since the
number of DOFs does not change. Therefore, this parameter could take any reasonable value
between 12 and 16.
%Difference between two consecutive runs
3.0
2.5
2.0
1.5
1.0
0.5
[0/90] s
& a/h=20
R z
=14, N r
=10, R r
=40
N
=30, S=6h
Nz=7 & Nz=6
Nz=8 & Nz=7
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 9. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
%Difference between two consecutive runs
2.0
1.5
1.0
0.5
[0/45] s
& a/h=20
R z
=14, N r
=10, R r
=40
N
=30, S=6h
Nz=10 & Nz=9
Nz=9 & Nz=8
Nz=8 & Nz=7
Nz=7 & Nz=6
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 10. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
40

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
%Difference between two consecutive runs
2.5
2.0
1.5
1.0
[0/30] s & a/h=20
R z =14, N r =10, R r =40
N =30, S=6h
0.5
Nz=10 & Nz=9
Nz=9 & Nz=8
Nz=8 & Nz=7
Nz=7 & Nz=6
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 11. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
%Difference between two consecutive runs
2.5
2.0
1.5
1.0
[0/60] s & a/h=20
R z =14, N r =10, R r =40
N =30, S=6h
0.5
Nz=10 & Nz=9
Nz=9 & Nz=8
Nz=8 & Nz=7
Nz=7 & Nz=6
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 12. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
Error between two consecutive runs
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
[0/90] s & a/h=4
R z =14, N r =10, R r =40
N =30, S=6h
Nz=10 & Nz=9
Nz=9 & Nz=8
Nz=8 & Nz=7
Nz=7 & Nz=6
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 13. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
41

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
%Difference between two consecutive runs
2.0
1.5
1.0
0.5
[0/90] s & a/h=100
R z =14, N r =10, R r =40
N =30, S=6h
Nz=8 & Nz=7
Nz=7 & Nz=6
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 14. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
%Difference between two consecutive runs
1.5
1.0
0.5
[0/90] s & a/h=20
N z =8, N r =10, R r =40
N =30, S=6h
Rz=14 & Rz=12
Rz=16 & Rz=14
0.0
0 10 20 30 40 50 60 70 80 90
degree
Fig 15. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N z parameter
B. Analysis of Parameters N r and R r
Based on the discussion in the previous section, N z is kept at value of 8 while R z is kept at 14
for all the forthcoming analyses. The effect of N r on the results was examined by assuming
different values for N r between 10 and 13 (Figs. 16-20). It is found that increasing N r would
enhance the results for all the considered stacking sequences and a/h ratios. This is because N r
participates significantly in the element size (in the plane of the plate) at the hole
neighborhood, which is a major factor affecting the accuracy of the results. In general we can
state that for thin and thick plies (a/h=100 and a/h=4) high number of elements and high N r
are required for convergence, while for moderately thick ones (a/h=20) fewer element and
low N r are needed. As an example, for a/h ratio equals 20, N r could be taken as 13 (Fig. 16),
while for a/h=4, N r equals 14 or 15 could be reasonable (Figs. 17 and 18).
The effect of the stacking sequence on Nr is shown in Figs. 16,19, and 20. The figures show
that the less the difference between the orientations (e.g. 0/45) of the two plies the less the value
of N r .
______________________________________
* Corresponding Author. Tel.: +971-3-7133516
E-mail : k.elsawy@uaeu.ac.ae
42

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
%Difference between two consecutive runs
4
3
2
1
[0/90] s & a/h=4
N z =8, R z =14, R r =40
N =30, S=6h
0
0 10 20 30 40 50 60 70 80 90
degree
Nr=11 & Nr=10
Nr=12 & Nr=11
Nr=13 & Nr=12
Fig 16. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N r parameter
%Difference between two consecutive runs
1.8
1.6
1.4
1.2
1.0
0.8
[0/90] s & a/h=20
N z =8, R z =14, R r =40
N =30, S=6h
0.6
0 10 20 30 40 50 60 70 80 90
degree
Nr=11 & Nr=10
Nr=12 & Nr=11
Nr=13 & Nr=12
Fig 17. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N r parameter
%Difference between two consecutive runs
2.0
1.5
1.0
[0/90] s & a/h=100
N z =8, R z =14, R r =40
N =30, S=6h
0.5
0 10 20 30 40 50 60 70 80 90
degree
Nr=11 & Nr=10
Nr=12 & Nr=11
Nr=13 & Nr=12
Fig 18. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N r parameter
43

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
%Difference between two consecutive runs
1.2
1.0
0.8
0.6
0.4
[0/45] s & a/h=20
N z =8, R z =14, R r =40
N =30, S=6h
Nr=11 & Nr=10
Nr=12 & Nr=11
0.2
0 10 20 30 40 50 60 70 80 90
degree
Fig 20. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N r parameter
%Difference between two consecutive runs
1.2
1.0
0.8
0.6
0.4
[0/60] s & a/h=20
N z =8, R z =14, R r =40
N =30, S=6h
Nr=11 & Nr=10
Nr=12 & Nr=11
Nr=13 & Nr=13
0.2
0 10 20 30 40 50 60 70 80 90
degree
Fig 20. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N r parameter
Based on his discussion, we may conclude that for thick and thin plates N r is advised to be
15, for medium ones N r =13 could enough. As the difference between the plies orientations
increases, one should add 2 to 4 elements to the above numbers.
As stated earlier in the analysis of R z , R r does not affect the solution speed and cost, and
then it can take any reasonable value between 40 and 60.
C. Analysis of Parameter N
The number of elements in the circumferential direction is usually dictated by the geometry of
the circular hole. This dictated number of elements is usually enough to produce accurate
stress results. The initial value of N was set to 30 elements. Increasing the value of N from
30 elements to 40 affects the results with %difference more than 1%, while increasing this
value from 40 to 50 elements affects the results slightly (Fig. 21) although it increases the
DOFs and solution time considerably. This applies for all stacking sequences and a/h ratios.
Therefore, we may conclude that N =40 is a reasonable choice that would provide accurate
stress results.
44

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
%Difference between two consecutive runs
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
[0/90] s & a/h=20
N z =8, R z =14, R r =40
N r =10, S=6h
-3.0
0 10 20 30 40 50 60 70 80 90
degree
N =40 & N =30
N =50 & N =40
Fig 21. Difference in stress z between two consecutive computer runs along the hole
boundary for the analysis of N parameter
D. Analysis of a Plate with Real Ply Sequence
The previous parametric study has been conducted for symmetrical four-ply laminated
composites having different stacking sequences and hole size to thickness ratio. The study
presented in this section targets the stress analysis of the same laminated composite plate with
the same material, but with a real number of plies and a real ply-sequence. The newly
examined symmetrical stacking sequence of the 18 plies is defined by 0/0/0/ 45/45/45/
90/90/90/ 90/90/90/ 45/45/45/ 0/0/0. The ratio of the hole size to the ply thickness (a/h) is
chosen to be 20.
Fig 22. Geometry of the Plate and the General Details of the Global Mesh (a/h=20)
As before, the sub-modeling technique is used and two FE models and meshes (i.e., global
and sub-models) have been used. The parameters used for the mesh for the zone of interest in
the global model are defined as follows:
N z = 2, R z . = 1
N = 10
N r = 5, R r = 15
45

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
Figure 22 shows the geometry and FE mesh of the zone of interest with the previously defined
parameters and the mesh for the rest of the global model.
Fig 23. FE mesh of the sub-model based on the chosen mesh parameters (a/h=20)
Figure 23 shows the geometry and the fine mesh for the sub-model where the following
parameters are used:
N z = 6, R z . = 10 (i.e. 3 elements on each half-thickness of the ply)
N = 40
N r = 15, R r = 50
An Intel Core2 Duo PC with 2.33 GHz processor has been used to perform all the
calculations. A hard disk space of about 160 MB was used to store all the vital files for the
analysis. The time required to analyze the global model for plate with 18 symmetrical plies
was less than one minute while the time required to analyze the sub-model was slightly less
than 10 minutes and required about 2.4 Giga Bytes of storage. This computation power
promises that a PC can handle a plate with number of symmetric plies more than 18, but with
more demand on the storage space.
E. Effect of Less Detailing of the Model through the Thickness Direction
Fig 24. The FE mesh of the sub-model with fine mesh around all plies (a/h=20)
46

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
Fig 25. The FE mesh of the sub-model with fine mesh around three plies only (a/h=20)
Fig 26. Stress zz along the ply thickness for a model with fine mesh for all plies (a/h=20)
Fig 27. Stress zz along the ply thickness for a model with fine mesh for the ply under
consideration and its sandwiching plies (a/h=20)
To reduce the problem size and the run time, the ply under consideration and its
sandwiching plies can be finely meshed while the rest of the other plies can be coarsely
meshed. Two analyses have been performed; the first for a sub-model mesh with N z =6 for all
plies (Fig. 24) and the second for a mesh with N z =6 for the ply under consideration and its
sandwiching ones and N z =2 for the rest of the plies (Fig. 25).
47

KHALED M. EL-SAWY, ABDUL HAMEED ARWANI
Figures 26 and 27 show the distribution of zz along the thickness of the considered ply. It
is clear that there is no considerable difference in the results from both analyses which proves
that fine meshing through the ply thickness is only required for the ply under consideration
and its sandwiching plies.
7. CONCLUSION
A study is performed to select the optimum values for the parameters controlling the mesh
used in the Finite Element analysis of laminated composite plates with holes. The mesh
parameters considered are N z , R z , N r , R r , and N . The recommended values for the mesh size,
arrangement, and parameters are as follows:
1) The number of elements through the ply thickness N z depends mainly on the ply
thickness. For thick plies (i.e. hole radius to ply thickness ratio a/h=4) N z can be taken as
10. For medium ones (i.e. a/h=20) N z =8 is enough, while for thin ones (i.e. a/h=100) N z =6
is sufficient.
2) The spacing ratio R z for the elements through the ply thickness could take any reasonable
value between 12 and 16.
3) The number of elements in the sub-model along the radial direction N r depends on the
hole radius to ply thickness ratio a/h. For a/h=20, N r could be taken as 13, while for
a/h=100 and a/h=4, N r may be reasonably selected between 14 and 15.
4) The spacing ratio for the element size in the radial direction R r can take any value
between 40 and 60 since it does not affect the size of the problem.
5) The number of elements along the hole circumference N may be reasonably chosen as 40.
6) To reduce the size and run-time of the composite plate problem, the ply under
consideration and its sandwiching plies can be finely meshed along the thickness direction
while the rest of the other plies can be coarsely meshed with 2 elements along the ply
thickness..
8. ACKNOWLEDGMENT
Support of this research was provided through a research project funded by the Dassault
Aviation, France, and the United Arab Emirates University.
9. REFERENCES
[1.] R. B. Pipes & N. J. Pagano. “Interlaminar Stresses in Composite Laminates Under
Uniform Axial Extension,” J. Composite Materials, Vol. 4, p. 204, 1970.
[2.] D. Pullman & J. R. Schaff. “Three Dimensional Spline Variational and Finite Element
Theory: A Comparison for a Laminate with a Circular Hole,” Structural Design and
Analysis Methods, 1996.
[3.] S. G. Lekhnitskii. “Theory of Elasticity of an Anisotropic Body,” Mir Publishers,
Moscow, 1963.
[4.] I. S. Raju & J. H. Crews, Jr. “Three Dimensional Analysis of [0/90]s and [90/0]s
Laminates with a Central Circular Hole,” NASA-TM-83300, April 1982.
48

OPTIMIZED 3D MESH FOR THE FINITE ELEMENT ANALYSIS OF LAMINATED COMPOSITE PLATE WITH CENTRAL HOLE
[5.] W. E. Lucking, S. V. Hoa & T. S. Sankar. “The Effect of Geometry on Interlaminar
Stresses of [0/90]s Composite Laminates with Circular Holes,” Journal of Composite
Materials, Vol. 17, pp. 188-198, 1984.
[6.] Y. Choi & E. S. Folias. “The 3D Stress Field in a Laminated Composite Plate with a Hole
Based on an h-r Finite Element Method,” AMD-Vol. 196. Failure Mechanics in
Advanced Polymeric Composites, ASME, 1994.
[7.] J. Y. Kim & C. S. Hong. “Three-Dimensional Finite Element Analysis of Interlaminar
Stresses in Thick Composite Laminates,” Computers & Structures, Vol. 40, No. 6, pp.
1395-1404, 1991.
[8.] T. Nishioka & S. N. Atlri. “Stress Analysis of Holes in Angle-Ply Laminates: An Efficient
Assumed Stress ‘Special-Hole-Element’ Approach and a Simple Estimation Method,”
Computers & Structures, Vol. 15, No. 2, pp. 135-147, 1982.
[9.] W. C. Hwang & T. C. Sun . “A Finite Element Iterative Approach for Analysis of
Laminated Composite Structural Elements,” Computer & Structures, Vol. 31, No. 1, pp.
55-62, 1989.
49

AHU J. of Engineering & Applied Sciences 3 (1) : 51-73 (2010)
© 2009 ALHOSN University
AFFORDABILITY, ENABLEMENT, AND SUSTAINABILITY OF
LOW-INCOME HOUSING
Mahbubur Rahman *
Universiti Teknologi Malaysia
International Campus, 54100 Jalan Semarak, Kuala Lumpur, Malaysia
ABSTRACT: Squatter settlements providing housing to 30-70% of the urban population in many developing countries
have grown due to widespread poverty and inadequate housing finance and land development systems. Governments
mostly assisted by the international aid agencies have improved environment, tenure security, income and resources in
many settlements, but could not eradicate the problems as benefits did not multiply due to lack in institutional
development, policy implementation, governance, participation etc. Moreover, the squatters’ capability to bring
affordable and sustainable solutions was ignored. This paper discusses the changing approaches to the issues of lowincome
groups housing worldwide in the above context, and examine the role of the World Bank in setting the trend.
Keywords: Affordability, Enablement, Housing Policy, Self-Help, Squatter Settlement, Sustainability, Urban
Development, World Bank.
1. INTRODUCTION
The emergence of cities as the unchallenged site of human development for the future and the goal
of sustainable development have pushed hundreds of them to seek better means to encourage a
synoptic view, act as levers for strategic change, and facilitate sustainable development (Holden
2006). Rapid urbanisation taking place in the developing world's cities needs to be critically
managed. There has been greater appreciation of their growing importance in the national
economies recently as development became dependent on the ability of the urban centres to meet
the essentials like housing. A rising standard of living and political ideologies have increased the
awareness of human needs and social values (UKGP 1998), putting the onus to provide them,
endorsed by the Habitat Conferences, on the government.
The developing world’s urban population is going to be doubled by 2030; this will be accompanied
by dramatic growth in the number of slums where a sixth of humanity lives today (UNCHS 2003).
Given the social, economic and political situation of these people, these ill-built and ill-served
houses were what most of them could afford (World Bank 1993), subject to exploitation due to
scarcity (Tipple 1994). The global community was falling short of the Millennium Development
Goal’s (MDG) target of significantly improving the lives of at least 100 million slum dwellers by
2020 (UN 2000). 1
This paper attempts to infer sustainable housing out of the concepts of sustainable development,
compare the low-income group’s housing in the developing world against that, and establish the
changing approaches to the low-income groups housing worldwide starting from the 1950s. It
particularly examines the role of the World Bank, and other international bodies in setting the core
1 Although slums are defined in a broader manner than inadequate housing, the UNHCS (2006) refers to it as the ‘shelter
dimension of urban poverty’. According to it, the number of slum dwellers in the world has increased from 715 million in
1991 to 913 million in 2001 and to 998 million in 2005, adding another 50 million in 2005-7. UNHCS (2007) projects that
the world will have 1.4 billion slum dwellers in 2020.
_____________________________________________________
* Corresponding Author. Tel.: +6017683404، Fax: +60326934844
E-mail : mmrahman417@gmail.com
51

MAHBUBUR RAHMAN
themes. It also highlights the advantages of self-built incremental in situ upgrading. The paper
builds the case supporting such developments as an affordable means of providing sustainable
housing to the low-income groups in the developing countries.
2. SUSTAINABLE DEVELOPMENT
According to the WCED (1987), sustainable development means meeting "the needs of the present
without compromising the ability of future generations to meet their own needs". The idea emerged
in the 1980s to bridge the gap between environmental concerns about the ecological consequences
of human actions and socio-political concerns about development issues. While the
conservation/preservation debate went on, the issues of population growth, pollution, and nonrenewable
resource depletion got prominence (Robinson 2004). 2 Government and private bodies
adopted the term sustainable development; academics used 'sustainability' reflecting the more
managerial and incremental but less radical approach of the Brundtland report. Concerned that
development was seen synonymous to growth, they focussed on the role of institutions,
governance, and social capital in the process (Lehtonen 2004). Following will be a brief resume of
approaches to sustainable development and its implication in housing.
2.1 Three Pillars:
Initially not distinct, later sustainable development became a combination of ecological, economic,
and social dimensions. This ‘triple bottom line’ constituted the basis for the generally accepted
definitions of sustainable development (Lehtonen 2004). The international organisations endorsed
the hierarchically equal, mutually interacting dimensions. While the importance of each pillar may
vary from situation to situation, the model did not attribute any priority; it rather placed them as
independent elements that can be analysed separately. Characteristics and logic specific to each of
them may conflict with others, with no basis to arbitrate between the conflicting objectives of
economic rationality, social justice and ecological equilibrium (Lehtonen 2004).
While the three-pillar concept has been criticised for legitimising the current goals of the society,
government agencies found favourable objectives to corroborate it. It also perpetuated the
‘economism’ and ‘productivism’ characters of modern societies. Yet this articulated a new set of
checks and balances beyond the basic efficiency-equity, and cost- benefit binaries of traditional
policy analysis. However, maintaining a positive balance in all three areas directly implies
privileging limits and precaution over growth and accumulation.
Philosophically, Norton (1999) defined sustainability as a social imperative, not ecological with
social and economic implications. Until recently, sustainable development perceived as
environmental issue not integrated into economic decision-making. The importance of social
dimension in the overall economy— considered the weakest ‘pillar’ of sustainable development in
its analytical and theoretical underpinnings, is only seen now (Forster-Kraus et al. 2009). Woolcock
(2001) attributed the social interest to the fall of communism, the ostensible difficulties of creating
2 Rather than a consensus-based understanding, the Brundtland report presents the term as a language truce about a set of
ideas, like democracy, freedom, or justice, over which wars could be waged (Mebratu 1998; Pezzoli 1997; Guha &
Martinez-Alier 1997). For international development scholars and practitioners in environmental sciences and policy
fields, sustainable development has become a universally integrative term, elevating the idea beyond urban planning and
policy domains.
52

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
market institutions in transitional economies, the financial crises in Latin America, East Asia, and
Russia, and the persistent problems of unemployment and social marginalisation.
The three-pillar model strengthens the idea that the economy can be detached from the social
context that embodies all human activities, separating the ‘social’ from the ‘economic’ and denying
the relations between human societies and their environment (Passet 1996). Moreover, though
always prevalent, these are not qualitatively equal; disagreements over their hierarchy have
prevented a model from being widely adopted. However, the social dimension that is often difficult
to quantify cannot be analysed through the same framework and tools as the ecological or
economic one (Empacher 2002; Dubois & Mahieu 2002). Passet (1996) suggested an alternative
model where economic activities should serve all while also safeguarding the biophysical systems
for their existence; the social would be in the command of the economic, and also to the ultimate
environmental constraints.
The hierarchy does not mean that the environment would always be the most important and
relevant dimension. In a given situation, social or economic aspects may be more relevant as long
as the operation of socioeconomic systems does not conflict with the environmental framework
(Norgaard 1994).
2.2 Sustainable Housing:
Sustainability focussed on social and economic conditions in developing countries, their connection
to environmental degradation, and the people’s coping ability. Policies advocated for sustainable
development see urban development as that promotes economic growth, maintains social inclusion,
and minimises environmental impact. Some adopters of sustainability frameworks, e.g. the European
Commission (2001), created more directed approaches for building professionals: “economic growth
[that] supports social progress and respects the environment, social policy [that] underpins economic
performance, and environmental policy [that] is cost-effective” (Roseland 2000).
Sustainability is a political act based on human decisions and ways of life, not a scientific concept
(Robinson 2004). According to Greider (1997), it “carries revolutionary implications” for urban
planning and management, “but sounds so wholesome that almost everybody can endorse it”. Initially
conceived as a term related to macro economic development, sustainability has been applied to
housing only recently (Choguill 1999). Given the phenomena of urbanisation and severe housing
problems in these cities where the greatest resource use occurs and which generate the most waste
products and pollution, the concept has relevance to urban housing with economic, environmental
and social implications.
The complexity accompanying this shift is a reason for this delay. To remain meaningful, the concept
of sustainable human settlements must stay within the absorptive capacity of local and global waste
(Foy & Daly 1992), the achievement of the sustainable use of renewable (Daly 1992) and replenishable
(Rees 1996) resources, the minimisation in the use of non-renewable resources (El Sarafy 1989),
and meeting basic human needs (Hardoy et al. 1992). The last, distinguishing the definition from the
more general environmental approaches to sustainability, guides the housing issues.
The relationship between development and urban social conditions is complex. That social aspects
of housing were as important as environment and economic dimensions of sustainability is
increasingly realised (Forster-Kraus et al. 2009). The economic sector addressing the financial
aspects of social justice, accompanying the environmental sustainability, is an important element of
53

MAHBUBUR RAHMAN
it, while the environmental limits are an important constraining influence on economic growth.
Social sustainability refers to “policies and institutions that can integrate diverse groups and
cultural practices in a just and equitable fashion. As cities develop, the degree of social inequality,
cultural conflict and political fragmentation experienced within urban boundaries has increased”
(Stren & Polese 2000).
The “free-market” city attempts to overcome both environmental and social challenges as their
wealth increases. Thus growth not only increases the general wealth of society, it also increases
scope for protecting environment by trickling down. In order to be sustainable, housing initiatives
must be economically viable, socially acceptable, technically feasible and environmentally
compatible; housing policies should target to achieve these aims. The ‘brown agenda’ laid down
guidelines for sustainable urban development, followed by the UN calling the local governments to
mobilise broad-based, participatory, and sustainable environmental improvement. Implementing
the agreements reached at the 1992 Rio Summit required a concerted action at the local level
(Agyeman & Evans 1996), focussing on community, participation, partnership, accountability, etc.
3. HOUSING POLICY AND PRACTICE
In the decades after the WWII, public housing became a major mode in the developing countries
where meagre resources were spent on housing mostly for the government employees (Keivani &
Werna 2001). Aimed at eliminating the 'ugly ' informal settlements, these could not overcome the
escalating urbanisation, overcrowding, poverty, and informal settlement growth (Pugh 1995). Many
of them did not meet the low income populations’ demand; 90% of the urban population depended
on the private and informal sectors (Baross & van der Linden 1990; Drakakis-Smith 1981; Kevani
& Werna 2001). For example, Grimes (1976) found that families in Ahmedabad, Bogota, Hong
Kong, Madras, Mexico City and Nairobi could not afford to buy a ‘cheap’ subsidised government
house.
Critics of public housing, urban renewal, and modernist visions of urban development were
growing during the 1960s. Abrams (1964) and Turner (1968, 1972, 1976a, 1976b), who influenced
the theories and policies of low-income housing for decades, drew attention to the housing
production process inherent in informal settlements (Choguill 1999). They brought the gross
housing shortages and huge squatter settlements lacking in basic utilities in the cities in these
countries into the world’s notice in the 1950s. By focussing on incremental housing of the lowincome
groups, they initiated in situ upgrading of such settlements that was the main form for
improving the living conditions of these people (Abbot 2002).
Turner advocated sites-and-services and slum improvement schemes identifying the aspects of selffulfilment
of the slum-dwellers and their commitment to housing shown in things they value (Pugh,
2000). He defined an extended process by ‘freedom to build’ or as a ‘verb’. Furthermore, Turner
(1976b) supported owner-built homes, however modest, to well-built public housing because of
their differing “structure[s] of authority and control”, i.e. capability to participate (Harris 2003) as
they could best decide their needs (Choguill 2007). Such autonomy was fundamental to Turner
(1968), who argued that squatter settlements improved over time within means, fitting their
circumstances, is more affordable to both the dwellers and the government.
Therefore, squatter upgrading and self-help would be more sustainable compared to mostly
unaffordable typical public housing schemes that did not reach the target groups (Rahman 1999,
54

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
2004). These were advocated consistently by experts and international agencies since the late 1940s
(Harris 2003). It meant not only the investment of sweat equity, but also the processes of ownerdesign
and management. Turner saw these as solutions, not problems (Mangin 1967; Drakakis-
Smith 1981). Watts (1997) credited him for the “reorientation of shelter programs, away from slum
clearance to the environmental improvement of squatter settlements; towards the setting-up of
policies to assist individuals to solve their own problems.”
Aided self-help of the early 1960s too was more affordable than public housing though did not
provide control. The later sites-and-services housing aimed to assist and repeat the success of the
incremental house building and improvement process of the informal settlements to supply
affordable housing units to the urban low income households. The provision of core units was the
alternative that could be occupied relatively quickly by the household and extended when the
occupier would afford the resources. The World Bank championing urban project assistance in the
developing countries adopted Turner’s ideas; by the 1987 International Year of Shelter for the
Homeless, these were well established, and remained broadly valid throughout. The following two
subsections will summarize the approaches by the international agencies and governments to lowincome
group’s housing, and analysis of spontaneous developments in the informal sector.
3.1 International Community:
The international agencies gave direction to the consulting community, governments, and the UN
(Choguill 2007) as “a trendsetter for development thinking” (Huchzermeyer 1999). Abbott (2002)
identified the radical shift from public housing, mandating self-help through sites-and-services and
in situ slum upgrading when “the housing policies of almost all developing nations were devoted to
complete houses for the poor" (Peattie & Doebele 1973). 3 Realising that this could not solve
housing problems, international bodies and governments started assisting the poor to build their
own houses in a cost-effective way after the 1960s.
1972–82, 1983–93, and post-1993 were the phases when the international agencies providing
financial assistance modified their housing development strategies (Pugh 2000), drifting from a
focus on self help to a holistic development, relating housing to other development sectors
(Kessides 1997). Full cost recovery was essential in the World Bank’s funding of sites-and-services
schemes since 1972 as large-scale subsidies were infeasible (Choguill 2007; World Bank 1972,
1973).
Early planning evolved to manage housing development on vacant plots that had many political,
professional, funding and management advantages (Abbot 2002). Still unaffordable to at least 20%
of people, it was found that sites-and-services schemes could not be sustained (Swan et al.1983);
the site-based projects did not multiply benefits, address the subsidy issues, eradicate poverty, or
increase land ownership to enhance the Poor’s access to housing, and hence were not replicable.
Also, planning and building regulations based on western ideals inhibited an expansion of
affordable mass housing, with scope for extension and remodelling, employment generation, etc.
3 Geddes had endorsed assisted self-help as early as 1918 and by the late 1950s its variations by Abrams (1964), Atkinson
(1961) and Weissman (1960) were familiar. So, too, had major international agencies (Harris 2003) that incorporated
economic arguments in favour of self-help in their wisdom. During the 1950s, they justified it for policies of assistance to
owner-builders. The World Bank in its first country report acknowledged the economic importance of self-help in
Colombia. By the early 1960s the economic benefits of self-help were part of conventional wisdom (Frieden 1965), taken
for granted by many of the contributors to the collection written for a US Senate subcommittee in 1963.
55

MAHBUBUR RAHMAN
The 1976 Vancouver Habitat Conference recommended that the governments intervene to supply
housing resources that involves large infrastructure and investment. The 1996 Istanbul Conference
furthered through public-private partnership involving the stakeholders to identify and transform
priorities into action plans: creating institutions for urban environmental improvement, and building
capacities to participate and cooperate. The policy shift towards sustainability accepted informal
settlements to bring a lasting development (World Bank 1991).
The same standard of service delivery in settlement upgrading as in formally planned ones was
ineffective. Thereafter, leaving low-income infrastructure provisions to the community to plan,
build and manage themselves led to “progressive improvement model” (Choguill 1999). Similar to
settlement upgrading, service provision could be gradually upgraded, and thus be affordable and
sustainable. This sought to match the level to the ability of either the community or the local
authority to manage in a sustainable way. Open and wide ranging in interventions, progressive
infrastructure was successful in the Sri Lanka million houses program, and in other South Asian
cities, owing to community decision-making.
Sustainability cannot be brought without making the economy, environment and society parts of an
overall development (Barbier 1988). Institutional reform had to be backed by comprehensive urban
policies to sustain programs in the long run (World Bank 1983). Yet project-oriented theory,
practice and policy of self help schemes that were to be translated into overall changes remained
strong due to available funding, and instant benefits for all (Rahman 1999). In some cases only, e.g.
in Karachi and Indonesia, large-scale upgrading became an essential part of the physical planning
process of the urban areas affecting all informal settlements (Saleem 1983; Silas 1983).
Redirected policies in the 1990s focussed on institutional reform. Gradually withdrawing from direct
physical involvement, agencies generated and channelled funds in social housing with self-help
components through structured finance and purpose-built institutions, the NGOs, 4 and CBOs. This
allowed faster fund disbursement, better chance to reach target groups, and increase recovery by
involving small groups. 5 The projects recovered the cost by setting target by affordability, and thus
remained financially sustainable. Following the motto of affordability-accessibility-replicability
(Choguill 1987), 6 this was more successful in countries with already developed housing finance
systems (Rahman 1999). 7
4 In addition, the NGOs can assist squatter settlements by helping to develop appropriate community organisations and
mobilisation, provide technical and organisational skills for aiding self help house building and increasing their access to
housing finance through creating financial co-operatives, provide loans directly or acting as intermediaries with formal
commercial banks as guarantors for low income borrowers (Mitlin & Satterthwaite 1992; Igel & Srinivas 1996; Desai
1996; UNCHS 1996a, 1996b; Rahman 1999).
5 The Grameen Bank in Bangladesh developed credit and technical advice for women's enterprise, for housing, and for
transforming social development among the poor; international agencies are supporting such initiatives (Rahman 1999).
6 Accordingly, capital costs were to be set by the target group’s ability and willingness to pay, not by planning ideals and
design standards. But cost recovery formula followed old convention where all components had to be paid for. It
matched the Bank's priority for recovery so that it could repay the international credit and make economically and socially
responsible uses of grants (Rahman 2004; Pugh 2000). At the end, the successful projects identified, improved and
mainstreamed (Abbot 2002) could be replicated in similar situations elsewhere (Choguill 1987; Pugh 2000).
7 In Bangladesh as international project grants were coming after the 1987 and 1988 floods, the government established a
rural employment generation foundation to disburse and monitor them, instead of the central bank. An allocation of US$
250 million by the World Bank in 1988 in India helped the Housing Development Finance Corporation to extend its
credit coverage down the income ladder and stimulate local housing finance institutions. With Bank loans Chile
introduced housing vouchers for sites-and-services schemes. In Brazil's Parana Market Towns Improvement Project in
56

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
However, costs could be seldom recovered, project sites were remote from employment
opportunities and unaffordable, institutional capability and expertise to implement and monitor was
often weak, and therefore gentrification compounded the problems (ADB 1983), corruption was
often rampant, inhibiting accessibility, and the projects did not lead to wider socio-economic
development of the entire urban areas (Pugh 1990; Nientied & van der Linden 1985; Skinner et al.
1987; Turner 1980). Moreover, outputs failed to make any qualitative or quantitative impact
(Rahman 1999).
Thus residents’ housing and non-housing choices do not need to be constrained to encourage
affordable housing. It rather has the potential to improve the social sustainable environment in
communities and cities (Forster-Kraus et al. 2009).
Such holistic approach is set to develop finance more, reduce the backlogs, increase inadequate
infrastructure, reform negative land management and land policy, introduce financial transparency
to accelerate supplies in low-income housing, increase competitiveness of the construction industry,
and establish or reform institutions (Pugh 2000). Social effectiveness was brought by decentralising
responsibility of maintenance and cost recovery through the beneficiaries’ groups. Compared with
the earlier shelter-oriented approach, approaches of the 1980s and 1990s prioritised broader and
deeper institutional reforms and development, and creating a strong base for reorientation of
policies in the future. Also funding through the local government bodies proved appropriate in the
backdrop of weak financial markets in many developing countries.
Recent thoughts on development policy, instigated by Stiglitz (1998) and Wolfensohn (1999), drifted
away from the above concept, which were followed by many developing countries that favoured
opening up the market. The Bank (1999a) emphasised on broader urban issues in the late 1990s to
enhance and sustain economic growth and modernisation. Thus improvement of living qualities,
poverty reduction, job creation and production, environmental sustainability, and enhancement of
agglomeration economies were included as strategies to strengthen a balanced urban development
(World Bank 1999). This gave priority in program finance and management to improve squatter
settlements.
3.2 Housing Process:
As low income groups are unable to acquire good formal sector housing, unconventional or
informal housing remains a necessary part of urban growth in the developing countries (UNCHS
1996c; Gilbert 1990; Drakakis-Smith 1981). Appearing disorganised and inadequate initially, these
with their own logic can be gradually upgraded as needs are felt and resources, including fund,
materials, labour, etc. are available (Angel & Benjamin 1976). Many squatters go on to consolidate
and improve their housing over time (Drakakis-Smith 1981; Turner 1976a), slowly shaping
communities. 8 Low-income people can live in incomplete shelters to be improved later. Thus
1983–88, a few towns created a municipal fund out of World Bank’s seed fund. The credit terms required sustainable
finance through cost recovery achieved by skilled management of transactions. Yet, the local governments and
households could have their own sub-projects, select the price according to their need, priority and affordability, through
participation (World Bank 1994).
8 This is not applicable to all squatter settlements or all in such settlements. In many Asian and African squatter settlements,
the squatters do not move on from their initial stage of occupation (Shakur 1987). In Latin America where the squatters
consolidated through improvement, there is a distinction between dynamic settlements and static and stagnant ones
(Drakakis-Smith 1981). Owner occupation, strong group cohesion and the emergence of political leaders and CBOs
57

MAHBUBUR RAHMAN
compared to formal sector that ignores the needs of survival and flexibility of the low and
intermittent income of the poor (Smets 1999), incremental upgrading distributes the affordable
consumption and saving over time. This increases sustainability.
Lower income dwellers can house themselves respectably at much lower than the formal sector
cost as a large degree of the construction work would be carried out by the occupants using
unconventional materials and techniques. The ‘process’ rather than the 'product' approach can halve
the initial requirements for building compared to formal construction costs (Benjamin & McCallum
1985), in exchange for social obligations to be met by the family over a period. The selfmanagement
replaces up to a third of the labour cost (Payne 1983); 9 participatory environmental
improvement is a saving too (Pugh 1994). In reality though most self help projects involve some
degrees of paid labour and contract building; 92% of households in the sites-and-services projects
in Lusaka used hired labour (Tipple 1994).
A house is the largest commitment of a family (Maliene & Malys 2009). Despite a humble result,
house building is an apocalyptic event for a low-income family, co-opting family members,
marshalling all physical and monetary resources, collecting debts, and calling upon the community
and the family for assuming new debt and obligations. Extended over time, this is the beginning of
a longer commitment to constantly make improvements and additions. Thus these turn out to be
more acceptable and suitable to the socio-economic needs of lower-income people (Turner 1976a),
and so are more affordable and sustainable.
Informal sector self-help and self-built provide housing to 30–70% of urbanites in many developing
countries (Keivani & Werna 2001). Self-management where skilled crew and hired labourers do
most works, popular among the middle and upper-income groups, works with incremental building
process too. The least common way is by the contractors for the wealthier people or organisations;
better-off is a society more of such houses emerge (Peattie 1987).
Secure or expected occupancy rights motivate expression of built form in squatter settlements.
While extending shelter, occupiers mark own aesthetic identity on it. This becomes sustainable due
to a commitment of low-income people to place and home (Turner 1976a), perceiving the
improvements as a part of wider resident activities are. Although the resourcing and organisation of
improving infrastructure and making personal investments to squatter houses have been discussed,
enhancement of aesthetics and cultural amenities was less discussed. Against such functional
dominance, Marcus (1995) focused on the personal meaning residents attach to their home that
leads them to improve housing.
Such deep human expression in self-help improvement is instigated by tenure form and length
(Rahman, 1999; Pugh, 2000). 10 The mark of environmental change, local culture, and design and
construction knowledge shows colour, adaptation, and space for rituals and festivals; and create
specific and varied living environments (Rapoport 1988). A make-shift shack— the outcome of
which induce investment in housing and influence national political leaders for recognition and facilities were the
distinguishing features of the former. The latter was characterised by high levels of renting, low group cohesion and
community representation (UNCHS 1996c; Gilbert 1990). the level of owner occupation that induces both personal
investment and group cohesion was the determining factor in this situation (UNCHS 1996c; Gilbert 1990).
9 Materials cost is reduced by buying recycled and used items in informal sector. Family labour is usually free; skilled labour
can be bartered for. Gerrul (1979) calculated that in lower-income housing, 35% labour is self-help; another 60% is semiskilled.
10 For example, San Miguel, Mexico, or Klong Toey squatter settlement with 70,000 families in Bangkok's Port Authority land.
58

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
rational thinking utilising limited available resources, reveals beliefs, aspirations, and the worldview,
simultaneously impacting the political, the visual, and the cultural. As it is ‘architectural’
same way as a ‘designed’ building, designers and planners could learn from these spontaneous
open-ended, multi-sensory, semi-fixed settlements adding on elements, as architecture is about
human drive, vision, interest and the identification of place (Unwin 1997).
Most housing solutions focused on price reduction to close to the households’ ability to pay. These
ignored the low-income families’ willingness to make extra effort to match financial requirements.
With ownership prospect, household can readily devote more of their meagre resources, both
monetary and commitment of non-monetary like spare time (Ward 1984a; Rahman 1999). Though
low-income families have little savings, some of them could amass housing resources (Keare &
Jimenez 1983), often by selling essential items (Rahman 2008). They improve affordability by
using allocated space (CIVIS 2003), for example by renting out or using as workshops, often
involving more family members (Setshedi 2006; Mai & Shamsuddin 2008). 11
4. POVERTY AND SUSTAINABILITY
Poverty is about the lack of capability to expand social opportunity in markets, in state policy, and
in households; these development requisites focus on the freedom of individuals to choose values
and lives worthy to them (Drèze & Sen 1995; Sen 1999; Sen & Wolfensohn 1999). Personal
commitments and appropriate human bondage generated in low-income housing through freedom
and control over the process can lead to poverty reduction. Fogel (1994) argued that improved
nutrition, the advancement of medical knowledge, and the qualities of housing increase health and
economic productivity over long-term development transitions. The principles are evident in the
low-income housing requiring social co-operation to improve environmental conditions in a
sustainable way.
Environmental regulation combined with market based town planning, can play a part in
ameliorating the slum problems if sufficient resources and powers are vested. The market
sensitivity to building performance and impact is usually inhibited by price inflation linked to
excessive demand for social, economic and other benefits. Integrating poverty alleviation and
environmental improvement, Brundtland concluded that “the problems are entwined such that
ecological sustainability cannot be achieved if the problem of poverty is not addressed” (Robinson
2004). This linked the vast and complex issue of environmental deterioration with the equally vast
and complex issue of human development, both affected by poverty.
The mainstream now views environment differently from when related problems were regarded as
minor, technical, soluble, and politically uncontentious in the 1980s. Economic growth and social
progress could duly solve these, as increasing wealth created the resources including technology.
Today, it is acknowledged that environmental problems require significant social and economic
changes, not merely technical solutions. Sustainable development is the vehicle for this change,
addressing the conflict between environmental protection and economic growth. It accepts that the
former requires fundamental change in the direction of economic progress and institutional policy.
But it argues that this is compatible with continued economic growth in a (regulated) global
capitalist system.
11 Projects would often be designed to integrate productive use enhanced by incentives like space and credit in Cairo,
Mexico, Nairobi and Senegal (Ward 1984b), or use future income sources to assess affordability (Lee 1985).
59

MAHBUBUR RAHMAN
Favouring the human-centred nature of the Brundtland Report, the reformists suggested that the
answer to both over- and under-consumption, i.e. the environmental concerns, lay in promoting
more of sensitive human development that requires improvements in technology and efficiency.
Devoid of spiritual values, or individual responsibility, it rather focused on collective institutional
responses and social responsibility, and embodies an anthropocentric approach around human
needs.
The WCED recognised poverty as a major source of environmental degradation. Yet development
agenda revolved around macro-economic stabilisation through relentless export-led growth and
market liberalisation of the early-1980s. The 1997-98 financial crises exposed the limitations of
such approaches followed by many developing countries— the absence of a broad basis in the
politics of socio-economic development. Countries like Thailand and Indonesia succumbed to
international speculation and bad market governance (Pugh 2000).
Stiglitz (1998) favoured medium-term strategic development policies to alleviate poverty through
socio-economic transformation. He advocated holistic societal changes understood as development
transitions, for example, improving the environmental and health dimensions, the changing
volumes and characteristics of poverty, etc., not isolated development of individual sectors. In an
overall context, such development policies could use transitions in a combination of varying
emphasis on different sectors, based on the pragmatic context-based realities, and socio-economic
opportunities.
The WECD downplayed the extent to which poverty alleviation and environmental improvement
result from wealth (Roseland 2000). Trainer (1990) was dismayed that it chose economic growth
and attendant social and environmental impacts (e.g. tendency to exploit labour and the
environment) over a consciously appropriate development strategy for the developing world (e.g.
adequate housing and clean water rather than industry and export). NTFEE (1987) stated that
“sustainable economic development does not require the preservation of the current stock of natural
resources or any particular mix of human, physical and natural assets. Nor does it place limits on
economic growth, provided that is both socially and environmentally sustainable.”
The oxymoron that comes from combining sustainability with its greatest threat, development, is
paralleled by calls to combat exclusion of the low-income group (Clark 2001), and restore a
broader social and economic purpose of housing equity. However, the aspiration is not matched by
the ideological commitment to provide the resources and powers necessary to change the status
quo; politicians and bureaucrats may adopt language and sentiment without the will or means, or
empowering the poor (Rahman 1999). Only a process of political mobilisation and mass education
can change assumptions and behaviour. Under ideal conditions acquisition and incorporation of
intelligence replaces rhetoric, informs policy and leads to a more measured approach to current
anxieties. But most of the time we deal in a less scientific or certain way as rhetoric and vision
influence priorities and decisions.
Sen (1987, 1999) criticised the way neo-institutional and ecological economics direct towards the
conventional economic theories, based on individual capabilities and the concept of ‘social capital’,
to address the social dimension of sustainable development. This meant the alternative
combinations of functionings a person can achieve that in turn denotes the various things he may
value— from elementary like shelter to complex like community participation (Sen 1999). Policies
should not focus on collective outcomes such as the distribution of income, but rather on building
60

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
individual capabilities, and ensuring that people have the freedom to convert economic wealth into
desirable outcome.
Ballet et al. (2003) defined socially sustainable development as one that “guarantees an
improvement of the capabilities of social, economic or environmental well-being for all, through
the aspiration of equity on the one hand, as intra-generational distribution of these capabilities, and
their transmission across generations on the other hand”. He extended the notion of capabilities
from individuals to cover also societies. The structure of capabilities expresses the adaptation of an
individual or a society to a number of external constraints.
The capability approach emphasises on the improvement of social conditions from one generation to
another, and on the interactions between the three spheres of development—social, environmental
and economic. In designing policies, not only the effects of economic and environmental actions on
the social dimension, but also decisions within the social sphere itself have to be looked into. This
expresses both an individualistic and a social point of view—on one hand, it looks at the capabilities
of rationally and responsibly acting individuals, and on the other hand, at the social capabilities of a
society and the roles of social actors. These are not necessarily in harmony with each other, since the
improvement of education, health, employment, etc. may have harmful effects on some groups and
thus even threaten social cohesion.
Social policies such as poverty reduction programs may in fact adversely affect certain capabilities,
with an increasing vulnerability of individuals and social inequalities as a result. Actors are thus
themselves to decide which capabilities are to be considered. Finally, freedom is a key element in
Sen's approach—all evidence seems to show that even the poor attach significant value to freedom,
they may not seem ‘rational’ from traditional economic perspective of maximising utility. Thus
arguments about self-help and identity in housing make more sense in terms of sustainability. The
following subsection will discuss another important issue of sustainable housing, i.e. of increasing
the capacity.
4.1 Enablement:
Reducing the state’s involvement in directly providing housing and expanding the role of the
private market was accepted in the 1980s, pursued by most developing countries irrespective of
ideology or political structures (Israel 1990). Governments for long have played the hopeless role
of ‘provider’ by constructing housing to directly solve the shortages in various countries. The
enablement strategy, given birth by an urgent need for scaling up housing supply, was aimed at
creating a congenial economic and social framework to enhance economic efficiency and social
effectiveness so that capability to solve own housing problems could grow. This could be met by
expanding the role of the private markets, rather than relying on project based approaches that
could not alleviate housing shortages.
Given the private sector dominates housing, enablement can expand housing production on a
national scale by public support for the activities of the formal/informal markets and for the self
help activities of the low income households. Attention was directed toward devising ways of
providing the financial, legal and institutional support (UNHSP 2005). Therefore, the international
agencies started to encourage the creation of an enabling environment emanating into deregulation
and institutional development of specially the land and housing markets to overcome the external
constraints (LaNier et al. 1987; Kimm 1987; Loh 1987).
61

MAHBUBUR RAHMAN
This shift resulted in the adoption of a new World Bank housing sector policy paper, and formed
the basis of the Global Strategy for Shelter to the Year 2000. This 1993 policy statement
emphasised enablement, the sector's contribution to macroeconomic development, and pro-poor
policies involving subsidies targeted only to the poor where required. It proposed the enablement of
private housing markets for scaling up housing production and developing the sector as a whole.
Similarly, UNCHS in June 1996 adopted ‘adequate shelter for all and sustainable human
settlements’ in the Habitat II, founding its shelter policies and recommendations on enabling
strategies for private markets (UNCHS 1996a).
Enablement would bring together technical know-how, capacity of development agencies, use
available resources, and recognise and define responsibilities of all stakeholders, through an
inclusive participatory approach among residents from all strata. The roles of each such partner in
the multi-institutional and multi-organisational environment could be guided by the underlying
socio-economic rationale. For example, private enterprises could contribute efficiency and
entrepreneurship, CBOs could mediate between households and government agencies that provide
urban management expertise, and the participants could provide various finances, self-help
resources, and localised relevance in the upgrading efforts (Pugh 2000). These represent a complex
process, with some risks of failure due to weak institutions, narrow coterie interests, corruption and
market manipulation found in the developing countries.
Such enablement framework was also relevant to new housing for other income groups too, where
it could supply housing, and builders could access development finance in a competitive market.
Implementation of proper land policies by coordinating with the infrastructure and utility agencies
could ensure adequate supplies of well-placed ready land at affordable price. The legal system
could protect property rights of all, developed finance institutions could introduce attractive
instruments to generate and manage funds. The overall policy and enablement framework could
have pro-poor and egalitarian elements for social-relevance and sustainability.
Enablement brought sustainability in sites-and-services schemes in India, and in the small loan
program in Sri Lanka, Hong Kong, Singapore, and Chile (Pugh 1997) that alleviated their housing
crises. However, most developing countries could not achieve the effectiveness or the
comprehensiveness of Chile or Singapore due to poor finance sector and institutions. Most of the
countries not meeting the targets due to institutional incapacities could overcome the situation by
enablement.
Thus in the 1990s, it was evident that benefits could not be sustained without good governance
(Rahman 1999). Therefore, within institutional-led reform, ‘enablement’ put governance into
central positions in economic, education, health, environment, housing, urban and other sectors
with a focus upon state–market–society relations. This encouraged community-based, participatory
elements in upgrading the squatter settlements and owning community assets so that processes
were more transparent and accountable and people were enabled to improve themselves (Rahman
1999).
The project-linked participation of the Sri Lanka Million Houses program (Lankatilleke 1990) had
planning, design, implementation and maintenance stages. Hamdi and Goethert (1996) identified
planning as crucial for the community and the city to jointly take key decisions and define the
program. Crane (1950) viewed self-help as part of a process of community development, requiring
and encouraging cooperation. The Recife Declaration on community control over decision-making
_____________________________________________________
* Corresponding Author. Tel.: +6017683404, Fax: +60326934844
E-mail : mmrahman417@gmail.com
62

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
stressed the importance of the integration of the informal city. However, the support to community
participation in the 1980s was lost later (Abbot 2002).
The enabling concept has been subject to much debate and criticism for its over-concentration on
the private sector to work, and shunning other modes of housing (Keivani & Werna 2001).
5. CONCLUDING DISCUSSIONS
Despite efforts by all, continuously growing squatter settlements occupy a large section in many
cities. 12 These have shown sustainability amidst squalor and disease. The locales for the life's
drama and contributions in huge urban and socio-economic transformations, this has remained a
dominant form of dwelling for many decades. These are necessary and important in terms of both
product and process, and socio-economic and physical utility (Kellett & Napier 1995). The people
living in these settlements are set in specific institutional conditions and processes that influence
their housing and social status. 13 The intricacy, variety, accomplishment, and resource efficiency in
squatter settlements and built forms also have social, cultural, economic, political, and architectural
implications (Pugh 2000).
Varying in theoretical, economic and technical characters, the role of spontaneous and informal
housing in the developing countries urban development was recognised late (Lawrence 1997; Ling,
1997). 14 Assisting self-help programs from 1950s, the funding agencies changed their methods from
site-specific projects to programs mediated through formal institutions, and subsequently to
developing policies, cooperation and participation. Some of the projects have also added economic
and often aesthetic value to urban assets. 15
Crane (1950) attacked the paternalistic thinking, opposing claims that only architects and allied
professionals can deal intelligently with housing. Turner (1976b) too condemned architect-designed
low-income public housing for being expensive and authoritarian. Back (1962) and Safa (1964)
found that such regimentation and lack of choice were disliked by occupants, whereas self-help
12 Slum population in India has more than doubled in the past two decades; in 2001, 54.1% of Mumbaians lived there
(NIHFW, 2006). Dharavi, Asia's second largest slum (Jacobson 2007) in central Mumbai houses 800,000 people (Davis
2006). A quarter of Sao Paolo population lives in poor conditions. With more slums dwellers, Kolkata has a higher slum
density. Based on water and sanitation access, 99% of Afghan and 94% in Central African Republic people live in slum
condition; even a third of the Argentines experience the same. China had nearly 175 mil living in slum conditions, 158 mil
in India, 42 mil. in Nigeria and 36 mil in Pakistan (UNFPA 2007). 327 million people live in slums in Commonwealth
countries - almost one in six of its citizens (Comhabitat 2006). In 11 African, 2 Asian and 1 Pacific countries urbanizing
rapidly, over two third urbanites live in slums.
13 The gradually transformed and consolidated Santa Marta settlement, Colombia, is recognised and accepted into the formal
sector. In Durban, spontaneous settlements are juxtaposed near formal settlements, being impermanent and temporarily
linking kith and kin.
14 Self-help construction existed in Roman towns, medieval Europe, and in early periods in the developed countries. Murals
were found in Pompeii on this. The Swedish ‘magic houses’ were built by families in their free times; the local government
provided land, materials and technical advice. In ancient Sri Lanka, self-help housing was the norm in urban settlements,
e.g. Anuradhapura founded in 483 BC.
15 Cross-subsidisation in land pricing and allocation enabled sites-and-services and squatter improvement programs to reach
the poor in the 1970s and 1980s in Chennai. Housing investment and wealth of all income groups increased; the
contracting between the World Bank, state government, and the project authorities blended state, market, and household
self-help roles (Pugh 1990a, 1997). In the Kampung Improvement Program, the World Bank provided US$ 439 million in
project loans, to improve living conditions, housing investment, incomes, and health. Some of its lessons led to wider
community participation and deeper institutional reforms (World Bank 1995).
63

MAHBUBUR RAHMAN
provided a variety of social, psychological, and economic advantages that offered a breakthrough in
housing problem (Frankenhoff 1967).
Though self help, household economics, affordability, and home sense could describe the roles of
individuals and households, the economists studied only market exchange value, the design and
impact of subsidies, or social questions of poverty and inequality (Pugh 1997; Stretton 1976). They
ignored non-economic resources as socio-economic assets, like time and energy used for home
building, domestic chores, income generation, physical improvements, childrearing and human
capital formation, and personal and community activities; most of these important aspects of lowincome
housing are in deed measurable.
Self help goes beyond the construction and management of housing and the environment. The
domestic sector is more sustainable as it uses own resources to produce home-based goods and
services, depending less on imported materials and technology. Those without proper housing lead
diminished and socially excluded lives, unable to participate fully in the community; their further
development depends on access to services and the security of a safe and healthy environment. Basic
housing is an ‘entry fee’ providing foothold to subsequently access such other benefits (Peattie 1987).
Market forces and official guidance are not enough for developments to meet contemporary
aspirations, rhetoric, commitment or technical possibilities; matters of environmental justice are
also involved as proper housing brings social and economic goods (Clark 2001). The domestic,
commercial, and the public sectors are interdependent in bringing overall socio-economic
development of the low-income groups. This raises the importance of domestic economics in
sustainable development by way of affordable housing and environmental improvement, and
supplements other areas by contributing to the human and labour development.
The state is forced to tolerate some illegal and irregular housing (UNCHS 1996c; Gilbert 1990;
Drakakis-Smith 1981); it denies better housing to maintain the status quo. While a few accept
aesthetic of squatter settlements, the authorities loath such view (Peattie 1987), which have
influenced housing policies and actions like the demolition of shelters and destruction of
communities. Though all communities have rights to improve their settlement, the legal,
professionals and participatory processes may not concur on squatter upgrading. 16
Some politicians, planners, and intellectuals had been sympathetic to the idea that settlements of the
poor were communities and deserved respect. For example, Jacobs (1961) articulated the functional
aspects of the communities that planners and politicians liked to label slums; it was “slums of
hope” to Stokes (1963) and others. With the acceptance of in situ improvement and regeneration,
self help can stand alongside modern technology in cultural–aesthetic form (Rahman & Mai 2010).
Housing and environment of low-income groups can be improved by recognising of squatters
housing rights, stable growth of income, and the development of social capital and empowerment.
Thereon, social, ethical, and aesthetic expressions cover the full range of living, and encompass
environmental, social, economic and political facets that encourage people to value lives. Allowing
one to participate and express attachment in a varied socio-political context results in more
commitments to affordable and sustainable improvements— incremental upgrading of low-quality
makeshift shacks into wholesome structures.
16 For example, social groups in Jordan strongly contended priorities and access to political and economic power (Raed
1998). The government's attempts at urban renewal in Delhi by the late 1950s ran into resistance and had engendered a
growing recognition that low-income communities had intrinsic value (Clinard 1966).
64

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
It is essential to implement housing through participatory democracy (UKGP 1998). Local
governments are the only elected representative making accountable planning and development
decisions for the community; this is critical in moving toward sustainable communities (Roseland
2000). Social homogeneity, community leadership, prior social co-operation experience, visible
outcome, prospective ownership and the affordability can help to achieve consensus regarding
development objectives and means (Rahman 1999). Environmental improvements can be converted
into action plans and partnership by distributing responsibilities, attribution of costs and self help,
and participatory and transparent management. In essence, both the process and the project need
good governance, organisation, management, and policy (Pugh 2000).
If legitimised and assisted, the large informal sector in most developing countries providing for the
low-income groups can contribute in socio-economic developments (de Soto 1989; Fernandez &
Varley 1998), in conserving economy, construction, environment, and health, and hence beckon for
sustainable improvement. The regeneration schemes are sustainable as these improve living
conditions, providing social opportunities for millions to add more socio-economic and
environmental values than high-profiled projects. This potential varies with the characters of the
squatter settlements. 17 Improvement can take place spontaneously where a form of tenure security is
envisioned. Political skills and pressures often influence the selection of improvements and the
distribution of costs and benefits among households. State-assisted regenerations often involve
redesigning and re-aligning lay-outs (Potter & Lloyd-Evans 1998), which disrupts socio-economic
network and identity, instead of preserving things of value to the target group that is possible only as
part of overall housing development and urban macro-spatial planning and stockholders participation.
Unprecedented urban growth in the face of increasing poverty and social inequality means that
there will be about 2 billion people living in the slums in 2030 (UNCHS 2007); the MDG urged the
international community to improve the lives of a part of them. Urbanisation of poverty (Whelan
2004) is increasing number of underprivileged in urban areas. It is not possible to scale up housing
production to meet the requirements of a billion people living in severe housing conditions through
project based policies predominant in developing countries (Tipple 1994; UNCHS 1996a).
International agencies want to replace them, which in themselves were not very successful, with
sustainable housing policies that can work on a national level.
Nobel economic laureates advocated state's welfare roles covering institutional reform, social and
private property rights, and governance quality. They focused on the way the economic ethics and
quality of institutions defined as norms, property rights, compliance procedures, etc. influenced
performance in long-term growth. The World Bank prioritised institutional reform, at the heart of
modern policy interest in governance, in current urban development and policy agendas in the
developing countries (Pugh 2000). Their policies were driven by the drive to balance the tensions
between economic growth, the environment, and social impacts. The three-pronged approach
conceptualised urban development that promoted economic growth, maintained social inclusion and
minimised environmental impact. Most of the focus in the literature has been placed on economy and
environment, though sustainable housing is more than just that. The economic sector addressing the
financial aspects of social justice, accompanying the environmental sustainability, is an important
17 In cases, the populations have expectations of imminent redevelopment, whereas in others de facto occupancy rights seem
secure. Sometimes a settlement generates its own leadership and organizational structures which can be used for
negotiating with politicians and bureaucracies for installing infrastructure. Other settlements either have apathy or
powerlessness.
65

MAHBUBUR RAHMAN
element of housing, while the environmental limits are an important constraining influence on
economic growth.
Sustainability in housing, viewed as an end in itself, has no meaning. Robinson (2004) argued to
integrate the social dimensions of sustainability with the biophysical dimensions; solutions
addressing environmental, social or economic concerns only are insufficient. This required a transdisciplinary
thinking that actively creates synergy, not just summation. Sustainability must also be
integrated across sectors or interests; governments alone have neither the will nor the capability to
accomplish it. Hence the private sector must be involved, supplemented by the monitoring, capacity
building, and alternative service delivery roles by the NGOs.
Sustainable development implies a shift in the capacity of individuals, companies and nations to
use rightful resources under favourable legal and economic arrangements. Although conventional
analyses recognise the need for changing the practices, few realise that moving toward a sustainable
society requires more than minor adjustments. Sustainable community development initiatives are
not only interventions, but also learning processes for mobilising positive changes. Changed social
behaviour can prevent a host of environmental and related social disasters, to create healthy,
sustainable more pleasant and satisfying communities than today. It emphasises on the efficient use
of urban space, minimising the consumption of essential natural capital, multiplying social capital,
and mobilising citizens and their governments toward these ends.
Sustainability, “an attack on conventional thinking and practice” (Gibson 2001), and a framework
for thinking about urban futures, provides an alternative with optimism. The global audience has
pinned its hopes on sustainable development to solve the urgent environmental and societal
problems (Roseland 2000). However, the right to affordable secure housing must be severed from
property speculation and economic competition. Though, policies for sustainable housing in
isolation may not overcome the urban problems, without them no solution can be found.
6. REFERENCES
[1.] Abbott, John (2002) An analysis of informal settlement upgrading and critique of existing
methodological approaches. Habitat International Volume 26, Issue 3, September 2002, pp.
303-315
[2.] Abrams, C. (1964) Housing in the Modern World. Faber & Faber, London.
[3.] ADB (1983) Regional Seminar on Financing Low-Income Housing- summary report. Asian
Development Bank, Manila.
[4.] Agyeman, J., & Evans, B. (Eds.) (1996) Statement of Purpose. Local Environment 1(1).
[5.] Angel, S. & Benjamin, S.N. (1976) Seventeen Reasons why the Squatter Problem cannot be
Solved. EKISTICS vol. 41, no. 242, pp. 20-26.
[6.] Atkinson, G.A. (1961) Jobbing builders or Self-help for African housing. Journal of African
Administration 13, pp. 46–49.
[7.] Awaal, S. (1982) A Study of the Influence of the Overseas Workers Remittance on the Land
Market of Dhaka. Unpublished Masters Thesis, Human Settlement Division, Asian Institute of
Technology, Bangkok.
[8.] Back, K.W. (1962) Slums, projects, and people. Social psychological problems of relocation in
Puerto Rico, Duke University Press, Durham, NC.
66

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
[9.] Ballet, J., Dubois, J.L. & Mahieu, F.R. (2003) Le développement socialement durable: un
moyen d'intégrer capacités et durabilité. Paper Presented at the Third Conference on the
Capability Approach, University of Pavia, 6–9 September, 2003.
[10.] Barbier, E (1988) The Economics of Environment and Development. Elgar, Cheltenham.
[11.] Baross, P. & van der Linden, J. (Eds.) (1990) The Transformation of Land Supply Systems in
Third World Cities. Avebury, Aldershot.
[12.] Benjamin, S.N. & McCallum, D. (1985) Low-Income Housing in the Third World- broadening
the economic perspective. Institute of Technology, Bandung.
[13.] Choguill, C.L. (2007) The search for policies to support sustainable housing. Habitat Int. Vol.
31, Issue 1, pp. 143-149
[14.] Choguill, C.L. (1999) Sustainable Human Settlements: some second thoughts. In: A.F. Foo &
B. Yuen (Eds.), Sustainable cities in the 21st century. The National University of Singapore,
Singapore, pp. 131–142.
[15.] Choguill, C. (1987) New Communities for Urban Squatters- lessons from the Plan that failed in
Dhaka, Bangladesh. Plenum Press, NY.
[16.] CIVIS (2003) The Enabling Environment for Housing Finance in Kenya: Shelter Finance for
the Poor Series. Cities Alliance issue IV, April, 2003.
[17.] Clark, M. (2001) Domestic futures and sustainable residential development. Futures Volume
33, Issue 10, pp. 817-836.
[18.] Clinard, M. (1966) Slums and community development,experiments in self-help. Free Press,
NY.
[19.] Comhabitat (2006) Briefing paper produced for the Commonwealth Civil Society Consultation.
Marlborough House, London.
[20.] Crane, J. L. (1950). How technical assistance in housing can help to achieve the purposes of
international cooperation. International Office, HHFA, Washington, DC, 28 December, 1950.
[21.] Daly, H.E. (1992) Steady state economics (2nd ed). Earthscan, London.
[22.] Davis, M. (2006) Planet of Slums. La Découverte, Paris.
[23.] de Soto, H (1989) The Other Path: the invisible revolution in the Third World. Harper & Row,
NY.
[24.] Desai, V. (1996) Access to Power and Participation. Third World Planning Review 18 2, pp.
217–242.
[25.] Drakakis-Smith, D. (1981) Urbanization, housing and the development process. Croom Helm,
London.
[26.] Drèze, J. & A. Sen (1995) India's Economic Development and Social Opportunity. Clarendon,
Oxford.
[27.] Dubois, J.L. & Mahieu, F.R. (2002) La dimension sociale du développement durable: reduction
de la pauvreté ou durabilité sociale? In: J.Y. Martin and L. Guillaume (Eds.), Développement
Durable: Doctrines, Pratiques, Evaluations. IRD, Paris, pp. 73–94.
[28.] El Sarafy, S. (1989) The proper calculation of income from depletable natural resources. In:
Y.J. Ahmad, S. El Serafy and E. Lutz (Eds.), Environmental accounting for sustainable
development. The World Bank, Washington, DC.
[29.] Empacher, C. (2002) Die Sozialen Dimensionen der Nachhaltigkeit—Vorscläge zur
Konkretisierung und Operationalisierung. Vortrag au der ordentlichen Mitgliederversammlung
67

MAHBUBUR RAHMAN
des Doktoranden-Netzwerk Nachhaltiges Wirtschaften, 26 April. Köln. Institut für sozialökologische
Forschung. http://www.isoe.de/ftp/kerpen.pdf.
[30.] European Commission (2001) A sustainable Europe for a better world- a European Union
strategy for sustainable development. European Commission, Brussels.
[31.] Fernandez, E. & Varley, A. (1998) Law, the city and citizenship in developing countries: an
introduction. In: E. Fernandez & A. Varley (Eds.), Illegal Cities: Law and Urban Change in
Developing Countries. Zed Books, London, pp. 3–17.
[32.] Fogel, R. (1994) Economic growth, population theory, and psychology: the bearing of longterm
processes on the making of public policy. American Economic Review June, pp. 369–395.
[33.] Foy, G. & Daly, H.E. (1992) Allocation, distribution and scale as determinants of
environmental degradation: Case studies of Haiti, El Salvador and Costa Rica. In: A.
Markandya & J. Richardson (Eds.), Environmental Economics. Earthscan, London.
[34.] Frankenhoff, C. (1967) Elements of an economic model for slums in a developing country.
Economic Development and Cultural Change 16, pp. 27–36.
[35.] Frieden, B.J. (1965) Search for housing policy in Mexico City. Town Planning Review 36
(1965), pp. 75–94.
[36.] Forster-Kraus, S., Reed, R. & Wilkinson, S. (2009) Affordable Housing in the Context of
Social Sustainability. ISA International Housing Conference, 1- 4 September 2009, Glasgow.
[37.] Gerrul, B. (1979) Economic Effects of Housing in Developing Countries; Vol. III & IV- types
of low-income residential areas in Dhaka. Technical University of Darmstadt, Darmstadt.
[38.] Gibson, Robert (2001) Specification of sustainability-based environmental assessment decision
criteria and implications for determining significance in environmental assessment. Available
from: .
[39.] Gilbert, A. (1990) The Costs and Benefits of Illegality and Irregularity in the Supply of Land.
In: P. Baross, P. & J. van der Linden (Eds.), The Transformation of Land Supply Systems in
Third World Cities. Avebury, Aldershot, pp. 17–36.
[40.] Guha R. & Martinez-Alier, J. (1997) Varieties of Environmentalism: Essays North and South.
Earthscan, London.
[41.] Greider, W. (1997) One World, Ready or Not: The Manic Logic of Global Capitalism. Simon,
New York.
[42.] Grimes, O.F. (1976) Housing for Low-Income Urban Families. John Hopkins University Press,
Washington, DC.
[43.] Hamdi, N. & Goethert, R. (1996) Action planning for cities, a guide to community practice.
Wiley, Chichester, NY.
[44.] Hardoy, J.E., Miltin, D. & Satterthwaite, D. (1992) Environmental Problems in Third World
Cities. Earthscan, London.
[45.] Harris, Richard (2003) Learning from the past: international housing policy since 1945. A
double irony: the originality and influence of John F.C. Turner. Habitat International Vol. 27,
Issue 2, June 2003, pp. 245-269
[46.] Holden, Meg (2006) Urban indicators and the integrative ideals of cities, Cities Volume 23,
Issue 3, June 2006, pp. 170-183
[47.] Huchzermeyer, M. (1999) The exploration of appropriate informal settlement intervention in
South Africa: Contributions from a comparison with Brazil. unpublished Ph.D. thesis,
University of Cape Town, South Africa.
68

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
[48.] Igel, B. & Srinivas, H. (1996) The Co-Option of Low-Income Borrowers by Informal Credit
Suppliers. Third World Planning Review 18 3, pp. 287–305.
[49.] Israel, A. (1990) The Changing Role of the State: Institutional Dimensions. World Bank Policy,
Research and External Affairs Working Paper # 495, Washington, DC.
[50.] Jacobs, J. (1961) The Death and Life of Great American Cities. Vintage, NY.
[51.] Jacobson, M. (May 2007). Dharavi: Mumbai's Shadow City. National Geographic, National
Geographic Society. http://ngm.nationalgeographic.com/2007/05/dharavi-mumbaislum/jacobson-text;
retrieved on 2009-04-28.
[52.] Keare, D.H. & Jimenez, E. (1983) Progressive Development and Affordability in the Design of
Urban Shelter Projects. Working Paper # 560, The World Bank, Washington DC.
[53.] Keivani, Ramin & Werna, Edmundo (2001) Modes of Housing Provision in Developing
Countries. Progress in Planning Vol. 55, Issue 2, February 2001, pp. 65-118
[54.] Kellett, P. & Napier, M. (1995) Squatter architecture? A critical examination of vernacular
theory and spontaneous settlement with reference to South America and South Africa.
Traditional Dwellings and Settlements Review 6 2, pp. 7–24
[55.] Kessides, C. (1997) World Bank experience with the provision of infrastructure services for the
urban poor: preliminary identification and review of best practices. The World Bank, TWU-
OR8, Washington, DC.
[56.] Kimm, P.M. (1987) Housing Progress in Developing Countries, in Proceedings of the Second
International Shelter Conference and Vienna Recommendations on Shelter and Urban
Development. National Association of Realtors, Washington, DC, pp. 72–75.
[57.] LaNier, R., Oman, C.A. & Reeve, S. (1987) Encouraging Private Initiative. USAID,
Washington, DC.
[58.] Lankatilleke, L. (1990) Community Action Planning: A case study from Sri Lanka. Trialog
23/24, pp. 24–27.
[59.] Lawrence, R. (1997) Space and Text. In: R. Lawrence & A. Wallace-Hadrill (Eds.), Domestic
Space in the Roman World: Pompeii and Beyond. Journal of Roman Archaeology
Supplementary Series No. 22, pp. 7–14.
[60.] Lehtonen, Markku (2004) The Environmental–Social Interface of Sustainable Development:
capabilities, social capital, institutions. Ecological Economics Volume 49, Issue 2, 1 June 2004,
pp. 199-214.
[61.] Ling, R. (1997) The Insula of the Meander at Pompeii, Vol. 1. The Structure Clarendon,
Oxford.
[62.] Loh, P. (1987) Policy and Programs for Urban and Shelter Lending: The Next Decade. In:
Proceedings of the Second International Shelter Conference and Vienna Recommendations on
Shelter and Urban Development. National Association of Realtors, Washington, DC, pp. 67–
71.
[63.] Maliene, Vida & Malys, Naglis (2009) High Quality Housing- a key issue in delivering
sustainable communities. Building & Environment Vol. 44, Issue 2, pp. 426-430.
[64.] Mangin, W. (1967) Latin American Squatter Settlements: a problem and a solution. Latin
American Research Review 2, pp. 65–98.
[65.] Marcus, S. (1995) House as a Mirror of Self: explaining the deeper meanings of home, Conari
Press, Berkeley.
69

MAHBUBUR RAHMAN
[66.] Mitlin, D. and Satterthwaite, D. (1992) Scaling-Up in Urban Areas. In: M. Edwards and D.
Hulme (Eds.), Making a Difference: NGOs and Development in a Changing World. Earthscan,
London, pp. 169–179.
[67.] Nientied, P. & van der Linden, J. (1985) Approaches to low-income housing in the Third
World: some comments. International Journal of Urban and Regional Research 9, pp. 311–329.
[68.] NIHFW (2006) Slum population in million plus cities. Slums in India - an Overview. Urban
Health Resource Unit, National Institute of Health and Family Welfare, Ministry of Health and
Family Welfare, Government of India, New Delhi.
[69.] NTFEE (1987) Report of the National Task Force on Environment and Economy. Canadian
Council of Resource and Environment Ministers, Ottawa.
[70.] Norgaard, R.B. (1994) Development Betrayed: The End of Progress and a Coevolutionary
Revisioning of the Future. Routledge, London.
[71.] Norton, Bryan (1999) Pragmatism, Adaptive Management, and Sustainability. Environmental
Values 8, pp. 455-466.
[72.] Passet, R. (1996) L'Economique et le vivant, 2 édition. Payot, Paris.
[73.] Payne, G. (1983) Low-Income Housing in the Developing World- the role of sites-and-services
and squatter upgrading. John Wiley & Sons, NY.
[74.] Peattie, L.R. (1987) Shelter development and the poor. In: L. Rodwin (Ed.), Shelter, Settlement
and Development. Allen & Unwin, Boston, pp. 263–280.
[75.] Peattie, L. & Doebele, W.A. (1973). Review of J.F.C. Turner & Robert Fichter (Eds.), Freedom
to Build. Journal American Institute of Planners, 39, pp. 66–67.
[76.] Pezzoli, K. (1997) Sustainable development: a transdisciplinary overview of the literature.
Journal of Environmental Planning & Management 40 (5), pp. 549–574.
[77.] Potter, R. & Lloyd-Evans, S. (1998) The City in the Developing World. Longman, Harlow.
[78.] Pugh, C. (1990) Housing and Urbanization: A Study of India. Sage, New Delhi.
[79.] Pugh, C. (1994) The development of housing finance and the global strategy for shelter. Cities
11, pp. 384–392.
[80.] Pugh, C. (1995) The role of the world bank in housing. In: B. Aldrich and R. Sandhu (Eds.),
Housing the urban poor: Policy and practice in developing countries, Zed Books, London.
[81.] Pugh, C. (1997) Poverty and progress? Reflections on housing and urban policies in developing
countries, 1976–96. Urban Studies 34 10, pp. 1547–1596
[82.] Pugh, C. (2000) Squatter Settlements: their sustainability, architectural contributions, and socioeconomic
roles. Cities Vol. 17, Issue 5, October 2000, pp. 325-337
[83.] Raed, H. (1998) Public participation and community organisation in the low-income housing
production: the Jordanian experience. Unpublished Ph.D. thesis, University College London.
[84.] Rahman, M.M. (1990) Seventeen Facts on Dhaka’s Bastees. EKISTICS, Vol. 57, No. 342-343,
May-August, pp. 172-180.
[85.] Rahman, M.M. (1999) Role of the NGOs in Urban Housing for the Poor. Robert McNamara
Fellowship Report; World Bank, Washington DC.
[86.] Rahman, M.M. (2004) Housing Affordability- ability or willingness? Symposium of Housing
II- affordable dwelling, March 27-30, High Commission for the Development of Arriyadh,
KSA.
70

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
[87.] Rahman, M.M. & Mokhtar Mai (2010) World Bank’s changing approach to Sustainability of
Housing Affordability in Developing Economis. In: MHA Samad et.al (Eds.), Issues on Global
Energy Crisis and its Impact on Design, USM, Penang, pp. 118-135.
[88.] Rapoport, A. (1988) Spontaneous Settlements as Vernacular Design. in Patton C. (ed)
Spontaneous Shelter: International Perspectives and Prospects, Temple University Press,
Philadelphia, pp. 51–57.
[89.] Rees, W. (1996) Revisiting carrying capacity: Area-based indicators of sustainability,
Population and Environment 17 (3).
[90.] Robinson, John (2004) Squaring the circle? Some thoughts on the idea of sustainable
development. Ecological Economics Volume 48, Issue 4, 20 April 2004, pp. 369-384
[91.] Robinson, J. and Tinker, J. (1997) Reconciling ecological, economic, and social imperatives: a
new conceptual framework. In: T. Schrecker (Ed.), Surviving Globalism: Social and
Environmental Dimensions, Macmillan, London, pp. 71–94.
[92.] Roseland, Mark (2000) Sustainable Community Development: integrating environmental,
economic, and social objectives. Progress in Planning, Vol. 54 Issue 2, pp. 73-132.
[93.] Safa, H.I. (1964) From shanty town to public housing. A comparison of family structure in two
urban neighbourhoods in Puerto Rico. Caribbean Studies 4 (1964), pp. 3–12.
[94.] Saleem, S. (1983) Land Management in the katchi abadis of Karachi. In: S. Angel, R.W.
Archer, S. Tanphiphat and E.A. Weglin (Eds.), Land for Housing the Poor, Select Books,
Singapore, pp. 144–155.
[95.] Sen, A.K. (1999) Development as Freedom. Anchor Books, NY.
[96.] Sen, A.K (1987) Commodities and Capabilities, Oxford India Paperbacks, Oxford Univ. Press,
Oxford.
[97.] Sen, A.K. & Wolfensohn, J. (1999) Development: a coin with two sides, World Bank,
Washington DC.
[98.] Setshedi, Gift Phalatse (2006) Architecture, Literature and the Advancement of Township
Enterprises. In: M. Rahman (Ed.), Society, Architects and Emerging Issues, CAA-IAB,
London-Dhaka, pp. 85-92.
[99.] Shakur, M.T. (1987) Approaches towards Housing Low Income Communities in the Third
World: a Literature Review. University of Liverpool, Working Paper, p. 30.
[100.] Silas, J. (1983) Spatial structure housing delivery, land tenure and the urban poor in Surabaya,
Indonesia. In: S. Angel, S., R.W. Archer, S. Tanphiphat and E.A. Weglin (Eds.), Land for
Housing the Poor, Select Books, Singapore, pp. 211–233.
[101.] Skinner, R., Taylor J. and Weglin E.A. (Eds.) (1987) Shelter Upgrading for the Urban Poor:
evaluation of Third World experience, Island Publishing House, Manila.
[102.] Smets, P. (1999) Housing finance trapped in a dilemma of perceptions: affordability criteria for
the urban poor in India question. Housing Studies 14, pp. 821–838.
[103.] Stiglitz, J. (1998) Towards a new paradigm for development: strategies, policies and processes.
9th Raul Prebisch Lecture, UN Conference on Trade and Development, Geneva.
[104.] Stokes, C. (1963) A theory of slums. Land Economics 38 (1963), pp. 187–197.
[105.] Stren, R., and Polese, M. (2000) Understanding the new socio-cultural dynamics of cities:
comparative urban policy in a global context. In: M. Polese and R. Stren (Eds.), The Social
Sustainability of Cities, University or Toronto Press, Toronto, pp. 3–38.
71

MAHBUBUR RAHMAN
[106.] Stretton, H. (1976) Capitalism, Socialism and the Environment, Cambridge University Press,
London.
[107.] Swan, P.J. et al., (1983) Management of sites and services housing schemes: The Asian
experience, Wiley, Chichester (1983).
[108.] Tipple, G.A. (1994) A Matter of Interface: the Need for a Shift in Targeting Housing
Intervention. Habitat International 18 4, pp. 1–15.
[109.] Trainer, T. (1990) A Rejection of the Brundtland Report. ifda dossier 77, 71–84, May/June
1990.
[110.] Turner, A. (1980) The Cities of the Poor. Croom Helm, London.
[111.] Turner, J.F.C. (1976a) Housing by People- towards autonomy in building environments.
Marion Boyars, London.
[112.] Turner, J.F.C. (1976b) Approaches to government-sponsored housing. Ekistics 41 (1976), pp.
4–7.
[113.] Turner, J.F.C. (1972) The Re-Education of a Professional, and Housing as a verb. In: J.F.C.
Turner and R. Fichter (Eds.), Freedom to Build, Macmillan, NY.
[114.] Turner, J.F.C. (1968) Uncontrolled urban settlements. Problems and policies. International
Social Development Review 1, pp. 107–130.
[115.] Turner, J.F.C. (1967) Barriers and channels for housing development in modernising countries.
Journal of the American Institute of Planners 33 3, pp. 167–181.
[116.] UK Green Party (1998) Manifesto, policy HO101,
http://www.brighton.co.uk/greenparty/housing.htm, accessed on 10.01.09.
[117.] UNCHS-Habitat (2007) UN-Habitat Sustainable Urbanization: local action for urban poverty
reduction, emphasis on finance and planning. 21st Session of the Governing Council, Nairobi.
[118.] UNCHS (2006) State of the World's Cities Report 2006/7, Third World Urban Forum,
Vancouver.
[119.] UNCHS (2003) The Challenge of Slums: new Global Report on Human Settlements, UN-
Habitat, Nairobi.
[120.] UNCHS-Habitat (1996a) The Istanbul Declaration and Habitat II Agenda, Istanbul.
[121.] UNCHS-Habitat (1996b) The Future of Human Settlements: Good Policy Can Make a
Difference, Istanbul, A/CONF.165/7.
[122.] UNCHS-Habitat (1996c) An Urbanizing World: Global Report on Human Settlements 1996,
Oxford University Press, Oxford.
[123.] UN Dept. of Economic and Social Affairs (1957) Financing of housing and community
improvement programmes, UN (ST/SOA/32), NY.
[124.] UNFPA (2007) State of World Population 2007, United Nations Population Fund, NY.
[125.] Unwin, S. (1997) Analysing Architecture, Routledge, London.
[126.] Ward, P.M. (1984a) Self-Help Housing- a critique, Mansell, London.
[127.] Ward, P.M. (1984b) Mexico- beyond Sites and Services. In: G. Payne (Ed.), Low-Income
Housing in the Developing World, John Wiley and Sons, NY, pp 149-159.
[128.] Watts, K. (1997) Outwards from home- A planner's odyssey, The Book Guild, Sussex.
[129.] Weissman, E. (1960) Mutual aid in low-cost housing. Annals of the American Academy of
Political and Social Science 329 (1960), pp. 107–114.
72

AFFORDABILITY, ENABLEMENT AND SUSTAINABILITY OF LOW-INCOME HOUSING
[130.] Whelan, S. (2004) One third of the world’s urban population lives in a slum. International
Committee of the Fourth International, www.wsws.org
[131.] Wolfensohn, J. (1999) A proposal for a comprehensive development framework. Speech to the
Board, Management and Staff of the World Bank Group, World Bank, Washington, DC.
[132.] Woolcock, M. (2001) The place of social capital in understanding social and economic
outcomes. The contribution of Human and Social Capital to Sustained Economic Growth and
Well-Being. International Symposium Report. Human Resources Development Canada and
OECD, pp. 65–88.
[133.] World Bank (1999) A Strategic View of Urban and Local Government Issues: Implications for
the Bank. World Bank, Washington, DC.
[134.] World Bank (1995) Indonesia impact evaluation report. Enhancing the quality of life in urban
Indonesia: the legacy of the Kampung improvement program. Operations Evaluation
Department, Report # 14747-IND, World Bank, Washington, DC.
[135.] World Bank (1994) Twenty years of lending for urban development, 1972–92, World Bank
Operations Evaluation Department, Report # 13117, Washington, DC.
[136.] World Bank (1993) Housing: Enabling Markets to Work, World Bank, Washington, DC.
[137.] World Bank (1991) Urban policy and economic development: An agenda for the 1990s, World
Bank, Washington DC.
[138.] World Bank (1983) Learning by Doing, World Bank, Washington, DC.
[139.] World Bank (1973) Housing Enabling markets to work, World Bank, Washington, DC.
[140.] World Bank (1972) Urbanization, World Bank, Washington, DC.
[141.] World Commission on Environment & Development (1987) Our common future, Oxford
University Press, Oxford.
73

AHU J. of Engineering & Applied Sciences 3 (1) : 75-85 (2010)
© 2009 ALHOSN University
DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061
ALUMINUM USING DESIRABILITY FUNCTIONS
Mahmoud I. Awad 1* , Elsayed Orady 2 , Chenmaya Dandekar 2
1
Industrial Engineering Department, ALHOSN University, P.O. Box: 38722, Abu Dhabi, UAE
2 Industrial and Manufacturing Systems Engineering Department, University of Michigan-Dearborn, Dearborn, MI. USA
ABSTRACT: This paper investigates the effect of drilling process parameters, such as cutting speed, feed, and peck
angle on the drilled hole quality characterized by surface roughness and burrs’ formation. Empirical models
describing process performance which can be used for prediction are presented. The models for average
performance and variability were used to optimize the roughness and burr using desirability functions.
Keywords: Drilling process, Robustness, Desirability functions
1. INTRODUCTION:
Due to its good mechanical properties such as machinability and low density, Aluminum
is commonly used in a wide range of industries and constitutes about 40% of all metal-cutting
operations (Hamade and Ismail (2005)). Drilling is a cutting process in which a hole is originated
or enlarged by means of a multipoint, fluted, end cutting tool typically aided by cutting fluids.
As the drill is rotated and advanced into the workpiece, material is removed in the form of chips
that move along the fluted shank of the drill. Chips are produced within the workpiece and move
in direction opposite to axial movement of the drill. Although long spiral chips usually result
from drilling, adjustment of the feed rate can result in chips with a range of shapes and sizes.
Consequently, chip disposal in drilling and the effectiveness of cutting fluids are important.
The most two important performance measures used to determine drilling quality are surface
roughness and burr height. Two surface roughness measurements are very common: R
a
and R
q
where R
a
is the arithmetic mean of the departures of the roughness profile from the mean line,
while R
q
measures the maximum departure (See Figure 1). R
a
and R
q
are highly correlated, so
minimizing one of them will lead to minimizing the other. R
a
is expressed mathematically as
where l m
denotes cut-off length.
R
l m
1
R
a
f ( y)
dx,
(1)
l
q
q
m
0
R is the RMS parameter corresponding to
1
l
m
l m
0
2
f
( y)
dx,
R
a
by
(2)
____________________________________
* Corresponding Author. Tel.: +971 4070585
E-mail : m.awad@alhosnu.ae
75

MAHMOUD I. AWAD, ELSAYED ORADY, CHENMAYA DANDEKAR
Figure 1: Surface roughness
Burr is plastically deformed projected material, generated on the part edge during drilling.
Both roughness and burr cause reliability problems and performance degradation especially in precise
assemblies. Burr formation requires additional manufacturing operations with additional cost and
time for disassembly and de-burring. The geometry and drilling process parameters of classical
drilling have been well studied over the years and gained more importance lately. For example Pande
and Relekar (1986) have found that feed rate is the most significant factor in the drilling process.
Lauderbaugh (2009) used experimentally verified simulation to predict burr height, force and
temperature for 2024-T351 and 7075-T6 aluminum. He concluded that feed rate, chisel edge to drill
diameter ratio, drill diameter, yield strength and point angle are the most significant parameters
affecting burr height. Pena et al. (2005) proposed a monitoring method, based on internal signals from
spindle torque, to detect non-desired burr formation during drilling operations.
Other researchers investigated tool wear as a function of cutting speed and coating material,
see Nouari et al. (2005) and List et al. (2005).
Most of the engineering applications, including drilling process, would comprise of multiple
responses such as surface roughness and burr height. Moreover, variability of process or product has
to be minimized and mean has to be optimized. Optimizing multiple responses simultaneously is a
difficult task that has been tackled by many researchers. One of the early methods proposed by
Taguchi (1987) used for multi-response optimization advocates the use of signal-to-noise ratios
(S/N). This methodology specifically utilizes both experimentation and optimization methods to
determine the system optimum operating conditions to minimize performance variability and
deviation from target value of interest. Many detractors such as Box (1988) and Nair (1992)
challenged S/N ratio methodology and question Taguchi choice of experimental designs and methods
of statistical analysis. Khuri and Conlon (1981) used the generalized distance approach to find the
optimal settings that minimize the distance function over the experimental region. Pignatielo (1993)
used a quadratic loss function to solve the multi-response problem. In this article, a desirability
function method (Deringer & Suich, 1980; Harington, 1965), which transforms the multi-response
problem into a single response problem by maximizing the combined desirability, is used. More
details about desirability method are provided in section 2.2.
In the next section, we review background information on experimental design and optimization
model using desirability function. In section 3, we utilize the above methods in optimizing the
drilling process relative to surface roughness and burr height and its variability’s. Finally, we
conclude the paper with a discussion of our findings.
2. BACKGROUND: EXPERIMENTATION AND DESIRABILITY OPTIMIZATION
2.1. Experimental designs
Experimental design is a scientific approach to planning an experiment using statistical
techniques (Montgomery, 1993). The process of planning an experiment is performed so that the
76

DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061 ALUMINUM USING DESIRABILITY FUNCTIONS
experimental data collected can be analyzed in order to obtain valid and objective conclusions. In
general, experimental data are subject to experimental errors and statistical approach is the only
objective approach to analyze such data. Three basic principles of experimental design that must
be emphasized when designing an experiment are the principles of replication, randomization,
and blocking.
Replication is the repetition of an experiment. It allows the experimenter to estimate the
experimental error, which is used as a basis for determining whether observed differences in
the data are really statistical differences. In addition, experimenter is usually interested in
estimating the effect of different levels of experimental factors. More importantly, replication
enables the experimenter to estimate and model variability in terms of standard deviation or
variance.
Randomization is an important concept underlying the use of statistical methods in
experimental design. It should be applied to two aspects of the experiment, which are 1) to
randomly allocate the experimental units (e.g. workpiece, material, etc.) to different
experimental settings, and 2) to randomly determine the order in which the individual runs
are to be performed. Statistical methods usually require that the experimental errors be
independently distributed random variables. Randomization helps assuring that this
assumption is met. Furthermore, properly applying randomization would balance out the
effect of extraneous factors that may be present.
Blocking is a technique used to increase the precision of an experiment’s result. A block is a
portion of the experimental units that is more homogenous than the entire set of experimental
units. Blocking involves making comparisons among the conditions of interest in the
experiment within each block.
2.2 Desirability Models
Derringer and Suich (1980) developed a procedure which can be very useful when several
responses are involved. The method utilizes a desirability function to be maximized in which
practitioners can assign their own priorities and desires on the response values and build it into
the optimization procedure. The method makes use of an objective function, D, called the
desirability function and transforms an estimated response into a scale free value (d i ), called
desirability as shown in 3.
Where n is the number of responses, and w i is importance of the response. The desirable scale
ranges from zero (least desirable) to one (most desirable) and have different forms depending on
the goal of responses as shown in Equations 4-6:
(4)
77

MAHMOUD I. AWAD, ELSAYED ORADY, CHENMAYA DANDEKAR
(5)
(6)
Where wt i is a weight factor range from 0.1 to 10 and controls the shape of the desirability
function. For example, if wt i =1, desirable scale d i will vary from 0 to 1 in a linear fashion. In this
paper, and without any general loss, the weight factor will be assumed one for all responses.
Taking into account all n responses, clearly one wishes to choose the conditions x for the process
parameters to maximize D. A value of D close to 1.0 implies that all responses are in a desirable
range simultaneously.
3. DRILLING EXPERIMENT
Machining processes, in general, have many parameters that significantly affect the
performance of the processes. The influences could be on the machine performance, the tool used
to perform the processes, and/or the products produced. In any machining process, like any
production machine, the ultimate goal is to produce high quality products with minimal cost.
The primary objective of the drilling experiment conducted in this paper is to establish a
relationship between the quality of the machined holes and machining variables. This is because,
to the best of the authors’ knowledge, there is no relationship (i.e. empirical or physics models)
that has been established for predicting the quality of drilled holes by using the machining
parameters.
In general, machining processes have numerous parameters. Some parameters related to the
machine such as machine dynamics and accuracy, other parameters related to the tool such as the
ability of the tool to perform a cut with certain cutting conditions, and finally parameters related
to the workpieces such as the mechanical and physical properties of the workpiece materials, and
the geometry of the workpiece. For the drilling process in particular, there are many methods
used to produce a hole. The first step is to machine a center for the hole using combination center
tool. The depth of the center is a function of the diameter of the hole. Depending on the diameter
and length of the hole, the procedure to produce the hole varies. The diameter of the hole could
be machined using a drill with the same diameter, or the hole could be machined in two or more
78

DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061 ALUMINUM USING DESIRABILITY FUNCTIONS
drilling sequences by using drills with different sizes to reach the desired diameter of the hole.
On the other hand, depending on the depth of the hole, it could be drilled in one sequence or one
tool feed, or using peck procedure.
3.1 Experimentation Design
In order to limit the number of experimental factors to a manageable number, three cutting
parameters were selected as controllable experimental factors. These include cutting speed, feed,
and peck. Regarding the range of the input variables, the cutting speed and feed for the selected
tool and workpiece pair materials were determined using the Machinery’s Handbook.
Several experimental designs were considered such as full factorial, fractional, Taguchi
orthogonal arrays, and regression models. Since our prime goal is to construct an empirical
model, regression designed experiments with number of levels equal to number of runs were
chosen. The other designs are rigid in terms of number of levels. More levels tested empower the
model in capturing any non-linearity if exist.
Six of 6"
4" 1"
workpieces were cut from a 6-foot long 6061 Aluminum bar. Each workpiece
was drilled 15 times. Since the objective of the experiment is to construct an empirical prediction
model, each drilling parameters was not treated as fixed effect factors, but was treated as
continuous variables. The drilling parameters (speed, feed and peck length) were randomly
generated from its input range using the Random function in Excel “rand()” to drill 15 holes in
each workpiece. Two blocks of parameter values for the drilling process were randomly
generated as shown in Table 1.
Table 1: Two blocks of parameter values for the drilling process
Standard Order Block Cutting speed (S)
(fpm)
Feed (f)
(ipr)
Peck depth (P)
(in.)
1 1 364 0.007 0.75
2 1 329 0.004 0.086
3 1 252 0.003 0.066
4 1 308 0.005 0.1
5 1 208 0.008 0.1
6 1 287 0.006 0.4
7 1 275 0.008 0.79
8 1 377 0.007 0.55
9 1 366 0.001 0.64
10 1 253 0.004 0.07
11 1 364 0.006 0.33
12 1 361 0.010 0.85
13 1 315 0.006 0.5
14 1 341 0.009 0.9
15 1 225 0.001 0.05
16 2 278 0.005 0.138
17 2 214 0.005 0.65
18 2 273 0.008 0.53
19 2 216 0.003 0.72
20 2 336 0.006 0.16
21 2 204 0.005 0
22 2 391 0.002 0.35
79

MAHMOUD I. AWAD, ELSAYED ORADY, CHENMAYA DANDEKAR
23 2 358 0.007 0.53
24 2 370 0.003 0.6
25 2 208 0.0075 0.23
26 2 331 0.009 0.1
27 2 305 0.0052 0.38
28 2 274 0.0068 0.1
29 2 239 0.0066 0.67
30 2 294 0.003 0.5
The Six workpieces were randomly divided into two sets of three workpieces, where each set
received a random selection of a parameter set, and the three workpieces in each set served as
replications for each parameter set. Two CNC programs were developed and loaded in the
machine (Fadal TRM Vertical milling machine) to perform the drilling process using six-half
inch Coated-HSS twist drills.
3.2 Experimentation Results
The response variables are measured using a Mitutoyo Surftest 402 and a Surftest Analyzer.
The surface roughness measurements Ra
were obtained using proper scale and resolution. The
burr heights were measured using a dial indicator with 0.001 inch resolution. For each hole, two
measurements were carried out for each of the surface roughness parameters and four
measurements were taken for the burr heights. Results are shown in Appendix I. Both means of
Ra and Burr (Ra mean , Burr mean ), along with standard deviations (Ra STDE , Burr STDE ) are calculated
for each run. The responses were consequently expressed in form of regression equations as
follows:
(7)
The analysis of variance (ANOVA) is summarized in tables 2-5:
Table 2. Response Surface Regression: Ra mean
Term Coefficient SE Coef T P
Constant 329.1 51.14 6.435 0.000
f 2470.7 3947.5 0.626 0.538
S -0.6 0.18 -3.291 0.003
P -300.4 101.40 -2.963 0.007
P 2 100.6 65.46 1.537 0.138
f P -13522.0 7292.43 -1.854 0.077
S P 1.0 0.33 3.041 0.006
80

DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061 ALUMINUM USING DESIRABILITY FUNCTIONS
Table 3. Response Surface Regression: Ra STDE
Term Coefficient SE Coef T P
Constant 92.5 31.53 2.934 0.007
f 5581.3 2388.15 2.337 0.028
S -0.4 0.11 3.251 -0.003
P -107.0 52.52 -2.038 0.053
f P -13949.3 4349.02 -3.207 0.004
S P 0.6 0.20 3.215 0.004
Table 4. Response Surface Regression: Burr mean
Term Coefficient SE Coef T P
Constant 0.050 0.013 3.978 0.001
f -8.325 2.633 -3.162 0.004
P 0.008 0.010 0.880 0.388
f 2 319.409 163.113 1.958 0.062
f S 0.010 0.008 1.262 0.220
f P -1.557 1.623 -0.959 0.347
Table 5. Response Surface Regression: Burr STDE
Term Coefficient SE Coef T P
Constant 0.00364 0.00409 0.889 0.382
f -0.97956 0.31322 -3.127 0.004
S 0.00003 0.00001 2.027 0.053
P -0.01071 0.00972 -1.102 0.281
P 2 0.01267 0.00964 1.315 0.200
Although the main effect of feed (f) is not significantly impacting Ra, the interaction of feed
and peck (fP) and the interaction of speed and peck (fS) are significant. All variables and their
interactions impacted the standard deviation of Ra.
On the other hand, feed was the only factor impacted burr mean significantly. Both feed and speed
impacted the standard deviation of burr.
3.3 Desirability Optimization
In this multi-objective optimization problem, four different objectives were selected to be
optimized: Ra mean, Ra standard deviation, Burr height mean, and Burr height standard deviation.
The goal of all of the four objectives is the lower the better (LTB), i. e. minimization. In multiresponse
optimization desired importance is given to all the responses. It’s the author’s point of view
that Burr is more important than surface finish. Table 6 shows the range of the responses, goals,
weights, and importance assigned to each response.
Table 6. Response goals
Term Lower Target Upper Weight Importance
Ra mean (micro inch) 120 120 200 1 1
Ra STDE (micro inch) 0 10 20 1 1
Burr mean (inch) 0 0 0.04 1 8
Burr STDE (inch) 0 0.001 0.005 1 1
81

MAHMOUD I. AWAD, ELSAYED ORADY, CHENMAYA DANDEKAR
In addition to response goals, there are some constraints that limit the search space in the
considered optimization problem to the allowed ranges for the process parameters given by the
validity range of the experimental models.
Minitab Software was used to search for the optimal solution using desirability functions
considering the process parameters constraints, the optimal solution is summarized in Figure 2.
Optimal
D
High
Cur
0.81476 Low
Composite
Desirability
0.81476
Ra_mean
Minimum
y = 127.0951
d = 0.91131
Burr_mea
Minimum
y = 0.0094
d = 0.76497
Ra_sigma
Targ: 10.0
y = 9.8466
d = 0.98466
f S P
0.0107 391.0 1.0
[0.0096] [324.8889] [0.5688]
0.0011 204.0 0.0513
Burr_sig
Targ: 0.0010
y = 0.0010
d = 0.99822
Figure 2: Optimization results
The results indicate that feed should be set at a value closer to the maximum value (0.0096),
high speed (324.9), and intermediate peck angle value (0.57). Setting the drilling process at these
values is expected to provide an acceptable results relative to the desired targets set in Table 6. For
example, the optimal settings will result in a surface roughness of 127micro inch which is acceptable
compared to the target which is 120 micro inch. Similarly, the optimal settings resulted in a burr
mean of 0.0094 inch which is acceptable compared to the zero burr height target.
4. DISCUSSION AND CONCLUSION:
To verify the results obtained in the previous section, several confirmation runs were
conducted with the optimal settings. Figures 3 and 4 depict the results of the confirmation runs of Ra
and burr height in the form of a boxplot respectively with confidence intervals of the means. A t-test
of t Ra mean was conducted where (H 0 : Ra_ave = 127, vs H a : Ra_ave 127). The results indicated
that there is no sufficient evidence to reject the null hypothesis with a p-value of 0.124. Similarly, a t-
test for burr height was conducted (H 0 : Burr_ave = 0.0094, vs H a : Burr_ave 0.0094) and indicated a
lack of evidence to disprove H 0 with a p-value of 0.053. Although the roughness and burr means
were close to the predicted values, the standard deviations of both responses were off. The standard
deviation of the confirmation runs indicated a standard deviation of 22 micro inch (predicted =10
82

DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061 ALUMINUM USING DESIRABILITY FUNCTIONS
micro inch) for Ra and 0.0033 inch (predicted =0.001inch) for burr height. The inaccuracy of
variability prediction for Ra and burr height is due to the nature of the responses.
Figure 3 Boxplot of Ra confirmation runs with 95% confidence interval
Ho
_
X
0.0050
0.0075
0.0100
Burr height
0.0125
0.0150
0.0175
Figure 4 Boxplot of burr height confirmation runs with 95% confidence interval
Other variables may influence the quality of drilled hole such as tool wear, cutting forces (torque
and thrust force) and machine and fixture rigidity. These factors were not included in the
conducted experiments and can be added in future research. In this study two measures of the
hole quality were included, namely; burr height and surface roughness. However, there are other
quality measures that can be included in the model to describe the hole quality in more
comprehensive manner. These measures may include circularity (roundness), cylindericity, and
center position.
The optimization scheme proposed in this research simultaneously optimizes both the mean and
standard deviation for multiresponse system design problems. The proposed method incorporates
noise factors within the experimentation phase in order to determine the settings of the control
factors to create response models that are insensitive to internal or external noise. The proposed
approach also uses a nonlinear programming technique to model a multiresponse problem as a
single objective function using a desirability index. Highest desirability could be obtained at
high level of feed, high speed, and an intermediate peck. Using desirability functions, both targets
were achieved simultaneously and different importance levels were assigned to responses.
5. REFERENCES
[1.] Box, G. (1988). “Signal-to-noise ratios performance criteria, and transformations.”,
Technometrics, Vol. 30(1), pp. 1–17
[2.] Coleman, D. E., and Montgomery, D. C. “A Systematic Approach to Planning for a Designed
Industrial Experiment (with Discussion).,” Technometrics. Vol. 35, pp. 1-27 (1993).
[3.] Derringer, G., & Suich, R. (1980). “Simultaneous optimization of several response variables.”,
Journal of Quality Technology, Vo. 12(4), pp. 214–219
83

MAHMOUD I. AWAD, ELSAYED ORADY, CHENMAYA DANDEKAR
[4.] Furness, R. J., Wu, C.L., and Usloy, A.G., (1992). “Statistical Analysis of the Effects of Feed
Speed and Wear on Hole Quality in Drilling.” Sensors and Signal Processing for Manufacturing,
PED-Vol. 55, pp. 97-112,
[5.] Gallowaym, D. F. (1995). “Some experiments on influence of Various Factors on Drill
Performance.” Transaction of the ASME, pp. 191-231
[6.] Hamade, R.F., Ismail, F. (2005). “A case for aggressive drilling of aluminum.”, Journal of
Materials Processing Technology, Vol. 166, pp. 86-97
[7.] Khuri, A. I., & Conlon, M. (1981). “Simultaneous optimization of multiple responses represented
by polynomial regression functions.”, Technometrics, Vol. 23(4), pp. 363–375
[8.] Kuehl, R. O. (1999). Design of experiments: Statistical principles of research design and analysis,
Duxbury Press, New York
[9.] Lauderbaugh L. Ken (2001). “An Exit Burr Model for Drilling of Metals.” Transaction of the
ASME, Journal of Manufacturing Science and Engineering, Vol. 23, pp. 562-66
[10.] Lauderbaugh L. Ken (2009). “Analysis of the effect of process parameters on exit burrs in drilling
using a combined simulation and experimental approach.”, Journal of Materials Processing
Technology, Vol. 209, pp. 1909-1919.
[11.] List, G., Nouri, M., Gehin, D., Gomez, S., Manaud, J.P., Petitcorps, Y. L., Girot, F. (2005).
“Wear behavior of cemented carbide tools in dry machining of aluminum alloys.” International
Wear, vol. 259, pp. 1177-1189.
[12.] Montgomery, D. C., Design and Analysis of Experiments (2000), 5 th Ed., Wiley, New York,
[13.] Myers, R. H., and Montgomery, D. C., Response surface methodology (2002). Process and
product optimization using designed experiments, Wiley, New York,
[14.] Machinery’s Handbook, 25 th Edition, 1996 Industrial Press Inc., New York (1996).
[15.] Madhavan Nair, R., Durairajan, B. and Bahr, B., (2002). “High Speed Drilling of Al-2024-T3
Alloy.” SAE Technical Paper 2002-01-1516
[16.] Nair, V. N., Abraham, B., MacKay, J., Nelder, J. A., Box, G., Phadke, M. S., et al. (1992).
“Taguchi’s parameter design: A panel discussion.”, Technometrics, Vol. 34(2), pp. 127–161..
[17.] Nouri, M., List, G., Gehin, D., (2005). “Effect of machining parameters and coating on wear
mechanisms in dry drilling of aluminum alloys.” International Journal of Machine Tools and
Manufacturing, vol. 45, pp. 1436-1442.
[18.] Pan, H. L., Chen, Y., Orady, E., and Yan, F. (1995). "Tool Wear Monitoring by Acoustic Analysis
Through a Solid Path.," International Journal of Production Research. Vol. 33, No. 9, pp. 2455-
2470
[19.] Pan, H.L., Orady, E, and Chen, Y. (1994). "Tool Useful Life Evaluation and Monitoring by
Thermal EMF," Proceedings of the International Symposium on Manufacturing Science and
Technology for the 21st Century, Beijing, China,
[20.] Pande, S.S., Relekar, H.P., (1986). Investigations on reducing burr formation in drilling.
International Journal of Machine Tool Design Research, Vol. 26 (3), pp. 339–348.
[21.] Pande, B., Aramendi, G., Rivero, A., de Lacalle, L.N., (2005).” Monitoring of drilling for burr
detection using spindle torque”. International Journal of Machine Tools & Manufacture, Vol. 45,
pp.1614-1621.
[22.] Pignatiello, J. J. (1993). “Strategies for robust multiresponse quality engineering”, IIE
Transactions, Vol. 25(3), pp. 5–15
[23.] Taguchi, G. (1987). System of experimental design: Engineering methods to optimize quality and
minimize costs. Dearborn, MI: American Supplier Institute.
[24.] Thangaraj, A. and Langenstien, M., (1990). “Investigation into the Relationship between Hole
Quality and Drilling Forces.” Proceedings of Manufacturing International ’90: Advances in
Materials and Automation, pp. 149-157
84

DRILLING PROCESS ROBUST OPTIMIZATION FOR 6061 ALUMINUM USING DESIRABILITY FUNCTIONS
APPENDIX I:
Ra (micro inch)
Burr Height (inches)
Standard
Order
Hole
No.
Drilling
order 1 2 3 Ra_mean Ra_STDE 1 2 3 Burr_M Burr_STDE
1 1 7 140.0 142.5 113.0 131.8 16.4 0.01575 0.01880 0.03050 0.02168 0.00779
2 2 9 146.0 134.5 145.0 141.8 6.4 0.01075 0.01338 0.01238 0.01217 0.00132
3 3 5 173.0 153.5 126.5 151.0 23.4 0.00750 0.00975 0.01250 0.00992 0.00250
4 4 15 220.0 205.0 175.0 200.0 22.9 0.04613 0.03113 0.04200 0.03975 0.00775
5 5 14 133.0 129.5 131.5 131.3 1.8 0.00763 0.01240 0.01325 0.01109 0.00303
6 6 2 138.0 122.0 156.0 138.7 17.0 0.00875 0.01700 0.00625 0.01067 0.00563
7 7 11 163.0 151.0 159.0 157.7 6.1 0.01925 0.01425 0.04850 0.02733 0.01850
8 8 6 123.0 131.0 131.0 128.3 4.6 0.00950 0.01038 0.01250 0.01079 0.00154
9 9 1 115.0 110.0 185.0 136.7 41.9 0.01625 0.03400 0.04975 0.03333 0.01676
10 10 8 171.0 165.0 157.0 164.3 7.0 0.01425 0.00893 0.01163 0.01160 0.00266
11 11 10 145.0 129.5 123.5 132.7 11.1 0.00800 0.00875 0.01595 0.01090 0.00439
12 12 3 115.0 128.0 121.5 121.5 6.5 0.01125 0.01300 0.01100 0.01175 0.00109
13 13 4 135.0 142.0 119.0 132.0 11.8 0.00950 0.01613 0.01425 0.01329 0.00341
14 14 12 135.5 127.0 153.0 138.5 13.3 0.00775 0.00775 0.01195 0.00915 0.00242
15 15 13 152.5 119.5 113.5 128.5 21.0 0.01425 0.02425 0.01750 0.01867 0.00510
16 1 9 167.0 117.0 146.0 143.3 25.1 0.01500 0.00375 0.01063 0.00979 0.00567
17 2 6 143.0 152.0 167.0 154.0 12.1 0.01513 0.01350 0.02063 0.01642 0.00373
18 3 5 180.5 210.0 210.0 200.2 17.0 0.02150 0.03263 0.01775 0.02396 0.00774
19 4 10 201.0 200.0 220.0 207.0 11.3 0.01950 0.01288 0.01175 0.01471 0.00419
20 5 11 148.0 350.0 270.0 256.0 101.7 0.00988 0.00963 0.00725 0.00892 0.00145
21 6 8 199.0 148.0 144.5 163.8 30.5 0.00988 0.01050 0.01175 0.01071 0.00095
22 7 3 143.0 142.0 148.0 144.3 3.2 0.00550 0.00625 0.01050 0.00742 0.00270
23 8 2 138.0 107.0 133.0 126.0 16.6 0.01675 0.00850 0.01213 0.01246 0.00414
24 9 7 159.0 185.5 245.0 196.5 44.0 0.02638 0.01700 0.02618 0.02318 0.00536
25 10 12 159.0 153.0 176.0 162.7 11.9 0.02080 0.01900 0.02100 0.02027 0.00110
26 11 1 157.5 193.0 177.5 176.0 17.8 0.00925 0.01250 0.01400 0.01192 0.00243
27 12 14 149.0 152.0 153.0 151.3 2.1 0.00713 0.00775 0.00900 0.00796 0.00095
28 13 4 130.5 139.0 133.0 134.2 4.4 0.00963 0.01200 0.01300 0.01154 0.00173
29 14 13 132.5 155.0 151.0 146.2 12.0 0.03038 0.00675 0.00750 0.01488 0.01343
30 15 15 165.5 202.0 186.5 184.7 18.3 0.02970 0.03850 0.01563 0.02794 0.01154
85

AHU J. of Engineering & Applied Sciences 3 (1) : 87-105 (2010)
© 2009 ALHOSN University
NEIGHBORHOOD DESIGN AND WALKABILITY:
A SYNTHESIS FROM PLANNING, DESIGN,
TRANSPORTATION AND
ENVIRONMENTAL HEALTH FIELDS
Mohamad Kashef *
Department of Architecture, ALHOSN University, P.O Box 38772, Abu Dhabi, UAE
ABSTRACT: This study develops a nuanced understanding of issues related to neighborhood design and
walkability, which received increased attention over the last two decades. It examines a broad range of
environmental health, transportation, and design/planning studies. The study argues that the bulk of published
empirical research relied on statistical tools that measured the design ingredients rather than the morphological and
spatial structures that enhance walkability. The qualitative design aspects of built environments were hardly
translated to quantifiable variables in much of the reviewed empirical research. This paper highlights critical
configurational and space syntax dynamics that complement the current understanding of neighborhood design and
walkability. It opens the door for further cross disciplinary research aimed at understanding and potentially
optimizing the relationship between urban form and pedestrian movement.
1. INTRODUCTION
This study engages the planning and urban design literature as well as transportation and
environmental health fields to develop an interdisciplinary understanding of issues related to
neighborhood design and walkability. It examines intellectual diversion and/or conversion
between environmental health studies, designers/planners narratives, and transportation analyses.
Over the last four or five decades, various design and planning publications called for reviving
traditional built forms and reintroducing mixed uses to enhance walkability and urban vitality
(Jacobs [53]; Alexander [4]; Rowe et al [81]; Krier [58]; Trancik [90]; Barnette [10]). The New
Urbanists also expressed a renewed interest in traditional spatial configurations and building
typologies. They contended that fine-grain connected urban blocks, mix of uses, and attention to
architectural details such as sidewalks, porches, and visual stimuli can encourage walking,
bilking, and civic engagement. The presumption is that technological and functional imperatives
of the dispersed metropolis did not change the basic human desire for well-designed
neigborhoods that support residents’ daily needs within a five-minute walking distance (Duany et
al [24]; Katz [56]; Dutton [26]). A host of recent empirical (statistically-supported)
investigations from the public and environmental health fields, on the other hand, pointed to the
lack of causality between neigborhood design and walkability. Most of these studies noted that
certain neigborhood types with design features such as connected sidewalks, architectural variety,
etc. are positively correlated with walking and biking. They generally explained such associations
in terms of the market self-selection process, i.e., people who like walking and biking commonly
choose to live in places that support their way of life (Cervero et al [19]; Atkinson et al [5]; Reed
et al [79]; Oaks et al [72]; Lovasi et al [65]; Nagel et al [70]).
This paper argues that the bulk of published empirical research relied on statistical tools that
measured the design ingredients rather than the morphological and spatial structures that enhance
____________________________________
* Corresponding Author. Tel.: +971 4070511
E-mail : m.kashef@alhosnu.ae
87

MOHAMMED KASHEF
walkability. The paper highlights shortcomings of statistically-supported research in capturing
the essence of qualitative design values and transforming them into measurable study variables.
Critical configurational and syntactical dynamics (Hiller [51], [49], [48]; Bafna [7]; Peponis et al
[74]; Baran et al [9]) were not considered in measuring the cause and effect relationship between
design and walkability. The paper is organized into three major parts: the first part reviews the
evidence from environmental health fields; the second part encompasses the results from
planning, design and transportation studies; the paper concludes with a discussion that
synthesizes the evidence and provides an interdisciplinary perspective that potentially benefits
public policy debates regarding land use and subdivision regulations that support walking and
biking.
2. ENVIRONMENTAL AND PUBLIC HEALTH STUDIES
These comprise a broad range of empirical investigations within health related disciplines such as
public health, nutrition, exercise physiology, epidemiology, and preventive medicine, among
several others. They addressed the nexus between walking and neigborhood design from a variety
of perspectives that consider objective and perceived built environments, personal attitudes and
behavior, safety (real and perceived), environmental constraints, purpose (leisure or
work/shopping related nonmotorized travel, as well as health and lifestyle characteristics). The
presence of sidewalks is often touted by the New Urbanism as an important environmental
resource that supports walking in neigborhoods (Calthorpe et al [16]; Fulton [35]; Duany et al
[24], [25]; Katz [56]). The evidence from health studies regarding the impact of sidewalks on
physical activity and walking in general is mixed. The following discussion articulates the
findings of health studies under three themes or headlines: walking behavior and design,
sidewalks, and street connectivity. These themes are not mutually exclusive; there is a great deal
of overlap and such articulation is merely offered as an organizational framework for the large
amount of data addressed below.
2.1 Walking Behavior and Design
In an attempt to predict walking behavior, Ryan Rhodes et al [80] integrated personality, the
perceived environment, and time planning into a framework called “TPB” (theory of planned
behavior). The study included various measures for personality traits, actual walking,
neigborhood geography (sidewalks, architecture, aesthetics, etc.), and attitudes towards outdoor
physical activity for leisure purposes. Research instruments were administered to 358 residents of
British Columbia, Canada. The study model provided evidence that leisure-time walking is
largely intention based behavior with an additional independent contribution from physical
proximity and accessibility of neigborhood retail and recreation facilities. As shown in the
adapted model diagram (Fig. 1), neigborhood design and aesthetics impact walking but only
through mediating intention based variables that include attitudes and other social cognitive
constructs. In that sense, the environment provides infrastructure support but is not the cause for
walking per se. Notable among the results of this study is the impact of proximity of parks to
homes. Individuals who are habitually active and lived at a reasonable walking distance from
recreation venues were more inclined to walk (Rhodes et al [80]). Proximity of homes to exercise
facilities was also positively correlated with the frequency of exercise among San Diego residents
(Sallis et al [82]). These results are of particular significance to community planning practice. It
88

NEIGHBOURHOOD DESIGN AND WALKABILITY
shows that adding destinations of interest and recreation venues can promote walking among
active residents and help close the intention-walking/activity gap.
Mixed-Use
Affective
Attitude
Design
Elements
Intention
Walking
Instrumental
Attitude
Aesthetics
Figure 1. Perceived environment and theory of planned behavior model to predict walking
(Adapted from: Rhodes et al [80])
2.2 Sidewalks
Sidewalk availability did not figure as a prominent indicator of physical activity among residents
of different neigborhoods in Washington State. Objectively measured built environment
characteristics including sidewalks within a mile from research participants’ homes accounted for
only one percent of residents’ total physical activity (Lovasi et al [65]). Research conducted in
Portland, Oregon (fifty-six neigborhoods) yielded similar results. Findings did not suggest that
built environment plays a significant role in the walking and physical activity of older adults.
Results were consistent with previous studies in which sidewalk presence and neighborhood
aesthetics were associated with increased physical activity among active residents who habitually
walk (Nagel et al [70]; Atkinson et al [5]). In a stark contrast with previous results, the findings
of some other studies pointed to positive associations between the perceived presence of
neighborhood sidewalks and physical activity behaviors (Ainsworth et al [2]; Addy et al [1]; Lee
et al [61]). These findings were significant because research participants were classified as
irregular walkers, namely, white older adults and African American women. This, in a sense,
substantiates the New Urbanists’ claims regarding the impact of neigborhood design on walking
behavior.
Various studies reported that both objective and subjective characteristics of the built
environment were correlated with walking, cycling, and outdoor physical activity in general.
Objectively measured correlates included the physical quality of walking routes and the presence,
type, and geographic proximity of destinations such as retail shops, parks, and other recreation or
cultural facilities. Subjective variables included the perception of safety (traffic and crime
related), the convenience of neigborhood services, and the perceived visual quality of
neigborhood landscape and architecture (Moudon et al [69]; McCormack et al [67]; Boehmer et
89

MOHAMMED KASHEF
al [11]; Lesliea et al [62]; Geus et al [36]). Using a multi-stage sampling design and existing GIS
databases, Li et al (2005) surveyed all neigborhoods in Portland, Oregon (521, 121 people) and
reported positive associations between urban form and walking activity. Places with mixed-use
and higher levels of green and open spaces were significantly associated with increased walking
activity. The results of the latter surveys were noteworthy because findings showed that variation
in walking activity across neigborhoods was accounted for by environmental characteristics.
2.3 Street Connectivity
Street connectivity also emerged as an important attribute of walkable neigborhoods. This is a
measure of the accessibility achieved by having smaller blocks (less than 1000 feet long) and
ease of movement between developed blocks without being obstructed by dead-ends. A
connectivity index is commonly derived by dividing the number of roadway links by the number
of roadway intersections and nodes (cul-de-sacs included). A higher number indicates a wellconnected
street system (Ewing, 1996). Data from Metropolitan Atlanta’s region, Oakland, New
Zealand, and Perth, Australia showed a significant correlation between increased levels of
outdoor physical activity among adults and well-connected streets, continuous pedestrian
sidewalk network, and mix of land uses (Frank et al [33]; Badland et al [6]; Chin et al [20]).
Proximity of workplace and commercial destinations was highlighted by some studies as
significant contributors to total transport-related walking. It was estimated that individuals who
regularly walk to work would approximately accumulate 166 weekly minutes of walking, which
meets health guidelines for adults (Cerin et al [18]). Residents of mixed-use high walkability
neigborhoods reported 70 more minutes of physical activity per week than did residents of singlefamily
low-walkability neigborhoods. They also had lower obesity prevalence after adjusting for
individual demographics (Saelens et al [83], [84]).
Neighborhood design and aesthetics figured prominently in other studies as conducive to active
aging, enhanced walkability, and higher perception of safety and sense of community (Michael et
al [68]; Leyden [63]; Wood et al [92]). This again substantiates some of the New Urbanists’
claims regarding the impact of Physical elements such as building aesthetics, streetscapes, vistas,
porches, etc. on to the livability and vitality of neigborhoods. Such elements are frequently cited
in the New Urbanism literature as the building blocks of communities. They have the potential of
enhancing social interaction and hence igniting the culture of community. The New Urbanists
generally contend that reinstating the traditional morphology of street and civic architecture can
contribute to restoring the lost sense of place and reviving many long-gone communal practices.
At present, suburban patterns are completely dependent on the private automobile and lack the
basic ingredients conducive to collective social and cultural dynamics (Duany [24; Katz [56];
Calthorpe [16], [17]; Kelbaugh [57]). Rob Boer et al [58] adopted the Smart Scorecard developed
by the New Urbanism (Fleissig et al, 2002) as a research instrument to assess the impact of
neigborhood design on walking trips in ten U.S. metropolitan areas. The Smart Scorecard
evaluates various built environment measures such as density, mix of uses, street network
connectivity, block sizes, diversity, landscape, etc. Based on data analysis from the 2000 U.S.
census, the study indicated that only three of the ten critical components of the Smart Scorecard
were found to be associated with walking, namely: accessibility and mobility choices
(connectivity), compactness (density), and mix of uses (Boer et al [12]).
90

NEIGHBOURHOOD DESIGN AND WALKABILITY
2.4 Summarizing the Results of Health Studies
It seems that health-related studies offer more evidence on the positive side of association
between built environment and walkability. However, the findings of these studies were largely
based on cross-sectional data. Most of the positive associations were derived from respondents
perceived or subjective characterizations of the built environment. Almost none of the reviewed
health studies claimed to have identified causal linkages between environmental attributes and
physical activity; study results could not attest to whether the criteria of time precedence and nonspuriousness
were established. Self-selection always figured in study limitations and caveats as a
possible explanation for the positive associations of neighborhood design and walkability; people
who walk regularly choose neighborhood environments that promote walkability and not vice
versa.
Two very recent studies with quasi-longitudinal design provided contradictory evidence. Handy
et al [45] pointed to a causal linkage between neigborhood environment and physical activity.
Wells et al [91], on the other hand, asserted that neigborhood features alone such as sidewalks,
front porches, small set-back distances, and aesthetics may not be enough to affect walking.
Residents of New Urbanism neigborhoods did not walk significantly more than those living in
conventional suburbs (Wells et al [91]). Research findings from the Health field did not provide a
definitive answer as to the relationship of walkability and neighborhood design one way or the
other. The only possible recommendation out of health studies is that neighborhood design does
affect the opportunities for walkability; it has the potential to enhance nonmotorized forms of
travel, especially for a select group of physically active individuals. The following part of the
paper examines the evidence from the design, planning, and transportation fields. While studying
similar constructs, the approach in most of these studies varies from health studies.
Transportation and some planning studies generally adopt quantitative methodologies with a
variety of measuring devices that include structural equations, systems analysis, computer
modeling and simulation, etc. Other planning and design studies are qualitative in nature, at times
intuitive, and provide rich historiography and in-depth narratives about cities, communities, and
the forces that shape them.
3. TRANSPORTATION AND PLANNING/DESIGN STUDIES
There is a great deal of overlap between planning, design, and transportation studies, especially in
regard to the impact of design on physical activity. Transportation studies focus more on travel
related issues such as modal split, traffic management and safety, as well as street network
optimization. Planning and design fields, on the other hand, capitalize on the data produced by
transportation studies to establish connections between neigborhood architecture, street network
integration, walkability, and sociability. Transportation and planning issues are inextricably
linked, which may not allow addressing them separately. The following discussion articulates the
findings under three themes: land use and transportation, walkability and place attachment, and
space syntax and connectivity. The issues discussed under these categories are overlapping, and
such articulation is mainly offered as an organizational framework for the large amount of data
addressed below.
91

MOHAMMED KASHEF
3.1 Land Use and Transportation
Coordinating transportation and land use emerged as a critical component of viable economic
development policies, infrastructure planning and management, and more significantly for this
paper, as a potential modifier of travel behavior. Zoning and land use are used exchangeably to
denote an approach to planning urban and suburban developments. Traditional zoning, often
referred to as “Euclidean Zoning” after the approach was approved in the Supreme Court
decision in 1926, began as a political response to the obnoxious effects of industry on human
settlements. Zoning regulation progressed further and enacted exclusive zones for office use,
commercial, and different housing categories. The rationale behind land use regulation was to
promote the public’s health, safety, and welfare. It was also meant to stabilize property values,
encourage stable and congenial neigborhoods, and regulate traffic flow. Over the past several
decades, zoning was subject to regular criticism and commonly described as inflexible and
exclusionary. It codified the relocation of commercial and office functions to suburbs, causing a
spatial mismatch between lower classes in downtowns and suburban employment locations. The
single-use areas prescribed by zoning resulted in sprawling suburban patterns with substantial
distances and voids between related urban functions. The private automobile was the most logical
choice, and other modes of transport such as public transit, walking, or biking became unviable.
There is a heated planning debate regarding the role that zoning played in shaping the American
urban/suburban form. Some planning studies portrayed zoning as a cultural artefact that reflected
long-standing societal choices, which continue to be asserted in buying decisions and outright
preference for auto-dependent, single-use suburban locations. These referred to the efficiency of
market-driven policies and the economic rationale of bid-rent curves which show how the highest
and best use of land is determined. They maintained that zoning bylaws and land use regulations
do not have as much effect on decentralizing American cities as consumer preferences and tastes;
developers are market-conscious and do not usually risk building an unacceptable product
(Kashef [55]). Other planning and design studies referred to zoning as the culprit behind the
dispersion and fragmentation of the North American built form. In that sense, suburban sprawl
was promoted by a combination of zoning regulations, mortgage policies, and massive highway
and road building programs that were aided by the federal and state governments. These policies
helped instate the private automobile as the one and only viable mode of transport within most
American cities. Most people had limited choices and were sold on a highly publicized and
government-subsidized ideal (the American dream) that shaped the development market over the
past several decades (Kashef [55]; Ewing [28]; Duany et al [25]). The New Urbansits have
always emphasized the need to replace Euclidean zoning with flexible built form controls that
provide more options and create socially and physically integrated communities and
neigborhoods. The latter should promote mix of uses and built forms and coordinate land use and
transportation, thereby allowing a rich modal split between motorized (car/public transit) and
non-motorized (walking/biking) transportation. The New Urbanist agenda is predicated on the
hypothesis that changing neighborhood land use configurations and increasing their density
would influence travel behavior and promote public transit, walking and biking (Duany et al [24],
[25]; Calthorpe et al [16]; Katz [56]; Dutton [26]).
The findings of some recent transportation and planning studies cast doubts on the potential of
neigborhood design on travel behavior. Bagley et al [8] examined the relationship of residential
92

NEIGHBOURHOOD DESIGN AND WALKABILITY
neigborhood type to travel behavior, incorporating attitudinal, lifestyle, and demographic
variables. Data were collected from five neigborhoods in the San Francisco Bay area with singleand
mixed-use configurations. The results showed that people with different travel propensities
select themselves into neigborhoods that support their lifestyle and attitudinal predispositions.
Mixed land use configuration in and of itself did not affect travel behavior (Bagley et al [8]).
Other Bay Area studies have shown that mixed land uses and close proximity to retail services
may induce nonmotorized transport. However, the data analysis revealed that land use on its own
had a modest and sometimes statistically insignificant effect on walking and biking. Various
exogenous factors, such as topography, darkness, and rainfall had far stronger influences than
land use. Demographic characteristics of trip makers were also far stronger predictors of walking
and biking rather than land use factors (Cervero et al [19]; Schwanen et al [86]). Evidence from
the RESIDE project in Australia also contradicted the New Urbanism hypothesis. The RESIDE
project aimed at evaluating the impact of land use mix among other urban design factors on
walking for recreation and transport. It was a five-year longitudinal study that used a baseline
survey completed by home owners before and after moving from single-family estates in monozoned
suburbs to mixed-use neighborhoods. The results underscored the process of self-selection
and provided no evidence that residents became more physically active in mixed-use
neigborhoods than other types of estates (Pikora et al [75]).
Reid Ewing, on the other hand, argued that development patterns and especially land use mix
affect travel patterns. By operationalizing various aspects of the built environment, Ewing
compared the extremes of sprawl and compact mixed use developments. He emphasized that
relatively higher densities and balanced mix of residential, commercial, and employment
activities would render active travel modes such as walking, biking, and transit more viable
(Ewing [27]). Such assertions were supported by various other studies that focused on
disaggregate data measuring walking activity on the household and individual levels. A study of
Austin neigborhoods in Texas showed that mixed-use developments with relative proximity of
retail, employment, and residential uses generated almost twice the walking activity of
conventional suburbs (Handy [37]). The findings of this and other studies suggested that
utilitarian trips to work, shopping, etc. were the source of overall differences in walking activity
between mixed-use and conventional suburbs (Handy [40]; Saelens et al [83]; Rajmani et al [77];
Handy et al [38], [39], [43]).
A general sense of unease, however, characterizes the presentation of findings of various
empirical investigations. Kevin Krizek [59] conducted a longitudinal study about land use and
travel patterns of 6,144 households in Seattle, Washington. The study analyzed the changing
patterns of travel behavior for families who moved from conventional suburbs to mixed-use New
Urbanist neighborhoods. The findings generally indicated that changes in urban form and land
use triggered changes in modal split. Overall vehicle miles traveled (VMT) were decreased after
moving to mixed-use environments. A closer look at the data revealed that such VMT reduction
was not universal and it was mainly generated by households who were previously active and
used transit even before they moved to New Urbanist communities (Krizek [59]).
93

MOHAMMED KASHEF
3.2 Walkability and Place Attachment
Walkability here refers to non-work related physical activity that does not necessarily rely on
mixed land use configurations that allow people to walk from home to work or major shopping
centers and vice versa. It is more about walking for leisure purposes, recreation (parks and
playgrounds), marginal shopping errands (buying a Popsicle), and/or exercise. Place attachment
is a complex and pertinent construct that figured in most design and planning literature as both a
by-product of and contributor to walking and overall human presence in the public sphere.
Walking in that sense does not contribute much to reducing VMT; it does not substitute for
automobile driving but may have positive impact on people’s daily physical activity levels and
overall health. Place attachment is often described as a “positive affective relationship between
people and place” that occurs because of people’s satisfaction and identification with a specific
place. Identification with a place results in emotional connections with its physical attributes that
people often associate with social events. Place and associated emotions and events become
inseparable or interchangeable in people’s minds. The meaning of place here encompasses built
environments as well as natural and cultural landscapes.
Designers have generally maintained that certain spatial configurations, built forms, and design
elements can play a significant role in promoting walkability, safety, and place attachment. The
most significant design theme arising from the design and planning literature and potentially
affecting neighborhood walkability and sociability is the interface between the private and public
realms (Kelbaugh [57]; Jacobs [53]; Alexander [4]; Duany et a [25]). It is essentially about the
articulation of the outer skin of buildings or homes and disposition of buildings in space. Various
studies indicate that increasing the interface between private and public realms tends to heighten
the level of activity around the edges of buildings. From a residential design perspective,
elements that contribute to increased private/public interface include windows of active living
spaces, transitional spaces such as entryways or porches, as well as the availability and
distribution of common spaces. As long as home privacy is maintained, the presence of such
elements creates opportunities for walking and social interaction among inhabitants (Duany et al
[25]).
The interaction referenced here can be characterized as passive or active. The passive social
interaction falls under a category that Bill Hillier calls ‘social or spatial co-presence.’ Hillier
contends that co-presence and therefore co-awareness amongst the individuals living in and
passing by an area can be considered a raw material for social interaction. Whilst co-present
individuals are not a community and may not know or acknowledge each other, they in fact form
a social resource for communal behavior. Co-presence is a very important psychological resource
in and of itself and can be brought about by design (Hillier [51]). Jane Jacobs’s ‘eyes on the
street’ and Newman’s ‘defensible space’ have earlier noted the significance of home windows
and visible entrance doors in providing surveillance and self-policing of public spaces (Jacobs
[53]). Active interaction is commonly perceived as a by-product of increasing the pattern of
‘probabilistic interfaces,’ a heightened sense of co-awareness or co-presence, leading to overt
social acknowledgment (Hillier [51]: 129-214). Transitional spaces such as home porches
provide a setting for conversation, a kind of ‘Shakespearean’ theatrical platform for social ‘play.’
Porches project the human presence within the house to passersby, creating a sense of ‘plot’; a
voluntary and potentially entertaining social exchange (Figs. 2, 3). The availability of the home
94

NEIGHBOURHOOD DESIGN AND WALKABILITY
porch is significant whether a conversational exchange takes place or not. The porch acts as a
precondition or a ‘stage’ for a social play that otherwise would not have been possible.
Figure 2. A single family home with a porch fronting a
neighborhood sidewalk, New Bern inner suburbs, North
Carolina (built in the 1940s).
Figure 3. A single family home in a modern suburb,
New Bern, North Carolina (built in 2006).
Another important dimension of the notion of walkabilty and sociability of built environments,
and a factor in the private/public interface, is the idea of closeness or compactness. The New
Urbanism called for shrinking private space and positioning homes closer to the curbside.
Reducing the distance between home porches and sidewalks contributes to the ‘Shakespearean’
plot and may enhance passive (co-awareness) and potentially active social exchanges between
residents and passersby. New Urbanism codes typically prescribe a 15-20 feet optimum distance
between sidewalks and porches to facilitate social encounters (Duany [25]). Reducing the
physical footprint of buildings within a given development would also free more space for public
activities. Public places in the form of parks and/or civic facilities such as picnic shelters, park
gazebos, playgrounds, sports fields, etc. provide venues for chance encounters and serve as
symbols of places (Langdon [60]). The urban design literature in general, and the New Urbanism
praxes in particular, recognize sense of place as a function of sense of space which requires
heightened attention to spatial configurations, design details, and disposition of public functions.
The planners, designers, and New Urbanists’ case regarding the impact of spatial form and design
details on human behavior is theoretically supported by environmental psychology and/or
environmental sociology studies. Individuals associated with the latter strand of research aim to
develop a body of knowledge that would enable designers, engineers, and other professionals to
envision optimum physical environments that enhance human functions and social relations.
Some environmental sociology studies showed statistical evidence that design and spatial
allocation of homes within residential blocks can have an impact on residents’ social relations
and interactions. They asserted that facilitating people closeness by proper arrangement of private
and shared public spaces enhances social contact and group/place attachment. Emily Talen [89]
referred to housing studies in Chicago and Pittsburgh that highlight the role of the neighborhood
physical environment in shaping residents’ social networks and interactions. Other researchers
95

MOHAMMED KASHEF
have found a direct correlation between home design and family interaction. Strategically located
kitchens and living spaces tend to foster interaction and provide more opportunities for
socialization among family members.
John Peponis et al [74] have also pointed to direct and statistically supported evidence that office
layout can contribute to the density of different network interaction and productivity. Integrated
office layouts establish patterns of informal co-awareness and interaction, thus facilitating
communication, sharing of ideas, and the joint exploration of possibilities in the non-routine
phases of knowledge work. Other recent planning studies provided some credence to the New
Urbanism claims regarding the impact of neighborhood design on non-work walking activity.
Among the most important variables predicting a difference in walking was a change in the
attractiveness of built environment. All else equal, people walked more after moving to visually
pleasing neighborhood environments with variety of housing styles, well-maintained sidewalks
and landscape, better street lighting, and more opportunities for social interaction. Such changes
did not necessarily translate to less driving; it mainly impacted leisure-based walking (Handy
[42], [43]). These results were corroborated by various other studies inside the U.S. and other
cities in Australia. Design aesthetics, perceived safety levels, sidewalk treatments, trees, and the
overall quality and variety of buildings were conducive to increased walkability and sociability
(Rajmani et al [77]; Doyle et al [23]).
3.3 Space Syntax and Connectivity
Bill Hillier and Julienne Hanson probably coined the term ‘space syntax’ which delineates a set
of spatial concepts and principles that underlie human interaction with the built environment
(Hiller et al [52]). The aim of space syntax is to optimize the spatial allocations and connections
between human functions so as to produce the most effective configurations and circulation
networks. The assumption is that movement through space is a key organizing element in the
social functioning of human settlements. Hillier contends that socioeconomic forces shape the
city primarily through the relations between movement and the structural layout of local districts.
A mixed-use strategy that brings homes closer to retail cannot generate movement unless shops
are selectively located on integrating lines that carry the most movement in relation to
surrounding spaces (Hillier [50], [51]). Hillier’s work sparked the development of a large volume
of research around the world that used space syntax logic to examine spatial configurations of
homes, offices, neigborhoods, cities, metropolitan regions or movement networks of entire
countries (Baran et al [9]; Hillier [49], [48]; Bafna [7]; Jiang et al [54]; Nophaket et al [71];
Raford et al [76]; Peponis et al [74]). There are some contentious issues surrounding the
application of space syntax tools to the real world. Space syntax relies primarily on twodimensional
mapping and geometrical techniques that discard vertical and horizontal dimensions
as critical components of spatial structures (Ratti [78]; Hillier [46]). A detailed investigation of
the space syntax and related contentious issues is outside the scope of the current paper. The
objective here is to highlight the value of the space syntax approach in understating the dynamics
of street network design and walkability. A neighborhood space syntax analysis starts with an
axial representation of its street layout. The axial map shows the neighborhood streets as a series
of intersecting lines creating a network of links and nodes. Each street or line on the map has a
different level of connectivity with the network based on the number of other streets that intersect
it. Streets that generate more intersections hold higher value as elements of integration in the
96

NEIGHBOURHOOD DESIGN AND WALKABILITY
network. Important functions that tend to catalyze movement such as parks, shops, or other
public amenities should be located on streets with the highest integration value in the network.
Cumulatively, networks with more open (three- or four-way) street intersections create higher
connectivity values (a grid system is the optimum).
Connectivity indices measure the directness and availability of alternative routes between
different locations within the road network. They are commonly used to assess the walkability
potential of street networks. A connectivity index is commonly derived by dividing the number
of roadway links by the number of roadway intersections and nodes (cul-de-sacs included). A
higher number indicates a well-connected street system (Ewing [28]). Figures 4 & 5 juxtapose
two hypothetical subdivision layouts with different connectivity potential. A link in the
connectivity index is simply a stretch of road between two nodes. A node ranges from a dead-end
cul-de-sac to a three- or-four-way intersection. The more links available in comparison to the
number of nodes, the higher the connectivity and the potential of the street network to generate
more pedestrian movement (Schmidt et al [85]). Various studies indicated that smaller blocks
(less than 1000 feet long) and higher connectivity indices tend to generate more pedestrian
movement (Ewing [28]; Southworth [88]; Krizek [59]; Cerin et al [18]; Frank et al [33]; Badland
et al [6]; Chin et al [20]).
Figure 4.
Typical Subdivision Tract with Poor Connectivity
Connectivity Index: Roadway Links/Nodes: 8/9 0.89
Subdivision Statistics:
No. of lots 84; average lot size 10,000 square feet
No. of external access points 1; no. of dead-end
streets 5
Poor internal connectivity: Pedestrians have to
backtrack from dead-end streets to navigate
different areas of the subdivision. Roads do not
encourage walking or biking.
Poor external connectivity: There is a single
access point on the collector road and no
possibilities for alternative connectors with neighboring subdivisions.
Figure 5.
Typical Subdivision Tract Outfitted with Good
Connectivity
Connectivity Index: Roadway Links/Nodes: 15/10 =
1.5
Subdivision Statistics:
No. of lots 84; average lot size 10,000 square feet
11
97

MOHAMMED KASHEF
No. of external access points 4; no. of dead-end streets 1 (less than 200 feet road stretch)
Good internal connectivity: Pedestrians can navigate different areas of the subdivision along a
continuous loop and small block sizes that encourage walking and biking.
Good external connectivity: 1 main access along the collector road and 3 other alternative
accesses that enhance connectivity with neighboring subdivisions and encourage walking and
biking.
Increasing street connectivity would mean better intra- and-inter-neighborhood accessibility not
only for pedestrians and cyclers, but also for cars. In that sense, neighborhoods would be
connected by internal streets that allow low-speed travel across a large number of neigborhoods
in suburban regions. This could potentially reduce congestion along arterial roads and provide
more route options when some links are blocked by traffic. Despite concerns about safety and
increased car travel along suburban roads, transportation studies indicated that improved
connectivity in suburban regions would provide a much needed flexibility of movement,
especially for emergency vehicles. Building some redundancy in road networks is critical to their
long-term viability and ability to generate a rich modal split between motorized and
nonmotorized forms of transportation. Pedestrians and cyclers would be encouraged to travel
along low-speed internal suburban roads rather than venturing into high-speed regional
transportation corridors (Crane [21]; Ewing et al [29]; Handy et al [41]; Langdon [60]; Alba et al
[3]). The findings of the Atlanta and Central Puget Sound traffic studies showed that doubling
suburban road intersections would decrease regional vehicle miles traveled by approximately five
to ten percent. Such decrease was attributed to the additional route options, shorter distances
between origins and destinations as well as a better modal split between driving, walking, and
biking (SMARTRAQ [87]; LUTAQH [66])
3.4 Summarizing the Results of Planning, Design, and Transportation Studies
Some transportation and planning studies corroborated the findings of health studies regarding
the role of self-selection process in explaining the difference in walkability between conventional
suburbs and mixed-use neigborhoods; people seem to select themselves into neigborhoods that
support their lifestyle. These studies generally doubted the efficacy of the New Urbanism design
techniques in changing people’s attitudes and encouraging a more active lifestyle. Planning
studies, however, implicated zoning and government policies in entrenching the suburban ideal
and creating car-dependent, low-density suburbs. Design and planning narratives criticized the
exclusionary practices of Euclidean zoning. In that sense, zoning compartmentalized land uses,
reduced densities, and eliminated functional overlaps that used to define vibrant human
settlements. The ensuing metropolitan form was riddled with social and spatial mismatches
between housing, commercial, and employment activities. Zoning and its attendant financial
policies and infrastructure projects resulted in unsustainable urban growth with restricted
individual options and inflexible development patterns. Two significant themes figured
prominently in planning and design analyses. The first theme was the nexus between walkability
and place attachment. Statistically supported investigations within environmental psychology and
sociology asserted that good design can be used to bolster the experiential quality of built
environments. Design provides possibilities for social interaction that enriches people’s lives and
enhances their sense of place. Identification with a place results in emotional connections with its
physical attributes. Homes, windows, sidewalks, porches, landscape, parks, etc. become critical
98

NEIGHBOURHOOD DESIGN AND WALKABILITY
correlates of people’s presence in the public sphere. The second theme was space syntax and
connectivity. Spatial relations, not architecture per se, define the ability of built environments to
generate effective patterns of movement. Neighborhood walkability gets depleted once
movement networks are compartmentalized through dead-ends and reduced intersections in the
name of exclusivity and efficiency. Integrated spatial networks, higher street intersection
densities, and supportive architectural configurations are all crucial for creating walkable
neighborhoods. The crux of the planning and design analyses is that cities should create welldefined
interfaces between buildings and streets and between localized movement in
neighborhood streets and the more globalized pattern of movement in metropolitan areas.
4. SYNTHESIS AND FUTURE RESEARCH DIRECTIONS
Though pointing to positive associations between built environment and walkability, the
empirical evidence discussed in this paper is not sufficient to frame a theory of built environment
and human behavior. It did not provide a definitive answer one way or the other to the
relationship between walkability and built environment. The available evidence can hardly be
used as a firm basis for policy guidance. However, juxtaposing the results of studies from
environmental health and planning, design, and transportation fields revealed some
inconsistencies in the way attributes of built environment were operationalized and measured in
health and planning empirical investigations. The qualitative aspects of built environments
highlighted in the design literature were hardly translated to quantifiable variables in much of
these empirical studies. For example, the issue of sidewalks was mostly operationalized in a
binary ‘0/1’ mode, available or unavailable. The design and space syntax literature indicated that
there are critical configurational and spatial relationships that must exist before physical design
elements can achieve their desired behavioral objectives. Sidewalks need to be located near home
fronts to enhance the private/public interface. Home architecture in return should encompass
transitional spaces that project human presence to the public space, e.g. porches and visible
entryways. Windows of active spaces are also crucial to enhance these configurational and
sociospatial dynamics. Other important design values include visual interest, coherence,
enclosure, legibility, and variety. None of the reviewed statistically-driven studies translated such
qualities into measurable variables. It is worth noting that Ewing et al [30] developed operational
definitions and measurement protocols for such qualitative urban design values as imageability,
enclosure, human scale, transparency, and complexity. These measurement protocols were not
utilized or tested in the most recent empirical studies.
The connectivity index used in various empirical investigations is another example of such
reductionist statistical measures. The notion that street networks have more four-way
intersections as well as more links than nodes does not say anything about the integrity of the
road and block system, the distribution of key public functions within the street network, or the
overall synergy achieved between form and function. The connectivity index is a very effective
measure for studying the flexibility of road networks for motorized traffic but not an
encompassing measure of intention-based and/or environment-provoked walking behavior. It is
definitely a helpful measure, but may not supplant for a qualitative multi-dimensional
understanding of the integration between local and regional road networks and disposition of
buildings and uses in space. It may not also be used as a unique measure to establish a cause and
effect relationship between built environment and walking behavior.
99

MOHAMMED KASHEF
The New Urbanism literature points to the need for creating a synergy between various design
qualities in order to produce walkable and vital built forms. A piecemeal or reductionist approach
that utilizes some elements and removes others may not achieve the desired objectives. The
position taken here is empirically based and calls for internalizing qualitative urban design values
and translating them into measurable protocols. Reductionist measures employed in many
empirical studies should be replaced by rich protocols that benefit from the New Urbanism
approach to understanding human settlements. Implicit in many of the recent empirical
investigations is that the New Urbanism developments embody most or all of the good urban
design principles that may enhance walkability. The New Urbanists have certainly succeeded to
create that image. However, New Urbanist developments manifest a variety of approaches and
selective inclusion of design principles in primarily profit-seeking development projects. This is
not something that runs contrary to the movement mission. The New Urbanism aims to achieve
its design objectives through private market dynamics. Thus, including New Urbanist
developments for statistical and comparative analyses should be accompanied by in-depth,
critical evaluation and assessment of their design and spatial morphology.
Despite its current limitations, the space syntax approach holds a great potential for
understanding morphological structures and built configurations. It goes beyond the simplistic
binary characterizations of mixed-use/single-use, available/unavailable sidewalks, or
connected/disconnected built environments. Space syntax analysis was used in some empirical
investigations as a two-dimensional mapping technique to assess connectivity of road systems.
But Hillier’s analyses highlighted a comprehensive approach to understanding human settlements
that goes beyond two-dimensional mapping. It tackled movement through space, spatial
allocation of important public functions, and configurational relationships between buildings and
streets. Bringing homes closer to retail cannot generate movement unless shops are selectively
located on integrating lines that carry the most movement in relation to surrounding spaces. This
is critical for applying space syntax and understanding the nexus between built environment and
walkability. Mixed-use was incorporated in some of the reviewed empirical investigations as one
of the determinants of walkability. Critical evaluations of spatial relationships between homes
and shops were generally lacking. This paper takes the position that integrating two- and-threedimensional
metrics with space syntax analyses holds a great potential for future research aimed
at understanding the complex relationship between built forms and human living patterns. The
quest for understanding form and behavior would also require a concerted effort to develop
comprehensive measurement protocols for a variety of qualitative urban design values.
5. REFERENCES
[1] Addy CL, Wilson DK, Kirtland KA, Ainsworth BE, Sharpe P, Kimsey D. Association of
perceived social and physical environmental supports with physical activity and walking behavior.
Am J Public Health. 2004; 94:440-443.
[2] Ainsworth B. E., Wilcox S., Thompson W. W., Richter D. L., Henderson K. A. (2003) Personal,
social and physical environmental correlates of physical activity in African American women in
South Carolina. Am Journal of Preventive Medicine, 25 (3, Supplement 1): 23-29.
[3] Alba, C. A., and Beimborn, E. (2005) Analysis of the effects of local street connectivity on
arterial traffic, Transportation research Board Annual Meeting, www.trb.org.
100

NEIGHBOURHOOD DESIGN AND WALKABILITY
[4] Alexander, C. (1977) A pattern Language: Towns, Buildings, Construction. Oxford University
Press, New York
[5] Atkinson, J. L., Sallis, J. F., Saelens, B. E., Cain, K. L., and Black, J. B. (2004) The association of
neigborhood design and recreational environments with physical activity, American Journal of Health
Promotion, 19 (4): 304-309.
[6] Badland, H. M., Schofield, G. M., and Garrett, N. (2008) Travel behavior and objectively
measured urban design variables: Associations for adults traveling to work, Health & Place, 14: 85–
95.
[7] Bafna, S. (2003) Space syntax: a brief introduction to its logic and analytical techniques,
Environment and Behavior, 35 (1): 17-29.
[8] Bagley, M. N., and Mokhtarian P. L. (2002) The impact of residential neighborhood type on travel
behavior: A structural equations modeling approach, The Annals of regional Science, 36:279–297.
[9] Baran, P. K., Rodriguez, D. A., and Khattak, A. J. (2008) Space syntax and walking in a New
Urbanist and suburban neighborhoods, Journal of Urban Design, 13: (1) 5-28.
[10] Barnette, J. (1995) The Fractured Metropolis: Improving the New City, Restoring the Old City,
Reshaping the Region. Harper Collins, New York.
[11] Boehmer, T. K., Hoehner, C. M., wyrwich, K. W., Ramirez, L. K., and Brownson, R. C. (2006)
Correspondence between perceived and observed measures of neigborhood environmental supports
for physical activity. Journal of Physical Activity and Health, 3: 22-36.
[12] Boer, R., Zheng, Y., Overton, A., Ridgeway, G. K., and Cohen, D. A. (2007) Neighborhood
design and walking trips in ten U.S. metropolitan areas, American Journal of Preventive Medicine, 32
(4): 289-304.
[13] Bogart, W. (1998) The Economics of Cities and Suburbs, Prentice Hall: New Jersey.
[14] Bourne L. S. (1996), "Reinventing the Suburb: Old Myths and New Realities", Process in
Planning, 46:3.
[15] Burchell W. D., Listokin, D., Galley C. C. (2000) Smart growth: more than a ghost of urban
policy past, less than a bold new horizon, Housing Policy Debate, 11 (4): 821-879.
[16] Calthorpe P, Richmond H (1992) Land use and transportation planning: Sustainable growth. In:
Green M (ed) Changing America: Blueprints for the new administration. Newmarket Press,
New York, pp 699–710
[17] Calthorpe, P. (1993) The Next American Metropolis: Ecology, Community and the American
Dream, Princeton Architectural Press, New York.
[18] Cerin, E., Leslieb, E., Toit, L. D., Owen, N., and, Frank, L. D. (2007) Destinations that matter:
Associations with walking for transport, HEALTH & PLACE, 13: 713-724.
[19] Cervero, R. and Duncan M. (2003) Walking, bicycling, and urban landscapes: evidence from the
San Francisco Bay Area, American Journal of Public Health, 93 (9): 1478-1483.
[20] Chin, G. K.,Van Niel, K. P., Corti, B. G., and Knuiman, M. (2007) Accessibility and
connectivity in physical activity studies: The impact of missing pedestrian data, Preventive Medicine,
46: 41-45
[21] Crane, R., (2000) The influence of urban form on travel: an interpretive review, Journal of
Planning Literature, 15 (1): 3-23.
[22] DiPietro L. (2001) Physical activity in aging: changes in patterns and their relationship to health
and function. Journal of Gerontology and Biological and Medical Science, 56(spec no. 2):13–22.
[23] Doyle, C., Schwartz, A. K., Schlossberg, M., and Stockard, J. (2006) Active communities:
environments and health, Journal of the American Planning Association, 72 (1): 19-31.
[24] Duany, A. & Z. Plater (1991) Towns and Town-Making Principles, Rizzoli, New York.
[25] Duany, A. and Plater-Zyberk, E. (2000) Suburban Nation: the Rise of Sprawl and the Decline of
the American Dream, North Point Press, New York.
101

MOHAMMED KASHEF
[26] Dutton, J. A. (2000) New American Urbanism: Re-forming the Suburban Metropolis, Skira
Architecture Library, Milano.
[27] Ewing, R. (2004) Can the physical environment determine physical activity levels? Exercise and
Sport Sciences Reviews, 33 (2): 69-75.
[28] Ewing, R. (1996) Best development practices: doing the right thing and making money at the
same time, Planners Press (www.planning.org).
[29] Ewing, R. and Cervero, R. (2002) Travel and the built environment: synthesis, Transportation
research record 1780 (TRB), www.trb.org.
[30] Ewing, R., Handy, S., Brownson, R. C., Clemente, O., and Winston, E. (2006) Identifying and
measuring urban design qualities related to walkability, Journal of Physical Activity and Health, 3
(1): s223-s240.
[31] Gratz R. B. & Norman M. (1998) Cities Back From the Edge: New Life for Down town, John
Wiley & Sons, New York.
[32] Fleissig, W., and Jacobsen, V. (2002) Smart Scorecard for development projects, In
Collaboration with the Congress for New Urbanism and the U.S. Environmental Protection Agency
(http://www.cnu.org/cnu_reports/scorecard_exp.pdf)
[33] Frank, L. D., Schmid, T. L., Sallis, J. F. Chapman, J. and Saelens, B. E. (2005) Linking
objectively measured physical activity with objectively measured urban form: findings from
SMARTRAQ, American Journal of Preventive Medicine, 28 (2S2): 117-125.
[34] Friedman, B., Gordon, S. P. (1994) Effect of Neotraditional Neighborhood Design on Travel
Characteristics, Transportation Research Record 1466, TRB, National Research
Council, Washington, D.C., 63–70.
[35] Fulton W (1996) The new urbanism. Lincoln Institute of Land Policy, Cambridge, MA.
[36]Geus, B. D., Bourdeaudhuij, I. D., Jannes, C., and Meeusen, R. (2008) Psychological and
environmental factors associated with cycling for transport among a working population, HEALTH
EDUCATION RESEARCH, 23 (4) 697-708.
[37] Handy, S. L. (1996) Urban form and pedestrian choices: a study of Austin neighborhoods,
Transportation Research Record 1552, National Research Council, Washington,
D.C., TRB: 135-144,
[38] Handy, S. L., and Clifton, K. J. (2001) Local shopping as a strategy for reducing automobile
travel, Transportation, 28: 317-364.
[39] Handy, S. L., Boarnet, M. G., Ewing, R., and Killingworth, R. E. (2002) How the built
environment affects physical activity: views from urban planning, American journal of Preventive
Medicine, 23(2S): 64-73.
[40] Handy, S. L. (2004) Critical assessment of the literature on the relationships among
transportation, land use, and physical activity. Department of environmental Science and Policy,
University of California, Davis. Prepared for the Committee on Physical Activity, Health,
Transportation, and Land Use.
[41] Handy, S., Paterson, R. G., and Butler, K. (2004) Planning for street connectivity: getting from
here to there, Planning Advisory Service report 515, American Planning Association,
www.planning.org.
[42] Handy, S. L. (2005) Which comes first: the neighborhood or the walking? ACCESS, 26: 16-21.
[43] Handy, S. L. (2006) Questioning assumptions, Planning, January, 2006: 36-37.
[44] Handy, S. L., Cao, X., and Mokhtarian P. (2006) Self-selection in the relationship between the
built environment and walking, Journal of the American Planning association, 72 (1): 55-74.
102

NEIGHBOURHOOD DESIGN AND WALKABILITY
[45] Handy, S. L Cao, X., and Mokhtarian, P. L. (2008) The causal influence of neighborhood design
on physical activity within the neighborhood: evidence from Northern California, American Journal
of Health Promotion, 22 (5): 350-358
[46] Hillier, B. and Penn, A. (2004) Rejoinder to Carlo Ratti, Environment and Planning B; Planning
and Design, 31: 501-511.
[47] Hiller, B., and Netto, V. (2002) Society seen through he prism of space: outline of a theory of
society and space, URBAN DESIGN International, 7: 181-203.
[48] Hillier, B. (2002) A theory of the city as object: or, how spatial laws mediate the social
construction of urban space, URBAN DESIGN International, 7: 153-179.
[49] Hillier, B. (1999) The hidden geometry of deformed grids, Environment and Planning B, 26:
169-191.
[50] Hiller, B. (1997) Cities as movement economies, in: P. Droedge (Ed.) Intelligent Environments,
Elsevier, Amsterdam.
[51] Hillier B. (1996) Space is the Machine: A Configurational Theory of Architecture, Cambridge
University Press, Cambridge.
[52] Hillier B. and J. Hanson (1984) The Social logic of Space, Cambridge University Press, London
[53] Jacobs, J. (1961) The Death and Life of Great American Cities, Random House, Inc. New York
[54] Jiang, B. and Claramunt, C. (2002) Integration of space syntax into GIS: new perspectives for
urban morphology, Transactions in GIS, 6 (3): 295-309.
[55] Kashef, M. (2008) Architects and planners approaches to urban form and design in the Toronto
region: a comparative analysis, Geoforum 39 (1): 414-37
[56] Katz P. (1994) The New Urbanism: Toward an Architecture of Community, MacGraw-Hill, New
York
[57] Kelbaugh D. (1997) Common Place: Toward Neighbourhood and Regional Design, University
of Washington Press, Washington.
[58] Krier, R. (1979) Urban Space, Rizzoli, New York
[59] Krizek, K. J. (2003) Operationaliz.ing neighborhood accessibility for land use-travel behavior
research and modeling. Journal of Planning Education and Research, 22(3): 270-28.
[60] Langdon, P. (2004) Street connectivity: emerging as a new municipal goal, New Urban News,
New Urban Publications, www.newurbannews.com, 21-24.
[61] Lee, C. and Moudon, V. (2006) Correlates of walking for transportation or recreation Purposes,
Journal of Physical Activity and Health, 3 (Suppi 1): S77-S98.
[62] Lesliea, E., saelensb, B., Frankc, L., Owen, N., Baumand, A., Coffeee, N., and Hugo, G. (2005)
Residents’ perceptions of walkability attributes in objectively different neighborhoods: a pilot study,
Health & Place, 11: 227-236.
[63] Leyden, K. M. (2003) Social capital and the built environment: the importance of walkable
neigborhoods, American Journal of Public Health, 93 (9) 1546-1551.
[64] Li, F., Fisher, J. K., Brownson, R. C., Bosworth, M (2005) Multilevel modeling of built
environment characteristics related to neighbourhood walking activity in older adults, Journal of
Epidemiology and Community Health, 59: 558-564.
[65] Lovasi, G. S., Moudon, A. V., Pearson, A. L., Hurviz, P. M., Larson, E. B., Siscovick D. S.,
Berke, E. M., Lumley, T. and Psaty, B. M. (2008) Using built environment characteristics to predict
walking for exercise, International Journal of Health Geographics, 7 (10).
[66] LUTAQH Project (2006), Central Puget Sound, www.metrokec.gov.
[67] McCormack, G. R., Cerin, E., Leslie, E., Du Toit, L., and Owen, N. (2008) Objective versus
perceived walking distances to destinations: correspondence and predictive validity, Environment and
Behavior, 40 (3): 401-425.
103

MOHAMMED KASHEF
[68] Michael, Y. L., Green, M. K., and Farquhar, S. A. (2006) Neigborhood design and active aging,
HEALTH & PLACE, 12: 734-740.
[69] Moudon, A. V., Dr. es Sc.; Lee, C., Cheadle, A. D., Garvin C., Johnson, D. B., Schmid, T. L.,
Weathers, R. D. (2007) Attributes of environments supporting walking, American Journal of Health
Promotion, 21 (5): 448-459.
[70] Nagel, C. L., Carlson, N. E., Bosworth, M., and Michael, Y. L. (2008) The relation between
neigborhood built environment and walking activity among older adults, American Journal of
epidemiology Advance Access.
[71] Nophaket, N., and Fujii, A. (2004) Syntactic and network pattern structures of city: comparison
of grid and meandering street patterns in Kyojima and Honjo, Journal of Asian Architecture and
Building Engineering, 349-356.
[72] Oaks, M. J., Forsyth, A., Shmitz, K. H. (2007) The effects of neigborhood density and street
connectivity on walking behaviour: the twin cities walking study, Epidemiologic perspectives &
Innovation, 4 (16).
[73] O'Sullivan, A. (1996) Urban Economics, Irwin, Chicago.
[74] Peponis, J., Bafna, S., Bajaj, R., Bromberg, J., Congdon, C., Rashid, M., Warmels, S., Zhang, Y.
and Zimring, C. (2007) Designing space to support knowledge work, Environment and Behavior, 38:
815-840.
[75] Pikora, T., Corti, B. G., Knuiman, M., Niel, K. V., Bulsara, M., Bull, F., Shilton, T. (2007) The
impact of urban design on walking and cycling: the RESIDE project, Planning and Transportation
research Center (PATREC).
[76] Raford, N., and Ragland, D. R. (2003) Space syntax: an innovative pedestrian volume modeling
tool for pedestrian safety, Institute of Transportation Studies, UC Berkeley traffic Safety Center, UC
Berkeley.
[77] Rajmani, J., Bhat, C. R., Handy, S., Knaap, G., Song, Y. (2003) Assessing the impact of urban
from measures in nonwork trip mode choice after controlling for demographic and level-of-service
effects, Transportation Research Board (TRB), Annual Meeting CD-ROM.
[78] Ratti, C. (2004) Space syntax: some inconsistencies, Environment and Planning B: Planning and
Design, 31: 487-499.
[79] Reed, J. A., Wilson, D. K., Ainsworth, B. E., Bowels, H., and Mixon, G. (2006) Perceptions of
neigborhood sidewalks on walking and physical activity patterns in a south eastern community in the
US, Journal of Physical Activity and Health, 3: 242-253.
[80] Rhodes R. E., Courneya K. S., Blanchard C. M., and Plotnikoff R. C. (2007) Prediction of
leisure-time walking: and integration of social cognitive, perceived environmental, and personality
factors, International Journal of Behavioral Nutrition and Physical Activity, 4 (51).
[81] Rowe C. and Koetter, F. (1978) Collage City, MIT, Cambridge
[82] Sallis J. F., Melbourne H, and Hofstetter C. R., et al (1990) Distance between homes and
exercise facilities related to frequency of exercise among San Diego residents. Public Health
Rep, 105:179-185.
[83] Saelens, B. E., Sallis, J. F., and Frank, L. D. (2003) Environmental correlates of walking and
cycling: findings from the transportation, urban design, and planning literatures, Environment and
Physical Activity, 25: 80-91.
[84] Saelens, B. E., Sallis, J. F., Black, J. B., and Chen D. (2003) Neighborhood-based differences in
physical activity: and environment scale evaluation, American Journal of Public Health, 93 (9):
1552-11558.
[85] Schmidt, S. J., and Wells, J. S. (2005) Transit village monitoring research: connectivity
measures, Alan M. Voorhees Transportation Center (VTC), Rutgers.
104

NEIGHBOURHOOD DESIGN AND WALKABILITY
[86] Schwanen, T., and Mokhtarian, P. L. (2007) Attitudes towards travel and land use and choice of
residential neigborhood type: evidence from the San Francisco Bay Area, Housing Policy Debate, 2
(1): 171-207.
[87] SMARTRAQ Project (2204), Atlanta, www.smartraq.net.
[88] Southworth, M (2005) Designing the walkable city, Journal of Urban Planning and
development, 131 (4): 246-257.
[89] Talen, A (1999) Sense of community and neighborhood form: an assessment of the social
doctrine of New Urbanism, Urban Studies, 36 (8): 1379-99.
[90] Trancik, R. (1986) Finding Lost Space: Theories of Urban Design, Van Nostrand Reinhold, New
York.
[91] Wells, N. M., and Yang, Y. (2008) Neighborhood design and walking: a quasi-experimental
longitudinal study, American Journal of Preventive medicine, 34 4): 313-319
[92] Wood, L., Shannon, T., Bulsara, M., Pikora, T, McCormiack, G., and Corti, B. G. (2007) The
anatomy of the safe and social suburb: an exploratory study of the built environment, social capital
and residents’ perceptions of safety, HEALTH & PLACE, 14: 15-31
105

106

AHU J. of Engineering & Applied Sciences 3 (1) : 107-118 (2010)
© 2009 ALHOSN University
QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
P. Rengaraju, C-H. Lung 1 , A. Srinivasan 2 , R.H.M. Hafez 1*
1 Department of Systems and Computer Engineering, Carleton University, Ottawa, Ontario, Canada
2 EION Inc. Ottawa, Ontario, Canada
ABSTRACT: In the last few years the usage of multimedia applications and VoIP using mobiles over the
internet has drastically increased. Providing Quality of Service (QoS) for the end user applications has become
an important consideration for the upcoming fourth generation WiMAX and LTE networks. Research is
continuing in Physical (PHY) layer, Medium Access Control (MAC) layer and cross layers in order to improve
the QoS for both WiMAX and 3GPP-LTE networks. This paper discusses fundamental issues in QoS
improvements in WiMAX networks in each layer.
Key words: IEEE 802.16; WiMAX; QoS; QoE
1. INTRODUCTION
WIMAX stands for Worldwide Interoperability for Microwave Access. It is the technology
aimed to provide broadband wireless data access over long distances [5]. It is based on
Institute of Electrical and Electronics Engineers (IEEE) 802.16 standards [1], [2] and the
standard defines the PHY and MAC layer functionalities. The technology provides basic
Internet Protocol (IP) connectivity to the user and the communication is connection oriented.
The initial IEEE 802.16d standard is specific to fixed and nomadic users. Based on increasing
mobile users requirements and to increase the coverage region, the standard released the IEEE
802.16e and IEEE 802.16j [2] versions. Now the IEEE 802.16m [3], [4] task group are
working to support the International Mobile Telecommunications (IMT) - Advanced
requirements. One of the most controversial requirements of IEEE 802.16m, during the
development of its system requirement document [4] was full backwards compatibility and
interoperability with the legacy systems.
The MAC and PHY functions can be classified into three categories, namely data plane,
control plane and management plane. The data plane comprises functions in the data
processing path such as header compression as well as MAC and PHY data packet processing
functions. A set of L2 control functions is needed to support various radio resource
configuration, coordination, signaling and management. This set of functions is collectively
referred to as control plane functions. The individual data and control plane functionalities of
PHY and MAC layer is shown in Fig.1. A management plane is also defined for external
management and system configuration. Therefore, all management entities fall into the
management plane category [5]. Some of the MAC layer functional blocks are discussed in
section III, MAC layer QoS. improvements. Relay functions and advanced air interface
support functions are specific to 802.16j and 802.16m standard.
________________________
* Corresponding Author.
E-mail : hafez@sce.carlton.ca
107

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
The different types of PHY carriers are Single Carrier (SC) for 11 to 66GHz, Single Carrier
access (SCa) for less than 11GHz, Orthogonal Frequency Division Multiplexing (OFDM) and
Orthogonal Frequency Division Multiple Access (OFDMA). Most of the IEEE 802.16e based
nodes, IEEE802.16j and IEEE 802.16m nodes uses only OFDMA. The QoS improvements on
OFDMA carriers are sub channelization, Multiple Input Multiple Output (MIMO), Adaptive
Antenna System (AAS) and Space Time Coding etc.
The term QoS refers to the probability of the telecommunication network meeting a given
traffic contract. The parameters that can be used to measure it quantitatively are throughput,
transmission delay or packet delay, delay jitter, percentage of packets lost etc. On top of this,
the International Telecommunication Union (ITU) recently released the Quality of Experience
(QoE) framework ITU-T (2004) in which an explicit distinction has been made between QoS
and QoE. The QoS expresses the ‘degree of objective service performance’ and QoE the
‘overall acceptability of an application or service, as perceived subjectively by the end user’.
From the last few years, many researches are on going to increase the capacity and improve
the QoS of the network [5]. This paper analyses the QoS improvements in each layer and the
cross layer.
The rest of the paper is organized as follows. Section II and III surveys the QoS
improvements in PHY and MAC layer. Section IV surveys end-to-end (E2E) QoS and cross
layer improvements. The last section V is the conclusion of the Paper.
2. PHY LAYER QOS IMPROVEMENTS
There are two ways to improve the PHY layer QoS; 1. Increase the spectral efficiency of
available spectrum by advanced OFDMA and MIMO techniques. 2. Re-use of available
frequency spectrum to increase the capacity. OFDMA demonstrates superior performance in
non-line-of-sight (N-LOS) multi-path channels with its relatively simple transceiver structures
and allows efficient use of the available spectrum resources by time and frequency sub
channelization. On the other hand MIMO technology multiplies throughput without requiring
additional frequency bandwidth, enhances link reliability through spatial diversity and
enlarges the coverage area by increasing the transmission range [5]. The Partial Frequency
Reuse (PFR) [11], [12] and Full Frequency Reuse (FFR) [10] can be applied to WiMAX
network to increase the capacity of the network.
108

QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
MIMO Techniques
Examples of MIMO techniques include single-user MIMO (SU-MIMO), multiuser MIMO
(MU-MIMO), and cooperative relay [7]. Multiuser interference becomes a bottleneck for
increasing multiuser MIMO throughput at high SNRs or for a relatively larger number of
multiuser MIMO users. This places a more stringent requirement on channel state information
(CSI) accuracy for multiuser MIMO than that for point-to-point MIMO. To acquire highlyaccurate
CSI without incurring excessive overhead, limited feedback can be integrated with
multiuser diversity. MIMO schemes can be broadly divided into two categories; the ones
which have and use CSI at the transmitter which is called closed-loop (i.e., beam forming
and/or pre-coding) and the other without CSI at the transmitter (i.e. space-time coding (STC)
and spatial rate multiplexing (SM)) which is called open-loop [6].
Space–time Coding [6]: This technique is used to transmit multiple copies of a data stream
across a number of antennas and to exploit the various received versions of the data to
improve the reliability of data-transfer. It combines all the copies of the received signal in an
optimal way to extract maximum information as possible from each of them and does not
require CSI, but extracts transmit diversity. Multiuser diversity refers to the degrees of
freedom due to independent fading in different users’ channels [5]..
Spatial Multiplexing [6]: It multiplexes more than one symbol per channel use. The number
of symbols transmitted per symbol use determines the spatial rate. This scheme allows
transmission of multiple independent data streams on a set of transmit antennas without any
CSI. However, the maximum number of streams that can be transmitted is limited by (Nr, Nt),
where Nr is the number of receive antennas, Nt is the number of transmit antennas. In
WiMAX, the spatial rate
Beam forming [6]: This scheme uses CSIT to put energy only into the Eigen modes of the
channel. It extracts both diversity and array gains. Since this scheme requires CSIT, it should
be ensured that the delay between the CSIT estimates and its usage be smaller compared to
the quasi-static interval of the channel (i.e., the period over which the channel remains
constant).
A comparison between the STC Alamouti, SM and SIMO is simulated in [6] for the same
spectral efficiency. At 1% block error rate (BLER) the STC scheme has a SNR advantage of
approximately 1.0 dB and the spatial multiplexing has an advantage of approximately 1.5 dB
compared to SIMO. In [8], the MIMO schemes are simulated for equal mean packet data call
duration (EMPC-D) and equal mean packet data volume (EMPC-V). The spectral efficiency
(h) and throughput performance for pedestrian B–3 model are given in the table 1.
109

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
MIMO for 802.16j and 802.16m: A more efficient solution to extend coverage and enhance
throughput in 802.16j and 802.16m is to deploy low-cost RSs in the network [5]. In relaybased
wireless network, the transmission has two phases, namely from source to relay (SR)
and from relay to final destination (RD). In this case the access RS or end user may receive
the signal either an opportunistic fashion or from more than one superior RS in co-operative
fashion [7].
Frequency reuse
The PFR takes the advantage of sub-channelization from OFDMA where FFR takes the
advantage of cell saccharisation.
PFR scheme: The key issue in PFR is how efficiently the sub-channel segmentation and the
sub-channel are reused. In PFR [11], the Cell Edge Users (CEU) operate on a sub-frame zone
with a fraction of all available sub-channels, while the inner cell users operate on a zone with
all sub-channels available. Usually, CEU are operated with frequency reuse=3 (called the R3)
zone and the cell center users (CCU) are operated in a R1 sub-frame zone. The transmission
frame structure with time and frequency resource allocations for the R1 and R3 zones is
shown in Fig.2. All the cells/sectors operated in the network have common boundaries. In
PFR [11] scheme, the interference margins at the cell edge are effectively controlled for larger
R3 zone sizes. In [12], there are five different algorithms (with or without frequency reuse)
that are compared with either uniform Fixed Channel Allocation (FCA) or non uniform FCA.
The main principle in the channel allocation in BS is the channel borrowing from CEUs to
CCUs and from lower priority services to higher priority services of CEUs and CCUs. From
their results [12], the performance of uniform-FCA scheme is much worse than non uniform-
FCA.
Figure 2 PFR-Transmission Frame [11]
FFR scheme: In cell saccharisation, the BS, Omni directional antenna is replaced by several
directional antennas each radiating within a specified sector. With hexagonal cell patterns, the
most common sectoring schemes employ 120° or 60° sectors. Splitting each cell into three
sectors multiplies the cell capacity by three, but since a separate channel is allocated to each
sector, three channels are needed to serve the entire cell and the spectral efficiency remains
unchanged. However, in practice, the radiation diagrams of the sector antennas used are
exceeding the sector boundaries and there is a significant overlap region between adjacent
sectors, which prohibits FFR.
110

QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
Figure 3- Cell Sectoring 120-deg hexagonal and 90-deg circle [10]
One way to handle this problem is to reserve a group of subcarriers to serve users located in
overlap regions, those are called CEU. The remaining subcarriers are reserved for the nonoverlap
regions and they are reused in all sectors. Since only one part of the subcarriers is
reused, this scheme ensures PFR. The FFR technique uses the existing MIMO concepts to
serve the users located in the overlap regions. More specifically, while two users located in
the overlap region of two sectors, say A and B, are served by these two sectors, the third
sector C serves a user located in its overlap free region Thus, three users are simultaneously
used and full frequency reuse is achieved. In [10], each cell is partitioned in to four sectors
with different reuse (half rate code and full rate code) is shown in Fig.3. For interference
reasons, 120° sectors are commonly used in hexagonal cell patterns. However, the number of
sectors per cell does not have to follow any rule when all of the sectors use the same
frequency.
3. MAC LAYER QOS IMPROVEMENTS
Even though the spectrum efficiency and throughput capacity is improved by PHY layer,
MAC layer plays the important role to improve the end user QoS/QoE. The standard [1, 2, 3
and 4] defines only the functionalities of the MAC functions but the algorithm
implementation in each block is vendor specific. The recent researches pay more attention to
some of the blocks like scheduler, Radio Resource Management (RRM), Handover and multichannel
HARQ IEEE 802.16 standards defines five different service classes and its QoS
parameters, which are Unsolicited Grant Service (UGS), Extended real time Polling Service
(E-rtPS), real time Polling Service (rtPS), non-real time Polling Service (nrtPS) and Best
Effort (BE) services. Any end user application follows any one of the above five service
classes. The MAC control function algorithm considers the type of service and its QoS
parameters.
A. Radio Resource Management (RRM)
One of the main functions of the RRM is the Call Admission Control (CAC). Possible
input parameters for the CAC module are: the connection type (downlink/uplink and
new/handover), service class and QoS. The CAC implemented in [13, 14, 15, 16 and 17] are
111

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
considered for our analysis. Performance metrics of the CAC algorithm is Call Blocking
Probability (CBP), Call Dropping Probability (DP) and Bandwidth Utilization (BU) [15].
The simple Quadra-Threshold Bandwidth Reservation (QTBR) CAC scheme is introduced
in [16], which uses the different threshold for different service class. In that real time
applications are more prioritized. An adaptive CAC implemented in [17] used the minimum
bandwidth requirement Bmin and the maximum bandwidth requirement Bmax input
parameters and a simple Dynamic Guard Channel scheme. In this scheme the bandwidth is
reserved for handoffs by defining two levels of threshold, thmin and thmax, where thmin <
thmax < B (total bandwidth). In [14], Adaptive Fuzzy-Based Admission Controller is used
which considers the buffer occupancy based algorithm and the connection type. Token Bucket
based CAC scheme is used in [15] and they compare the performance of existing handover
prioritized CAC algorithm and Service class specific CAC algorithm. Their algorithm follows
token bucket algorithm equation (1). Higher priority is given to the handover calls.
Where, f = Duration of the frame in msec. CrtPS and CNRT = BW allocated to rtPS and
nrtPS connections. ri and bi = Token arrival rate and bucket size in Kbps. mi = d/f, mi
(integer); d = Maximum latency for rtPS in msec.
In [12], the CAC considers the E2E delay as one the important parameter for call admission.
Since the current network is based on multi-hop communication, E2E delay should be
considered for any new CAC algorithm.
Mobility management
It scans neighbour BSs and decides whether the MS should perform a handover operation.
In general, the handover can be broadly classified in to hard handover or soft handover. The
handover may occur with the same WiMAX network, called as horizontal handoff or it may
occur with different network, called as vertical handoff. The QoS parameters considered
during the handover are handover latency and packet drop.
IEEE 802.16e [1] defines 3 types of handover, the hard Handover scheme which may be
applied in the case of a low speed and it uses the mechanism break-before-make. Other two
types are Soft Handover, Macro Diversity Handover (MDHO) and Fast Base Station
Switching (FBSS) which may be applied in the case of a high speed and use the make-beforebreak.
In our study, the Mobile IP (MIP) based handover performs well to provide more
security and less delay [37]. In FBSS MS maintain the group of BSS nearer to it and select
one of them as an Anchor BS among BSS, and communicates all traffic with only the Anchor
BS. The mobile has the right to change Anchor BS when it wants as long as it is connected
with. But these soft handover can be applied only when the cell size is small to maintain more
than two BSS in the MS diversity set or relay networks to maintain more than two relays in
the MS diversity set.
Handoff with other wireless networks: The fourth generation network is based on All-IP
and the emergence of multiple standards, such as WiMAX, 3GPP (LTE), Wi-Fi, Mobile-Fi
(IEEE 802.20) and IEEE 802.15, it has become necessary to define a new interworking
standard IEEE 802.21, which is being designed to facilitate handover among any wireless
access technologies. IEEE 802.21 defines the Media Independent Handover (MIH) functions
which facilitates messaging among IP and the various wireless link layer technologies, to
select an appropriate network for handover and to set up new wireless links. In MIH, QoS at
the IP layer can be interworked with QoS at the wireless link layers to provide the QoS
112

QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
required by the user application. Fig 4 shows the MIH function protocol stack. The three type
of messages used are event notification from the same layer or other layer like signal
threshold command service for processing the handover with any network and finally the
information service to provide information to other layers.
Figure 4- MIH Functions (Protocol Stack) [25]
In [23] network initiated handover based on IEEE 802.21 is simulated. Both the Mobile
Terminal Initiated Handover (MIHO) and Network Initiated Handover (NIHO) are simulated
in [24]. The comparative analysis of alternative approaches that have been developed by IEEE,
3GPP and others for handover of PSTN and Internet sessions among multiple wireless
technologies including UMTS, WiFi, WiMAX, Mobile-Fi and WPANs discussed in [25]. The
software architecture which supports the seamless handover between WiMAX and UMTS
network is discussed in [30].
Scheduler
It is one of the major blocks to ensure the QoS and it performs the scheduling of MAC
control and data packets. The WiMAX schedulers can be classified into two main categories:
channel-unaware schedulers and channel-aware schedulers. The channel-unaware schedulers
are quite simple for design, which assumes that the channel is error free for packet
transmissions. Basically, the channel-unaware scheduler makes the scheduling decision
without any channel state information. However, in wireless environment, there is a high
variability of radio link such as channel attenuation, fading and noise interference. The
channel-aware schedulers, on the other hand, consider the channel state information while
scheduling the packet.
Since there is no single homogenous scheduling algorithm meets all the QoS requirements
for different service classes, researchers are trying to find hybrid scheduling algorithms to
satisfy QoS for different service classes. The hybrid scheduling algorithm combines more
than one legacy homogeneous schedulers to satisfy the QoS. Three tiers scheduling also
proposed in [20], but the complexity is increased.
Channel-aware schedulers are kind of an opportunistic schedulers, whereby each MS is
assigned a priority based on its channel quality and service status. In [18], some of the
channel aware schedulers are analyzed for different QoS objective functions. The cross layer
scheduling is analyzed in the next section.
Hybrid automatic repeat-request (HARQ)
It is critical for IEEE 802.16e OFDMA network, as it can significantly improve the
reliability of wireless link. The IEEE 802.16e [1] standard defines the basic HARQ support.
Since the data transmission is frame based and different OFDMA sub-channel are allocated
for different users, multi-channel HARQ scheme is introduced to improve the performance. In
113

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
order to adapt to various link conditions and to efficiently utilize available bandwidth, in the
present design of multi-Channel (or N-process) HARQ operation, the MAC Service Data Unit
(MSDU) as well as the PHY Service Data Unit (PSDU) is often divided into several
fragments. With Multichannel HARQ, the transmission and retransmission of the fragments
are taken care of in multiple HARQ channels that run in parallel as shown in Fig.5
Figure 5- Multi-channel HARQ [26]
IEEE 802.16m draft versions support the multi-channel HARQ for the IMT advanced
WiMAX network. The maximum number of parallel HARQ channels defined in the current
system will render a potential bottleneck and preponderantly limit the overall system capacity
of next generation WiMAX network, wherein the wireless link is anticipated to support much
higher data rate. The shared multichannel HARQ for IEEE 802.16j and IEEE 802.16m
networks is an on going research. In order to enhance the throughput performance and
improve the preparedness of multichannel HARQ protocol for adoption by future 802.16j and
802.16m system, adaptive fragmentation and ACID field expansions are introduced in [26].
Network coding
HARQ is adopted to transmit data packets reliably. However, it sacrifices resilience in time
varying channels, and it may under-utilize the wireless medium in the cases of multipath and
multi-hop transmissions. On the other hand, random wireless network coding (NC) has been
shown to be effective towards improving throughput in multi-hop wireless networks. NC can
be implemented with multi-channel HARQ to improve the performance of transmission
efficiency, which is called as shared multi-channel HARQ.
Header Compression [27]
Robust Header Compression (ROHC), a standard header compression scheme proposed by
IETF, Bidirectional Optimistic-O mode provides higher compression efficiency and
robustness compared with Payload Header Suppression (PHS) which is an optional header
compression scheme for mobile WiMAX. However, if consecutive packet losses occur
because of a Shadow Region of Mobile WiMAX, header decompression failure rate of the
ROHC-O mode increases. Therefore, a complementary mechanism is proposed in [27] using
the dynamic adjustment of an optimistic parameter, which has an effect on both the
compression efficiency and the robustness, for improving the performance of ROHC - O
mode.
114

QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
Framing
The WiMAX uplink frame contains the random access channel for ranging and bandwidth
request for BE service, Fast feedback channel to inform the channel conditions and user data
channels. A proper frame allocation also leads to effective bandwidth usage of the system and
some of it concentrates on the proper frame allocations to improve the QoS. To improve the
system access success probability a mathematical basis for an accurate model is proposed in
[28] ¸ that can be efficiently used to determine the resource allocation for each class. It also
provides a useful analytical tool for contention systems that utilize random access for network
entry, resource acquisition, and/or contention-based data transmission
4. E2E AND CROSS LAYER QOS IMPROVEMENTS
To assure the QoS/QoE for end user application it is not necessary to satisfy the QoS of
access network alone, the backhaul traffic in service provider’s connectivity network should
satisfy the necessary QoS requirements. Here the E2E network considered for the analyses are;
Homogeneous network and heterogeneous network.
A. Homogeneous E2E QoS
IEEE 802.16 networks provide mechanisms for QoS support at MAC level, but E2E QoS
issues are not addressed in detail in the standards. Some kind of cross-layer work is needed to
satisfy both the network and MAC layer QoS. Number of enhancements have been proposed
to enable the offering of different levels of QoS in IP networks including the integrated
services (IntServ) architecture, and the differentiated service (DiffServ) architecture [29].
IntServ is implemented by four components: the signalling protocol (e.g. RSVP), the
admission control, the classifier and the packet scheduler. Furthermore, some rules are
prescribed to classify DiffServ IP packets into different priority queues based on QoS
indication bits in IP header. Therefore, the QoS architecture of WiMAX access network can
support both IntServ and DiffServ. So the mapping rules are created between WiMAX MAC
layer and IP layer.
In traditional way, RSVP signalling message can be classified into a special high priority
queue, and be transmitted in the second management connection. So the E2E resource is
reserved for each connection to provide the network QoS. In the cross-layer approach, QoS
mapping is created and the WiMAX MAC connection request message, Dynamic Service
Addition (DSA) request carries the RSVP-PATH specific message. If the BS supports both
MAC and network specific QoS, accepts the connection and informs it in the DSA response
message. Since the communication in WiMAX network is connection-oriented; the
application must establish the connection with the BS before data transmission. The message
exchange for DSA and DSC can be deployed to carry QoS parameters of IntServ services for
E2E resource reservation. For DiffServ services, on the other hand, a number of per-hop
behaviors (PHBs) for different classes of aggregated traffic can be mapped into different
connections directly.
B. Heterogeneous E2E QoS
Users shall be able to access its services through different access technologies, such as
WLAN, WiMAX, UMTS and DVB technologies, from the same or different network
operators, and to seamless move between different networks with active communications [31].
The recent standardization efforts in the IETF for a new extensible IP signalling protocol suite
(NSIS) describes the network layer signalling to provide network QoS. So the recent
115

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
researches are trying to integrate both WiMAX E2E architecture defined by the WiMAX
forum and NSIS signalling architecture. The QoS NSIS Signalling Layer Protocol (QoS-
NSLP) is used to signal the application QoS requirements and request the resource allocation
along the full path of data flows. The NSIS protocol supports the interoperability between
different QoS-enabled domains defining distinct QoS-Models (QOSM) depending on the
underlying network technologies. The QoS-Model defines the QoS parameters, the traffic
descriptors and the methods to provide the desired QoS through the Resource Management
Function (RMF) specification for the NSIS nodes of a specific domain. The QoS parameters
and the traffic descriptors are encapsulated in an object called QSPEC and are interpreted by
NSIS nodes according to the QOSM. Two different types of QSPECs are defined: Initiator
and Local QSPEC. The former is created by the QoS NSIS Initiator (QNI) and the later is
created at the edge node between two QoS enabled domains through a coherent mapping of
the Initiator QSPEC parameters.
Other cross layer improvements
In multiuser scenario, channel gains of a specific subcarrier vary from one wireless terminal
to another wireless terminal. At any given time in a large network for each sub-channel there
is a high probability that there is a user whose OFDMA subcarrier assigned to the users who
see good channel gains on them. Maximization of system performance can be achieved by
allocating the OFDMA subcarriers to the users with good channel condition. The performance
of the MAC functions such as CAC, sub-channel allocation, schedulers and HARQ modules
depends more on the PHY channel condition.
In [33], priority function is defined for each connection admitted in the system, and BS
update it dynamically depending on the wireless channel quality, QoS satisfaction and service
priority across layers. The connection with the highest priority is scheduled each time. The
occupation of physical resource is an important feedback for the CAC to admit the new
connection. In [34] the cross layer based CAC and sub-channel assignment is simulated,
which takes the PHY layer channel information. In [35], channel aware WF2Q+ based
scheduling for point-to-multipoint (PMP) network is simulated, where the weights of the
service class is dynamically varied based on the channel condition. On the other hand, Blind
channel aware schedulers leads to fairness issue since those schedulers result in optimal
network resource utilization, it does not take into account each user’s QoS perspective. To
avoid this problem utility function is used to strike a balance between system capacity and
fairness among users is called proportional fairness (PF). In [32] channel aware PF schedulers
are simulated.
The new WiMAX networks are based on MIMO and RS nodes. So the diversity and multihop
communication should be considered for the upcoming researches. In [7], the scheduling
and radio resource allocations performed at BS is based on the PHY channel conditions. The
BS uses a look-up table developed for point-to-point flat fading links with a given SNR
conditions. For each sub-channel and for each user, the BS calculates the post processing
SNR with relays and calculates the E2E throughput for each cooperative diversity
communication. Finally it selects the best one and uses the same for data transmission.
Based on wireless channel condition, adaptive fragmentation is applied for multi-channel
HARQ to improve the performance of normal multi-channel HARQ [26].
In adaptive random network coding [36], the channel state information can be used to
dynamically construct the packets that could be better matched to the fluctuating channel
conditions, so as to obtain a higher throughput.
116

QOS IMPROVEMENTS IN MOBILE WIMAX NETWORKS
5. CONCLUSION
This paper provides a comprehensive discussion of recent works in WiMAX networks to
support the required QoS for multimedia applications. In traditional communication networks,
the open systems interconnection (OSI) layer architecture has been widely adopted, and the
performance optimization is conducted largely within each individual protocol layer. This
may work well for the wired and fixed wireless networks. On the other hand due to high
mobility and high data rate support nature of 4G networks, the conventional OSI architecture
may not work well to support the real time applications. This leads to the Cross-layer design
approaches for efficient utilization of the scarce radio resources with QoS provisioning. At the
same time Cross-layer design approaches are critical in nature. This paper summarized in one
place most of the efforts being conducted in the cross-layer designs and QoS provisioning in
4G networks.
REFERENCES
[1] IEEE 802.16 -2009 “IEEE Standard for Local and metropolitan area networks Part 16: Air
Interface for Broadband Wireless Access Systems”. IEEE press 2009.
[2] IEEE 802.16j -2009 “IEEE Standard for Local and metropolitan area networks Part 16: Air
Interface for Broadband Wireless Access Systems Amendment 1: Multiple Relay Specification”.
IEEE press 2009.
[3] IEEE 802.16m TGm “IEEE 802.16m System Description Document, 80216m-08-0034r1”, IEEE
press 2009.
[4] IEEE 802.16m TGm “IEEE 802.16m System Requirements Document, 80216m-07_002r9”,
IEEE press 2009.
[5] M.D. Katz, F.H.P. Fitzek;"WiMAX Evolution - Emerging Technologies and Applications", A
John Wiley and Sons, Ltd, Publication.
[6] S.P. Alex and L.M.A. Jalloul, “Performance Evaluation of MIMO in IEEE802.16e/WiMAX”,
IEEE journl of selected topics in signal processing, vol. 2, 2008, pp 181-190.
[7] Q. Li, X. Lin, J.Zhang and W.Roh, “Advancement of MIMO technology in WiMAX- from IEEE
802.16d/e/j to 16m” IEEE Comm. Mag. vol. 47, issue 6, 2009, pp 100-107.
[8] C.F.Ball, E.Humburg, S.Eder and L.Lacinak, “WiMax Capacity Enhancements introducing
MIMO 2”, 16th IST mob. and wireless comm. summit, 2007, pp 1-5.
[9] B.Can, H.Yanikomeroglu, F.A Onat, E.De Carvalho and H.Yomo, “Efficient Cooperative
Diversity Schemes and Radio Resource Allocation for IEEE 802.16j”, Wireless Comm. and Net.
Conf. 2008, pp 36-41.
[10] H.Sari and S.Sezginer, “MIMO Techniques and Full Frequency Reuse in mobile wimax network”,
Int’l Conf. on Telecomm., 2009, pp 8-12.
[11] M.Hunukumbure, B.Upase, L.Sarperi, and S.Vadgama, “Advanced Techniques for Improving the
QoS of the WiMAX Cell Edge User”, IEEE GLOBECOM, 2008, ECP-1082, pp 1-5.
[12] Y. Qi, X. Zhong and J. Wang, “A Downlink Radio Resource Allocation Algorithm with
Fractional Frequency Reuse and Guaranteed Diverse QoS for Multi-Cell WiMAX System”, 3rd
Int’l Conf. on Comm. And Net., 2008, pp 289-294.
[13] S.Lee, G.Narlikar, M.Pal, G.Wilfong and L.Zhang, “Admission Control for Multihop Wireless
Backhaul Networks with QoS Support”, IEEE WCNC 2006, pp 92-97.
[14] S.Ghazal, Y.H Aoul, J.Ben Othman and F.Nait-Abdesselam, “Applying a Self-Configuring
Admission Control Algorithm in a New QoS Architecture for IEEE 802.16 Networks”, IEEE
Symp. on Comp. and Comm., 2008, pp 1023-1028.
[15] S.Kalikivayi, I.S Misra, and K.Saha, “Bandwidth and Delay Guaranteed Call Admission Control
Scheme for QOS Provisioning in IEEE 802.16e Mobile WiMAX”, IEEE GLOBECOM 2008,
ECP-246, pp 1-6.
117

P. RENGARAJU, C.H. LUNG, A. SRINIVASAN, R.H.M. HAFEZ
[16] H.Y Tung, K.F Tsang, L.T Lee and K.T Ko, “QoS for Mobile WiMAX Networks Call
Admission Control and Bandwidth Allocation”, 5th IEEE Consumer Comm. and Net. Conf. 2008,
pp 576-580.
[17] B.Shafaq and K.Ratan, “Adaptive Connection Admission Control and Packet Scheduling for QoS
Provisioning in Mobile WiMAX”, IEEE Int’l Conf. on signal proc.and Comm., 2007, pp 24-27.
[18] So-In.Chakchai, R.Jain and A-K.Tamimi, “Scheduling in IEEE 802.16e Mobile WiMAX
Networks: Key Issues and a Survey” IEEE Journl on Selected Areas in Comm., 2009, pp. 156 -
171.
[19] R.Perumalraja, C-H. Lung and S.Anand " QoS Assured simple Hybrid Uplink Scheduler for
WiMAX networks", Paper submitted in Int’l Conf. on Comm. 2010.
[20] M. Maode, L. Jinchang, S.K. Bose and C-N. Boon, “A Three-tier Framework and Scheduling to
Support QoS Service in WiMAX”, 6 th Int’l Conf. on Info., Comm. and Signal Proc., 2007, pp. 1-5.
[21] J.Jin, B.Li and T.Kong, “Is Random Network Coding Helpful in WiMAX?”, 27th IEEE Conf. on
Comp. and Comm., 2008, pp 2162-2170.
[22] T.Bchini, N.Tabbane, S.Tabbane, E.Chaput, and A.L Beylot, “QoS in IEEE 802.16e for VoIP and
Video with mechanism of Soft Handover – FBSS in Highway”, 8th Int’l Conf. on Networks,
2009, pp 88-93.
[23] J-Y.Baek, D-J.Kim, Y-J.Suh, and Y-D.Chung “Network-initiated Handover Based on IEEE
802.21 Framework for QoS Service Continuity in UMTS/802.16e Networks”, IEEE VTC Spring
2008, pp 2157-2161.
[24] V.Jesus, S.Sargento and R.L Aguiar, etc., “Mobility with QoS Support for Multi-Interface
Terminals Combined User and Network Approach”, 12th IEEE Symp. on Comp. and Comm.,
2007, pp 325-332.
[25] D.J Wright, “Maintaining QoS during Handover among Multiple Wireless Access Technologies”,
Int’l Conf. on Management of mobile Business, 2007, pp 1-10.
[26] T.Zhifeng, A.Li, J.Zhang and T.Kuze, “Performance Improvement for Multichannel HARQ
Protocol in Next Generation WiMAX network”, IEEE WCNC, Ap.2008, pp 2009-2014.
[27] Zhuoqun Li, Walter Featherstone and Stavros Tzavidas "Shared Multi- Channel HARQ" IEEE
C802.16m-09/1335.
[28] J Kim, H Woo, H Lee and M Lee “Dynamic adjustment of optimistic parameter of ROHC for
performance improvement”, Int’l Conf. on nfo Net., 2009, pp 1-3.
[29] J Chen, W Jiao and Q Guo, “An Integrated QoS Control Architecture for IEEE 802.16 Broadband
Wireless Access Systems”, IEEE GLOBCOM, Vol. 6, 2005, pp 3330-3335.
[30] Anand S.V, “QoS based Handover layer for a Multi-RAT Mobile Terminal in UMTS and Wi-
MAX Networks”, 3rd Comm. Sys. Soft. And middleware workshop, 2008, pp 472-479.
[31] S Sargento and V Jesus, etc., “Context-Aware End-to-End QoS Architecture in Multi-technology
Multi-interface Environments”, 16 th IST mob. and wireless comm. summit, 2007, pp 1-5.
[32] A Ukil and J Sen, “QoS Aware Cross-Layer Optimized Resource Allocation in WiMAX
Systems”, 1st Int’l Conf. on wireless Comm. Veh. Tech., Info. theory and Areo space., 2009, pp
818-822.
[33] Q Liu, X Wang and G.B Giannakis, “Cross-Layer Scheduler Design with QoS Support
forWireless Access Networks”, Proc. of 2nd Int’l Conf. on QoS in Het. Wired/Wireless Networks,
2005.
[34] B.Xie, W.Zhou and J.Zeng, “A Novel Cross-Layer Design with QoS Guarantee for WiMAX
System”, 3rd Int’l Conf. on pervasive computing and App., 2008, pp 835-840.
[35] A.Lera, A.Molinaro and S.Pizzi, “Channel-Aware Scheduling for QoS and Fairness Provisioning
in IEEE 802.16-WiMAX BWA systems”,IEEE Network Magzine, 2007, pp 34-41.
[36] J.Jin and B.Li, “Adaptive Random Network Coding in WiMAX”, IEEE Int’l Conf. on Comm.,
2008, 2576-2580.
[37] R.Perumalraja, C-H. Lung and S.Anand " An Analysis on mobile WiMAX Security", IEEE
toronto Int'l Conf. on Science and Tech. for humanity, 2009, pp 439-444.
118

AHU J. of Engineering & Applied Sciences 3 (1) : 119-128 (2010)
© 2009 ALHOSN University
THE SOLAR-WIND-EV PARADIGM: A STRATEGIC POLICY
FOR A SUSTAINABLE ENERGY FUTURE
Zuwei Yu 1 , Tarik Aouam 2*
1 Choren USA and College of Engineering, Purdue University
2 Department of Industrial Engineering, ALHOSN University, P.O. Box: 38722, Abu Dhabi, UAE
ABSTRACT: The main idea of the proposed strategic policy is to speed up the deployment of Electric
Vehicles (EVs) that can be fueled by solar and wind power. The solar-wind-EV marriage is a bright
paradigm for solving major energy and environmental problems thanks to recent EV technology
breakthroughs. To quantify the benefits of such a policy, we formulate a leader-follower gaming model,
whose solution results characterize a significant decrease of oil prices, when there is a high penetration of
EVs. As a result, US oil imports would be reduced by nearly 60% from current levels, energy security
would be much improved, and CO 2 emissions would be reduced significantly. The economic impact of the
proposed policy is significant.
Keywords: Solar; wind power; EV; oil; energy security; gaming.
1. INTRODUCTION
Oil consumption and oil dependence have caused severe energy security, environmental, and
economical problems. The U.S. generated 50% of the CO 2 emissions in the category of passenger
cars and light trucks, with a huge trade deficit, and reduced energy security [40]. Therefore, there
is an urgent need for effective policies. One potential policy is the fast deployment of EVs with
batteries that can be charged using solar and wind power.
The California Air Resources Board (CARB) mandated a set of targets in the early 1990s for
adopting EVs. However, these targets were either altered or disbanded due to technical and
political barriers. CARB voted, in March 2008, to give EVs and HFCVs (hydrogen fuel cell
vehicles) parallel treatment. Hydrogen vehicles can also help achieve the desired results but they
are costly and may contribute to ozone damage due to leakages. The U.S. hydrogen initiative has
been questioned even since it was promoted in 2003 [1]. The argument has been that it may be
too expensive to deploy hydrogen passenger cars and light trucks. Unfortunately, policy makers
never paid attention to the argument. Today, the hydrogen infrastructure is still badly lacking
even with heavy subsidies. Should we wait for the hydrogen economy to come with a huge cost
burden? No. Even if the cost of such an economy is not too prohibiting, a hydrogen economy
would still require many years to develop. The HFCVs have much lower energy efficiency than
the EVs. On the other hand, EVs have become a reality after several breakthroughs in 2007,
including advanced batteries with a range of up to 250 miles per charge, a much reduced recharge
time and the battery safety features [2].
The paper will quantitatively analyze the impact of large scale deployment of EVs on oil
demand, prices, and energy security. A gaming model is formulated to simulate oil price
movements and energy security improvements when various EV penetration rates are assumed.
____________________________________
* Corresponding Author. Tel.: +971 4070562
E-mail : t.aouam@alhosnu.ae
119

ZUWEI YU, TARIK AOUAM
Figure 1. The annual average solar power in the continental U.S. (courtesy of NREL).
An area of 100 square miles in the red zone of Figure 1 could provide the total US energy needs.
In fact, such an area can provide 160,000 2 ×8kWh = 204,800,000 MWh/day of electricity. According
to the 2004 EPRI-DOE Handbook: “Supplement of Energy Storage for Grid Connected Wind
Generation Applications,” the U.S. peak demand was 808,000 MW in 2003, which may be converted
to a peak electricity demand of no more than 820,000×24 = 19,200,000 MWh in 2008. Hence, the
solar power in the 100 square mile area could provide the total U.S. energy needs.
Wind resources are significant in the Great Plains; North Dakota alone has a wind power potential
of 415,000 MW [13]. Some offshore areas also have rich wind resources but with costs relatively
high, [14]. The Great Plains can provide wind power at a price range around $60-80/MWh. Although
wind is intermittent, wind power can be stored in large scale by the CAES (compressed air energy
storage) technology [15]-[16]. Aquifer CAES systems can be used for efficient air compression, as
pressure could be kept relatively constant to reduce excessive heat generation.
2.2 EVS, PHEVs and HFCVs
High performance zero emission EVs have been the dream cars for decades. California state
legislators created the ZEV (zero emission vehicle) mandate back in 1990 after General Motors
promised to mass-produce its battery-powered EV1 [17]. Unfortunately, battery development
seriously lagged. Consequently, CARB repeatedly trimmed the quotas and allowed
manufacturers to build a larger number of the so-called “partial zero-emissions vehicles.”
The ZEV regulation came to a halt in 2003 when GM led a lawsuit arguing that CARB's ZEV
mandate was regulating fuel efficiency, a power granted to the federal government. CARB settled
the case by giving up the mandated EVs penetration rates and allowing automakers to research
and commercialize HFCVs, beginning with the demonstration of fuel-cell cars by 2008. Partly
due to the pressure from the automakers, CARB allowed fewer credits for EVs than for HFCVs
until March 2008, when it voted again to give EVs and HFCVs parallel treatment.
120

THE SOLAR-WIND-EV PARADIGM: A STRATEGIC POLICY FOR A SUSTAINABLE ENERGY FUTURE
The first breakthrough related to EVs was the development of li-ion and li-ion polymer
batteries that have a range around 100 - 250 miles per charge. The second is that recharge times
are about 10 minutes and the third was the safety features in batteries to guard against
temperature rise and explosion [2]. In addition, the nano ultracapacitor technology has reached
the demonstration stage, which may support a drive range of 500 miles per charge with a charge
time less than 10 minutes [18] – [19]. In fact, the Honda FCX fuel cell vehicle has a small
ultracapacitor installed to boost the power for acceleration. Several EV models are summarized in
Table 1. Even though the sporty Tesla is priced at $98,000, several firms have been designing
mid-size EVs that can be sold at prices from $20,000 to below $50,000. As a contrast, there have
been no HFCVs promoted for sale in this price range.
Table 1. Selected EV Models [24].
Model Range Charge Energy Battery Country Launch
(miles) time (min.) (kWh)
Cleanova-II 130 ~30 30 Li-ion France 2008
Dodge Zeo 250 n/a 64 Li-ion U.S. n/a
F6e 64 ~10 n/a Iron China 2009
GM Volt 40 n/a 8 Li-ion U.S. 2010
Hotzenblitz 218 n/a ~40 PLIB Switzerland 2008
Lighting GT n/a 10 35 Li-ion UK 2009
Phoenix SUT 130 10 35 Li-ion U.S. 2008
250 10 70 (nano)
Regane/Mixim n/a n/a n/a Li-ion France/Japan 2010
Subaru G4e 125 15 n/a Li-ion Japan 2008
Tesla 220 210 53 Li-ion U.S. 2008
PLIB – Polymer Li-ion battery, also used as Li-Po sometime.
PP – peak power
Table 2 shows convincingly the efficiency advantages of EVs over HFCVs. It is assumed that
hydrogen is produced at renewable power generation sites so electricity losses are zero, which is
quite optimistic. The data is from various sources (e.g. [20] - [23]) and modified based on our
best knowledge (e.g. 9.3% of electricity losses instead of 10% in the U.S.) The efficiency
comparison is based on the percentages of the same amount of electricity being delivered to the
wheels of different vehicles. EVs are clearly twice as efficient as HFCVs. Notice that our
estimate of the HFCV efficiency is consistent with other estimates. For example, a target
efficiency of 42% is set for the fuel cell hybrid vehicles in the Toyota FCHV BOOK.
Table 2. Estimated Energy Efficiency for EVs and HFCVs (%).
Vehicle Power H 2 Power Battery Fuel cell Drive train Overall
to H 2 * Compression
T&D**
or liquefaction
EV 89-92.4 85-92 90-95 68-81
HFCV 75-85 90-92 87-90 *** 55-60 90-95 29-40
(gas)
HFCV 75-85 65-70 94 *** 55-60 90-95 22.7-26
(liquid)
* Electrolysis is assumed and power supply losses are included. The upper limit is the target in the long run while 75% is attainable now.
** In power T&D (transmission & distribution), the average efficiency is 90.7% and the average loss is about 9.3% .
*** Most HFCV models also have small batteries installed to improve efficiency. The batteries also have charge/discharge losses which are not
considered in this table due to their smaller sizes.
121

ZUWEI YU, TARIK AOUAM
Based on the results above, EVs would only need about half of the solar and wind power
compared to the HFCVs for the same driving distance and same weight. That is, HFCVs would
be twice as expensive as the EVs in terms of the renewable power infrastructure. As shown in
Table 1, typical EVs (e.g., Tesla and Phoenix SUT) may need 0.25 – 0.3 kWh for driving one
mile. If they are charged overnight, the cost of 1.0 kWh may be no more than 10 cents, resulting
in a cost of about 3 cents per mile of EV driving. The cost of gasoline driven vehicles for the
same one mile would then be 360 cents/21 = 17 cents, given that the gasoline price is $3.6/gallon
and the average MPG is 21 (21 is the national average for cars and light trucks).
Other advantages of EVs may include: much lower maintenance time and cost, overnight
charging to smooth out the power load, greater safety factor than the HFCVs, minimization of
water use, no water vapor produced, etc. Results from the French postal service La Poste EV test
program show that EVs can reduce operating costs significantly [25]. As a result, La Poste
decided to replace its old fleet with EVs. However EVs have notable disadvantages, including
their susceptibility to power outages and currently high investment cost. The first issue may not
be an important concern because most families can buy one EV and one PHEV, while the second
issue can be resolved by mass production and battery standardization. It should be noted that EVs
are more reliable than PHEVs in the use of electricity from renewable sources, as there is no
guarantee that PHEVs will be plugged-in due to various reasons including inconvenience and the
possibility of gasoline price going down. In this paper, PHEVs are used for hedging the risks of
power outage rather than a reliable means to reduce oil consumption.
3. THE LEADER-FOLLOWER GAMING MODEL
The model is an extension of what we have developed previously for electricity analysis [10]. It
is a 2-level leader-follower gaming model where the leaders’ Cournot gaming is in Level-1 while
the fringe producers’ profit maximization is in Level-2, as described below.
To facilitate the presentation, we first adopt the notations below:
f fringe producer index
s leader producer index
t time period (in year)
g strategic storage owner index (usually governments)
i oil production field (or plant) index
maximum oil production limit
availability of oil production
cost of oil production
intercept of the linear (inverse) demand function
slope of the linear demand function
oil price
profit of producer s
oil produced by plant i
oil sale by s
Pmax( t, s, i)
avail( t, s, i)
Cv( t, s)
a0( t)
b( t)
Price( t)
( s)
op(, t s,)
i
q(, t s)
122

THE SOLAR-WIND-EV PARADIGM: A STRATEGIC POLICY FOR A SUSTAINABLE ENERGY FUTURE
The mathematical model for the leaders’ individual profit maximization in Level-1 can be
summarized below:
Minimize - ( s)
a0() t b()( t qts (, ) qt (, f)) qts (, ) Cvtsioptsi (, ,) (, ,), st
,
s f
i
(1)
optsi (, ,) availtsiP (, ,) max(, tsi ,), tsi
, ,
(2)
qts (, ) optsi (, ,),
ts
,
(3)
i
op( t, s, i) 0 t, s,
i
(4)
qts ( , ) 0 ts ,
(5)
The Cournot strategy. (6)
where q ’s are quantities from the fringe producers. Notice that (6) is not a regular constraint but
rather a requirement that the Cournot quantity strategy will be used for developing a set of
pseudo Karush-Kuhn-Tucker (PKKT) first order conditions in the MCP formulation.
The fringe producers’ profit maximization problem in Level 2 is:
Max price() t q(, t f ) Cv(, t f ,) i op(, t f ,), i t,
f
(7)
i
s.t.
op(t, f,i) avail(, t f ,)Pmax(, i t f ,), i t, f , i
(8)
qt (, f) opt (, f,) i,
t,
f
(9)
i
op( t, f , i) 0 t, f , i
(10)
qt (, f) 0 t,
f
(11)
Notice that price(t) is calculated from the solution of the Cournot-Nash gaming in Level-1 and
fed to (7) as a fixed quantity.
The solution procedure: An iterative method is developed that iterates between the leaders’
MCP problem and the followers’ simple profit maximization problem. The MCP method by itself
is verified to be an efficient solution procedure as describe in [12] – [13]. The method involves
the conversion of the leaders’ problem (level-1) into a set of MCP conditions (or the PKKT
conditions, as mentioned earlier), which can be found in our earlier studies and will not be
repeated, [12]. The solution to the MCP equations can be obtained through the use of an MCP
solver, for example, PATH or any others that work. The details of these MCP solvers can be
found in reference [41]. The nodal prices, price(i), are calculated after the MCP equations are
solved. The fringe producers’ problem is solved using a regular profit maximization solver. The
whole solution process is an iterative one that can be summarized below.
Step 1: Set fringe producers’ quantities qt (, f) 0.8 Pmax(, t f,)
i.
i
Step 2: Solve the leaders’ problem using PATH.
Step 3: Calculate price(t).
Step 4: Check for convergence. If price(t) is within a preset tolerance from the price(t) of the
last iteration, stop. Else, continue.
Step 5: Solve profit maximization for all fringe producers. Go to Step 2.
123

ZUWEI YU, TARIK AOUAM
Step 6: Stop.
In this study, the convergence of the proposed iterative method for solving the 2-level model
is very fast, often requiring only three iterations or so to reach the equilibrium nodal prices.
4. CASE STUDIES
In the first case study, California is assumed to lead the nation in adopting EVs with a 1%
penetration in 2010 and a 2% in 2011, followed by a 3% nationwide EV penetration in 2012. The
second case study examines a 5% EV penetration nationwide plus a 3% PHEV penetration. The
third case, where OPEC and Eurasia form a coalition and no EVs and PHEVs are deployed, will
serve as a reference scenario. Deployment of EVs could occur prior to large scale solar-wind
power projects in regions where coal power is limited, such as California. Natural gas power
plants can be used for charging EVs and reducing CO 2 emissions compared to conventional
gasoline vehicles. The overall CO 2 reduction could be around 27% if PHEVs are charged by the
existing power systems [26]. Conventional oil capacity expansion is taken from recent EIA
studies including new fields discovered. The U.S. ethanol production is assumed to be in line
with the U.S. House Bill HR-6 passed in late 2007 [27]. Oil from the Canadian oil sands is
assumed to be in line with the projection by the Canadian Association of Petroleum Producers
[28]. The CTL fuels such as the CTL diesel, gasoline, naphtha, DME (dimethyl ether) and
methanol are based on the data in [29]. Biofuels of other countries are from [30].
4.1 Two leaders with a 3% annually EV penetration
In this section, the reference “oil demand growth” is taken from recent EIA studies [31]-[32]
but with some modifications. In the EIA studies, the annual oil demand growth rate of the U.S. is
about 0.7% in 2000 – 2007, and about 5-6% for China (may decline to about 2.9% after 2020). It
is assumed that the U.S. oil demand growth is about 0.2% in 2008 – 2012 due to economy slow
down. We modify China’s growth rate so that it declines from 5% in 2010 to about 1% in 2022
with equal annual decrements. We use lower oil demand growth rates than the EIA data in later
years due to our belief that persistently high oil prices would slow down oil demand in the long
run. EVs and PHEVs are considered separately as extra oil demand reduction.
We assume a policy of 1% EV penetration in the sales of new cars and new light trucks in
2010 and 2% in 2011 in California, followed by a 3% penetration in the U.S. and other OECD
countries starting from 2012. We assume that the fast industrializing countries (e.g. China and
India) would do the same but with a 3-year delay. The 2008 U.S. new car and light truck sales are
around 16 million, as forecasted by the industry, about one million less than in 2005. We forecast
that this number will be around 18 million in 2012 due to population increase. According to [33]-
[34], the average driving distance for U.S. cars and light trucks is about 12,000 MPY (miles per
year), with each consuming about 570 gallons per year (GPY) of gasoline on average. All of the
factors are considered in the construction of the linear demand functions for the planning horizon
starting from 2009. Notice that the fuel use of EVs is 0, which equals gasoline (or diesel) savings
of 570 GPY in 2010.
The results are summarized in Figure 2. One can see that if OPEC and Eurasia act like cartels
and there is only an annual 3% EV penetration in the selected regions, oil prices would remain
relatively high for the next 10 years or so. The reasons are multiple: high demand growth in
developing countries for at least the next 8 years, limited capacity expansion from non-cartel
124

THE SOLAR-WIND-EV PARADIGM: A STRATEGIC POLICY FOR A SUSTAINABLE ENERGY FUTURE
players, and little capacity expansion from the leaders. This case shows that a low EV penetration
rate is not enough to bring down oil prices in the short run. As a result, oil demand remains
relatively high, and energy security is still an issue for the U.S. for a while.
4.2 Two leaders with 5% EV and 3% PHEV penetration
In this case, we assumed an additional annual 3% PHEV penetration on top of the 5% EV
penetration as described in case III.2. According to [33], each PHEV may consume 226.8 GPY of
gasoline by 2012, with an estimated fuel savings of 570-226.8 = 343.2 GPY. An extra 3% fuel
reduction per year is also assumed for PHEVs beginning from 2012. For example, if 20 million
cars and light trucks are sold in the U.S. in 2018, the gasoline savings from the 3% annual PHEV
penetration would be 20million*(1.03^7-1)*343.2*(1.03^6)= 122,899 bpd (barrels per day) of
gasoline. The results are illustrated in Figure 2. The U.S. cumulative oil savings from both a 5%
EV and a 3% PHEV penetration are illustrated in Figure 3. The total U.S. gasoline saving would
be nearly 7.6 mbpd in 2024, which translates to at least 8.0 mbdp of crude oil saved. The savings
would be about 60% of the current U.S. oil import (13.5 mbpd). Based on [40], one gallon of
gasoline emits about 5.3 pounds of CO 2 in the U.S. The CO 2 reduction from reduced oil import
and consumption is then nearly 800,000 tons of CO 2 by 2024. Notice that we assume a 75%
PHEV plug-in rate when oil prices are below $70/bbl (i.e., 25% of the PHEVs would not be
plugged-in). Therefore, oil savings from PHEVs would be reduced somewhat after 2016.
We can see that due to the additional PHEV penetration in new car and light truck sales
starting from 2012, oil prices are further reduced. By the end of the study horizon, prices would
be around the lower $50/bbl (or even below $50/bbl by 2022). The prices at this level would be
welcomed by the consumers, who would also be assumed to have a positive attitude toward fuel
efficiency (i.e., most of the consumers would be assumed to purchase EVs and PHEVs around
2020).
4.3 Coalition between OPEC and Eurasia without EVs and PHEVs
In this case study, we assume that OPEC would form a coalition with Eurasia so that the two
behave like a single cartel. We assume that there would be no EVs and PHEVs deployment. This
scenario will serve as a reference and will show characterize extreme oil prices as depicted in
Figure 2. We can see that if OPEC and Eurasia form a single cartel to coordinate their output, and
if there is no EV and PHEV penetration while hybrid penetration is at the current level, the world
oil prices would be over $160/bbl by 2022.
US oil sales.
125

ZUWEI YU, TARIK AOUAM
$/bbl (2007$)
180
160
140
120
100
80
60
III.1
III.2
III.3
40
20
0
2008 2010 2012 2014 2016 2018 2020 2022
Year
Figure 2. Plot of the simulated prices (The 2008 price is from EIA [32]).
8000000
bpd (oil import reduction)
7000000
6000000
5000000
4000000
3000000
2000000
1000000
0
2010
2011
2012
2013
2014
2015
2016
2017
Year
Figure 3. U.S. oil savings in case III.2.
2018
2019
2020
2021
2022
2023
2024
More case studies and scenarios can be done. However, except for extreme cases such as the
closure of the Hormuz Straight, most other likely scenarios would fall between Cases III.1 and
III.3. The price movement curves would fall between the curves of Cases III.1 and III.3.
5. CONCLUSION
U.S. policy makers have failed to include EVs in the national energy policy. As a result, federal
support for EV research and development has been almost non-existent. Fortunately, private
sectors have been leading R&D and production of EVs through venture investments over the
126

THE SOLAR-WIND-EV PARADIGM: A STRATEGIC POLICY FOR A SUSTAINABLE ENERGY FUTURE
past several years. It is time to revisit national and international energy policy in the era of global
warming, increasing oil demand and production gap, and energy security. Several policy issues
should be reevaluated.
First, we notice that the infrastructure for EVs is not as massive as for HFCVs. The majority
of the current power network can be used to charge EVs overnight due to a much lower power
demand during the night than during the day. Parking lot charging facilities may be constructed
near places such as offices, shopping centers, sport and game areas, and even gas stations. These
actions require regulatory reforms in the utility sector.
Second, policy makers should help facilitate the research, demonstration and deployment of
EVs. The U.S. may need to redirect some funding from HFCVs to EVs for project demonstration.
Third, biofuel quotas should be reduced. In fact, biofuel production has been causing food
prices to rise globally without solving much of the U.S. oil problems. Some recent studies show
that biofuel may damage the environment. Expanded land cultivation causes carbon dioxide to
escape from the soil into the atmosphere, leaving a carbon deficit from 17 years to more than 420
years depending on the land cleared for ethanol production [37]. Several OECD countries have
now either cut back or been considering cancelling subsidies to some biofuels due to new
findings on biofuel-related problems [39].
Finally, government should subsidize EVs. One of the issues policy should address is how to
overcome the car market entry barriers.
6. REFERENCES
[1] D.W. Keith, A.E. Farrell, “Rethinking of Hydrogen Cars,” Science, Vol. 301, no. 5631, July 18, 2003.
[2] D. Schneider, “Who’s Resuscitating the Electric Car?” American Scientist Magazine, Sept-Oct., 2007.
[3] J. Grinffin, “OPEC Behavior: A Test of Alternative Hypothesis,” The American Economic Review,
Vol. 75 1985, pp. 954-963.
[4] S. Gülen, “Is OPEC Cartel? Evidence from Cointegration and Causality Tests,” The Energy Journal,
April 1996.
[5] S. Shaffer, “Stable Cartel with a Cournot Fringe,” Southern Economic Journal, Vol. 61, 1995, pp. 744-
754.
[6] E. Diamantoudi, “Stable Cartel Revisited,” Economic Theory, vol. 26(4), 2005, pp. 907-921.
[7] A. Tuzuner, Z. Yu, “A Stochastic Gaming Framework for Analyzing Market Potentials and Risks of
CTL,” in Proceedings of the Pittsburgh Coal Conference, Sept. 2006.
[8] Z. Yu, B. Bowen, “Integrated Energy Resource Planning Approach to the U.S. Energy Security
Issues,” Oil, Gas, Energy & Law, Nov. 2007.
[9] Z. Yu, F.T. Sparrow, T. Morin, G. Nderitu, "A Stackelberg Price Leadership Model for Deregulated
Electricity Markets," Proceedings of the IEEE PES Winter Meeting, Singapore, January 2000.
[10] Z. Yu, P. Preckel, F.T. Sparrow, D. Nderitu, “A Strategic Power Market Model with CO2 Limits,”
International Journal of Electric Power & Energy Systems, Vol. 23, no. 6, 2001.
[11] B. F. Hobbs, “Linear complementarity models of Nash–Cournot competition in bilateral and
POOLCO power markets,” IEEE Trans. Power Syst., vol. 16, no. 2, pp. 194–202, May 2001.
[12] Penn Well, “Solar Power - Utility-scale sun power,” Power Engineering, June 2007.
[13] Pacific Northwest National Laboratory, “An Assessment of the Available Windy Land Area and
Wind Energy Potential in the Contiguous United States,” a reported prepared for the Department of
Energy (DOE), 1991.
[14] PACE, “Assessment of Offshore Wind Power Resources,” prepared for Long Island Power
Authority, August 22, 2007.
127

ZUWEI YU, TARIK AOUAM
[15] P. Vadasz, “Compressed Air Energy Storage: Optimal Performance and Techno-Economical
Indices,” Int. J. Applied Thermodynamics, Vol.2, No.2, pp. 69-80, June-1999.
[16] J. Greenblatt, S. Succar, D. Denkenberger, R. Williams, “Toward optimization of a wind/compressed
air energy storage (CAES) power system,” Electric Power Conference, Baltimore, MD, March 30-
April 1, 2004.
[17] P. Fairley, “California to Rule On Fate of EVs,” IEEE Spectrum, Nov. 2007.
[18] J. Schindall, “The Charge of Ultracapacitors,” IEEE Spectrum, Nov. 2007, pp. 42-46.
[19] A. Westervelt, B. Brun, “Two Batteries,” published by The Sustainable Industries, Jan. 4, 2008.
[20] U. Bossel, “Efficiency of Hydrogen Fuel Cell, Diesel-SOFC-Hybrid & Battery Electric Vehicles,”
European Fuel Cell Forum, Oct. 20, 2003.
[21] K. Wipke, et al., “Learning Demonstration Interim Progress Report –Summer 2007,” NREL/TP-560-
41848, July 2007.
[22] Stephen Eaves, and James Eaves, “A cost comparison of fuel-cell and battery electric vehicles,”
Journal of Power Sources, Vol. 130, Issues 1-2, May 3 2004, pp. 208-212.
[23] GAO, “Hydrogen Fuel Initiative,” Government Accountability Office: Report to Congressional
Requesters, GAO-08-305, Jan. 2008.
[24] Z. Yu, “Wind power and transportation: A bridge to energy security,” Proceedings of the IEEE PES
eit conference, May 18-21, 2008, Iowa State University.
[25] D. Guthier-Villars, “Electric Cars Gather Speed,” The Wall Street Journal, March 4, 2008.
[26] M. Kintner-Meyer, K. Schneider, R. Pratt, “IMPACTS ASSESSMENT OF PLUG-IN HYBRID
VEHICLES ON ELECTRIC UTILITIES AND REGIONAL U.S. POWER GRIDS,” a report prepared
for DOE, Pacific Northwest National Laboratory, Nov. 2007.
[27] HR 6, “The Energy Independence and Security Act of 2007,” the U.S. House of Representatives,
Dec. 2007.
[28] CAPP, “Canadian Oil Sands Outlook,” a report prepared by the Canadian Association of Petroleum
Production, Sept. 2007.
[29] Z. Yu, T. Aouam, “A Monte Carlo Gaming Model for Analyzing the 21C World Oil Markets with
PHEVs,” a book chapter accepted by the NOVA Science Publishers, to appear in 2008.
[30] “Biofuel Market Worldwide,” a report prepared by Biotechnology, July 2006.
[31] EIA, “The 2007 Annual Energy Outlook,” released in June 2007.
[32] EIA, “The Short-Term Energy Outlook 2008,” released in Jan. 2008.
[33] EPRI (Electric Power Research Institute), “Environmental Assessment of Plug-In Hybrid Electric
Vehicles,” a report prepared for DOE, July 2007.
[34] NREL (National Renewable Energy Laboratory), “Field Operations Program - Overview of
Advanced Technology Transportation,” a report prepared for the U.S. DOE, July 2000.
[35] J. Olsen, “Electric Car System Planned in Denmark,” The Associated Press, March 27, 2008.
[36] T.K. Tromp, R. Shia, M. Allen, J.M. Eiler, and Y.L. Yung, “Potential Environmental Impact of a
Hydrogen Economy on the Stratosphere,” Science 13 June 2003: Vol. 300. no. 5626, pp. 1740 – 1742
[37] J. Fargione, J. Hill, D. Tilman, S. Polasky, P. Hawthorne, “Land Clearing and the Biofuel Carbon
Debt,” Science online publication, Feb. 7, 2008.
[38] P.J. Crutzen, A.R. Mosier, K.A. Smith, and W. Winiwarter, “N 2 O release from agro-biofuel
production negates global warming reduction by replacing fossil fuels,” Atmospheric Chemistry and
Physics Discussion, Vol. 7, 11191-11205, 2007.
[39] E. Rosenthal, “Europe, Cutting Biofuel Subsidies, Redirects Aid to Stress Greenest Options,” The
New York Times, Jan. 22, 2008.
[40] J. DeCicco, F. Fung, “Global Warming on the Road – The Climate Impact of American’s
Automobiles,” Environmental Defense, 2006.
[41] GAMS – The Solver Manuals, GAMS Development Corporation, 2004.
128

ALHOSN UNIVERSITY JOURNAL
OF ENGINEERING AND APPLIED SCIENCES
GENERAL INFORMATION
AHU journal of Engineering and Applied Sciences
is a refereed journal, published bi annually by
ALHOSN University. The aim of of the journal
is to emphasize the use of engineering design and
analysis, to publish a high quality research, to
strike a balance between research and application
and to promote international dialogue and
collaboration. The journal welcomes scholarly
submissions of original and innovative articles
in all fields of engineering and applied sciences.
These include mutually but not exclusively,
articles pertaining to the reporting of advances in
theory, techniques, methodologies, applications,
short communication, review articles and book
reviews.
The journal is published bi annually in June
andDecember each year. Special issues may be
published from time to time. A guest editor may
be invited to edit a special issue and he/she will be
responsible for the quality of all papers selected
for publication.
Submission of manuscript for publication will be
considered to imply a tacit understanding that it
has not been previously published is not currently
considered for publication and will not be sent for
publication elsewhere.
3. Articles should be served in word format and
emailed to the editor with a cover page stating
name, email address and where applicable,
academic affiliation. If a submission includes
graphics and pictures, this should be served
in jpg format and sent as a separate attachment
along the submission.
4. Articles must be doubled spaced and in a
single column format.
5. Instructions on the preparation of tables,
figures, references and abstracts should
follow the IEEE guidelines for publication.
at: http://www.ieee.org/publications_standards/
publications/authors/authors_journal
6. The author of articles will be supplied with
one free hard copy of the journal in which his/
her paper has appeared by airmail soon after
the journal is published
129

130