Journal of Civil Engineering

and Architecture
Volume 6, Number 4, April 2012 (Serial Number 53)

David Publishing

David Publishing Company

www.davidpublishing.org

Publication Information:
Journal of Civil Engineering and Architecture (ISSN 1934-7359) is published monthly in hard copy and online by
David Publishing Company located at 9460 Telstar Ave Suite 5, EL Monte, CA 91731, USA.
Aims and Scope:
Journal of Civil Engineering and Architecture, a monthly professional academic journal, covers all sorts of researches
on structure engineering, geotechnical engineering, underground engineering, engineering management, etc. as well
as other issues.
Editorial Board Members:
Dr. Tamer A. El Maaddawy (Canada), Prof. San-Shyan Lin (China Taiwan), Dr. Songbai Cai (China), Dr. Xiaoyan Lei
(China), Prof. Vladimir Patrcevic (Croatia), Dr. Sherif Ahmed Ali Sheta (Egypt), Prof. Nasamat Abdel Kader (Egypt),
Prof. Mohamed Al-Gharieb Sakr (Egypt), Prof. Olga Popovic Larsen (Denmark), Prof. George C. Manos (Greece), Dr.
Konstantinos Giannakos (Greece), Pakwai Chan (Hong Kong), Dr. K. Muthukkumaran (India), Chiara Vernizzi (Italy),
Prof. Michele Maugeri (Italy), Dr. Giovanna Vessia (Italy), Prof. Valentina Zileska-Pancovska (Macedonia), Dr. J.
Jayaprakash (Malaysia), Mr. Fathollah Sajedi (Malaysia), Prof. Nathaniel Anny Aniekwu (Nigeria), Dr. Marta Sowik
(Poland), Dr. Rafael Aguilar (Portugal), Dr. Moataz A. S. Badawi (Saudi Arabia), Prof. David Chua Kim Huat
(Singapore), Dr. Vail Karakale (Waiel Mowrtage) (Turky), Dr. A.Senem Deviren (Turkey), Dr. Yasemin Afacan
(Turkey), Dr. Ming An (UK), Prof. Ahmed Elseragy (UK), Prof. Jamal Khatib (UK), Dr. John Kinuthia (UK), Dr.
Johnnie Ben-Edigbe (UK), Dr. Yail Jimmy Kim (USA), Dr. Muang Seniwongse (USA), Prof. Xiaoduan Sun (USA), Dr.
Zihan Yan (USA).
Manuscripts can be submitted via Web Submission, or E-mail to civil@davidpublishing.org. Submission guidelines
and Web Submission system are available at http://www.davidpublishing.org.
Editorial Office:
9460 Telstar Ave Suite 5, EL Monte, CA 91731, USA
Tel: 1-323-984-7526 Fax: 1-323-984-7374
E-mail: civil@davidpublishing.org; shelly@davidpublishing.org.
Copyright2012 by David Publishing Company and individual contributors. All rights reserved. David Publishing
Company holds the exclusive copyright of all the contents of this journal. In accordance with the international
convention, no part of this journal may be reproduced or transmitted by any media or publishing organs (including
various websites) without the written permission of the copyright holder. Otherwise, any conduct would be
considered as the violation of the copyright. The contents of this journal are available for any citation. However, all
the citations should be clearly indicated with the title of this journal, serial number and the name of the author.
Abstracted / Indexed in:
Database of EBSCO, Massachusetts, USA
Chinese Database of CEPS, Airiti Inc. & OCLC
Cambridge Science Abstracts (CSA)
Ulrichs Periodicals Directory
Summon
Subscription Information:
$520/year (print) $360/year (online) $680/year (print and online)
David Publishing Company
9460 Telstar Ave Suite 5, EL Monte, CA 91731, USA
Tel: 1-323-984-7526 Fax: 1-323-984-7374
E-mail: order@davidpublishing.org; shelly@davidpublishing.org.

DAVID PUBLISHING

David Publishing Company

www.davidpublishing.org

Journal of Civil Engineering

and Architecture
Volume 6, Number 4, April 2012 (Serial Number 53)

Contents
Technical Papers
391

Evaluation of the Policies for Seismic Retrofit of Buildings

Shoichi Ando

403

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case

Study
Basil Kasim Mohamed

415

Modeling Bridge Condition Levels in the United States

Denver Tolliver and Pan Lu

433

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

Ayaho Miyamoto and Akito Yabe

444

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces
Valerio Cutini

457

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

Claus Bech-Danielsen

470

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities
Semih S. Tezcan and Zuhal Ozdemir

479

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion
Technique
Jun Teng, Ting Zhang and Wei Lu

489

Static Requirement and Types Complexity in the Diagnostics Phase

Antonella Guida and Antonello Pagliuca

496

Characteristics of User Redesign Process: A Study of Changes Made by Users in

Architect-Designed Housing
Nghai Ezekiel Suleman

Technical Reports
504

Irradiation Damage of Nano-C2S Particles Studied by In-Situ Transmission Electron Microscopy

Hongfang Sun, Mahir Dham, Eric A. Stach and Carol Handwerker

509

Architecture, Faith and Culture Antagonism or Harmony?

Hassib Rehailia

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 391402

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Evaluation of the Policies for Seismic Retrofit of

Buildings
Shoichi Ando
International Institute of Seismology and Earthquake Engineering (IISEE), the Building Research Institute (BRI), Tsukuba,
Department of Urban Engineering, Graduate School of Engineering, the University of Tokyo, Japan
Abstract: The Japanese government has established a law to promote seismic retrofitting of buildings immediately after the Great
Hanshin-Awaji Earthquake in 1995. This paper evaluates the effectiveness, efficiency, administrative feasibility and technological
incentives of the policies related to the law. The data shows that the policy target of seismic safety of existing buildings will be
achieved in 2018 if the current trends of improvement will be continued. In the field of school buildings, national government
supports the school retrofit works that are carried out by the local governments, using the guideline for school retrofit. However,
there are still significant issues to make all buildings safe. One of the key challenges is how to persuade the elderly who would not
invest their money to improve their old houses. Another challenge is to make owners understand the importance and have priority in
improving the seismic safety of buildings. Currently many efforts are taken by the local governments, such as holding seminars for
local communities, preparing financial support schemes, providing consultancy for seismic assessment and making earthquake
hazard maps. This paper also provides comments on the improvement of the current policies for promoting seismic retrofit based on
some international experiences in retrofit of buildings.
Key words: Seismic retrofit, school retrofit, seismic assessment, hazard map, financial support, policy.

1. Introduction
After the Great Hanshin-Awaji Earthquake in 1995,
the Japanese government has established around 20
legal systems including the Act for Promotion of the
Earthquake Proof Retrofit of Buildings (Retrofit
Promotion Act) that has been established in 1995 as
one of new legal systems. Moreover, many Japanese
local governments become to provide various support
systems in order to promote seismic retrofit conducted
by owners and the private sector. The national
government also provides new subsidy systems such
as the regional housing grants and the community
renovation grants based on the Retrofit Promotion Act.
Furthermore a tax reduction system of loans for
seismic retrofitting has started from the fiscal year
2006.

Corresponding author: Shoichi Ando, PhD, professor,

research fields: urban disaster. E-mail: ando@kenken.go.jp.

Why so many public assistance systems for housing

seismic retrofitting exist, though houses are private
assets? Originally, this argument arose immediately
after the Great Hanshin-Awaji Earthquake. Has the
government decided not to appropriate tax revenue
(public assistance) for the reconstruction of individual
houses, then?
One of the reasons why such a policy change has
made, may result from the establishment of the Act
concerning Support for Reconstructing Livelihoods of
Disaster Victims in 1998 and its revision in 2004.
This Act is legislation at the instance of house
members on the basis of several disasters after the
Great Hanshin-Awaji Earthquake. According to this
Act, in case of completely or partially destroyed
houses, a certain amount of public assistance can be
provided to owners of such private assets. For instance,
the owner of a completely destroyed house can allow
one million yen for purchasing household effects and

392

Evaluation of the Policies for Seismic Retrofit of Buildings

two million yen for reconstruction of the house, i.e., in

total three million yen will be granted.
In addition, collapse of houses often causes streets
blockade and this may bring about crucial obstacles to
escape, fire fighting and/or relief activities, when an
urban fire occurs as the experience at the Great Kanto
Earthquake in 1923 and in Nagata ward of Kobe city
at the Great Hanshin-Awaji Earthquake. Namely,
seismic retrofitting of buildings including house, is
indispensable in order to secure entire urban safety.
Therefore, public assistance can be provided, even
though houses are private assets.
Taking a view of the estimated tolls by the Tokai
Earthquake published by the Cabinet Secretary in
2005 as a reference, the maximum number of death
toll will reach to approximately 9,200 persons by the
assumed ocean-type Tokai Earthquake. And around
85% of death toll, i.e., approximately 7,900 dead
persons will be due to the collapse of buildings and
like. At the same time, the Japanese Cabinet Secretary
announced a target to reduce those tolls by half in the
Earthquake Disaster Mitigation Strategy for Tokai
Earthquake. For that purpose, a detailed target to
improve housing seismic retrofitting ratio was set up
from current 75% to 90% within 10 years (till 2015).
The Earthquake Disaster Mitigation Strategy was
also created for coming Tohnankai and Nankai
(South-east Ocean and South Ocean) Earthquake in
2005 and the main targets consist of housing seismic
retrofit and tsunami disaster prevention measures.
These are the backgrounds of movement that many
actors established new supporting measures for
housing seismic retrofit in all parts of Japan from the
establishment of the Strategy in 2005 till now.

2. Retrofit Systems in Japan

2.1 Technical Background
The 1995 Great Hanshin-Awaji Earthquake
disaster revealed the weakness of reinforced concrete
(RC) structures that were designed in accordance with
the pre-1981 design code. Failure of RC columns was

primarily in the lack of lateral reinforcement. Larger

deformation capacity may be attained by enhancing
the capacity by
(a) Jacketing RC columns with steel plates and
(b) Wrapping RC columns with fiber reinforced
plastics (FRP).
The use of FRP sheets has merit of easy
construction work and of light material weight.
Placement of bracing structures (structural walls or
steel braces) is effective limiting the response
deformation of the structure, thus avoiding the failure
of brittle members.
The occupancy of a building during the retrofit
work should be considered in selecting retrofit works.
For example, the strengthening work of RC columns
normally requires the removal of mortar and other
finishing materials (tiles) from the concrete surface.
The noise, vibration and dust during the retrofit work
will not allow occupants to stay in the building.
If advanced technology is affordable, especially in
hospitals for post-earthquake medical treatment of the
injured, the earthquake induced forces may be reduced
by placing isolation devices at the base. The response
of a structure may be reduced by installing dampers or
energy dissipating devices are available. The failure of
foundation piles was reported after the 1995 Kobe
earthquake disaster. In some structures, the failure of
pile foundation is said to reduce the earthquake
ground motion input to the structure and limit the
damage in the super-structure. However, the cost of
damage investigation of foundation as well as the
repair work of damaged foundation is expensive. It is
normally desired to provide the foundation structure
with higher resistance. [1]
The need of retrofit relates to the revision of
building code especially up-grade of seismic code. If
building code revised in 2005 and as shown in Fig. 1,
when small portion of building follow new code, there
is not so much demand of retrofit of old buildings.
However when large portion of building stock follow
the new code, owners of existing building may require
retrofit. (by Prof. Shunsuke Otani)

Evaluation of the Policies for Seismic Retrofit of Buildings

Table 1
level.

393

Numbers of buildings under (1981) seismic code

Houses

Fig. 1

Relation of Design Code for New Construction [2].

2.2 Retrofit Promotion Act and Its Support Systems

The Retrofit Promotion Act was enforced in 1995
immediately after the Great Hanshin-Awaji
Earthquake in January 1995, since the lessons learnt
from the devastated disaster urged quick response of
the government and policy makers to secure the safety
of urban built-environment that are mainly consisted
of houses and buildings. Key components of the new
Act are:

Obligation of owners to make best efforts to

assess and retrofit the buildings that are utilized many
people;

Exemption of retroactive application of building

code except for seismic related code to approved
retrofit works;

Guidance, advices and instructions from the

responsible governmental agency.
With regard to new obligation for the owners of
building that is utilized by many people, such building
is defined as, more than 3 stories and more than 1,000
m2 of floor area and its use is in line with designated
one such as school, hospital, department store, office,
shop, hotel, care facility for the elderly and so on.
When the Retrofit Promotion Act was established,
the Japanese Ministry of Land, Infrastructure,
Transport and Tourism (MLIT) surveyed the
conditions of houses and buildings in terms of
applicability to seismic code (in 1981, the current
level of seismic safety). The updated result of the
survey is indicated in the Table 1. The target of
applicable level of seismic code is also shown in the
Fig. 2.

Total houses in Japan

Total numbers of houses
47 million units
Under seismic code units
11.5 million units
(percentage of under level)
Around 25%
Estimated based on the data in 2003
Non-residential Building
Total number
3.4 million buildings
Under seismic code units
1.2 million buildings
(Percentage of under level)
Around 35%
Estimated by the Ministry of Land, Infrastructure,
Transport and Tourism in 2002.

Fig. 2 Trend estimation of safe houses (seismic code)

(Estimated by the Ministry of Land, Infrastructure,
Transport and Tourism in 2003).

The Act was established to introduce retrofit of

houses and buildings. Therefore, in addition to
regulatory measures some economic measures have
been prepared at the same time by the national
government. Those economic measures are available
only in the local government that established Plan for
Retrofit Promotion. Table 2 shows the number and
ratio of the local governments that have established
the Plan for Retrofit Promotion. Table 3 shows the
number and percentage of the municipalities that have
prepared the subsidy systems for seismic assessment
and retrofitting.
Those tables show the fact that even in 2009, one
quarter (25%) of municipalities have prepared the
subsidy system for condominium to assess/evaluate
the vulnerability to earthquake. In case of detached
houses, almost 2/3 of municipalities have established
subsidy system by 2009 for assessment and almost

Evaluation of the Policies for Seismic Retrofit of Buildings

394

half of them have prepared financial support for

retrofitting of houses. Figures of the Table 3 in the
entre- parenthesis indicate the number of percentage
of the April 2008.
Within only one year, the percentage has been
significantly improved because of the following
policy measures, while a limit of financial support to
the retrofit can be observed especially in the field of
condominium and non- residential buildings.
The following policies are described as the MLIT
has reported as below;
(1) Formulation of municipal Retrofit Promotion
Plans;
(2) Establishment of prefectural/municipal subsidy
system;
(3) Promotion of retrofitting of public buildings;
(4) Securing engineers who manage assessment/
retrofit;
(5) Utilization of tax incentives for business use
buildings;
(6) Preparation of seismic hazard mapping;
(7) Best practices for seismic assessment/retrofit;
Table 2 Number of Local Governments that have retrofit
promotion plans (as of 1 April 2009).
Have plans
Prefecture
Total (%)
Municipality
Toal (%)

-2009.9

-2010.3

-2010.4

47
47
100.0%
1.193

50

185

70

1.193

1.243

1.428

1.428

66.3%

69.1%

79.3%

83.2%

Table 3 Subsidy system for seismic assessment and

retrofit (as of April 2009, Japan).
Applicable local Govt. for
reftrofit subsidy
Building type Project type
Number of Ratio: Applicable
local Govt.
Govt.
Assessment
1,227
68.2% (62.7%)
Detached house
Retrofitting
857
47.6% (37.2%)

Assessmen
Retrofitting
Assessmen
Non-residential
Retrofitting
Condominium

450

25.0% (19.0%)

321

17.8% (12.1%)

310

17.2% (13.2)

154

8.6% (5.9%)

(8) Model projects on seismic safety houses and

buildings.
Those measures are comprehensive. Socioeconomic measures, information measures, technical
measures, and institutional measures are included to
promote the policy for seismic assessment and retrofit
of buildings. [3]
2.3 Retrofit of Schools in Japan
Because of the Great Hanshin-Awaji Earthquake
in 1995, school buildings were also severely damaged
by the shake. [4] According to a report provided by
the Ministry of Education, Culture, Sports, Science &
Technology
(MEXT),
approximately
4,500
educational facilities were structurally or nonstructurally damaged, though there were fortunately
no death tolls resulted from damaged schools since the
Great Hanshin-Awaji Earthquake occurred in early
morning at 5:46 a.m.. After the strike of the
Earthquake, 390 schools took the role for evacuation
shelter
and
these
schools
accommodated
approximately 180,000 evacuated people.
Furthermore, at the time of recent major
earthquakes such as Niigata-Chuetsu Earthquake in
October 2004 and Iwate-Miyagi Nairiku Earthquake
in June 2008, while many school buildings were
damaged, non-damaged schools accommodated many
evacuated people. On the basis of these experiences, it
is critical to ensure that school students are safe and
school facilities are fit to serve as evacuation shelters
for local populations. MEXTs policies on structural
and nonstructural retrofitting of school buildings are
introduced. Since school buildings have the following
crucial roles, it is indispensable to assure the safety of
school buildings against earthquakes.
(1) Place for educating children: school buildings
are the place where many children study and live most
part of their days. It is, therefore, vital to keep school
buildings in safer and healthier environment.
(2) Place for cultural and sporting activities: school
is a well-known building to the people who live near
the school. School buildings are, therefore, often

Evaluation of the Policies for Seismic Retrofit of Buildings

395

utilized for the cultural and sporting events for the

local population.

Table 4

(3) Place for evacuation: school often becomes an

evacuation shelter when a major natural disaster
occurs. To this end, it is important that school
buildings accommodate necessary functions for
evacuation shelter.
The Building Standard Law of Japan was revised in
1981 and new seismic resistant design methods were
adopted. According to the revised law, the buildings
constructed based on the new design would have no
damage in the case of middle class earthquakes (about
JMA 5 upper scale). Moreover, there would be no
casualties in these buildings and no severe collapse of
these buildings even in the case of major earthquakes
(about JMA 6 upper) (Table 4).
In order to evaluate the seismic capacity of an
existing school building, the seismic capacity index of
structure (Is) is used in Japan based on the regulation
of the Retrofit Promotion Act. The law regulates that a
building has low risk of collapsing if the Is of the
building is more than 0.6. However, in consideration
of the importance of school building, MEXT
recommends that the Is of school building should
surpass 0.7 after retrofitting.
Is (Seismic Capacity Index of Structure) (Table
5): An index to define the seismic capacity of an
existing reinforced concrete building
IsEoST
Eo: A basic structural seismic capacity index
calculated by the elements of Strength index (C),
Ductility index (F) and Story Index (St)
Eo = CFSt
S: A reduction factor to modify Eo index, which is
based on the structural balance in both plan and
elevation.
T: A reduction factor to modify Eo index, which is
graded by time-dependent deterioration.
A survey carried out by MEXT in April 2002
showed that public school buildings had not been
satisfactorily retrofitted. It emerged from the 2002
survey that seismic assessment/diagnosis was carried

Medium
scale Larger
scale
Type of earthquake
earthquake (about 5 earthquake (over
(JMA scale)
upper)
6 upper)
Old seismic resistant
No major damage
Not verified
design (until 1981)
New seismic
No major damage Will not collapse
JMA Scale: Scale indicating the strength of seismic motion,
which was formed by JMA (Japan Meteorology Agency);
5 Upper: Many people are considerably frightened and find it
difficult to move;
6 Upper: Impossible to keep standing and to move without
crawling.

Difference between new and old seismic resistant

design.

Table 5

Seismic capacity index of structure (Is).

Is 0.3

There is high risk of collapsing

0.3 Is 0.6

There is risk of collapsing

0.6 Is

There is low risk of collapsing

out on only 30% of buildings built based on the

pre-1981 Old Seismic Resistant Design, and only
about 45% of public primary and junior high school
buildings had been retrofitted.
In this connection, a council called Co-operators
Meeting for the Survey and Study of the Promotion of
Earthquake-Resistant
School
Buildings
was
established by MEXT in October 2002. The outcomes
of the councils discussions were submitted to MEXT
in April 2003 in a report entitled the Promotion of
Earthquake-Resistant School Buildings. Based on
this report, the Guidelines for the Promotion of
Earthquake-Resistant
School
Buildings
was
stipulated by MEXT in July 2003.
Chapter 1 of this guideline describes the basic
concept of the earthquake-resistant school building
and Chapter 2 outlines the methods for devising
earthquake-resistant promotion plans, the points to
bear in mind, and the suggested methods for
determining the urgency of earthquake resistance
projects.
The basic principles pointed out in this guideline
are:
(1) to prioritize earthquake resistant measures for
school buildings at high risk of collapse or severe

396

Evaluation of the Policies for Seismic Retrofit of Buildings

damage; (2) to implement seismic resistant capacity

evaluation promptly; (3) to develop a plan for
promoting earthquake resistance promptly; (4) to
disclose the results of the seismic resistant capacity
evaluation and the plans for promoting earthquake
resistance; and (5) to check and take measures for the
earthquake resistance of non-structural elements. [5]
MEXT has been urging municipal governments,
which are responsible for school buildings, to promote
school buildings retrofitting based on the
above-mentioned guideline. In addition, as the
following figure shows, MEXT has a subsidy system
regarding public school buildings (Table 2). In line
with the Sichuan Earthquake in China in May 2008,
MEXT has raised the subsidy rate for vulnerable
school buildings (Is < 0.3) from a half to two thirds in
June 2008.
By utilizing the above-mentioned subsidy system,
the retrofitting of school buildings has been
implemented in Japan. The data shows the status of
earthquake resistance on elementary and lower
secondary public schools in Japan as of April 1, 2008.
Approximately 48,000 of school buildings, or 38% of
school buildings were found lacking needed
earthquake resistance or needed further assessment.
Above all, 10,000 of these buildings were estimated to
be at high risk of collapse in expected large scale
earthquakes. A commitment was made to reinforce all
of these buildings at high risk within 5 years. In
addition, as mentioned, the subsidy rate for vulnerable
school buildings has been raised in June 2008.
Moreover, in order to accelerate the 5 years
retrofitting program into 4 years, MEXT has added an
additional national fund (114 billion JPY) to the
regular budget of fiscal 2008 (115 billion JPY, total
229 billion JPY) in the supplementary budget of fiscal
2008 of Japanese government in October 2008 based
on the Table 6.
Even though structural parts of school buildings
such as columns, beams and walls are enough
retrofitted, if non-structural members such as ceiling

Table 6

Subsidy rate for public school building.

Type of construction

Subsidy rate from MEXT

New construction

1/2

Reconstruction

1/3. 1/2 (Is 0.3)

Renovation

1/3

Seismic rehabilitation
1/3, 1/2 (Is 0.3)
Budget of fiscial 2008:229 billion JPY.

materials, various fixtures and furniture are not

sufficiently retrofitted, these non-structural members
may fall or topple when a major earthquake occurs.
Children and evacuated local people can be killed or
injured by these vulnerable non-structural members.
Therefore, the retrofitting of non-structural members
of school building is extremely important.
In order to urge municipalities to implement nonstructural seismic retrofitting of school buildings, the
National Institute for Educational Policy Research of
Japan (NIER) published a reference book on
non-structural seismic retrofitting of school building
in December 2005.
The following case is an example in this reference
book. [6]

3. Issues of Retrofit Works

3.1 Technical Issues [7]
It should be noted that these countermeasures may
not be the same from a country to another because the
expected performance (minimum required strength
and acceptable damage) of buildings varies from a
country to another. Each country has different levels
of
(a) Seismic risk,
(b) Hazard tolerance,
(c) Economic background, and
(d) Technical development (construction practices).
Most building codes in the world explicitly or
implicitly accept structural damage to occur in a
building during strong earthquakes as long as the
hazard to life is prevented. Indeed, many earthquakes
caused such damage in the past. Then, what

Evaluation of the Policies for Seismic Retrofit of Buildings

percentage of buildings suffered heavy damage in

major earthquakes. The Architectural Institute of
Japan (AIJ) collected damage statistics in Mexico City
and Lazaro Cardenas after the 1985 Mexico
Earthquake, Baguio after the 1990 Luzon, Philippines
Earthquake, Erzincan after the 1992 Turkey
Earthquake, and Kobe after the Hyogo-ken Nambu
1995 (Great Hanshin-Awaji) Earthquake. A heavily
damaged area was first identified in each city, and the
damage level of all buildings in the identified area was
assessed by structural engineers and researchers.
From damage statistics (Table 7 and Fig. 3), the
importance of identifying the small percentage of
those buildings possibly vulnerable in future
earthquakes can be easily realized. Therefore a simple
procedure is desirable to examine the vulnerability of
all existing buildings in a region, spending a few
hours at most for a building, and screen out the
majority of safe buildings. A more detailed and
sophisticated procedure, spending a few weeks, may
be utilized to those buildings identified as vulnerable
by the simple procedure.
Table 7 Damage Statistics from Major Earthquakes [8].
City, year of earthquake Operational Heavy
Collapse Total
damage damage
Mexico City, 1985
4251
194
87
4.532
(93.8%) (4.3%) (1.9%)
Lazaro
Cardenas,
137
25
2 (1.2%) 164
Mexico, 1985
(83.5%) (15.2%)
Baguio
City,
138
34
9 (5.0%) 181
Phillippines, 1990
(76.2%) (18.8%)
Erzincan City, Turkey,
328
68
28
424
1992
(77.4%) (16.8%) (6.6%)
Kobe
(pre-1981
1186
149
158
1493
construction), 1995
(79.4%) (10.0%) (10.6%)
Kobe
(post-1982
1733
73
38
1844
construction), 1995
(94.0%) (4.0%) (2.1%)

Fig. 3

Damage distribution of Mexico City.

397

In a screening procedure, for example, dimensions

of columns and structural walls per floor areas may be
used to roughly estimate lateral load resistance. The
lateral load strength is not a single index to represent
the safety of a building, but gives some idea if the
structure has a sufficient capacity to resist earthquake
motions by strength. Those buildings, identified as
questionable by the simple procedure, must be
analyzed by more sophisticated procedure
The following development and application of
technology are needed to mitigate earthquake disaster
from construction point of view: i.e.:
(a) Effective earthquake resistant building codes for
new construction,
(b) Earthquake vulnerability assessment methods
for existing buildings,
(c) Seismic strengthening technology for vulnerable
buildings,
(d) Seismic damage evaluation methods for
damaged buildings after an earthquake,
(e) Technology to repair damage for immediate
occupancy, and
(f) Technology to rehabilitate damaged buildings
for permanent use.
3.2 Socio-Economic Issues
An examination of allocation of government
resources, in terms of financial and human, for
pre-disaster and post-disaster disaster programs
reveals that more resources are devoted recovery than
to disaster prevention. They commit a large amount of
financial and human support after a disaster. This is
particularly true for the international community.
Disaster prevention programs attract little attention.
However, because emergency operations take place
after a disaster when many lives have already been
lost, only few lives can be saved. In contrast, the
implementation of disaster preventative measures can
potentially save many more lives.
Hence, we must shift our done on seismic resistance
and isolation technologies of high rise buildings, but

398

Evaluation of the Policies for Seismic Retrofit of Buildings

little research is done on conventional houses. But

over 80% of total stock in the world is non-engineered.
Because these unsafe buildings are occupied by
humans, we cannot reduce disaster losses unless we
improve the safety of non-engineered buildings (in
case of Japan, wooden houses). However, they attract
little attention and research funds. Similarly, when
spending habits for new and existing houses are
compared, people tend to spend generously for new
houses but not as much for the maintenance. However,
many more lives can be saved by improving the safety
of existing houses.
Another essential aspect is cost reduction. There are
several ways to achieve this, for instance,
technological development and government subsidies.
It is also necessary to train masons and carpenters on
available techniques.
The political commitment is also crucial. The
rea1493son is that many individual house owners
would pay to reinforce their houses if they understand
the need; but not all house owners would. Everybody
dies eventually. Considering this, the probability of
death from an earthquake, which chance of occurrence
is 40 percent in every 30 years, might seem negligible.
Just like the fact that many smokers wouldnt stop
smoking even if they are told to do so, not all
individual house owners would reinforce their houses.
[9]
One of the key challenges is how to persuade the
elderly who would not invest their money to improve
their old houses. Another challenge is to make the
housewives understand the importance and have
priority in improving the seismic safety of houses.

4. Retrofit Examples in the World

This section shows four retrofitting examples in the
world. In addition to the cases in Japan, the author had
relation with some retrofitting cases in Nepal,
Indonesia, Uzbekistan and China through projects at
the UNCRD Hyogo Office and the Building Research
Institute (BRI).

4.1 Example 1 (Houses in Nepal)

Nepal faces a variety of disaster risks owning both
to its natural characteristics and human induced
factors. Nepal has experienced several major
earthquakes: The Bihar Earthquake in 1934 which
measured 8.3 on the Richter scale killed 4,300 people,
and destroyed 20% of all structures (Earthquake and
Megacities Initiatives, 2005). Three earthquakes of
similar size occurred in Kathmandu Valley in the 19th
century: in 1810, 1833, and 1866. In 1988, there was
another earthquake, which caused to loss of 709 lives
(The National Society for Earthquake Technology
NSET- Nepal).
United Nations Centre for Regional Development
(UNCRD) has carried out a training project in Nepal
in 2007. The training was organized with technical
support from NSET to train practical measures that
can be applied at the house level. 20 female members
from target communities participated in the training.
In the workshop, they learnt the basic science of
earthquakes, importance of disaster risk reduction, and
how to apply non-structural risk mitigation measures
in their homes. For example, they visited several
houses to learn practical ways of securing refrigerators
and shelves by using brackets and props.
After the initial training, follow-up evaluation
meetings were held with the participants. 19
participants reported that they have applied
non-structural measurements in their homes within
one or two weeks after the training by themselves (13
people) and/or with male members in the family (16
people), while there was one person who hired a
handyman. 17 participants reported having talked
about the training with relatives and/or friends, and 15
participants had showed their relatives and/or friends
what they had done in their homes to secure their
furniture.
Furthermore, 14 participants answered that they
know relatives/friends who have implemented such
non-structural risk reduction measures in their homes
after observing their examples. The result showed that

Evaluation of the Policies for Seismic Retrofit of Buildings

there was a strong potential for using womens

network and communication to disseminate disaster
risk reduction strategy. [10]
Though the project has not aimed at retrofitting
itself, fixture of nonstructural part of a house is
recognized as the first step to raise awareness of
community people, especially for housewives who
must play an important role as decision makers for
house maintenance. And this case study shows the
communication
network
among
housewives
effectively works to disseminate the seismic measures
for houses like fixture of furniture (as shown in Fig. 4
and Fig. 5).
4.2 Example 2 (Schools of Indonesia)
Earthquakes in the

past

have

exposed that

Fig. 4
Household assessment of non-structural part
(NSET).

Fig. 5 A trainer of NSET secures furniture using anchors

in a community participants home in Kathmandu (by
UNCRD).

399

vulnerability of school buildings is disproportionately

high compared to the other infrastructures. For
instance, in the 1999 Chi-Chi Earthquake, Taiwan 43
schools in Nantou and Taichung area were completely
destroyed and a total of 700 schools nationwide were
damaged to different extent. The 2001 Gujarat
Earthquake in India caused damages to over 11,600
schools. The 2005 Kashmir earthquake resulted in
collapse of 6,700 schools in North-West Frontier
Province and 1,300 in Pakistan-administered Kashmir
[11]. Recently in May 2008, Wenchuan Earthquake in
China killed about 7,000 students trapped in damaged
school buildings. When an earthquake hit Spitak area
of Northern Armenia during school hours in 1988,
many children lost their lives due to collapse of school
buildings. For example, 285 children out of 302 in
total died at one school. This resulted in almost 2/3 of
total deaths of 25,000 were children and adolescents.
UNCRD and the Center for Disaster Mitigation
(CDM) of the Institute of Technology Bandung (ITB)
conducted a collaborative project to reduce the
vulnerability of existing school buildings in the
corridor of the School Earthquake Safety Initiative
(SESI) project. Two school buildings, SD Cirateun
Kulon II and SD Padasuka II both in Bandung County,
were selected for this project due to the dire needs of
improvement and severe deficiencies of earthquake
resistant systems. The project included retrofitting and
strengthening of school buildings, and other activities
to improve school community preparedness regarding
earthquake.
Prior to conducting any physical work to the
structure, the locations and building layouts were
checked to ensure that the buildings could be
retrofitted. The existing structures were investigated to
determine the type and quality of materials used, as
well as the existing lateral resisting system. Then, the
retrofitting was designed based on the structural
deficiencies/weak parts and their accessibilities,
weighing in retrofit factors on buildings life time,
earthquake resistance capacity, their function, and

400

Evaluation of the Policies for Seismic Retrofit of Buildings

appropriate retrofit strategy/ techniques. The design of

retrofit strategy also considered factors of continuation
of normal function, availability of materials and
skilled construction workers, needs of upgrades for
non structural components, and total costs.
The retrofitting project was first conducted at the
school SD Cirateun Kulon II. The school buildings
consisted of two buildings made of RC frames and
masonry walls. Each building has four rooms. Based
on results from survey and tests, structural analyses
were performed on the existing structures using the
actual material and structural components. Earthquake
risks were introduced to the buildings by applying
loads based on potential seismic risks and local soil
conditions. The analysis showed that both buildings
were considered likely to behave poorly under seismic
loadings, thus required retrofitting. The physical
works were then conducted to improve the structural
quality and reduce the earthquake vulnerability.
Building I which was considered to have lower
quality was retrofitted by adding adequate RC frames
with mat footings. Anchorage was provided to
connect walls with columns and beams. Building II
which was in better condition was retrofitted using
wire mesh for strengthening wall elements (as shown
in Fig. 6). Double tie beams were added adjacent to
the existing one for better foundation system. For both
structures, proper detailing was applied to roof truss
systems, and repair was carried out for nonstructural

Fig. 6

Retrofit work of School Building I (ITB ).

elements such as doors/windows and ceilings.

Finishing/cosmetic repair and improvement of
sanitary facilities were also conducted for both
structures.
4.3 Example 3 (Schools of Uzbekistan)
In Tashkent city, there are more than 360 schools.
Nearly 20% of school buildings have had deficiencies
of different level at present. Preliminary analysis of
seismic risk for Tashkent city showed that more than
25% of school buildings may be completely destroyed
and 30% may be heavily damaged in case of design
earthquake.
The buildings of schools in Tashkent are
represented mainly by two construction systems:
bricks and RC frame-panel consisting major portion of
school building stock and a few buildings made up of
from adobe bricks. Nearly 35% of school buildings
were constructed before Tashkent earthquake of 1966
for design intensity 7 by MSK scale. Since 1966, half
of school buildings were constructed using assembled
RC frames of IIS-04, which are inherently weak in
seismic resistance. The weakness of this construction
type was revealed Spitak (1988), Kairakkum
earthquakes (1985) and also confirmed by through the
engineering analysis of earthquake consequences.
Many school buildings in Tashkent are located in
the zone with slumping soils, and as a result many
buildings, both brick walled and frame panel type are
likely to be damaged. The survey showed that typical
structures used for school building in Tashkent
basically consist of brick works up to 4 stories in old
construction, and reinforced concrete frame-panel for
the more recent buildings (as shown in Fig. 7).
Recurrent structure typologies for school buildings are
categorized in the following three groups:
(1) Mixed type of brickwork and reinforced
concrete or wood reinforcing frame year of
construction 40s;
(2) Brickwork structures, frequent typology used
until 60s;

Evaluation of the Policies for Seismic Retrofit of Buildings

Fig. 7

Retrofit of RC panel school in Tashkent (UNCRD).

(3) Frame-panel, widely used in the modern

construction.
In order to establish an effective and recognizable
linkage to the local professional practice in the Central
Asian region, and to follow the standard analysis
procedure, it is ensured that the characterization is in
compliance with the previous study on the Risk
Assessment of Tashkent city in the framework of
IDNDR RADIUS project in 1990s. [12]
4.4 Example 4 (Buildings of China)
Ministry of Housing and Urban-Rural Development
of Peoples Republic of China (MOHURD) has held
several investigations on the disaster after the
Wenchuan Earthquake. Upon the analysis and
research, MOHURD has constituted Technical Guide
for Appraiser and Strengthening of Earthquake
Affected Buildings on July 23rd, 2008, and promptly
issued Seismic Technical Specification for Building
Construction in Town and Village on June 13th,
2008, and implemented on October 1st, 2008.
The Chinese Standard for Classification of
Seismic Protection of Building Constructions
GB50223-2008 has been enacted on July 30th, 2008.
It has been partially amended on the original
Standard for Classification of Seismic Protection of
Building Constructions GB50223-2004. In addition,
Code for Seismic Design of Buildings
GB50011-2008 (enacted on July 30th, 2008) has been

401

amended on the original Code for Seismic Design of

Buildings GB50011-2001.
Wenchuan
earthquake
provided
abundant
experiences for seismic project constructions. In order
to improve the capability of disaster areas
reconstruction and Chinese project construction
disaster-prevention, MOHURD and experts from
relevant departments have researched and analyzed
earthquake experiences, and have summed up the
seismic engineering research results.
On the above amendments, MOHURD has adjusted
some seismic standard, especially enhancing the
public building seismic standard on secondary &
primary schools and hospitals, extending the seismic
prevention scope of public seismic buildings including
retrofitting works. Besides, MOHURD has also
amended the design code and criteria from the aspects
of specific technologies.
To learn from Japans accumulated seismic building
technology and further promote the application of
seismic technology, to improve the capability of
earthquake and hazardous prevention, and to increase
the technical level in China, Financing & Foreign
Affairs Dept. of MOHURD and Japan International
Cooperation Agency (JICA) has reached an agreement
on the China-Japan Seismic Training Program on
May 12th, 2009. This program will last 3 years and be
fully supported by Ministry of Land, Infrastructure,
Transport and Tourism (MLIT) of Japan.
The International Institute of Seismology and
Earthquake Engineering (IISEE) of the Building
Research Institute (BRI), Tsukuba, Japan is the
implementing agency of the training on assessment,
retrofit and design for seismic building since 2009.

5. Conclusions
The Japanese government has established a law to
promote seismic retrofitting of buildings immediately
after the Great Hanshin-Awaji Earthquake in 1995.
This paper evaluates the policies related to the Retrofit
Promotion Act. The data shows that the policy target

402

Evaluation of the Policies for Seismic Retrofit of Buildings

of seismic safety of buildings (90% of buildings

follow the 1981 seismic code) will be achieved in
20152018 if the current trends of improvement will
be continued. In the field of school buildings, national
government supports the school retrofit works that are
carried out by the local governments, using the
guideline for school retrofit.
However, there are still significant issues to make
all houses safe. One of the key challenges is how to
persuade the elderly who would not invest their
money to improve their old houses. Another challenge
is to make the housewives understand the importance
and have priority in improving the seismic safety of
houses.
Currently many efforts are taken by the Japanese
local governments, such as holding seminars for
communities, preparing financial support schemes,
providing consultancy for seismic assessment and
making earthquake hazard maps. This paper also
provides comments on the improvement of the current
policies for promoting seismic retrofit based on some
international experiences of house and building
retrofit.
Some international cases will help to
develop an effective Japanese policy for promoting
retrofit, such as the case of Nepal shows the role of
females to disseminate the seismic safety of
non-structural part of houses.
The Japanese experiences on seismic retrofit will be
able to contribute to the disaster risk reduction in the
other seismic-prone countries in the world. The
lessons from Kobe and lessons from all over the world
should be disseminated to find solutions towards
challenges to promote policies and actions of seismic
assessment and retrofitting.

References
[1]

Shunsuke Otani, Another problem: Seismic protection of

existing buildings, in: International Symposium on
Earthquake Safe Housing, Proceedings, National
Graduate Institute for Policy Studies (GRIPS), Building
Research Institute (BRI) and United Nations Centre for
Regional Development (UNCRD), 28 Nov., 2008.
[2] The Ministry of Land, Infrastructure, Transport and
Tourism (MLIT), Japan, Promotion of assessment and
retrofitting, power-point presentation, July 2009, Housing
Bureau, MLIT.
[3] Takayuki Nakamura, From model projects to country
-wide intervention, School Earthquake Safety Initiative
(SESI); Reducing Vulnerability of School Children to
Earthquake, 2009, pp. 6879, UNCRD Disaster
management Planning Hyogo Office.
[4] Available online at: http://www.nier.go.jp/shisetsu/pdf/
e-taishinsuishin.pdf.
[5] Available online at: http://www.nier.go.jp/shisetsu/pdf/
e-jirei.pdf.
[6] Shunsuke Otani (2008), GRIPS, BRI, UNCRD
Proceedings of International Symposium 2008, pp.
4047.
[7] Kenji Okazaki, Culture of disaster prevention in the
context of housing and urbanization, in: Proceedings of
UNCRD 2007 International Symposium, 2007, pp.
4547.
[8] UNCRD, Disaster Management Planning Hyogo Office,
Yoko Saito, Edward Sumoto et al., Gender Perspectives
in Community Based Disaster Management (CBDM),
2008.
[9] UNCRD Disaster Management Planning Hyogo Office,
Jishnu Subedi, Ayako Fujieda, School Earthquake Safety
Initiative (SESI); Reducing Vulnerability of School
Children to Earthquake, 2007.
[10] UNCRD Disaster Management Planning Hyogo office,
Shoichi Ando, Jishnu Subedi, Hayato Nakamura et al.,
Reducing Vulnerability of School Children to Earthquake,
2009.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 403414

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Sustainability Impact on Construction Management in

Developing Countries: Iraq as a Case Study
Basil Kasim Mohamed
School of Civil and Construction Engineering, Oregon State University, USA
Abstract: Sustainable development contributes to the economic and social advancement of construction projects. Ignoring
sustainability during the cost estimating process associated with evaluation phase develops many problems of conventional decision
methodology used in construction industry in the most of developing countries. This research focuses on sustainability which is the key
factor in all decision making in design and construction processes. Therefore, Iraqi construction industry should respond to
sustainability as a leading indicator of its infrastructure rebuilding. The construction management and sustainability could help to
manage the rebuilding of the destroyed public infrastructure during the disaster wars periods passed thru thirty years ago and also to
modernize of private buildings and plants. Field study has been carried out by adopting personal interviews and questionnaires
conducted with experts, experienced consultants and designers working in many design organizations. The data collected has been
analyzed and the results obtained show a clear weakness in the field of sustainability understanding, practicing and implementation.
Moreover, the design work in Iraq is suffering from a lack of scientific techniques and a lack in experience in the domain of
sustainability field. This will take an integrated approach to study the impact of sustainability on construction management which
encompasses both public and private sector perspectives for building and engineering construction, and the roles played by all the
participants in the construction team (owners, contractors, design professionals, and other supporting professionals).
Key words: Sustainability concepts, construction, management, engineering, developing countries, Iraq.

1. Introduction
Sustainability is defined by the ISO 15392
international standard as the state in which components
of the ecosystem and their functions are maintained for
the present and future generations (ISO Standard 2008).
Buildings and structures enabled mankind to meet their
social needs for shelter, to meet economic needs for
investment and to satisfy corporate objectives.
However, the satisfaction of these needs usually comes
with a high price, i.e., an irreversible damage to our
environment. This lead to a growing realization around
the world to alter or improve the conventional way of
development into a more responsible approach which
can satisfy the needs for development without harming
the world we live in.
Corresponding author: Basil Kasim Mohamed, PhD,
professor, research fields: project management, quality
management systems, sustainability practices, engineering
education. E-mail: basilkasim555@yahoo.com.

The ideas and strategies initiated by these world

events have prompted positive actions and plans by
many countries to implement this approach in their
industries. Delivering sustainable construction requires
action from all engaged in constructing and
maintaining the structure or building including those
providing design, consulting and construction services
[1]. It requires willingness to explore new fields in
construction approach and prepare to adopt new ideas
and practices [2]. As global interest on sustainability
has steadily blooming, developing countries should not
fell short in its attitude on sustainability and sustainable
construction.
The concept of sustainability in building
construction has evolved over many years. The housing
is now universally recognized as a human right and that
effort to implement this right must be strengthened and
accelerated in the developing countries. Sustainability
in construction is not clearly recognized in most of
developing countries included Iraq. The success and

404

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

progress of human society depends on physical

infrastructure, and a nations economic strength is
reflected in its infrastructure assets [3]. The traditional
design process and construction in these countries
focuses only on cost, quality and time, while
environmental objective could add to these criteria
minimization of resource depletion, reducing harmful
emissions and maintaining a healthy environment
together with conserving natural areas and biodiversity.
That can be shown in Fig. 1 as a new paradigm to
construction. In addition, economic, social and cultural
dimensions of sustainability become very present
especially in the global context [4].

2. Sustainability in Construction
The concept of sustainability in construction has
initially focused on issues of limited resources
especially energy, and on how to reduce impacts on the
natural environment with emphasis on technical issues
such as materials, building components, construction
technologies and energy related design methods. The
appreciation of the significance of non-technical issues
has grown, giving recognition to economic and social
sustainability concerns as well as cultural heritage of
the built environment as equally important [5, 6].
Sustainable construction could be defined as a
construction process which incorporates the basic
themes of sustainable development. Such construction

Fig. 1 Sustainability new approaches [4]

processes
would
thus
bring
environmental
responsibility, social awareness, and economic
profitability objectives to the fore in the pre project
evaluation of built environment assets [7]. The green
building movement is gaining momentum around the
world. One of the biggest challenges facing contractors,
designers and planners is how to ensure that our cities
are developed and regenerated to be sustainable for the
future. Building and construction companies in both
public and private sectors should encourage the
adoption of sustainable and environmental policies and
practices yet there is evidence that in general it is a
concept that is still misunderstood and unsupported by
many project owners [8].
Sustainable construction is a process whereby, over
time, sustainability is achieved. Sustainability should
be applied into construction industry to influence the
manner in which a project could be conducted to strike
a balance between conserving the environment and
maintaining prosperity in development. Attaining
sustainability does not mean the eradication of adverse
impact, which is an impossible vision at present, but
rather the reduction of it to a certain reasonable level
[2].
The direction of the construction industry is now
shifting from developing with environmental concern
as a small part of the process into having the
development process being integrated within the wider

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

context of environmental agenda. Thus, the activities

of construction industry should work and comply with
the needs to protect and sustain the environment.
Previously, the concern on environment is relatively
a small part of most of construction development.
However, with the growing awareness on
environmental protection due to the depletion of
nonrenewable resources, global warming and
extremity of destruction to ecology and biodiversity
impact, this issue have gain wider attention by the
construction practitioners worldwide. The direction of
the construction industry is now shifting from
developing with environmental concern as a small part
of the process into having the development process
being integrated within the wider context of
environmental agenda. Thus, the activities of
construction industry must comply with the needs to
protect and sustain the environment [9].

Presently, the concept of sustainable construction

governs three main pillars: environmental protection,
social well-being and economic prosperity. Fig. 2
shows the tree diagram of these three headings and
their areas of concern [10] where it covers the main
sustainability aspects as indicated in the international
standard ISO 15392, and ISO 16813. Environmental
protection concerns on the built environment and the
natural environment. The built environment refers to
the activities within the construction project itself,
which may, if not handled effectively, have a serious
adverse impact on the environment. Environmental
sustainability is also concerned with the extraction of
natural resources. Although builders have little
influence over the extraction of natural resources, they
can help discourage this activity by demanding less

Sustainability in Construction

Environmental protection

Social Well-being

Environmental protection

Built Environment &

Workers Well-being &

Micro Benefits &

Natural Resources

Community/Users Benefits

Macro Benefits

Location and Land

Health and welfare

Whole life cost

utilisation

Saftey issues

Image/business

Material selection

User comfort/satisfaction

Energy conservation

Accessibility

Legislation compliance

Water quality

Aesthetics/visual

Cost efficiency

Waste minimisation

Nuisance to neighbours

Risk assessment

Pollution control

Social involvement

Biodiversity and
ecology

Fig. 2

The diagram of sustainability in construction [10].

405

enhancement

406

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

non-renewable natural resources, more recycled

materials, and efficient use of energy and mineral
resources [11]. Social well-being concerns with the
benefits of the workers and the future users. Basically,
this aspect is concerned with human feelings: security,
satisfaction, safety and comfort and human
contributions: skills, health, knowledge and motivation
[12].

3. Infrastructure Reconstruction Problems in

Iraq
After the war 2003 in Iraq the infrastructure has been
largely destroyed (about 70%) roads, bridges, oil
refineries, gasoline storage tanks, electric power plants,
health and educational institutes, industrial facilities,
public and residential buildings, communication
system, even village water tanks. After the wars end,
only 20% of Iraqs pre-war electrical generating
capacity had been restored and daily blackouts were a
fact of life. With the destruction of Iraqs power
generating capacity, Iraq was no longer able to run its
water-pumping stations and sewage treatment plants,
depriving the Iraqi people of a hygienic source of water.
In preparation for the October 2003 Madrid Donor
Conference, the joint United Nations/World Bank team
conducted an assessment of funding needs for
reconstruction in Iraq during the period 20042007.
The resulting report identified 14 sectors and
associated funding needs of US$36 billion. The
Coalition Provisional Authority estimated an
additional US$20 billion in need including US$5
billion for security and police and US$8 billion for oil
industry infrastructure [13].
Furthermore, a comprehension housing problems
occurred and in its totality has resulted in a large
proliferation of slums and unauthorized public building
in all Iraqi cities and specially the capital Baghdad.
Also, there were and till now many of unemployed
educated, the destitute, senior citizens and working
women, which when combined, comprise almost 1/4 of
Baghdads urban population [14]. However, the

government hadnt spent the large budget gained from

potential resources on projects for the benefit of these
groups. The demands are growing at a faster rate than
the delivery system. Such disorientation puts severe
environmental constraints upon cities dwellers. The
governments efforts to solve this problem were not
adequate by increasing investments in urban areas,
leading to inadvertent migration resulting in these high
densities, unliveable cities. This problem seems till
now without apparent feasible solution in sight, rolling
along like the veritable snowball, ever increasing in
size [15].
In developing countries it is often not necessary to
implement completely new building methods and
materials in order to provide a safe life at a suitable
human level. Local building practices in Iraq have not
been assessed yet and weaknesses and strengths
identified considering the local type and recurrence of
natural hazards. Simple and expensive buildings were
constructed during the past few years combined with
poor quality construction methods. The materials used
in these constructions were produced in the same
traditional poor technologies without required care that
must be taken to ensure that an adequate skills base
exists for their use, or that training is provided, in order
to avoid increased poor construction. The design of
critical facilities and infrastructure that are essential for
relief and recovery purposes in the event of a disaster
are not given special consideration. The adoption of
hazard-proof criteria set out in codes of practice for
normal structures are not adequate in these buildings
and the non-operation of these facilities was not
socially acceptable. New developments should be
implemented in construction performance based design
of critical facilities to allow for the lower level of
socially acceptable risk [16].

4. Justifications and Objectives

Developing countries wish to see all construction
companies integrate sustainability into their
development approaches to create more sustainable

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

building projects. This will have important

implications for construction companies involved in
social rebuilding and will require placing sustainability
instead of the traditional construction practice. Iraq
looks to achieve economic development to secure
suitable standards of living for people. It is well known
that the consideration of sustainability early in a
projects evaluation processes can result in less of an
increase in capital costs than those made at a later stage.
Therefore, the involved construction companies should
be aware of the importance of early stage of potential
project design decisions.
The developing cities in Iraq like Baghdad is having
a rapid annual growth in construction sector which
contributes about 6 million people. The existing multi
story buildings and many public buildings are
destroyed during the war 2003 and the civil war
20062008. Therefore, the reconstruction campaign
becomes essential to construct new buildings by the
real estate contractors and international construction
companies which can create a significant stress on
water, energy and waste management.
This paper is an attempt to address the sustainable
construction approach and how to be accepted in the
developing countries based on existing socio-economic
and environmental condition. Construction sector in
Iraq has considered its projects to have long life cycle
value especially during infrastructure rebuild. After the
war Iraqi engineers realized that they should work on
processes, services to the users of facilities. Quite
recently infrastructure and housing are stated to be the
main users of resources and the main producers of
environmental burdens. The quality of the built
environment is also claimed to contribute to the quality
of life. Therefore, Iraqi construction sector should
without delay develop its construction processes
towards sustainable direction to remedy the built
environment.
The main research objectives are:

Describing a comprehensive sustainability

concept understanding, practicing, and implementing.

407

Identifying current approaches to building

construction practices.

Identifying the sustainability ideas, evaluation and

continuous improvement in Iraqi companies.

Exploring the reality of sustainability impact on

project construction in Iraq as a developing country.

5. Field Survey and Methodology

The field survey was conducted by participation of
many Iraqi construction companies. These companies
have business contracts and actual involvement in
rebuilding projects in Iraq. A questionnaire was used to
cover the qualitative implementation of sustainability
principles, various concepts, evaluation and continuous
improvement of Iraqi infrastructure. The main focus
was to identify the sustainability impact on the
construction management in the Iraqi companies.
The questionnaire was prepared according to the
main sustainability principles indicated in the literature
review. The survey has been carried out in Iraq to study
the reality of construction management relative to
sustainability impact and to find how the Iraqi
engineers respond to sustainability concepts and
principles, and was conducted for a statistical sample
of 20 (public and private) construction companies in
Iraq.
The research sample included 50 personnel who
have significant role in construction field such as
project managers, company managers, contractors,
professional engineers, social, and economical
professionals. These participants have an active
experience in construction field between 10 to 30 years.
The majority of them have been involved with more
than 20 projects with project cost within the range of 1
to more than 10 million dollars. The respondent
profiles for the field survey for the construction
companies are 70% public sector and 30% private
sector. The questionnaire lists distributed were 70 and
the collected was 20 not complete thus ignored and the
50 complete respondents profile is demonstrated in
Table 1.

408

Table 1
No.
1

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

The participants profile.

Participants
Companies Managers

Numbers

Percentages

10%

Project Managers

12%

General Contractors

12%

Social Professionals

8%

Economy Professionals

8%

Architects

14%

Civil Engineers

14%

Mechanical Engineers

8%

Electrical Engineers

6%

10

Environment Engineers

8%

50

100%

Totals

6. Results and Discussion

The literature review shows that the sustainability
concept is not well defined, recognized, and
implemented in the construction industry in developing
countries. To show how far sustainability concept has
penetrated the construction industry, a field survey has
been conducted among construction companies in Iraq
which focuses on the research sample identifying this
subject and whether they have incorporate knowledge
within their current projects.
The data and information collected from
participating companies were analyzed to recognize the
reality of sustainability concept understanding,
practices, future perspective for implementation, and
the impact of sustainability on various aspects in Iraqi
rebuilding projects.
The data gathered was analyzed qualitatively and
quantitatively. The statistical analysis was used to
calculate straightforward totals, percentages and
averages. Qualitative technique was applied to make
sense of meanings. Interviews were used to connect
statements, opinion and comments to provide a
complete picture. The results are discussed as follows:

7.
Sustainability
Implementation

Understanding

and

The level of knowledge among research sample is

demonstrated in Fig. 3 with regards to sustainability

concept, the respondents were asked to rate their

knowledge on this subject matter and what they
perceived, based on their experience. As shown in Fig. 3
only 18% of the respondents have a moderate
sustainability understanding relies on written materials
or very limited internet websites to improve their
knowledge about sustainable construction. Other
sources of knowledge are through education, seminars
and conferences. Meanwhile the majority of
respondents have 42% of very poor knowledge about
sustainability. The learning level indicates that most of
the companies should promote sustainability concept
within their organizations.
Aligned with the level of knowledge revealed by the
respondents, the level of implementation is also at very
poor level 50% and 35% considered it low as shown in
Fig. 4, and none of the respondents considered that the
implementation of sustainable practices is excellent.
The respondents were requested to give their opinion
on the prospect of sustainable construction application
in Iraq in the next 5 years depending on current
construction projects applications. Fig. 5 shows clearly
that 45% of the respondents believed that in the next
five years, there will not be much change and the level
is still at low level. About 15% of respondents believed
that it will get a moderate level, while 10% of the
respondents considered it would be good, while the
whole respondents believe that no way to get excellent
level.

8.
Traditional
Approach

Construction

Design

The sustainable construction concepts are not clearly

recognized in Iraqi construction field as shown in Fig. 3.
The traditional design approach is still the main
practice in the construction companies in Iraq which
focuses on cost, quality and time. Fig. 6 shows the
importance of each factor. The majority of the survey
respondents 60% considered that cost is the major
factor to be considered during design stage of
construction projects. Regarding to time factor there

Susttainability Imp
pact on Cons
struction Man
nagement in Developing
D
C
Countries:
Iraq as a Case Study
S

was only 255% of the reespondents who

w indicate that
project scheeduling and maintaining the total prooject
duration is very imporrtant in projject construcction
planning. Thhe respondentts didnt focuus on quality as
a an

50%
%

portant factorr to achieve the projects successfullyy

imp
wheere only 15%
% of them consider it important too
con
nform to stanndards and sppecifications stated in thee
projject design.

42%

40%
%
26%

30%
%

18%

20%
%

14
4%

10%
%

0%

0%
%
vverypoorlowmoderate
egood
dexcellent
Fig. 3 Levells of understan
nding sustainab
bility concept.

50%
50%
%
35%

40%
%
30%
%
20%
%

10%
5
5%

10%
%

0%

0%
%
vverypoorlowmoderattegood
dexcelllent
Fig. 4 Impleementation Levvels of Sustain
nability Practicces in Iraq.

50%
%
40%
%

45
5%
30%

30%
%
15%

20%
%

10
0%

10%
%

0%

0%
%
ve
erypoorlowmoderategoodexcelle
ent
Fig. 5 Prosp
pect of Sustainaability Practices in 5 Years in
n Iraq.

4099

410

Susttainability Imp
pact on Cons
struction Man
nagement in Developing
D
C
Countries:
Iraq as a Case Study
S

Requireements

Concepttual

Detailed

Cost

Assesssment

Designn

Design

E
Estimating

Plan
nning
an
nd
Constrruction

Feasibillity

Start
S

Studyy

Up
U

Qualitty
15%
%
Time
25%

Cost
60%

Fig. 6 Tradiitional design approach.

a

9.
Susttainable
Approach
h

Constructtion

Dessign

The research samplee was askedd simultaneoously

about the im
mpact of the sustainability
s
y approach onn the
constructionn managemennt practices to verify their
t
responds tow
wards this neew issue. As shown clearlly in
Fig. 7, the majority
m
of thhe respondennts 55% conssider
that the most important faactor is the ecconomy, whille 35%
of them connsidered the environmentt requirementts in
new sustainaable construcction design are
a important,, and
only 10% off the responddents takes thhe social factoor of
the sustainaable design approach innto consideraation
which indiccates clearly how the dissaster condittions
passed in Iraaq reflect a great
g
ignorancce for humann life
related to prroviding suitaable residentiial homes, Thhese
results show
w that the tradditional consttruction approoach
is still appliicable in the Iraqi construuction compaanies
while the susstainability im
mpact has nott essential bassis to

be practiced or implementedd in construcction design..

Theese companiees could not rrealize that th
he sustainablee
design is an integgrated designn where each component
c
iss
con
nsidered as a critical partt of the whole successfull
susttainable design comparedd with traditiional design..
Earrly decisions have the grreatest impacct on energyy
effiiciency, passsive solar deesign, day lighting,
l
andd
natu
ural cooling. Furthermore,, the sustainaable buildingss
do not have to
t cost morre, nor are they moree
com
mplicated thann traditional cconstruction.
Fig.
F 8 shows a comparisoon between th
he traditionall
and
d sustainabilitty concepts ssimultaneously where thee
bluee line referss to the dom
minant tradittional designn
app
proach whichh is upper than the sustainability
s
y
imp
portance red line. Fig. 9 shows the sustainability
s
y
design approachh where suustainability requirementss
cou
uld be implem
mented durinng the early stage of thee
design process and evaluateed at each design
d
phasee
thro
oughout the value
v
engineeering.

Susttainability Imp
pact on Cons
struction Man
nagement in Developing
D
C
Countries:
Iraq as a Case Study
S

Sustaiinability Req
quirements
Environmentt

Econo
omy

Socciety

411

Finnal
Design

Requuirements

Conceeptual

Preliminarry

Budget

Asseessment

Desiign

Design

Prrogramming

Construction
Plaanning

Vallue

Start
S

Engineeering

Up

Society
S
10%
En
nvironment
35%

Econom
my
55%

Fig. 7 Sustainable design approach.

a

70%
60%
50%

60%
55%

40%
30%
20%

3
35%
2
25%
15%
10%

10%
0%
Fig. 8 Comp
parison betweeen the tradition
nal and sustain
nability concep
pts importancee.

10. Sustain
nability Principles
The conceept of sustainnable construcction that govverns
environmenttal protectioon, social well-being and

d to illustratee
economic prospeerity has beenn investigated
how
w the construuction companies in Iraq
q respond too
thesse areas of concern and tto show the sustainability
s
y
imp
pact on currrent construcction player, where thee

412

Susttainability Imp
pact on Cons
struction Man
nagement in Developing
D
C
Countries:
Iraq as a Case Study
S

respondents indicate cleaarly that these companies could

be acting successfully to implem
ment sustainnable
constructionn by creatingg suitable buiilt environmeents,
and restoringg damaged annd polluted ennvironments..
They reallize that in orrder to achieve sustainabiility,
the construcction manageement policyy in Iraq should
change the processes
p
of creating
c
the buuilt environm
ments.
This could be achieveed by channging the liinear
processes too cyclic processes within the construcction
industry, annd it is shhown in Figg. 9 where the
respondents revealed thhat the of variious environm
ment
aspects have a significaant importance as follow
wing:
(location andd land utilizattion 15%,matterial selectioon 20%
energy consservation 16%
%, water quality 18%, waste
w
minimization 8%, pollution conttrol 13%, and
biodiversity and ecology 10%) for chaanging the waay in
which all thee constructionn activities arre undertakenn.

15%

EnvironmentalAspects
0%
%

Regarding
R
to the social facctors mention
ned in Fig. 2,,
the respondents show the connstruction co
ompanies thatt
invo
olved in real projects to reebuild the infr
frastructure inn
Iraq
q should takke into consideration th
he importantt
susttainability im
mplications of social requirementss
whiich are shownn in Fig. 10, w
where the research samplee
reallized the importance of heealth and wellfare by 20%,,
safeety 16%, user
u
comfort and satisfa
faction 18%,,
accessibility 144%, aesthettics 12%, nuisance too
neig
ghbors 8%, and social involvement 12%
%.
As
A shown preeviously in F
Fig. 6 that thee majority off
the respondents considered thhat cost and the economyy
are the major impportant factorrs whether in traditional orr
in sustainable
s
b
building
consstruction approaches. Thee
resu
ults shown inn Fig. 11 indiccated that thee respondentss
foun
nd the importtance of the eeconomical aspects as

20%
16
6%
18%
8
8%
13%
%
%
10%
2%

4%

6
6%

8%

10%

12%

14
4%

16%

18%

20%

Fig. 9 The im
mportance of the
t environmental aspects.

20%
SocialAspects

16
6%
18%
14%
12%
8
8%
12%

0%
%

2%

4%

6%

Fig. 10 The importance off the social aspects.

8%

10%

12%

14
4%

16%

18%

20%

Susttainability Imp
pact on Cons
struction Man
nagement in Developing
D
C
Countries:
Iraq as a Case Study
S

4133

EconomicalAspects

2
20%
25%
10%
30%
15%
%
0%
%

5%

10%

15%

20%

25%
%

3
30%

Fig. 12 The importance off the economicaal aspects.

following (w
whole life coost 20%, imaage and busiiness
enhancemennt 25%, legisslation complliance 10%, cost
efficiency 300%, and risk assessment 15%).
1

11. Conclu
usions
The main conclusions deduced
d
from
m the researchh can
be illustratedd as followinggs:
(1) There are many reespondents sttress the needd for
enforcementt through law
w and legislaation to incrrease
educational opportunitiees to have more
m
knowleedge
about sustaiinability concepts and im
mplementationn in
constructionn field.
(2) The respondents
r
highlighted the necessityy of
professionallly qualified players in the construcction
industry in Iraq
I
to improove the sustainnability practtices
and implemeentation propperly.
(3) The results
r
showeed that the respondents
r
h
have
poor know
wledge aboutt sustainabillity, where the
learning level of susttainability concepts
c
in the
constructionn companies should be promoted
p
wiithin
their organiizations. Thiis requires cooperation
c
w
with
of
constructionn experts for the development
d
appropriate educational and
a training.
(4) The Iraqi construuction comppanies could not
realize that the integrity of the succeessful sustainnable
mpared with traditional design, and the
design com
sustainable buildings
b
whhich do not haave to cost more,
m

nor are they more compplicated than

n traditionall
con
nstruction.
(5) Iraq is faccing a disaster conditionss and seriouss
neeed to accelerrate sustainaability implem
mentation inn
reco
onstruction of
o the infrastrructure in geeneral and too
raisse the livinng standardds of the populations,,
meaanwhile, reduucing energyy costs, pollu
ution and too
prom
mote the usee of renewabble materials and energiess
thatt should be thhe priority in aany action tak
ken to protectt
the environmentt.
(6) Many resspondents beelieved that sustainability
s
y
can
n be implemented succcessfully in
n the Iraqii
con
nstruction prrojects if thhe managem
ment policiess
orieented by goveernmental finaancial supporrt, where theyy
stilll suppose thaat sustainable practices inccrease projectt
costt.
(7) The responndents show a high consid
deration to thee
imp
portance of various
v
sustaiinability prin
nciples whichh
can
n improve the whole infrrastructure reeconstructionn
projjects in Iraq.
(8) There is a need for a better understaanding of thee
man
nagement peerformance aafter disasteer conditionss
wheere the infr
frastructure aand tradition
nal buildingg
matterials and technologies have to be
b promotedd
reseearch into design of structtures to reducce the effectss
of expected
e
hazaards.
(9) Post-disasster reconstruuction projeccts present a
reall necessity for hazaard-proof measures
m
inn

414

Sustainability Impact on Construction Management in Developing Countries: Iraq as a Case Study

construction and land use planning which sustainability

approach provided. Delegation of national,
international companies and humanitarian agencies
should coordinate in such conditions.

References
[1] WS Atkins Consultants, Sustainable Construction:
Company Indicator, CIRIA C563. London: CIRIA, 2001.
[2] Ofori, C. Briffett, G. Gang and M. Ranasinghe,
Impact of ISO 14000 on construction enterprises in
Singapore, Construction Management and Economics 18
(2000) 935947.
[3] K. Chaharbaghi and R. Willis, Study and practice of
sustainable development, Engineering Management
Journal 9 (1) (1999) 4148.
[4] P. Huovila, Managing the life cycle requirements of
facilities, Proceedings of 8DBMC Symposium,
Performance, Service Life Prediction and Sustainable
Construction, Vol. 3, Vancouver, May 30June 3, 1999,
pp. 18741880.
[5] S. Parkin, Context and drivers for operationalizing
sustainable development, Proceedings of ICE, vol. 138,
Nov. 2000, pp. 915.
[6] S. Parkin, Sustainable Development: the concept
and the practical challenge, Proceedings of the Institution
of Civil Engineers: Civil Engineering, vol. 138 (special
issue 2), 2000, pp. 38.
[7] N. Raynsford, Sustainable construction: The
governments role, Proceedings of ICE, Vol. 138, Nov.
2000, pp. 1622.

[8] D. G. Selmes, Towards sustainability: Direction for

life cycle assessment, PhD thesis, Heriot Watt University:
Edinburgh, UK, 2005.
[9] Hayles, The role of value management in the
construction of sustainable communities, The Value
Manager, vol. 10, No. 1, available Online at:
http://www.hkivm.com.hk /publications/04/TVM2004-.
[10] N. Zainul Abidin, Using value management to
improve the consideration of sustainability within
construction, Ph.D. Thesis, Loughborough University,
United Kingdom, 2005.
[11] B. Addis and R. Talbot, Sustainable Construction
Procurement: A Guide to Delivering Environmentally
Responsible Projects, CIRIA C571, London: CIRIA,
2001.
[12] P. L. Lombardi, Responsibilities towards the
coming generations: Forming a new creed, Urban Design
Studies 7 (2001) 89102.
[13] United Nations/World Bank, Joint Iraq needs
assessment, A report prepared by the Madrid Donor
Conference in October, 2003 UN/World Bank Joint Iraq
Needs Assessment October 2003.
[14] Homeless families and public buildings, A report
published by the Ministry of Immigration, Baghdad, Iraq,
2004.
[15] Housing Problems in Iraq after the War 2003, A
report published by Housing Humanitarian Society, NGO,
Baghdad, 2005.
[16] Agenda 21 for Sustainable Construction in
Developing Countries A Discussion Document, CIB &
UNEP-IETC Publication, CSIR Building and construction
Technology, South Africa, 2002.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 415432

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Modeling Bridge Condition Levels in the United States

Denver Tolliver and Pan Lu
Upper Great Plains Transportation Institute, North Dakota State University, Fargo, ND58105, USA
Abstract: The objectives of this paper are to (1) quantify the effects of age and other key factors on bridge deterioration rates, and (2)
provide bridge managers with strategic forecasting tools. A model for forecasting substructure conditionisestimated from the National
Bridge Inventory that includes the effects of bridge material, design load, structural type, operating rating, average daily traffic, water,
and the state where the bridge is located. Bridge age is the quantitative independent variable. The relationship between age and
substructure condition is a fourth-order polynomial. Some of the key findings are: (1) a bridge substructure is expected to lose from
0.52 to 0.11 rating points per decade as it ages from 10 to 70 years; (2) levels of deterioration increase significantly as the material
changes from concrete, to steel, to timber; (3) slab bridges have lower levels of deterioration than other structures; (4) bridges that span
water have lower condition ratings; (5) bridges with higher operating ratingshave higher condition ratings; and (6) substructure
condition ratings vary significantly among states.
Key words: Highway bridges, bridge condition levels, bridge deterioration rates, statistical forecasting models.

1. Introduction
Nearly 600,000 bridges located on public roads in
the United States are greater than 6.1 meters or 20 feet
in length [1]. As stipulatedin the National Bridge
Inspection Standards, bridges of these lengths are
inspected at least once every two years. During these
inspections, the conditions of the three major bridge
components (deck, superstructure, and substructure)
are rated using a standard scale developed by Federal
Highway Administration, which includes eight interim
levels between excellent and failure (Table 1). A
composite score is assigned by evaluating the severity
of the deterioration of individual bridge elements and
assessing the extent to which the deterioration or
disrepair is widespread throughout the entire
component. A rating of 4 or less indicates poor
condition or worse and typically results in a bridge
being classified as structurally deficient.
Bridge condition assessments and other vital
statistics are compiled in the National Bridge Inventory
(NBI). According to the NBI, approximately 14% of
Corresponding author: Pan Lu, PhD, research fields: GIS
applications in transportation; transportation infrastructural
management system; freight transportation; environmental
analysis; multi-mode transportation. E-mail: pan.lu@ndsu.edu.

the highway bridges in the United States longer than

6.1 meters are classified as structurally deficient (Fig. 1).
A bridge labeled structurally deficient is not
necessarily unsafe or likely to collapse. However, the
bridge will incur additional expenses for increased
inspection, spot maintenance, and repairs in order to
remain in service. Moreover, the maximum gross
vehicle weight may be restricted to less than the legal
limit. An additional 15% of highway bridges are
classified as functionally obsolete, due to substandard
lane or roadway widths and other geometric
shortcomings. Because of growingdeterioration and
obsolescence, the American Society of Civil Engineers
[2] has given Americas bridges a grade of D on scale
from A to E.
According to the NBI, more than one billion vehicles
utilize functionally obsolete or structurally deficient
bridges in the United States each day. However, this
utilization rate does not consider traffic that is diverted
from or routed around substandard bridges and the
additional distances and travel times required. The
average distance to bypass functionally obsolete or
structurally deficient bridges is 20 kilometers,
suggesting that posted, deteriorated, or substandard
bridges may be impeding commerce and mobility.

416

Modeling Bridge Condition Levels in the United States

Table 1 Condition ratings used in the national bridge inventory.

Code

Meaning

Description

Excellent

Very Good

No problems noted.

Good

Some minor problems.

Satisfactory

Structural elements show some minor deterioration.

Fair

Primary structural elements are sound but may have minor section loss, cracking, spalling or scour.

Poor

Serious

Critical

Imminent Failure

Failed

Advanced section loss, deterioration, spalling or scour.

Loss of section, deterioration, spalling or scour has seriously affected primary structural components.
Local failures are possible. Fatigue cracks in steel or shear cracks in concrete may be present.
Advanced deterioration of primary structural elements. Fatigue cracks in steel or shear cracks in
concrete may be present or scour may have removed substructure support. Unless closely monitored it
may be necessary to close the bridge until corrective action is taken.
Major deterioration or section loss present in critical structural components or obvious vertical or
horizontal movement affecting structure stability. Bridge is closed to traffic, but with corrective
action, may put back in light service.
Out of service -- beyond corrective action.

United States Department of Transportation: Recording and Coding Guide for the Structure Inventory and Appraisal of the Nations
Bridges, Washington, D.C., 1995, p. 38.

Fig. 1 Status of highway bridges in the United States.

Computed from the 2009 National Bridge Inventory, including bridges > 6.1 m in length, with culverts excluded.

The objectives of this paper are two folds: (1) to

quantify the effects of key factors on bridge
deterioration rates and (2) to provide bridge managers
with forecasting tools to help them project condition
ratings as a function of age, bridge type, material, and
geographic and jurisdictional location. In the United
States, the Federal Highway Administration has
developed the National Bridge Investment Analysis

System (NBIAS) to analyze and forecast bridge

conditions. The NBIAS analyzes deficiencies at the
level of individual bridge elements (such as beams,
trusses, girders, cables, etc.) using the National Bridge
Inventory. A probabilistic method of modelling bridge
deterioration is used in which transition probabilities
are used to project the likelihood that a bridge element
will deteriorate from its current condition state to a

Modeling Bridge Condition Levels in the United States

lower condition level during a future interval. The

NBIAS assumes that the probability of a bridge
element deteriorating from its current condition to the
next (lower) level is independent of age.
Muchof the research in the area of bridge
managementhas focused on predictions derived from
Markov Chains using transition probabilities. In these
models, the probability that a bridge (or bridge element)
will be in a certain condition at time t1 is a function of
its condition at time t0.While useful, Markov Chainsare
based on simple, often unrealistic assumptions. (1) The
future condition of a bridge is dependent only on its
current condition: i.e., the history of bridge
deterioration in previous periods is not taken into
account. (2) A bridge stays in acondition state for a
predefined time. At the end of each time period the
bridge either moves to a new state or stays in the same
state for another predefined period. (3) The transition
probabilities are constant over time.
Less emphasis has been given tolong-term predictive
models inwhich future bridge conditions are forecasted
from a set of explanatory variables. Nevertheless, in the
literature it is possible to find information that age is a
key determinant of deterioration [35] and
otheressential bridge features contribute significantly
to bridge deterioration [57].

417

Roughly half of the highway bridges in the United

States were constructed prior to 1970 (Fig. 2), so age is
a key consideration in bridge analysis. As noted by the
United States Department of Transportation [1] if a
bridges age (independent of current condition) has an
effect on its deterioration rate this could create a
situation in which bridge investment needs would be
clustered in certain time frames rather than distributed
more evenly over time. To the extent that such spikes
can be anticipated, such information would be very
useful in designing system-wide bridge management
strategies.Before describing amodelthat canbe used
forstrategic planning, some essentialbridge features are
defined: (1) design load, (2) material type, (3)
structural type, and (4) operating rating.

2. Bridge Characteristics
2.1 Structural and Material Types
The National Bridge Inventory classifies bridges
into 23 structural categories. As shown in Fig. 3, the
most common structural types are stringer (53%), slab
(17%), tee-beam (10%), and box-beam/girder (8%).
The remaining 19 categories (which include arch, truss,
suspension, frame, and floor-beam/girder structures)
comprise 12% of all bridges. Fig. 4 shows that concrete

Fig. 2 Year of construction of highway bridges in the United States.

Computed from the 2009 National Bridge Inventory, including bridges > 6.1 m in length, while excluding culverts.

418

Modeling Bridge Condition Levels in the United States

Fig. 3 Classification of highway bridges in the United States by structural type.

Computed from the 2009 National Bridge Inventory, including bridges > 6.1 m in length, while excluding culverts.

Fig. 4 Highway bridges in the United States by dominant material.

Computed from the 2009 National Bridge Inventory, including bridges > 6.1 m in length, while excluding culverts.

is the dominant construction material in more than 57%

of the highway bridges in the United States, while steel
is the primary material in nearly 37% of bridges.
2.2 Design Load
In the United States, the American Association of
State Highway and Transportation Officials (AASHTO)
has defined prototypical vehicles for analyzing bridge
loads. The alphanumeric codes associated with these
vehicles start with H or HS. A prefix of H denotes
a single-unit truck, whereas HS denotes a tractor
pulling a semitrailer. For H bridges, the numeric suffix

represents the gross weight in tons of a single-unit

truck. For example, H-10 denotes a truck with a gross
weight of 10 tons (9.07 metric tons), while H-20
signifies a 20-ton truck. In comparison, the numeric
suffix of HS vehicles represents the assumed weight on
the first two axle sets of the truck: e.g., the weight
distribution to the tractors steering and driving axles.
For example, HS-20 signifies a truck with a total of 20
tons (18.14 metric tons) on the tractors axles and
additional weight on the semitrailers axle. As shown
in Table 2, 49% of the bridges in the United States are
HS-20 or HS-20+ bridges.

Modeling Bridge Condition Levels in the United States

2.3 Operating Rating

The operating rating is the maximum permissible
live load that can be placed on a bridge without a
special permit. Although the maximum gross vehicle
weight on the interstate highway system of the United
States is 40 tons, bridge operating ratings may be
higher or lower than 40 tons. In fact, 85% of the
bridges in the United States (excluding culverts) have
operating ratings of less than 40 tons. Fifty percent of
bridges have operating ratings of less than 24.5 tons.
The operating rating provides unique information
not necessarily reflected in the design classification. As
illustrated in Table 3, H-10, H-15, and H-20 bridges are
distributed differently among rating classes. An H-10
bridge with an operating rating > 40 tons is superior to
an H-10 bridge with an operating rating 20 tons. In
this paper, bridge material, design load, structural type,
and operating rating are used to collectively describe
the quality dimensions of a bridge. However, due
Table 2 Percentages of highway bridges in the united states
designed for specific vehicle loads.
Design Load

Number of Bridges Percent of Bridges

H-10

11,275

2.42%

H-15

66,430

14.24%

H-20

35,701

7.65%

HS-15

10,127

2.17%

HS-20

173,581

37.20%

HS-20 Plus

54,131

11.60%

HS-25

19,728

4.23%

Other

95,628

20.49%

Computed from the 2009 National Bridge Inventory, including

bridges > 6.1 m in length, while excluding culverts.
Table 3 Distribution of operating rating by design load
category.
Operating Rating (Tons)

Percent of Bridges by Design Load

H-10

H-15

H-20

40.2%

6.7%

2.7%

> 20 and 30

19.9%

14.3%

4.9%

> 30 and 40

28.3%

21.8%

17.6%

> 40

11.5%

57.2%

74.9%

20

Computed from the 2009 National Bridge Inventory, including

bridges > 6.1 m in length, while excluding culverts.

419

to the categorical diversity within the other

structures group (which includes 19 different types),
and the limited number of observations in the other
material category (which includes only 1% of the
observations), the analysis focuses on the primary
bridge materials of concrete, timber, and steel and the
four main structural types of stringer, slab, tee-beam,
and box girder.

3. National Bridge Model

A statistical forecasting model of bridge condition
rating is presented in this section of the paper. First, the
model form, main effects, and variables are discussed,
followed by a summary of the results and parameter
estimates. After the estimates are analyzed, methods of
forecasting with the model are introduced.
3.1 Model Form and Main Effects
In a statistical analysis, separate models can be
estimated for each unique category of a dependent
variable. However, in doing so, an implicit assumption
is made. The assumption is that there are no
interactions among categories. While categories of
bridges do not interact physically, they are often
managed by the same agency; thus, all bridges within a
given jurisdiction are subject to a global budget cap.
The approach taken in this study anticipates that the
financial resources and attention given to various
categories of bridges within a state may not be
independent. Consequently, a multivariate model is
formulated includingeight effects: bridge type, design
load, structural type, operating rating, average daily
traffic (ADT group), whether the bridge crosses water,
whether the bridge has been reconstructed, and the
statein which the bridge is located. The values or levels
associated with these effects are summarized in Table 4.
The effects are defined as indicator or indicator
variables. To avoid singularity, only n1 indicator
variables are used to represent a given effect. Each
category has its own unique intercept. However, the
slope (or rate of change in substructure rating with age)
is the same after controlling for all effects.

420

Modeling Bridge Condition Levels in the United States

Table 4

Class level information: national bridge substructure model.

Class Variable

Levels

Values

Reconstructed

0, 1

Bridge Material

Concrete, Steel, Timber

Design Load

H-10, H-15, H-20, HS-15, HS-20, HS-20+, HS-25, Other

Structural Type

Stringer, Slab, Beam, Girder

Operating Rating

20 tons, 20.130 tons, 30.140 tons, > 40 tons

Water Span

0, 1

ADT

<= 100; 101-1000;1001-5000; 5001-19,000; > 19000

State

50

AK AL AR AZ CA CO CT DC DE FL GA HI IA ID IL IN KS KY LA MA MD ME MI
MN MO MS MT NC ND NE NH NJ NM NV NY OH OK OR PA RI SC SD TX UT VA
VT WA WI WV WY

3.2 Dependent Variable

Kim and Yoon [5] studied bridge decks and
superstructures and found that age is the most
significant contributor to the structural deficiency of
decks and bridge superstructures in cold regions,
followed by the structural characteristics of the bridge
and traffic volume. In a complementary way, this paper
focuses on bridge substructures which include
elements such as piers, abutments, footings, and
foundations that transfer loads from the deck and
superstructure to the ground.
Substructure condition (dependent variable of the
model) is treated as an integer-scaled variable using the
scale shown in Table 1. A change of one unit anywhere
on the scale has the same statistical effect.
The interpretation of bridge condition as an
integer-scaled variable is acceptable because the
purpose of this study is to forecast when condition
ratings will change, neglecting the seriousness of the
changes or the need for remedial actions. Nevertheless,
a corollary issue exists: the condition ratings may be
scored differently by different bridge inspectors. To a
certain extent the differences may be reflected in the
state indicator variables, which describe the general
location and administrative jurisdiction of the bridge,
because these differences change geographically
among states or are attributable to inspectors drawn
from different states.

3.3 Functional Relationship between Condition and

Age
The age of each bridge is computed as 2009 minus
the year of original construction. Theoretically, the rate
of bridge deteriorationcan be described by polynomial
function instead of linear function. This hypothesis is
based on two suppositions. (1) When the bridge is
deteriorated to fair condition, maintenance and repairs
are implemented to keep itin acceptable condition.
These

improvements

may

slow

the

rate

of

deteriorationover time. (2) Once a bridge is in serious

condition (described in Table 1) it may continue in
light service, limiting the traffic loadsand making spot
repairs.
A scatter plot of mean substructure ratings against
age exhibits a higher-order polynomial formas shown
in Fig. 5. Examination of the scatter plot in Fig. 5
shows: (1) bridge rating was highly dependent on age;
(2) bridge rating exhibits a clearly higher-order
polynomial relationship with age; and (3) observations
for bridges of approximately the same age are very
close or clustered with each other until around age 120.
After age 120, the observations begin to spread out.
The main reason for bridges that are over 120 year age
mark a wider condition range is the result of the bridges
being built before design rules of standardization.

421

Modeling Bridge Condition Levels in the United States

Rating
9

4
0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

Age

Fig. 5

Scatter plot of mean bridge substructure condition ratings and age.

While these bridges are important, the analysis is

restricted to bridges 120 years of age. Reasons for
this are: (1) bridges built under bridge design
standardizations should be researched separately from
bridges not built to design standards, (2) bridge design
standardizations effect on bridge deterioration is out
the scope of the analysis, and (3) the relative high
variation of bridge ratings for bridges that are more
than 120 years old will jeopardize the quality of the
forecasting model.
3.4 Model Properties
The overall results of the substructure regression
model are summarized in Table 5. F-Value and
Prob(F) statistics test the overall significance of the
Table 5 Regression
substructure model.

summary

for

national

bridge

Number of Observations Used

409,993

Degrees of Freedom (DF)

409,918

F-Value

4,899.6

F-Test for Model Fit (Prob(F). > F)

<.0001

R-Square

0.4694

Adjusted R-Square

0.4693

Coefficient of Variation

14.29

regression model. Specifically, they test the null

hypothesis that all of the regression coefficients are
equal to zero. If the model explains much of the
variation in condition rating, the F-Value will be large
and Prob(F) will be small.
The F-value equals Mean Square Model divided by
Mean Square Error (F = MSM/MSE). Where MSM
equals Sum of Square for Model divided by Degree of
Freedom for Model (MSM = SSM/DFM) and MSE
equals Sum of Square for Error divided by Degree of
Freedom for Error ( MSE = SSE/DFE). The F-Value
ranges from zero to an arbitrarily large number.
The value of Prob(F) is the probability that the null
hypothesis for the full model is true (i.e., all of the
regression coefficients are equal to zero). In this case,
an F-value is 4,900 and the Prob(F) value is <0.0001,
which indicates a less than 1in 10,000 chance that all of
the regression parameters are zero. This would imply
that at least some of the regression parameters are
nonzero and that the regression equation fits the data
with validity. The F-value does not indicate which
variables are most significant and if all of the effects
are statistically meaningful.

422

Modeling Bridge Condition Levels in the United States

3.5 Incremental Sum of Squares Tests for Effects

The bridge model must still be determined if the
theoretical effects are statistically significant in the
NBI dataset. Each variable effect is evaluated using
Type I and Type III sums of squares. Type I sums of
squares are called sequential sums of squares. They are
computed by sequentially adding variables to a model
(one at a time) and computing the reduction in the error
sum of squares attributable to a particular explanatory
variable. Thus, the Type I sums of squares indicate how
much the residual sums of squares are reduced by
adding the particular explanatory variable to the model
that contains all other variables. The corresponding
p-value can test the hypotheses:
H0: the specific regression coefficient = 0 given that
no other independent variables are included in the
model.
HA: the specific regression coefficient 0 given that
no other independent variables are included in the
model.
A p-value that is < 0.0001 indicates a less than 1 in
10,000 chance that the specific regression coefficient is
zero given no other independent variables are included
in the model.
Table 6

In comparison, Type III sums of squares are called

marginal sums of squares. They reflect the incremental
contribution of a specific variable when it is added to a
model that already includes all other variables. For this
reason, Type III effects are often referred to as partial
sums of squares. Thus, the Type III sums of squares
indicate how much the marginal sums of squares are
added when all the other independent variables are
already included in the model. The corresponding
p-value can test the hypotheses:
H0: the specific regression coefficient = 0 given that
all other independent variables are included in the
model.
HA: the specific regression coefficient 0 given that
all other independent variables are included in the
model
A p-value that is < 0.0001 indicates a is less than 1 in
10,000 chance that the specific regression coefficient is
zero given that all other independent variables are
included in the model.
As shown in Table 6, all the sum of squares tests are
highly significant for the all eight main effects with
p-values of <.0001, indicating these variables improve
the explanatory power of the model.

Incremental sum of squares tests for national bridge substructure model.

Source

Type I Sum of Squares

Type III Sum of Squares

Mean Square

F-Value

Pr. > F

MeanSquare

F-Value

Pr. > F

7,395

8,335

<.0001

3,088

3,481

<.0001

Bridge Material

26,387

29,742

<.0001

2,114

2,383

<.0001

Design Load

13,116

14,784

<.0001

727

820

<.0001

804

906

<.0001

29

32

<.0001

9,020

10,167

<.0001

5,127

5,779

<.0001

Water Crossing

740

834

<.0001

274

308

<.0001

State

861

970

<.0001

640

721

<.0001

ADT

1,140

1,285

<.0001

128

144

<.0001

Age

Reconstructed

Structural Type
Operating Rating

75,764

85,398

<.0001

6,332

7,137

<.0001

Age

15,926

17,951

<.0001

1,177

1,326

<.0001

Age

645

727

<.0001

546

615

<.0001

Age

401

452

<.0001

401

452

<.0001

Modeling Bridge Condition Levels in the United States

3.6 Coefficient of Variation and R-Square

The coefficient of variation (CV) is 14.29 (Table 5).
It is computed by dividing the standard error of the
regression by the mean value of the dependent variable
(CV = /) and multiplying by 100 to express this ratio
as a percentage. The CV is a key indicator of the
precision of a model. In this case, the CV is relatively
low, which bodes well for prediction.
The R2 is the ratio of the sum of squares explained by
the regression model to the total sum of squares. A
higher R2 is preferred, ceteris paribus. The R2 value of
0.4694 in Table 5 means that the model explains 47%
of the variation in substructure condition rating.
Although the effects attributable to bridge type, design
load, structural type, operating rating, traffic volume,
water, and general location (i.e., state) are captured in
the model, many individual bridge effects are not; This
could be a reason that the model only explains 47% of
the variation in substructure condition rating. The other
factors that might contribute to substructure condition
rating but are not included in the model can be fixed
bridge effects, such as quality controls and conditions
during initial construction, the inspectors, the
maintenance program used, the amount of maintenance
funds available, the frequency of unusual loadings,
de-icing practices (e.g., how frequently the bridge has
been exposed to chemicals), and effects of flooding.
The reason these other factors are not included in the
model is because they cannot be readily measured or
the data is not readily available.
3.7 Parameter Estimates and Standard Errors
The estimates from the substructure model and their
corresponding standard errors are shown in Table 7. As
shown in Column 3, the standard errors of most
variables are small in relation to the estimated values.
However, the standard errors may be suspect unless the
variance of the regression is consistent over the entire
range of the dependent variable.
In statistics, a collection of variables is
heteroscedastic if the variables do not have the same

423

variance. The issue of non-constant variance, or

heteroscedasticity, is common in regression analysis.
In most instances, the form of heteroscedasticity is
unknown and cannot be ascertained from the data. In
such cases, the variance is said to be inconsistent,
meaning it is not a function of an independent variable
and does not increase or decrease monotonically. The
regression coefficients (i.e., the parameter estimates)
are not biased by heteroscedasticity. However, there
are two potential issues. (1) Regression coefficients
estimated from sample data may no longer be efficient
(e.g., minimum variance estimators). (2) The standard
errors may be affected. As a result, hypothesis tests
may be unreliable.
The first issue is not really a concern for this study
because the parameters are estimated from population
data. Nevertheless, as recommended by Hayes and Cai
[8], heteroscedasticity-consistent errors are used to
assess thepotential effects of inconsistent variance.
These standard errors (shown in Column 6 of Table 7)
are computed under the assumption that the variance is
not constant.
The t statistic is computed by dividing the estimated
value of the parameter by its standard error (t =
parameter estimate/standard error). This statistic is a
measure of the likelihood that the actual value of the
parameter is not zero. The larger the absolute value of t,
the less likely that the actual value of the parameter
could be zero. The Pr. >|t| value is the probability of
obtaining the estimated value of the parameter if the
actual parameter value is zero. The smaller the value of
Pr.>|t|, the more significant the parameter and the less
likely that actual parameter value is zero.
The values shown in Column 2, 3, 4, and 5 are the
ordinary least square (OLS) regression statistics, which
assume
constant
variance
or
known
as
homoscedasticity. The values shown in Column 6, 7,
and 8 are the heteroscedasticityconsistent regression
statistics, which assume non-constant variance or
known as heteroscedasticityand reduce the effects of
heteroscedasticityon inference by employing a
heteroscedasticity-consistent standard error estimator

424

Modeling Bridge Condition Levels in the United States

Table 7 Parameter estimates and probabilities of national bridge substructure model.

Heteroscedasticity Consistent

Variable

Parameter Estimate

Standard Error

t Value

Pr. > |t|

Standard Error

t Value

Pr. > |t|

Intercept

8.37624

0.01311

639.09

<.0001

0.01192

702.76

<.0001

Reconstructed

0.28033

0.00475

59.00

<.0001

0.00509

55.06

<.0001

Water Crossing

-0.07514

0.00428

-17.56

<.0001

0.00382

-19.65

<.0001

Timber

-0.49083

0.00758

-64.73

<.0001

0.00957

-51.27

<.0001

Steel

-0.22225

0.00469

-47.34

<.0001

0.00451

-49.31

<.0001

Material

Concrete
Structure
Beam

-0.05950

0.00639

-9.31

<.0001

0.00618

-9.62

<.0001

Girder

-0.03618

0.00603

-6.00

<.0001

0.00555

-6.52

<.0001

Stringer

-0.02205

0.00500

-4.41

<.0001

0.00480

-4.59

<.0001

H_10

-0.20541

0.01072

-19.16

<.0001

0.01481

-13.87

<.0001

H_15

0.21019

0.00633

33.22

<.0001

0.00721

29.17

<.0001

Slab
Design Load

H_20

0.28143

0.00688

40.88

<.0001

0.00740

38.01

<.0001

HS_15

0.27502

0.01119

24.58

<.0001

0.01198

22.95

<.0001

HS_20

0.36023

0.00540

66.66

<.0001

0.00588

61.28

<.0001

HS_20+

0.30037

0.00698

43.03

<.0001

0.00703

42.71

<.0001

HS_25

0.39663

0.00951

41.71

<.0001

0.00911

43.55

<.0001

20 tons

-0.94360

0.00727

-129.76

<.0001

0.01051

-89.75

<.0001

30 tons

-0.35199

0.00704

-50.02

<.0001

0.00821

-42.89

<.0001

40 tons

-0.09150

0.00490

-18.69

<.0001

0.00515

-17.78

<.0001

Other
Op. Rating

> 40 tons
ADT Class
0-100

0.08893

0.00686

12.96

<.0001

0.00637

13.97

<.0001

101-1,000

0.13321

0.00629

21.17

<.0001

0.00560

23.77

<.0001

1,001-5,000

0.09758

0.00615

15.88

<.0001

0.00530

18.40

<.0001

5,000-19,000

0.03866

0.00588

6.58

<.0001

0.00494

7.83

<.0001

-0.08444

0.00099948

-84.48

<.0001

0.00093242

-90.56

<.0001

> 19,000
Age
Age

0.00139

0.00003811

36.42

<.0001

0.00003763

36.89

<.0001

Age3

-0.00001337

5.390358E-7

-24.80

<.0001

5.559425E-7

-24.04

<.0001

5.33248E-8

2.508156E-9

21.26

<.0001

2.68578E-9

19.85

<.0001

AK

-0.24860

0.03145

-7.90

<.0001

0.04357

-5.71

<.0001

AL

-0.08251

0.01271

-6.49

<.0001

0.01148

-7.19

<.0001

Age

State

To be continued

425

Modeling Bridge Condition Levels in the United States

Continued
AR

0.14248

0.01362

10.46

<.0001

0.01200

11.87

<.0001

AZ

-0.10406

0.01950

-5.34

<.0001

0.01701

-6.12

<.0001

CO

-0.19672

0.01126

-17.48

<.0001

0.00957

-20.56

<.0001

CT

-0.27745

0.01843

-15.06

<.0001

0.01595

-17.40

<.0001

DC

-0.27188

0.07556

-3.60

0.0003

0.06914

-3.93

<.0001

DE

-0.46873

0.04080

-11.49

<.0001

0.03183

-14.72

<.0001

FL

-0.09118

0.01255

-7.27

<.0001

0.00982

-9.28

<.0001

GA

-0.35534

0.01252

-28.39

<.0001

0.01153

-30.83

<.0001

HI

0.15139

0.03310

4.57

<.0001

0.02978

5.08

<.0001

IA

-0.32161

0.01029

-31.26

<.0001

0.01030

-31.21

<.0001

ID

-0.46769

0.01830

-25.56

<.0001

0.01977

-23.66

<.0001

IL

0.05499

0.00988

5.57

<.0001

0.00862

6.38

<.0001

IN

-0.30917

0.01058

-29.23

<.0001

0.00915

-33.80

<.0001

KS

0.16030

0.01057

15.17

<.0001

0.00974

16.46

<.0001

KY

-0.32335

0.01195

-27.06

<.0001

0.01112

-29.07

<.0001

LA

-0.28424

0.01203

-23.63

<.0001

0.01348

-21.09

<.0001

MA

-0.27870

0.01700

-16.39

<.0001

0.01595

-17.47

<.0001

MD

-0.47123

0.01785

-26.39

<.0001

0.01439

-32.75

<.0001

ME

-0.24353

0.02337

-10.42

<.0001

0.02342

-10.40

<.0001

MI

-0.22662

0.01227

-18.47

<.0001

0.01230

-18.42

<.0001

MN

-0.16077

0.01353

-11.88

<.0001

0.01372

-11.72

<.0001

MO

0.34109

0.01044

32.68

<.0001

0.00975

34.99

<.0001

MS

0.08470

0.01318

6.43

<.0001

0.01425

5.94

<.0001

MT

0.06077

0.01624

3.74

0.0002

0.01581

3.84

0.0001

NC

-0.28434

0.01138

-24.99

<.0001

0.01068

-26.62

<.0001

ND

0.04686

0.02071

2.26

0.0236

0.02401

1.95

0.0510

NE

0.51367

0.01188

43.24

<.0001

0.01220

42.09

<.0001

NH

0.17189

0.02415

7.12

<.0001

0.02517

6.83

<.0001

NJ

-0.30251

0.01504

-20.11

<.0001

0.01230

-24.59

<.0001

NM

-0.88481

0.02217

-39.91

<.0001

0.02152

-41.12

<.0001

NV

-0.26525

0.02359

-11.24

<.0001

0.01728

-15.35

<.0001

NY

-0.22312

0.01116

-20.00

<.0001

0.01116

-19.99

<.0001

OH

0.04124

0.00945

4.36

<.0001

0.00878

4.70

<.0001

OK

-0.72676

0.01076

-67.56

<.0001

0.01159

-62.72

<.0001

OR

-0.30992

0.01411

-21.96

<.0001

0.01427

-21.72

<.0001

PA

-0.79479

0.01156

-68.76

<.0001

0.01152

-69.00

<.0001

RI

-0.95154

0.04017

-23.69

<.0001

0.04058

-23.45

<.0001

SC

-0.40671

0.01322

-30.77

<.0001

0.01265

-32.16

<.0001

SD

-0.51615

0.01640

-31.48

<.0001

0.01621

-31.84

<.0001

TX

-0.41609

0.00912

-45.65

<.0001

0.00755

-55.15

<.0001

UT

-0.11995

0.02257

-5.31

<.0001

0.01986

-6.04

<.0001

VA

-0.34138

0.01213

-28.15

<.0001

0.01112

-30.71

<.0001

To be continued

426

Modeling Bridge Condition Levels in the United States

Continued
VT

-0.00301

0.02123

-0.14

0.8873

0.02206

-0.14

0.8915

WA

-0.07969

0.01416

-5.63

<.0001

0.01236

-6.45

<.0001

WI

-0.17566

0.01157

-15.18

<.0001

0.01031

-17.03

<.0001

WV

-0.29327

0.01507

-19.47

<.0001

0.01473

-19.91

<.0001

WY

-0.34206

0.02055

-16.65

<.0001

0.01841

-18.58

<.0001

CA

of OLS parameter estimates. A comparison of Columns

4 and 7 shows only modest differences between the t
values, suggesting mild inconsistency. Onlyone of the
hypothesis tests is affected, based ona probability
threshold of 0.05.
3.8 Probability Values and Inferences
The NBI database constitutes the inventory or
population of publicly-owned bridges in the United
States. Because an inventory is available, sampling
variability is not an issue. Nevertheless, it is beneficial
to envision the NBI as a large sample of bridges that do
(or could) exist. This visualization allows hypothesis
tests that provide intuitive insights concerning the
statistical significance of particular effects. For each
variable, the null hypothesis is that the partial effect
attributable to the variable is statistically insignificant.
This means that the intercept shift attributable to the
variable is not significantly different from zero. For
quantitative variables, the null hypothesis is that the
partial slope coefficients are not significantly different
from zero.
The probability values (or p-values) associated with
the t statistics are shown in Columns 5 and 8 of Table 7,
respectively. With two exceptions, the p-values in
Column 8 indicate that the estimated independent
variables parameters are highly significant and are not
equal to zero (i.e., values of < .01). The p-values of71
of the variables in the model are < 0.0001,
indicatingless than a 1in 10,000 chance of observing t
values as large as those observed. However, the large
p-value of 0.89 for Vermont (variable name is coded as
VT) indicates that the intercept for this state is not
significantly different from the model intercept, which

implies that the interceptsofVermont and California

(the state left out of the model) are not statistically
different. The p-value of 0.051 for North Dakota is
marginally significant (i.e., < 0.10).
As shown in Table 7, the parameter estimates of Age,
Age2, Age3, and Age4 are highly significant. If any of
the terms of the polynomial were unimportant, the
probability values associated with the tratios would be
much higher. The negative sign of age indicates that
bridge condition rating decreases with age. The
positive sign of age squared indicates that the rate of
deterioration is not linear and will slow with age i.e.,
when bridges reach advanced ages, such as 65. The
negative and positive signs associated with Age3 and
Age4 indicate that the regression surface turns up and
down again, as illustrated in Fig. 5.
3.9 Class Variable Effects
In this section, the interpretation of the estimated
parameters

associated

with

material,

structural

typestructural, design load, operation rating, and an

ADT class is introduced.
In the regression model, the effect of a class level is
interpreted in relation to the base level subsumed in the
intercept. As shown in Table 7, the parameter estimates
of steel, timber, and other bridges have negativesigns,
indicating that these bridges deteriorate at faster rates
than concrete bridges (ceteris paribus). Moreover, the
larger negative coefficient of timber indicates that the
expected condition ratings of timber bridges are lower
than the expected condition ratings of steel bridges.
These results are consistent with the findings of Dunker
and Rabbat [6].

427

Modeling Bridge Condition Levels in the United States

The coefficients of stringer, tee-beam, and boxgirder

bridges are negative in relation to slab bridges,
indicating that slab bridges are expected to have higher
condition ratings than other types of structures. These
results are consistent with the findings of Kim and
Yoon [5].
The coefficients of H-15, H-20, HS-15, HS-20,
HS-20+, and HS-25 bridges are positive, meaning that
substructure ratings should be higher over time for
these designs than for bridges included in the other
category. The greatest positive effects are associated
with HS-25, HS-20, and HS-20+ bridges, respectively.
The only negative sign is associated with H-10, which
indicates that these bridges (which have the lowest load
ratings) are expected to deteriorate at faster rates than
bridges included in the other category.
The signs of the operating rating variablesare
negative and must be interpreted in relation to bridges
with operating ratings > 40 tons. The negative signs
suggest that bridge substructure rating is expected to
decrease with operating rating. This is because the
operating rating is a reflection of the capability of a
bridge to accommodate modern truck traffic.The
relative magnitudes of the coefficients make sense too,
because the magnitudes of the effects increase as the
operating ratings decline. However, the difference
between bridges with operating ratings of 30.1 to 40
tons and bridges with operating ratings > 40 tons is
relatively small.
The signs of the ADT class variables (0100;
1011,000; 1,0015,000; and 500119,000 vehicles
per day) are positive in relation to the base level (>
19,000 vehicles per day), suggesting that traffic
contributes to loss of condition rating over time.
Similarly, the negative sign of waterspan indicates that
exposure of bridge piers and foundations to water
results in lower predicted values of condition. However,
the positive sign of reconstruction indicates that, after
reconstruction, a bridges condition rating jumps. In
the regression model, this effect is reflectedin the
intercept shift attributable to reconstruction.

3.10 Multicollinearity
Multicollinearity exists when one or more of the
independent variables are highly correlated with each
other. In multiple regression analysis, multicollinearity
is a question of degree. However, extensive
multicollinearity may create issues. (1) The standard
errors of the estimates may become inflated. As a result,
hypothesis tests may be unreliable. Because of inflated
errors, a variable that is actually important may fail a
hypothesis test. (2) The estimates of the parameters
may be conditional upon other variables. Consequently,
the parameter estimates of several variables may
change if a highly correlated variable is dropped from
or added to the model.
Multicollinearitycan be assessed throughthe
variance inflation factor (VIF), which is computed by
regressing one independent variable against all others
and using the R2 from that regression. Opinions vary
widely about in what degree ofmulticollinearity can be
tolerated. One suggestion is that variables with VIFs
greater than 10 may indicate problems. Other, more
conservative ideas, suggest that VIFs greater than 5.0
should be investigated. The VIF scores of the 70
independent variables associated with the eight effects
range from 1.01 to 4.18 with a median value of 1.48,
suggesting relatively low levels of multicollinearity.
As shown in Table 8, the parameter estimate of age is
relatively robust with respect to the dropping or
exclusion of effects.
As shown in Table 7, the coefficient of age is
-0.08444. For purposes of comparison, revised
estimates of age after dropping the main effects one at a
Table 8 Sensitivity of age to main effect variables.

State

Parameter Estimate Percent Change in

of Age
Estimate
-0.08430
-0.2%

Design Load

-0.08401

-0.5%

Bridge Material

-0.08342

-1.2%

ADT Class

-0.08402

-0.5%

Operating Rating

-0.08182

-3.1%

Structural Type

-0.08471

0.3%

Excluded Effect

428

Modeling Bridge Condition Levels in the United States

time are shown in Column 2 of Table 8. The percentage

change as a result of each dropped effect is listed in
Column 3. As the table shows, the largest change
would occur if the operating ratingis dropped from the
model. Perhaps this is because very old bridges tend to
have lower operating ratings. The second largest
change would result from dropping bridge material,
suggesting there is some collinearitybetween the age of
structures and the material used. This could be
attributable to a shift in construction materials over
time from timber to concrete and steel.
The main inferences from these illustrations are (1)
without controlling for operating rating and bridge
material, the parameter estimate of age would be biased,
and (2) multicollinearity is relatively low, making the
estimate of age somewhat robust. The practical
inference of this discussion is that none of the variables
should be dropped from the model because, in doing so,
information would be lost.In general, the omission of
relevant variables results in biased coefficient
estimates for the explanatory variables left in the model
[9].
3.11 Model Forecast
Because the predictions are ratio scaled, they include
fractional results. In effect, they provide information
about when a bridge is in transition from one condition
level to the next. For example, a bridge with a predicted
condition rating of 6.8 is likely to stay in satisfactory
condition for several years. In contrast, a bridge with a
predicted condition rating of 6.05 is on the verge of
transitioning from satisfactory to fair.
All of the indicator variables must be included in a
forecast because the effect attributable to a coefficient
(e.g., H-20 bridges) reflects all of the base levels of the
other indicator variables subsumed in the intercept
e.g., a concrete slab bridge in California with an
operating rating >40 tons that has not been
reconstructed and does notspan water. If the levels of
any of the indicator variables change from their bases,
these effects must be included in the forecast. For
example, if an H-20 bridge is built with steel, the

coefficients of H-20 and steel must be added to the

intercept to create an adjusted intercept. The
forecasting process is illustrated for an HS-20 box
girder steel bridge in New York that has never been
reconstructed, spans a body of water, has an operating
rating > 40 tons,and average daily traffic of 15,000
vehicles. The specific intercept is computed as
8.376240.07514 0.22225 0.03618+ 0.36023 +
0.03866 0.22312 8.22. Seven terms are reflected in
this calculation: the model intercept (8.37624), the
New York intercept shift (-0.22312), the material type
intercept shift (-0.22225), the bridge design intercept
shift (0.36023), the structure intercept shift (-0.03618),
the water crossing intercept shift (-0.07514) and the
ADT intercept shift (0.03866). There is no adjustment
factor or shift for operating rating. This is because
operating rating of > 40 tons serves as the base
operating rating of the model. Moreover, there is no
intercept shift for reconstruction.
In this process, the parameter estimate of each class
variable is added to the intercept to compute a specific
intercept for the type and design of bridge and traffic
class within the state of interest. In this example, a new
HS-20 steel bridge is predicted to have a condition
rating of 6.35 after 40 years. In other words, the bridge
is expected to be in satisfactory condition with only
minor problems.The condition ratings over time for
this example are shown in Fig. 6.
The change in substructure condition can be
estimated for small changes in age by taking the partial
derivative of the function. However, the rate of bridge
deterioration during any year is relatively small.
Instead of using the derivative, Fig. 7 shows the
projected rates of condition loss for bridges by decade
from 10 to 100 years, as estimated from the model. The
chart resembles a portion of the curve shown earlier in
Fig. 5. Because of the polynomial form, the rate of
condition loss varies with age. A bridge loses 0.52
rating points during the second decade of existence.
However, the rate of loss drops to 0.37, 0.27, and 0.21
in the third, fourth, and fifth decades, respectively. The
expected rates of decline are 0.17, 0.11, and 0.08 for

Modeling Bridge Condition Levels in the United States

429

Fig. 6 Substructure ratings versus age for HS-20 Box Girder Steel Bridge in New York.

Fig. 7

Rate of expected bridge substructure condition loss by decade.

bridges in their sixth, seventh, and eighth decades of

life, respectively.

4. State-Level Model
While the national model could be used to forecast
condition ratings within a particular state, its primary
value is to provide strategic information for policy and
funding considerations. More specific forecasts of
bridge conditions can be derived from a state-level
model, which is illustrated for Iowaa state located in
the north-central region of the United States.
As would be the case with any state, Iowa has some
unique characteristics that warrant adjustments to the

national model. Only four of Iowas bridges are in the

HS-25 category. These observations are deleted due to
insufficient numbers. Less than 0.5% of the
observations fall into the H-10 and H-15 categories.
For statistical reasons, these classifications are
combined into one category.The traffic levels are much
lower in Iowa than in more populatedregions.
Approximately 50% percent of the bridges carry less
than 60 vehicles per day. Given the large percentage of
bridges with low traffic volumes, ADT is highly
collinear with other variables and has a weak effect.
Therefore, it is dropped from the model. In addition,
the highway district variable is substituted for the state

430

Modeling Bridge Condition Levels in the United States

indicator. With these modifications, the remaining

variables are the same as in the national model.
Although the F-tests for the remaining effects are
highly significant, a third-order polynomial fits the
Iowa data better than a fourth-order model.
With an R2 of 0.637, the Iowa model explains nearly
64% of the variation in substructure condition rating
withinthe state. As shown in Table 9, all but two of the
variables in the model are statistically significant with
p-values of < 0.05, including the sub-state district

variables. The VIF scores of the effect variables range

from 1.02 to 3.09 with a median value of 1.79. The
exclusion ofADT from the model drops the R2 slightly,
from 0.6376 to 0.6370, indicating that traffic is not a
significant explanatory variable, given the high
proportion of bridges with low traffic volumes.
The signs of the design load categories are all
positive. As expected, HS-20 and HS-20+ bridges show
the greatest increases in condition ratings relative to
other bridges, and the coefficient of H-20 bridges is

Table 9 Parameter Estimates and Probabilities of Iowa Bridge Substructure Model.

Variable

Parameter Estimate

Standard Error

t Value

Pr. > |t|

Water Crossing

-0.12521

0.02969

-4.22

<.0001

Timber

-0.50655

0.02928

-17.30

<.0001

Steel

-0.24036

0.02237

-10.74

<.0001

Material

Concrete
Structural Type
Beam

-0.43045

0.03350

-12.85

<.0001

Stringer

-0.00281

0.02226

-0.13

0.8995

H-10/H-15

0.22460

0.02449

9.17

<.0001

H-20

0.39009

0.02722

14.33

<.0001

Slab
Design Load

HS-15

0.36686

0.11538

3.18

0.0015

HS-20

0.51128

0.02743

18.64

<.0001

HS-20+

0.71507

0.07222

9.90

<.0001

20 tons

-1.20902

0.03235

-37.38

<.0001

30 tons

-0.40276

0.02688

-14.98

<.0001

40 tons

-0.01122

0.02010

-0.56

0.5769

-0.10687

0.00243

-43.99

<.0001

0.00107

0.00005336

20.12

<.0001

-0.00000347

3.238061E-7

-10.72

<.0001

0.06306

0.02602

2.42

0.0154

0.14217

0.02596

5.48

<.0001

-0.17984

0.02606

-6.90

<.0001

-0.16648

0.02697

-6.17

<.0001

0.14587

0.02704

5.39

<.0001

Other
Op. Rating

> 40 tons
Age
Age

Age3
District
1

Modeling Bridge Condition Levels in the United States

greater than the coefficient of H-10/H-15 bridges. The

coefficients of timber and steel are similar to the
coefficients of those variables in the national model. In
addition, the signs of structural type and operating
rating are the same as in the national model; however,
stringer is not statistically significant, indicating that
the intercepts for stringer and slab bridges in Iowa are
essentially the same. Moreover, there is no significant
difference between bridges with operating ratings of
30.1 tons to 40 tons and bridges with operating
ratings > 40 tons.
In the national model, the state variable controls for
climatic, geographic, political, financial, and
jurisdictional differences among states. At the state
level, the highway district indicator variable serves a
similar purpose, controlling for variations within the
state. While temperature and moisture may vary within
a state, the differences are relatively small compared to
variations ona national scale. The advantage of the
state model is homogeneity of conditions.

5. Conclusion
A national model for estimating substructure
deterioration rates has been developed which has good
statistical properties: e.g., a large F-value with many
highly significant parameters, low multicollinearity,
very mild heteroscedasticity, and a relatively low
coefficient of variation. This model is appropriate for
system and subsystem planning, in which the objective
is to provide agency managers and policy makers with
strategic information: e.g., the expected condition
levels of subsets of bridges. However, a more specific
forecasting model can be developed for each state
using NBI data and the procedures described in this
paper. However, in doing so, some adjustments to
category levels and variables may be necessary.At the
state level, it is possible to supplement NBI data with
local climate data from nearby weather stations [5],
which may reflect local climatological differences
more specifically than the highway district indicator
variable.

431

The relationship between age and substructure

condition is a fourth-order polynomial. Some of thekey
findings are (1) a bridge substructure is expected to
lose from 0.52 to 0.11 rating points per decade as it
ages from 10 to 70 years; (2) levels of deterioration
increase significantly as the material changes from
concrete, to steel, to timber; (3) slab bridges have lower
levels of deterioration than other structures; (4) bridges
that span water have lower condition ratings; (5)
bridges with higher operating ratings have higher
condition ratings; and (6) substructure condition
ratings vary significantly among states.
In conclusion, it is important to summarize some of
the key assumptions and information not available for
this study. (1) The history and timing of maintenance
expenditures for the life of each bridge are unknown. (2)
The condition ratings are scored by different people.
As a result, human variations are reflected in the
evaluations.(3) Many individual bridge effects, such as
the frequency of unusual loadings, deicing practices,
and initial and extreme conditions are not reflected in
the models.(4) Because only current ADT values are
known, traffic is treated as a categorical (rather than a
continuous) variable. Nevertheless, the ADT
classification of a bridge may have shiftedover time as
a result of very large traffic changes. (5) The analysis
includes only stringer, tee-beam, slab, boxgirder,
concrete, timber, and steel bridges, while excluding
culverts. (6) The results are specific to the types of
bridges analyzed.

References
[1]

[2]
[3]

United States Department of Transportation (USDOT),

2008 Status of the Nations Highways, Bridges, and
Transit: Performance Report to Congress, Washington,
D.C., 2010.
American Society of Civil Engineers (ASCE), 2009
Infrastructure Fact Sheet, Washington, D.C., 2009.
M. Jiang and K. C. Sinha, Bridge service life prediction
model using the markov chain, Transportation Research
Records 1223, Transportation Research Board,
Washington, D.C., 1989.

432
[4]

[5]

[6]

Modeling Bridge Condition Levels in the United States

M. J. Kallen and J. M. Van Noortwijk, Statistical
inference for markov deterioration models of bridge
conditions in the Netherlands, in: P. J. Cruz, D. M.
Frangopol and L. C. Neves, Bridge Maintenance, Safety,
Management, Life-Cycle Performance and Cost:
Proceedings of the Third International Conference on
Bridge Maintenance, Safety and Management, 16-19 July,
Porto, Portugal, London, Taylor & Francis, 2006, pp.
535536.
Y. J. Kim and D. K. Yoon, Identifying critical sources of
bridge deterioration in cold regions through the
constructed bridges in North Dakota, Journal of Bridge
Engineering, October, 2010.
K. F. Dunker and B. G. Rabbat, Highway bridge type and
performance patterns, Journal of Performance of
Constructed Facilities 4 (3) (1990) 161173.

[7]

[8]

[9]

S. M. Madanat, M. G. Karlaftis and P. S. McCarthy,

Probabilistic infrastructure deterioration models with
panel data, Journal of Infrastructure Systems, March,
1997.
A.
F.
Hayes
and
L.
Cai,
Using
heteroskedasticity-consistent standard error estimators in
ols regression: An introduction and software
implementation, Behavior Research Methods 39 (4)
(2007) 709722.
W. H. Greene, Econometric Analysis. Pearson-Prentice
Hall, New Jersey, 2008.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 433443

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

State Representation Methodology (SRM) for Bridge

Condition Assessment in SHM
Ayaho Miyamoto1 and Akito Yabe2
1. Graduate School of Science & Engineering, Yamaguchi University, Ube, Japan
2. Seismic Engineering Department., KOZO KEIKAKU Engineering Inc. Tokyo, Japan
Abstract: This paper introduces a new concept of State Representation Methodology (SRM) which is a kind of bridge condition
assessment method for structural health monitoring system (SHM). There are many methods for system identification from the
simplicity comparison of damage index to the complicated statistical pattern recognition algorithms in SHM. In these methods, modal
analysis and parameters identification or many defined indices are common-used for extracting the dynamic or static characteristics of
a system. However, there is a common problem: due to the complexity of a large size system with high-order nonlinear characteristics
and severe environment interference, it is impossible to extract and quantify exactly these modal parameters or system parameters or
indices as the feature vectors of a system in damage detection in an easy way. The SRM considered a more general theory for the
non-parametric description of system state.
Key words: Bridge, damage detection, system state representation methodology (SRM), structural health monitoring (SHM), state
representation function (SRF), SRM Tool, Kernel function.

1. Introduction
In recently, structural health monitoring (SHM) has
received an increasing attention in the field of
deteriorating civil infrastructures, such as bridges,
highway networks, etc [1]. Meanwhile, the condition
assessment is becoming one of the most important
issues in this field. A large size structure such as
bridges produces a huge amount of monitoring data for
every day, every month and every year. It becomes also
an important thing that how to detect damages from
such huge number of monitoring data, that is a big
challenge for analysis to discover the damage
information of a target structure. Furthermore, there are
not only many kinds of sensor data, but also many
undetermined factors in the system, such as the
dynamic loads, ever-changing climate condition, and
severe noise interference. At the same time, we cannot

Corresponding author: Ayaho Miyamoto, PhD, professor,

research fields: bridge engineering, structural concrete
engineering, advanced systems engineering. E-mail:
miya818@yamaguchi-u.ac.jp.

exactly find a model to describe any structure and a

variety of data interrelationship in design stage or
during any operating period. In generally, the large
structure can be viewed as a system, and the large-scale
systems theory or any other new methods can be
employed to research the system. It is obvious that
those large systems are a high-order nonlinear system.
The nature exists a great of complex systems, many of
them have some similar or not. At least, the description
of a system state is an essential problem, and even a
philosophical methodology problem [2].
Fig. 1 shows that how important the condition
assessment in SHM research. It is easy to understand
from this figure that the practical condition assessment
is still a basic problem in SHM research, although there
are many literatures to study on the topics for SHM. In
this paper, it will be focusing on to try to easy
explanation of the analysis methods of condition
assessment in SHM with comparison between
traditional methods and a new proposed method which
called Sate Representation Methodology (SRM) [3,
4]. Then the main content of the paper is consisted

434

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

Test Methods
Local
test:
Local
nondestruc-tive evaluation (NDE)
Global Test: Global structural
health monitoring (SHM).

Research Topics
Sensors
Analysis methods
Application system

Detect Objectives
Level 1: Existence
Level 2: Locating
Level 3: Estimate extent
Level 4: Estimate Lifetime

Model-Based Methods
Damage indices
Model updating

Analysis Methods
Spatial domain
Modal domain
Frequency domain
Time domain

(A) Damage Index Methods

(A1)Damage
location
assurance
criterion (DLAC) and (Multiple) MDLAC
(A2)Modal assurance criterion (MAC),
and (Coordinate) COMAC
(A3) Mode shape curvature (MSC)
(A4)Modal strain energy index (MSEI)
(A5)Element energy quotient (EEQ)

(B) Model Updating Methods

(B1) Optimal matrix
(B2)Eigenstructure
assignment
methods;
(B3)Sensitivity-based
updating
methods;
(B4) Stochastic model
(B5) Regularization methods.

(C) Signal-Based Methods

Auto-Regressive Moving-Average (ARMA)
Empirical mode decomposition and Hilbert-Huang
transform (EMD-HHT)
Wavelet transform (WT)
Singular value decomposition (SVD)

Fig. 1 Data analysis methods for condition assessment in SHM research.

from the basic theory (concept) of Sate

Representation Methodology (SRM) to its
applications to practical Condition Assessment for
bridge structures.

2. SRM Concept Based on Kernel Function

Method
2.1 What is the SRM?
Here, we will be focusing on the structural health
monitoring system (SHMS). Especially, how do we
describe the probability state from the structure
monitoring data? In another view, can we know more
about whats happened & when the monitored signals
have been changing, and what is the relationship
between the system state and the sensors signals? All
of those about detecting and locating are very

important for a practical SHMS. In this paper, we will

be able to give a new idea and systemic methods to
describe and assess the structural state which called as
the system State Representation Methodology (SRM)
[2, 3]. Based on theoretical research of the SRM, we
developed a systemic method to describe damage
feature and state of structural system.
In a complex system such as bridge structures, it
comes some common and basic questions as follows:
(1) What is the present state of a system?
(2) How is it possible to build a model to describe
the system state?
(3) How is it possible to compare the state change of
the system?
(4) How is it possible to extract the state features of
the system?

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

435

(5) What kind of damages are occurring in the

system?
(6) How is it possible to locate a damage of the
system?

in time domain because it is obvious time-varying

2.2 What is the State of a System?

the procedure of feature extracting, in where x is called

The state of a system is interpreted as the overall

response to its internal and external factors which
essentially depend on the response of structure itself or
structural properties and nature environment. Then, the
quantitative assessment of the system state is the
description for system responding to the exciting
factors. If the responses satisfy the suitable values, the
system state is considered as normal state, otherwise as
abnormal (damaged) state. In usually circumstances or
under normal use conditions, the system is in stable
state, this means its state should be a constant, or
fluctuation near a steady state. In general sense, we
therefore assume that it is a steady random variable and
it is usually obeyed the normal distribution.
Assumption: The state of a system is a function of
the system response to environmental exciting.

is the system structure alias parameter. Therefore the

signal. Therefore, we usually translate into the

frequency domain or other transformation domain.
This process is called features extracting. Fig. 2 shows
the system feature vector,

is the state variable,

state of a system can be described by the variable

and it should be a function about the system features

that can be extracted from the responses to its various
factors. In fact, the successful use of the SRM depends
on the experimenters ability to develop a suitable
approximation for the system state function f () ,
which it is called State Representation Function
(SRF). The SRF can be written as,
f ( , x ) 1

2.3 System State Approximating

Fig. 3 shows a basic idea for state variable
expression by approximating methods. The following

Nevertheless, we cannot use the response data directly

Signal Space:
Response
Data

SRM tool

Feature Space:
Feature Vectors

x
SRM

Probability

f ( , x )
Normal

State
SRM
Assessment

Sate Statistic

State Variable:

SRM State

(1)

The SRM tools mentioned in Fig. 2 will be

introduced in Ref. [4].

Obviously, this assumption has so natural sense.

Fig. 2 SRM conceptual model.

436

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

Kernel Function
Expression

Non-Linear
Expression

Linear
Expression

f ( , x ) i k (hi ,x)
i 1

w( k ) x 2 ( k ) ...

k 1

w( k ) x ( k )
k 1

Fig. 3 Schematic diagram of approximating the state variable.

will analyze the main idea of the Kernel Function

expression.
Let w is constant vector in Rn, called as system state
support vector which is related to the system structure
parameters, we take the first-order model to
approximate the Eq. (1) as the following:

Then,

(2)

k 1

Use the Least Squares Estimators (LSE) method, it is

easy to get the weight vector w; if the system has only
one response feature vector h in its initial feature
space (IFS), i.e. we assume that h IFS R n , the
least squares solution of Eq. (2) should be:

(3)

However, we respectively take m feature vectors of

IFS, the least squares solution respectively should be:

wi

hi
hi

, i 1,2....m

(4)

Now let us consider every response feature vector

h IFS as a projector of the current system state
along with each feature direction, then we give them
with a weight: i 0 and

i ( 0 ,1 ),
m

i 1
m

i 1 ,

w i wi i hi , i
i 1

i 1

i 1

i 1

(5)

f ( , x ) i hi , x

(6)

i 1

Eq. (6) is called the first-order representation

function of system state, its a linear operator, where
vector is also called as system alias parameter or
system state parameter, which is relative with the
system structure. Note that Eq. (6) can be rewritten as,
m

f ( , x) i
i 1

One can define a function included parameter vector

, as:

1 w, x w( k ) x ( k )

1 w, x i hi , x i hi , x

x
h x
i ,

hi
hi x

i 1

1, i 0 (7)

If hi x is always assumed, then we have,

m

f ( , x ) i
i 1

hi x
,

hi x

i 1

1, i 0 (8)

Then,
(9)
f ( , x ) 1 for always
Let w k h k , then Eq. (8) can be rewritten
hk
k 1
as:
x
(10)

f ( w , x ) w ,
x
m

The inner product < . > can be replaced by a Kernel

i
hi

Function form as follows. In generally, we can redefine

s, x , s X , x X as,

437

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

s, x (s), ( x) k (s, x), s X , x X

where,

is

vector

space.

(11)
Here

k ( s, x ) ( s ), ( x ) is called a Kernel Function.

There are many choices for a Kernel Function k (,) .
Here,

the

following

Kernel

f ( , x) i k (hi , x)

Functions

are

recommended in [3, 5].

Eq. (14) can be illustrated as Fig. 4, where c is

constant for a system, and c eT G . We can get

m
the alias parameter of system structure, which is
also called as the support vector in SVM.

k ( s , x ) exp(

d ( s, x)

(12)

vector x, i.e., = f(,x). has different value with

here,
n
(s x )2
p
d ( s, x) i i
or d ( s, x) ( si xi )1 / p
si xi
i 1
i 1
n

k ( s, x ) ( s x ) d , ( s 1, x 1 )

Let the system be in state in current time with feature

(13)

Note that is called the SRM scale. In fact, the

difference between two any objects is absolute on any
scales, but their similarity is relative on any scale.
Therefore, scale concept is important to study the
difference between two states of a system.
Subsequently, we can get state function in the Kernel
Function form:
m

f ( , x) i k (hi , x)

different feature vector x because of the system

complexity. We may not know how many state values a
system has, but we usually assume to be a normal
probability distribution. Many statistics methods can
be performed to estimate the probability distribution.
Therefore, we can use in variety of statistic tests such
as F-test and t-test etc. to assess the difference between
virgin state and damaged state in the past time. By
using a laboratory bridge monitoring system (LBMS)
(see [6]) to verify the SRM, experiments show that the
SRM is available and steady to express the system
state.

In order to satisfy that state should be a constant, we

are able to define the following objective function:
min

(I

Subject to

1 T
ee )G
m

1,0 i 1, i 1,2,3,.... m

(14)

here,

2.4 How Differences between the Existing Methods

and the SRM?

k (h1 , h1 ), k (h1 , h2 ),....k (h1 , hm )

k (h2 , h1 ), k (h2 , h2 ),....k (h2 , hm )

G
............................................

k (h , h ), k (h , h ),....k (h , h )
2
m
m
m
m 1

Fig. 5 shows that traditional methods based on

frequency domain decomposition (FDD) are widely
recognized as a simple method [6]. However, it tends to
often lead to the loss of sensitivity and accuracy of
damage detection, because the power spectrum of the
measured responses could not be accurately estimated,
particularly for high-damped systems and systems with
severe modal interference and high noise.

Then, we can get state function as following:

Features Vector

xi , i 1,2...m

c f ( , x )
0 c 1
System

Alias

i , i 1,2...m

Fig. 4 Learning process in SRM.

Parameter

Output

SRM
Sate Learn

System Response
Data Set:
Training Data

438

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

Deteriorating

PSD

PSD
Frequency
Decrease

Frequency

Vibration Characteristic
Change

Frequency

Assessing health condition according to natural frequency, mode

shape and damping etc.
Fig. 5 Frequency domain decomposition (FDD) principle.

Meanwhile, there are some important questions in it:

The natural frequency, corresponding mode shape and
damping coefficient are usually changing very slowly
when the bridge system is deteriorating in the early
stage, in other words, the bridge health condition is
always not sensitive with those parameters. At the
same time, since a complex system include many
structural parts, it has many frequency components,
and they are interfered each other, so it is very difficult
to use FDD to analyze a system in damage detection
because it is impossible to get modal parameters
exactly. Then we proposed a new method based on an
overall view parameter to identify a system. In here, we
assume that the system state is a dynamic variable, and
a probability method with random factors was
introduced as a better method to describe its present
state because it is always in many random conditions
(factors).
Fig. 6 shows the main principle of the proposed
SRM. At first, it needs to change time domain data into
the transformation domain features in the SRM. Then
we are able to derive a system state variable in the

feature space. Furthermore, we need to make the

statistic probability distribution of the derived variable.
Finally, the question of condition assessment of the
present system becomes into a problem of state
assessment.
As stated above, among various damage detection
methods, the vibration-based one is most widely used,
the core idea is that vibration responses are directly
relative with the physical structure and properties, such
as mass, stiffness and boundary conditions, and those
changes of vibration responses can be used to
characterize the structural damage. Although many
intuitive and considerable research efforts have been
devoted on it during the past decades, assessing
structural damage in large-scale bridges still remains a
challenging task for civil engineers. A research [6]
shows that many factors can affect the assessed results,
such as the insensitive of modal properties to local
damage of bridge structures, uncertainty and
incompleteness in measurement data, modal variability
arising from varying operational and environmental
conditions, and modeling errors in the analytical model

439

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

Deteriorating

smaller

State
Moving

Normal State

SRM State
Probability

with

State Distribution

State

State Distribution

Moving

Observing with
bigger scale
State Features Change

SRM State
Probability

SRM State
Probability

Observing
scale

State Distribution

Assessing health condition according to SRM State

Representation Equation
Fig. 6 SRM principle.

etc. Use of physics-based damage detection or

model-based damage detection methods often lead to
update a large number of damage parameters,
especially when the structure has an abundance of
structural members. Therefore they have to reduce their
parameters. One problem which may arise with this
method is that parameter reduction may identify the
most damage-sensitive parameters but may fail to
locate the damage correctly. At the same time, these
same factors often lead to the ill-conditioning of model
updating and damage detection problems, where small
measurement noises could be magnified, this means
that noise or complex factors can always corrupt the
analysis accuracies. However, the sensitivity and the
stability of a method are simultaneously important to
the practical applicability. Meanwhile, many literatures
dont consider these facts that the model updating
condition is dynamic and our knowledge is an
embedded process as the time ongoing. Therefore, not

only requires the improving of approaches for model

updating and damage detection, but also our methods
should be built on a process of observation by means of
continuously accumulated data. The SRM gives the
method driven by monitoring data or by experience
data in inspection. The SRM directly helps to know
what is the difference between virgin state and
damaged state after an observed time. A
nonparametric
structural
damage
detection
methodology based on nonlinear system identification
approaches is presented for the SHM.

3. Application to
Monitoring System

Laboratory

Bridge

3.1 Outline of a Laboratory Monitoring System by

Bridge Model
The bridge model used in this study is a simply
supported girder bridge model with three main girders,
which is the minimum number of girders needed to

440

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

obtain necessary information such as load distribution

characteristics in the direction perpendicular to the
bridge axis. Fig. 7a shows the bridge model prepared
for the purposes of a laboratory bridge monitoring
system (LBMS). As a damage effect capable of
modeling the effects of many types of damage in an
idealized way, a decrease in flexural stiffness was
mainly considered [7]. Girder stiffness reduction was
introduced by reducing lower flange width such as Fig.
7b [no-damaged] and Fig. 7c [damaged]. A total of six
damaged girders with various damage conditions were

prepared for impact hummer dynamic tests. These

girders were replaced with sound (no-damaged) girders
to above mentioned damage conditions.
Single impact load as the dynamic test was applied
to flour points (B1-D3 in Fig. 8) of the bridge model by
using an impact hammer. Accelerometers were
installed in bottom flange of the main girders as shown
in Fig. 8 (acc1-acc9). To enhance the accuracy of the
transfer function, impact loads were applied 10 times at
each point.

(a) Overview of Bridge Model

(b) Main girder without damage

(c) Main girder with damage in lower flange
Fig. 7 Details of bridge model for laboratory bridge monitoring system.

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

acc7
D1

acc4
C1

acc1
B1

acc8
D2

acc5
C2

acc2
B2

acc9
D3

acc6
C3

acc3
B3

shown in Fig. 10 will be applied to get the state variable

as bridge condition assessment. In the algorithm, it is
necessary that the system features need to extract from
the complex responses observed data in the system. A
new time-frequency analysis tool, called Frequency
Slice Wavelet Transform (FSWT) [4] which is
implemented to transform the time domain data into the
time-frequency domain data will be able to powerfully
reveal a change of the characteristics in vibration signal.
Figs. 9b & 9c show an example of Fourier spectrum
and 2D map of FSWT with feature extracting grids on
an impact hummer test data, respectively.
Based on the SRM algorithm as shown in Fig. 10, we
will be able to get the distribution of state variable ,

Fig. 8 Details of bridge model for laboratory bridge

monitoring system.

3.2 Condition Assessment of Bridge Model Based on

SRM

F Spectrum

Accelerationg

Fig. 9a shows an example of acceleration response

of a main girder as the laboratory monitoring data.
Based on the time domain data, the SRM algorithm as
(a)

(b)

Hz
sec

(c)
sec

Timesec

FB100

FB21
FB20

FB1 FB2

SRM
Transform
Tool 2: FSWT
State
Assessment

Fig. 10 SRM algorithm.

Feature Space
Feature Vector
State Space
Determine
State Variable

SRM Tool 1: Compatible

Gradient Algorithm for Large
Scale Linear Problem

Hz
Fig. 9 SRM transform; (a) Original signal, (b) Fourier spectrum and (c) 2D map of FSWT.

Time Domain
Data

441

442

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

that is the state probability distribution, as shown in

Figs. 11 and 12. Here, Fig. 11 shows the results of
comparison between C2 damaged (symmetry state;
see Fig. 8) and no-damaged girders on the SRM state
probability distribution. As the same manner, Fig. 12
shows the results of comparison between C3
damaged (asymmetry state; see Fig. 8) and
no-damaged girders on the SRM state probability
distribution. It is clear that damages like girder

stiffness reduction tend to not only move away the

peak value of SRM state probability distribution from
their normal (original) state (ex. 0.0096 0.0036
for C2 damage & 0.0096 0.0048 for C3
damage) but also change the parameters related to
state variable. Then, it is found that based on these
distributions, we will be able to recognize the
condition changes between the current state and
previous state (normal state) in a deteriorating bridge.

C2 damage scale=0.125
normal state

0.25

C2 damage

0.2

p()

0.15

0.1

0.05

0.0384

0.036

0.0336

0.0312

0.0288

0.0264

0.024

0.0216

0.0192

0.0168

0.0144

0.012

0.0096

0.0072

0.0048

0.0024

Fig. 11 Comparison between C2 damaged and no-damaged (normal) girders on SRM state probability distribution.
C3 damage scale=0.125
normal state

0.25

C3 damage

0.2

p()

0.15

0.1

0.05

0.0384

0.036

0.0336

0.0312

0.0288

0.0264

0.024

0.0216

0.0192

0.0168

0.0144

0.012

0.0096

0.0072

0.0048

0.0024

Fig. 12 Comparison between C3 damaged and no-damaged (normal) girders on SRM state probability distribution.

State Representation Methodology (SRM) for Bridge Condition Assessment in SHM

4. Concluding Remarks

References

This paper introduced the details of a newly

proposed State Representation Methodology (SRM)
and its application to bridge condition assessment
based on the bridge monitoring data. The SRM is a
novel tool that can provide some ideas and algorithms
for data mining in the bridge monitoring system. The
state of a system such as bridge structure could be
obtained by a state variable that calculate from a

[1]

State Representation Equation (SRE). A Kernel

function method which plays an important role in the
Support Vector Machines (SVM) was applied to get
solutions of the SRE. In the computation of the SRE, it
needs to be changed into a Large-Scale Linear
Constraint Problem (LSLCP). And a new
time-frequency analysis tool, called Frequency Slice
Wavelet Transform (FSWT), was able to powerfully
reveal a change of the characteristics in vibration signal.
Finally, an application example in the laboratory bridge
monitoring system was presented so as to demonstrate
how to apply the SRM to practical problems.

[2]

[3]

[4]

[5]

[6]

[7]

443

A. Miyamoto, Usage management of civil structures, in:

C. Boller, F. Chang and Y. Fujino (Eds.), Encylopedia of
Structural Health Monitoring, A John Weley & Sons Ltd,
Publication, UK, Vol. 4, 2009, pp. 16351671.
Z. H. Yan and A. Miyamoto, An idea of state
representation methodology, in: Proceeding of the 2nd
Y-Y Joint Seminar On Strategic Lifetime Management
for Civil Infrastructures with the Latest Information
Technologies, Ube, Japan, Jan. 2008, pp. 2147.
Z. H. Yan and A. Miyamoto, SRM and experimental
study on bridge condition assessment method, in: A.
Miyamoto, S. Hamada and S. H. Lee (Eds.), Lifetime
Engineering of Civil Infrastructure, Research Center for
Environmental Safety (RCES), Yamaguchi University,
Ube, Aug. 2008, pp. 267282.
Z. H. Yan, A. Miyamoto and Z. W. Jiang, Frequency
Slice Algorithm for Modal Signal Separation and
Parameter Identification, Computers and Structures, 2010.
(in press)
G. W. Housner and L. A. Bergman et al., Structural
control: past, present and future, Journal of Engineering
Mechanics 123 (9) (1997) 897971.
H. Wenzel and H. Tanaka, SAMCO Monitoring Glossary:
Structural Dynamics for VBHM of Bridges, Austria,
2006.
A. Nishimura, M. Fujii, A. Miyamoto and T. Kagayama,
Sensitivity of Mechanical Behavior of Bridges for Their
Damage Assessment, Jour. of Structural Mechanics and
Earthquake Engineering 380 (I-7) (1987) 355364.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 444456

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Managing Accessibility The Configurational

Approach to the Inclusive Design of Urban Spaces
Valerio Cutini
Department of Civil Engineering, University of Pisa, Pisa, Italy

Abstract: This paper concerns the theme of the universal accessibility to urban spaces, proposing the contribution of configurational
approach to their inclusive design. Born in the first Nineties on the roots of the efforts and the studies for reducing material and
architectural barriers for disabled people, such discipline extends its sphere and purposes so as to mitigate the effects of spatial
exclusion of a wider and wider variety of human categories, up to virtually include any potential urban user. Several questions persuade
that a merely functional approach, essentially based on the position of the located activities, cannot be said satisfying: only a
configurational approach can account for the effects of the grid configuration on the actual accessibility of its spaces. Our research
focuses on Italian urban settlements, which appear as ideal case studies, due to the typical consistency of their urban spaces, which,
especially in the inner historical cores, are densely build, geometrically irregular and hence generally far from actually providing a
universal material accessibility. Here the impedance of space is much more than a theoretical hypothesis and concrete interventions are
generally requested to eliminate or reduce its barriers and to make space actually and safely usable. Rather than at indicating the
technical solution of single urban problems, which of course are strongly different each other, the research aims at defining a method
suitable for any local context. Such method, integrating into a reliable tool the configurational vision with the functional and
interactional approach, will provide a hierarchy of urban spaces with reference to the necessity of their universal accessibility.
Keywords: Accessibility, configurational analysis, inclusive design, town planning.

1. Introduction
The universal accessibility of the urban spaces is one
of the themes that in the recent years has gone
attracting a growing attention: the progressive ageing
of the population as well as the increasing expectation
of a normal fruition of buildings, spaces and functions
from people of all conditions involve the pressing need
of modifying spaces and paths, in order to upgrade their
actual level of material accessibility. The growth of an
exigency like that goes matching with the cultural
development around the approach to the theme of
disability and accessibility. Once, around the first 80s,
such matter did actually coincide with the theme of the
removal of the architectural barriers, and essentially
consisted in the efforts for reducing the material
obstacles that physically keep the disabled people from
Corresponding author: Valerio Cutini, PhD, professor,
research fields: town planning, urban modeling, spatial analysis,
configurational analysis. E-mail: valerio.cutini@ing.unipi.it.

the actual use of the space, so as to discriminate or even

to exclude them from social life and human relations. It
really was a philanthropic kind of approach, and, as a
matter of fact, it has gone providing remarkable results,
driving the governments of several countries all over
the world to issue laws and regulations aimed at
imposing a minimum threshold, consisting in an
acceptable level of accessibility of some categories of
buildings (such as public offices, public services,
restaurants and hotels). Among them, we can mention
the Americans with Disabilities Act (ADA) of 1990 in
the United States [1], recently modified by the
Americans with Disabilities Act Amendments Act
(ADAAA) of 2008, the Disability Discrimination Act
of 1992 in Australia and the Disability Discrimination
Act of 1995 in the United Kingdom. As time has gone
by, the theme has gone growing up and modifying this
kind of approach, also thanks to the overcoming of a
strict interpretation of the definition of disability. As a

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

consequence, the discipline has extended its sphere and

purposes from the mere protection of disabled people
towards the mitigation of the effects of spatial
exclusion of a wider and wider variety of human
categories: we obviously still refer to disabled people,
but also to elderly people, weak and sick persons (such
as, for instance, heart patients and obese persons), and
even to transitory invalids and occasionally impaired
persons (such as injured persons and pregnant women)
as well as children, up to virtually include any potential
urban user; what hence provides the term accessibility
with the attribute universal. The language in fact has
gone underlining this aspect, substituting in the
common speaking the previous terms barrier-free or
disabled accessibility with other notions (substantially
equivalent each other), such as inclusive design, design
for all, universal design, integral accessibility.
Summing all up, making a space universally accessible
means to make it usable without any limitation by
every kind of people, so that each person, regardless of
his age, sex, social class and specific condition, can be
free of any obstacle that may exist in buildings,
transportation terminals, sidewalks, paths, roads and
vehicles, and then normally behave and feel free from
any spatial restriction and discrimination in his daily
life and relations [2]. In the last decade several
operators, foundations and institutions have been
working on this issue, promoting cultural development,
scientific researches, information, and concrete
interventions on the theme.
Indeed this new kind of approach has given a
stimulus to a large amount of studies on the theme,
which have gone generating a wide set of architectural
solutions, suitable to solve (or, at least, to acceptably
mitigate) a large quantity of problems regarding the
accessibility of urban and architectural space. As an
example, for what specifically concerns the open urban
spaces, solutions have been worked out and largely
experimented for eliminating the steps in the sidewalks
and in the accesses to the buildings, for easily accessing
to parking and recreational areas, for safety crossing

445

vehicular roads, for getting over differences in height,

drops and slopes, for choosing the correct size and
paving materials of sidewalks and pedestrian paths, and
so on. A detailed treatment of such solutions, that
appear irrelevant to the purposes of the present research,
can easily be found in the wide specific literature (in
particular see Refs. [27]).
If then the question regarding the technical way of
enhancing the spatial accessibility can be said plainly
set up, at the moment what concerns the actual
realization of the interventions is instead far from an
acceptable solution. The efforts so far appear in fact
essentially concentrated within the most prominent
public activities, and are aimed at upgrading their inner
accessibility; in such sense, laws and regulations in
several countries are getting more and more strict for
what concerns the inner architectural space, extending
instructions and duties even to private residential
buildings. On the contrary, the grid of the paths which
connect them and all the open spaces mostly remain an
unexplored field, and its elements are still far from
being regarded as universally accessible. The question,
of course, is mainly an economic matter: the
intervention for modifying the public open space in
order to make it accessible would be extremely
expensive, due to the large extension of the whole grid,
to the large amount of any kind of obstacles and to the
frequent presence of constraints and limitations. In
Italy, these problems appear particularly difficult to
face, due to the irregular orographic consistency of a
large part of the urban settlements, which, often located
on the steep hills, can hardly be made actually
accessible. Beside, most of the Italian settlements have
grown around an inner ancient core, often still enclosed
within the original town walls, which is densely and
irregularly build and characterized by tortuous and
narrow paths: here, where most of the prominent urban
activities are located, the intervention for the inclusive
design of public open spaces would be very expensive
and subjected to several architectural duties and
limitations. Working on the theme of the physical

446

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

accessibility of urban space, Italian settlements hence

appear as ideal case studies, because here problems and
difficulties often result so magnified as to be assumed
as paradigmatic. Here the conditions of the context
hinder the pedestrian movement by several kinds of
material barriers and make the spatial impedance much
more than a theoretical assumption or a conceptual
notion; here the universal accessibility is a hard
challenge to face and then an expensive target to gain.
These obvious remarks therefore lead to assume the
universal accessibility of the whole urban grid as an
asymptote, and hence to regard each effort for reducing
exclusion and segregation as a step towards such ideal
goal. Basing on the consequent evidence that the whole
space of a settlement cannot be made all equally and
universally accessible, a fundamental question does
hence arise: which paths should be reclaimed from any
possible material obstacle? In other words, the matter
regards the way of selecting, within the whole grid of
an urban settlement, the parts that ought to be interested
by an intervention aimed at making them universally
accessible. More in general, the question is the way of
calibrating the actual level of physical accessibility
requested by each single part of a settlement, from the
lowest degree up to its highest grade, that is the
universal accessibility; and this question is anything
but obvious, as well as essential for avoiding to waste
resources and for the optimal distribution of the
available budget. In Italy, the first laws regarding the
removal of architectural barriers date back to 1989,
when two acts (L. nr. 13/1989 and DM nr. 236/1989)
essentially imposed a minimum standard of
accessibility to the inner space of new buildings. Later,
a specific kind of urban plan, named PEBA (Piano per
l'Eliminazione delle Barriere Architettoniche, or Plan
for Architectural Barriers Elimination) was established
by other acts (L. nr. 104/1992 and DPR nr. 503/1996),
aimed at extending to the urban spaces an acceptable
level of accessibility. Such plan was specified in
several Italian regions by the local laws, and in
Tuscany, one of the keenest regions on the theme of

disability, a specific act was passed in 1991, imposing

the PEBA to each local municipality. Nonetheless, at
present the guidelines for the construction of those
plans have not yet been passed, just few PEBAs have
been worked out, and the presented ones appear far
from being satisfying, in that they are limited at
monitoring and managing the actual degree of the
present material accessibility of the most used paths;
when, rarely, some concrete intervention is provided,
this is selected on the mere basis of a discretionary and
intuitive evaluation. Its then clear that an objective
method for calibrating the need of accessibility in the
urban grid is definitely advised and highly requested, in
order to make actual the respect of the existing laws
and to avoid the waste of the available resources.
Although obvious, it's worth pointing out that an effort
like this would be unnecessary if only the local
conditions and the economic resources could allow to
plainly extend a universal accessibility all over the grid;
what unfortunately everywhere, but in particular in
Italy, is still so far from happening.

2. The Methodological Approach

The easiest and most frequent approach to the
question is functional: since the most prominent public
activities ought to be located in accessible buildings,
also the paths that connect them (at least the shortest
ones or, even better, the most used and then crowded)
should hence be made universally accessible. Several
faults are hidden behind this apparent solution. First,
which activities along the paths of the grid such
accessible paths should connect? Beside, activities can
be easily and frequently moved or displaced: what then
about the effects of their possible shifting on the
accessibility of the connecting paths? And, if we
assume to preferably connect the most relevant
punctual activities, what about the diffused ones (for
instance shops, offices and minute services), that are
atomized and capillarity scattered along the paths of the
urban grid? And, finally, what about the residential
function, which in most cases appears homogeneously

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

distributed all over the settlement and does generate a

large amount of the origin destination movements to all
the other activities? All these questions are hard to
answer, and make hence evident that a merely
functional approach, essentially based on the presence
and the position of the located activities, cannot be said
satisfying, since it cannot reliably identify the most
used and crowded paths in order to assure them a
universal accessibility. In other words, and more in
general, the grid of the urban paths cannot be seen as
the mere inactive stage where activities mutually
interact; space does matter, in that its configuration
affects the way the space itself is actually used. This
remark induces to set aside the located activities and to
focus on the urban grid, to be seen as a machine for
maximizing contacts and interactions [8]; the analysis
of its configuration will then be expected to provide
significant results with reference to the likely use of the
urban space. Such results could possibly be integrated
with the results of a traditional interactional analysis,
based on the functional consistency of the settlement:
the urban grid with its configuration in a one hand and,
in the other hand, the activities with their position and
their consistency will then allow a calibration of the
relevance of the paths with reference to their actual use
and hence the opportunity of assuring them an
acceptable level of material accessibility.
In the last decade, a remarkable amount of
researches have been applied to test the actual
reliability of the configurational approach. Studies [9]
did verify the existence of a narrow correspondence
between the configurational indices (in particular the
integration value) with the pedestrian movement rates;
further studies [10] did then verify also a strong
correlation of the same configurational parameters with
the actual presence of the located activities. As a matter
of fact, these researches and others have proved the
techniques of configurational analysis as a reliable tool
for reproducing the distribution of the attractiveness
towards the activities, and hence the distribution in a
whole urban settlement of the level of centrality, if we

447

define it in terms of appeal for the location of activities.

Summing all up, on such basis we may assert the
integration value as an indicator of the pure
accessibility, meaning by the attribute pure the
independence of its value from the presence and the
position of the located activities. In other words, the
integration value represents in the configurational
vision what the economic-spatial potential is in the
interactional approach. The economic-spatial potential
of a location a is analytically described by expressions
like the following ones:
Ea = Kj j daj

or

Ea = j exp(-caj)
respectively according to a traditional gravitational
approach [11, 12] or to an entropic vision [13], where
j is the size of the j activity, daj its distance from the
location and caj the respective movement cost. The
economic-spatial potential is also commonly called
generalised accessibility, since this kind of
accessibility is here computed on the basis of the
activities which are actually present, without the
external and axiomatic assumption of a centre,
traditionally defined [14] as the most desirable location
for any kind of activity. Therefore it appears clear that
the generalised accessibility (i.e., the economic-spatial
potential) and the pure accessibility (i.e., the
integration value) respectively represent in the
interactional approach and in the configurational one
the specific indicator of the relevance of each part of a
settlement as an attractor (of movements and activities).
It is singular (and significant) the apparent recurrence,
with different meanings, of the notion of accessibility:
defined in classic modeling as the opposite of the
spatial impedance in the relations between activities,
and then depending on their distribution (generalized
accessibility), assumed in the configurational vision as
determined by the configuration of the urban grid (pure
accessibility) [15], appraised in the inclusive design as
the possibility of using the space without any
limitations (universal accessibility). Linking all up, the

448

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

idea at the base of this research may be summarized

saying that the distribution of both pure accessibility
and generalised accessibility, integrated and taken
together, can usefully indicate for each part of a
settlement the specific opportunity of providing it with
a material universal accessibility.

3. Our Case Study

Pisa has been selected as our specific case study, in
order to apply a configurational approach to support
and guide the construction of a PEBA. Several
particular features concur in making Pisa an excellent
urban laboratory for this kind of test. First, the size of
the settlement: Pisa has got a population of around
100,000 inhabitants, what make it big enough for the
significance of the results, and small enough for
exhausting the analysis and complete the processing of
the method. Second, the presence of a large (when
compared to the extension of the whole settlement)
historic core, whose origins date back to the XIII
century and is at present still substantially unchanged;
though a large part of the most prominent activities is
still within this core, yet its geometric and
morphological features make of course buildings and
open spaces full up with obstacles and then hardly
accessible. Besides, the presence of an enormous
(again, comparatively) University: side by side with the
local inhabitants, more than 50,000 students and about
2,000 teachers and researchers live in Pisa, giving life
to an articulated network of interactions and
movements to the University sites, which of course
ought to be made easily accessible to anybody; a large
amount of strong attractors does hence characterize the
settlement. Finally, the presence of some highly
relevant tourist points of interest, and among them
mainly the leaning tower and its monumental
adjacencies: millions visitors are annually attracted
here from all over the world, and of course also this
mighty attractor ought to be made universally
accessible.
With this purpose, we have focused on the present
urban grid of Pisa, as it can be drawn out of the CTR

(Carta Tecnica Regionale, the 1:2,000 planimetric base

representation of the regional territory provided by the
Tuscan regional administration) setting aside any other
material object except the mere perimeter of the public
spaces, that is those which can be used without any
limit by everyone. The urban grid was then processed
by Depthmap software, submitting its axial map,
consisting of around 1200 lines, to axial analysis. Such
processing did provide as a result a complete
representation of the distribution of the configurational
indices in the lines of the axial map, and, what's most
relevant for our purposes, the distribution of the
integration value all over the grid. As we have seen
above, in fact, integration reproduces the distribution
of the levels of attractiveness towards activities, and
hence the distribution of centrality within an urban
settlement. The distribution of global integration in the
lines of Pisa, as it results from Depthmap processing, is
here shown in Fig. 1.
Furthermore, the same analysis provided the
distribution of local integration (that is radius 3
integration) in the lines of the grid of Pisa. Of course,
those two analyses actually provided far different
results: the representation of global integration, in fact,
shows an evident centripetal distribution, with the
strongest integrators materialized by the prevalent
north-south axis, correspondent to the streets Borgo
Stretto and Corso Italia, and in several orthogonal
streets. These, as well as several less important streets
which

surround

them,

materialize

the

present

integration core of the whole settlement, which is here

represented in Fig. 5. A completely different layout
arises from the representation of the distribution of
radius 3 integration, presented in Fig. 2: here several
strong local integrator appear, variously scattered in
the urban grid from its inner core up to the far edge of
the settlement; on the contrary, the lines resulting main
global integrators in Fig. 1 still maintain high values of
radius 3 integration in Fig. 2 and hence their strength as
main local integrators.
With reference to the theme of centrality and hence

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

449

Fig. 1 Chromatic representation of the distribution of global (R = n) integration in the urban grid of Pisa.

Fig. 2 Chromatic representation of the distribution of local (R = 3) integration in the urban grid of Pisa.

to the functional pattern of the settlement, the

researches so far allow to interpret such a radical
discordance and the different significance of global and
local configurational values: the global integration, as
we have exposed above, was proved [16] able to

describe the level of centrality of each part of a small

size settlement; all the same, if we analyze the whole
urban area of a bigger settlement, we may notice that
correlation of activities versus global integrations does
weaken, while, within limited infra-urban areas, we

450

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

observe a narrow correlation of the same activities

versus the local integration value. In Pisa, we have
selected 120 lines and we have observed the activities
actually located in the correspondent streets; the results
of the analysis of the correspondence of radius 3
integration and density of activities shows that a strong
correlation connect them within local sub-systems;
assuming each of them as defined by the lines
gravitating towards a strong local integrator, in Pisa we
have identified 10 sub-system, and then we have
analyzed 10 different (but analogue) correlations. In
other words, the observed correlations seems to
reproduce the presence of several sub-systems of the
grid; the correlation of activities versus local
integration, that weakens if extended beyond each
sub-system, is on its turn strongly influenced by the
distribution of global integration. Summing all up, we
may observe that in each sub-system the local
integration value IR3 and the density of activities A
appear narrowly correlated according to the following
expression
A = K exp W IR3
where the coefficients K and W stands for the effects of
the global centrality (due to several elements: among
them, the global integration value, the presence of
strong global attractors, the demographic density of the
area and the presence of relevant utilities and amenities)
of the integrator of the sub-system on that correlation.
As a consequence, the expression above can be
represented by mean of several exponential curves, and
each of those curves is referred to the correspondent
sub-system, around its own local integrator. Other
conditions being equal, the higher is the global
integration value of that integrator, the higher in the
(IR3, A) graph we will find the correspondent curve.
Roughly speaking, a clear layout can most likely
account for this correspondence: an urban settlements
characterized by several local centres, with different
levels of global relevance, and hence of global
attractiveness; and within the area surrounding each of
them, a narrow correlation between radius 3 integration

and attractiveness, as materialized by the present

density of activities. In such view, different kinds of
activities are likely to populate the different level
centres: common activities (that is activities which are
not worth long distance movements, such as retail
shops) mainly directed to the peripheral centres, rare
activities (that is addressed to a wider market area and
characterized by a wider range) mainly centre oriented,
that's to say strongly attracted by to the central centre.
In Fig. 3, we may observe the representation of the
exponential curves local integration versus activities in
the defined subsystems, where each of them is
identified by a specific colour.
It can also be argued that pedestrian movements, as
mainly short and exhausting within small distances,
take place in a local urban context and then are more
narrowly described in their distribution by the radius 3
integration rather than by the global one. Down to the
matter of this research, and if we assume the location of
activities as influenced by the distribution of natural
movement and producing an attracted one, its evident
that the efforts for reducing any material obstacle and
for providing the public open spaces with a universal
accessibility ought to be oriented towards the most
attracting urban areas, as the most likely to generate
and receive to and from movement from all over the
grid. And such purpose should be pursued assuring
accessibility both to the global integration core of the
urban settlement and to the local ones.
The obvious limit of this approach is the presence of
non-configurational activities, which are activities that
do not necessarily follow the configurational pattern of
the grid. Among them, by far the most frequent and
relevant are the monopolistic activities, in that those
activities can be located setting aside the distribution of
the configurational parameters, since they wont suffer
any damage in case of a non-appealing location [17].
Many kinds of activities share this enviable feature:
among them, all the public services (schools,
universities, hospitals, administrative offices), but also
the public amenities (parks, museums, monuments and

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

451

Fig. 3 Correlation of local (R=3) integration value versus density of activities in the lines of Pisa. Each colour distinguish one
of the sub-systems.

tourist attractions). And just schools, universities,

hospitals, administrative offices, museums and
monuments are precisely what the town of Pisa is
mainly made of. As a provement, we may get a look at
the correlation integration versus activities mentioned
above, and notice a clear lack of correspondence in the
subsystem nr. 2 (where R2 = 0.546), due to values of
density of activities largely higher than the expectancy
resulting on the basis of integration values. As a matter
of fact, that area is characterized by the presence of
several faculties (Engineering, Chemistry, Pharmacy,
Medicine, Foreign Languages, Informatics: on the
whole more than 25,000 students) of the University of
Pisa, by the presence of an important regional hospital,
the Santa Chiara University Hospital, by the presence
of relevant military (paratroopers) barracks, and, above
all, by the presence of the leaning tower and of the
other monuments of the Piazza dei Miracoli (around
2,500,000 visitors in a year). All these activities,
undoubtedly monopolistic as lacking in competitors in
a wide area around Pisa, of course attract here a large
amount of movement and activities (mainly hotels,
restaurants, souvenirs shops, bookshops, stationery
stores, etc.) although the lines of the area do actually
result characterized by poor values of global

integration. This obvious evidence do persuade that a

PEBA merely oriented onto the configurational pattern
of the grid would inevitably disregard the presence of
the monopolistic activities, and more in general of the
non-configurational ones; what in Pisa should clearly
be dramatically wrong and unacceptable, for two main
reasons: first, those activities are a large and prominent
part of the urban functional structure; moreover the
universal accessibility to those activities, mainly to be
regarded as public and social services, is strongly
requested as a necessary requirement.
It is then clear that the results of the analysis above,
worked out by means of configurational techniques,
ought to be integrated by a further kind of analysis,
based on a functional approach, which is assuming as
an input the presence, the size and the specific position
of the located activities. More in detail, we decided to
integrate the results of configurational analysis, and in
particular the distribution of the pure accessibility
index, with the outputs of an interactional model,
providing the distribution of generalized accessibility;
pure accessibility and generalized one, taken together,
will then been assumed as a reliable indicator of the
crowding of paths and of the consequent necessity of
their universal accessibility. On such basis, we have

452

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

assumed as our input the distribution of economic

activities in the municipality of Pisa, as in Italy it is
provided with a ten-yearly frequency by ISTAT
(Istituto Nazionale di Statistica); we have used the data
of 2001 (the most recent available data), and the
selected values are expressed in the number of
occupied employers and spatially referred to the
singles census tracts. The whole municipality of Pisa is
covered by 918 census tracts, and that partition is
sufficiently minute so as to allow a systemic analysis of
the settlement by means of a gravitational model. On
the basis of the employments data in the census tracts,
we have then calculated the distribution of spatial
economic potential (the generalized accessibility
mentioned above) according to the usual expression
Ea = Kj j daj

what obviously implies the processing of a 918918

symmetric matrix. We have then organized the results
of the processed model by means of a GIS, using as
cells the same 918 ISTAT census tracts, thus obtaining
and presenting by a chromatic representation the
distribution of generalized accessibility in the whole
area of the municipality of Pisa, which is here shown in
Fig. 4.
The several and strong differences with regard to the
distribution of integration (pure accessibility, as we did
assert) can be observed by the comparison of Fig. 1 and
Fig. 4. The same differences may be easily noticed by
the comparison of the spatial economic core with the
integration core deducted by the analysis above. In Figs.
56, its possible to respectively observe the global
integration core and the local integration core of Pisa,
while in Fig. 7 is planimetrically represented the
economic spatial potential of the settlement; in all the
mentioned cases, the marked out areas result
characterized by values over 90th percentile.
More in detail, we may notice that high values of
generalized accessibility (red cells) characterize the
areas along the river Arno, as well as the whole
north-west area of Pisa, where we find Piazza dei

Miracoli, the Santa Chiara Hospital, the faculties of

Engineering, Medicine and Pharmacy; as it can be
easily understood, and could be easily predicted, such
high values depend on the presence of those prominent
activities, and are in contrast with the poor values of
integration. On the other hand, all the main global
integrators result located within high potential areas,
demonstrating that a favourable position on the grid
(from a configurational point of view) actually attracts
activities and makes the correspondent spatial potential
to increase. Its also worth saying that in the processed
model we could not take in sufficient account the role
of movement attractors of several activities and public
services, which result provided with a limited number
of employers with respect to their actual importance
and attractiveness: here we obviously refer to the
monuments of Piazza dei Miracoli, but also to the other
churches and monuments of Pisa, as well as to the
railway station and to the bus station. Therefore, in all
those cases we must consider underestimated the
respective spatial potential. At last, also in the
interactional approach, as we did above in the
configurational method, we have selected and pointed
out an accessibility core, which was assumed as
characterized by the areas over the 90th percentile of
the spatial potential value.
The two different pattern of accessibility can
therefore be integrated, so as to point out the urban
spaces that, either in a configurational vision or in an
interactional approach, result deserving interventions
for providing a universal accessibility. The result is a
grid of paths, extending from the inner core (where
they are, as predictable, far denser) up the edge of the
urban settlement, connecting both the configurational
sub-centres and the most prominent and attracting
located activities.
Obviously, a procedure like that can be regarded as
too rough, in that it assumes a strictly binary logic,
simply dividing all the urban spaces in two categories
and distinguishing them in universally accessible
spaces and inaccessible one. A slightly more detailed
procedure could instead graduate the several levels of

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

453

Fig. 4 Chromatic representation of the distribution of economic spatial potential in the whole area of the municipality of Pisa.

Fig. 5 Global integration core of Pisa.

454

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

Fig. 6 Local (radius 3) integration core of Pisa.

Fig. 7 Economic spatial potential core of Pisa.

material accessibility the urban paths ought to comply

with, according to their respective actual
configurational/interactional accessibility. As an
example, we could define three levels of spatial
accessibility requirement, which could respectively be
named universal accessibility requested, partial
accessibility requested and accessibility non
requested; here the expression partial accessibility

could refer to the mere elimination of the architectural

obstacles and barriers, so as to just allow disabled
people to safety move along the sidewalks (Gallagher,
Scott, 1996). In Fig. 8 we can notice the represented
result of a subdivision like that in the planimetric
layout of Pisa, where the universal accessibility is
requested for the paths with an accessibility value over
the 90th percentile and the partial accessibility is

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

455

Fig. 8 Proposed classification of the public open spaces of Pisa with reference to the required material accessibility: in blue
the universal accessible spaces, in light blue the partial accessible ones.

requested for the lines with an accessibility value from

the 50th up to the 90th percentile; in the exposed
representation, the universal accessibility is required
for the areas which result marked in blue, while the
only partial accessibility is required for those marked in
light blue. Obviously, the proposed subdivision and the
correspondent values are here presented as a mere
example, since they can largely vary according to the
specific features of the local context (that is the actual
presence of obstacles or territorial impediments), to the
available funds for the intervention as well as to the
political willingness to approach the asymptote of the
universal accessibility.

4. Conclusions
The configurational approach to the analysis of
urban spaces is here applied to support the working out
of accessibility plans in order to provide urban
planning with an objective decision making support
tool as well as to overcome the faults and the limits of
the traditional functional vision. Only a configurational
approach, in fact, can account for the effects of the grid
on the distribution of accessibility, and hence on the

distribution of pedestrian movement flows. In such

sense, the experimentation of the configurational
techniques to the urban case of Pisa, for several reasons
here presented as an ideal case study, has provided
outstanding results, as it allows to single out a
widespread network of paths that, due to their
centrality, are likely to get crowded with pedestrian
movement and then chiefly needy of intervention for
enhancing their material accessibility.
At the same time, those techniques make some
physiological limits of the configurational approach to
emerge, in that it cant account for the effects on
movement of the non-configurational activities, and
mainly, among them, of the monopolistic ones, which
in Pisa are numerous, conspicuous and strongly
attractive; it hence makes clear that the configurational
techniques can be suitably integrated by a gravitational
model, based on the interaction relations expected to
connect all the located activities. The results of the
configurational analysis, that is the distribution of pure
accessibility (local and global integration values), and
the results of the processing of a spatial interaction
model, that is the distribution of generalized

456

Managing Accessibility The Configurational Approach to the Inclusive Design of Urban Spaces

accessibility (economic spatial potential), taken

together, indicate the distribution of the actual demand
for the material accessibility of the urban spaces.
Furthermore, beyond the limits of the matter of the
inclusive design of urban spaces, two general results of
this research appear worth noticing. First, the concrete
application of the configurational techniques as a
decision making support tool in the operational urban
planning. Second, the integration of the configurational
techniques with the traditional interactional analysis,
so as to account for the effects of both grid and
activities on urban phenomena.

[7]

References

[13]

[1]

[2]

[3]

[4]
[5]

[6]

Switzer Jacqueline Vaughn, Disabled Rights: American

Disability Policy and the Fight for Equality. Washington,
DC: Georgetown University Press, 2003.
Clarkson John, Roger Coleman, Ian Hosking and Sam
Walter, Inclusive Design Toolkit, Cambridge: Cambridge
Engineering Design Centre, 2004.
Goldsmith Selwyn, Universal Design: A Manual of
Practical Guidance for Architects, Oxford: Oxford
Architectural Press, 2000.
Preiser Wolfgang and Elaine Ostroff, Universal Design
Handbook, New York: McGraw Hill, 2001.
Imrie Rob and Peter Hall, Inclusive Design: Designing
and Developing Accessible Environments, London: Spon.
2001.
Clarkson John, Roger Coleman, Simon Keates and Cherie
Lebbon, Inclusive Design: Design for the Whole
Population, London: Springer, 2003.

[8]
[9]

[10]

[11]

[12]

[14]
[15]

[16]

Burton Elizabeth and Lynne Mitchell, Inclusive Urban

Design: Streets for Life, London: Architectural Press,
2006.
Hillier Bill, Cities as movement economies, Urban
Design International 1 (1996) 4160.
Cutini Valerio, Urban space and pedestrian movement
A study on the configurational hypothesis, in: Cybergeo,
Revue Europenne de Geographie, n 111, 26.10.1999.
Cutini Valerio, Centrality and Land Use: Three Case
Studies on the Configurational Hypothesis, in: Cybergeo,
Revue Europenne de Geographie, n 188, 26.03.2001.
Batty Michael, Recent developments in land use
modeling: A review of British research, Urban Studies 9
(1972) 151177.
G. Alan Wilson, Urban and Regional Models in
Geography and Planning, London: Wiley, 1974.
G. Alan Wilson, Entropy in Urban and Regional
Modelling, London: Pion, 1970.
Alonso William, Location and Land Use, Cambridge:
Harvard University Press, 1964.
Bortoli, Michele and Valerio Cutini, Centralit e uso del
suolo urbano. Lanalisi configurazionale del centro
storico di Volterra, Pisa: ETS, 2001.
Cutini Valerio, Configuration and centrality: Some
evidence from two Italian case studies, in: Proceedings of

the Space Syntax 3rd International Symposium, Atlanta

711 May 2001, 32.1-32.11, Alfred Tauban College of
Architecture and Urban Planning, University of Michigan.
2001.
[17] Cutini Valerio, Working on the edge of town: The
periphery as a spatial pattern in computational science
and its applications, ICCSA, Berlin: Springer, 2008, pp.
2641.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 457469

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

The Kitchen: An Architectural Mirror of Everyday Life

and Societal Development
Claus Bech-Danielsen
Danish Building Research Institute, Aalborg University, Denmark
Abstract: This paper is part of a research project that analyses trends in housing architecture over the past 100 years. The research aims
toshow how changing norms and new forms of everyday life have altered our views on housing and have led to fundamental changes in
housing architecture. In this paper the analysis focuses particularly on the kitchen. A hundred years ago the kitchen of the bourgeoisie
and the middleclass was only used by servants and other employees. Accordingly, the design of the kitchen was not a task for architects
at all. However, during the 20th century the kitchen became an important architectural focal point. In the early part of the century
architects considered it a practical workspaceto beimproved through rational analysis. Later on the kitchen was seen as a space with
great social qualities, and the informal character of the kitchen was developed and exported to the rest of the dwelling. Today the
kitchen has become the central space in many dwellings, but as the dwelling is increasingly being rendered representative value,
modern kitchens are designed with emphasis on their aesthetic appearance. They are life-style kitchens, which demonstrate the good
taste of the residents and reflect their personalities.
Key words: Kitchen design, housing architecture, cultural trends.

1. Introduction
This paper is part of a research project that surveys a
series of momentous changes in residential architecture
in the last 100 years. This was done by focusing on a
number of specific functions of the dwelling,
specifically the kitchen. How did the actual work in the
kitchen and the view of kitchenwork change
throughout the 20th century and how did this affect the
spatial organization of the dwelling and the practical
design of kitchens?
Primary focus is on developments in Danish housing.
However, international developments have also been
included, if they have been important for changes in
Denmark. This means that the relationship between
international developments and Danish developments
varies in the description of the various periods. In
particular during the period 19201930, change was
primarily on the international stage rather than in

Corresponding author: Claus Bech-Danielsen, PhD,

professor, research field: housing architecture. E-mail:
cbd@sbi.dk.

Denmark, but in the post-war period, focus turned to

concentrate on developments in the kitchen in
Denmark, so this section focuses more on Danish
conditions.
The method included literature studies. On the one
hand these studies focused on societal and cultural
trends (health, hygiene, rationality, food culture,
gender, sustainability, consumerism etc.) and on the
other hand they focused on housing architecture,
changing residential conditions as well as kitchen
design. A number of typical house plans of the periods
have also been analyzed. These analyses primarily
focused on the spatial conditions of kitchens the size,
the location and the design of kitchens. The emergence
of conditions typical of each individual period was
often discovered through comparison across the
different periods.
This paper presents some results of the study. It will
be shown how kitchen design and its relation to
everyday life in the dwelling has changed over time.
With focus on the kitchen, it is shown that housing
architecture has undergone drastic changes over the

458

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

last 100 years. Thus it becomes obvious that the

concept of good housing has constantly been
changing. We may imagine that housing today will
meet future housing needs, but in the light of the
previous changes, we have to recognize, that the future
is also likely to bring great changes.

2. Once Upon A Time Kitchens 100 Years

Ago
In the early 1900s it was rare to find a woman from
the middle class or the bourgeoisie in her kitchen. The
American housing researcher Sudjic compares
contemporary kitchens with the engine room of a ferry:
it was a workspace for the crew and a place where
passengers were unwelcome [15]. The kitchen was the
domain of the servants, and the role of the housewife in
relation to work in the kitchen was that of an employer
[4, 15]. Her only contact with the staff was when the
cook or the housekeeper went upstairs to the living
rooms in order to discuss the menu for the day.
Thus, at this time having a nice kitchen was not
associated with status. Cooking was associated with
hard and dirty work: firewood was carried, meals were
prepared from scratch, smoke and dust came from the

cooking range etc., and the status related to the kitchen

was usually based on the number of servants employed
[3]. The servants work in the kitchen was not meant to
be seen, heard or smelled in the rest of the dwelling.
Neither guests nor masters were supposed to concern
themselves with the work in the kitchen, and they met
the servants as less as possible.
In contemporary housing the kitchen was therefore
located far from the primary spaces of the dwelling. In
this context it should be remembered that
contemporary housing had a strong representative
character. In villas the kitchen was typically located in
the basement, next to laundry, coke depot, servants
rooms, etc., and in multi-storey buildings the kitchens
were located in the private section of the dwelling and always oriented towards the backyard (see Fig. 1).
The advantage was that noise, smoke and smell did not
spread to the representative part of the house, but at the
same time it meant that the kitchen was located far
from the dining room. This was inconvenient for the
workflow around cooking and serving, but quite logical
in a period where the comfort of the gentry took
priority over convenient working conditions for the
servants.

Fig. 1 Danish Housing from 1898: The kitchens are oriented towards the backyard, they have scattered furnishing and no
coherent layout.

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

There was ample space in the houses of prosperous

citizens, and kitchens were also relatively spacious.
They served as both work and dining areas for the staff
[18], and a table was usually placed in the middle of the
kitchen around which employees could eat their meals.
A number of rooms were related to the kitchen: a cool
pantry for storing food was a necessary part of
contemporary kitchens and in the large residences there
was even a pantry where the food was arranged before
being carried into the dining room.
Since kitchens were spacious, since it was easy to
find cheap domestic servants, and since no one was
interested in the working conditions of servants, no one
focused on the kitchen as an important workplace.
Kitchen layout was dictated by other factors [14]. For
instance the position of the chimney dictated the
position of the stove, and as the chimney was supposed
to double as a heat source and exhaust duct for more
than one room, the chimney was usually located well
inside the house and away from the windows. Of
course this made it difficult to ventilate the kitchen for
smoke and heat. The sink, on the contrary, hung on the
wall facing the backyard, as the wastewater was
drained directly into the courtyard (until a sewerage
system was implemented).
Usually there were many doorways in a kitchen. One
door served as the entrance to the kitchen typically
from a corridor, another door led to a pantry, and often
other doors would lead to a storage area and a staircase.
The many doorways made it difficult to establish a
coherent layout for the kitchen and scattered furnishing
was emphasized by the fact that kitchen components
were still not arranged to form a cohesive whole in an
overall design [14]. The stove was one single element,
the kitchen sink hung by itself, and each piece of
furniture for storage might be located in different
places, and was not designed specifically for the
particular space.

3. Modernism and Kitchen Design in the

1920s and 1930s
When modernists in the 1920s and 1930s developed

459

modernistic architecture, not only housing for the

wealthy was on the agenda. The architects of
modernism would develop housing for the general
population, and since the general population had no
servants, the kitchen became an architectural topic and
a significant new venture for architects. Previously
architects had dealt with monumental buildings and
palaces for the upper class, and as kitchen work had
never been a part of their daily lives, kitchen design had
never been on the architectural agenda. Thus the task
was entirely new.
The architects of modernism were inspired by
engineers who developed new types of products: cars,
airplanes, bridges, ocean liners, etc. Engineers had
their rational way of working and, as they were more
familiar with new materials and mechanical production
methods, they had taken the lead in designing these
new products. Modernists based their design of new
homes on similar rational considerations, and in so
doing the kitchen was an obvious place to focus.
Specific functions were performed in the kitchen, and
the kitchen was therefore the perfect place to realize the
concept of modernism. It was an ideal place to
demonstrate rational methods and functional analysis
to streamline the workflow and optimize use of space.
The trend had already been developed for some time.
Christine Frederick [5], who was an American
co-editor of Ladies Home Journal, made herself the
protagonist of the work, and in 1919 she published the
book Household Engineering: Scientific Management
in the Home [14]. As the books title suggests,
Frederick found her model in Taylorism, which had
great success in rationalizing industrial enterprises.
Frederick considered the kitchen as a company with
only one employee [12, 14], and as a parallel to the
rationalization experts of the factories, she identified
the workflow in the kitchen in order to save time (see
Fig. 2). She created an alternative kitchen: An efficient
and time-saving kitchen meant only for cooking and
with direct access to the dining area [5, 12]. However,
Fredericks kitchen still consisted of separate elements
that were not pieced together into an overall design.

460

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

Fig. 2 On good and bad examples of floor plans, Christine Frederick [5] demonstrated how cooking could be implemented by
time and step-saving devices.

The ambition to rationalize and streamline work in

the kitchen had its base in a society where it was
becoming ever more difficult for the middle class to
hire servants [14]. Around the First World War, many
women found work in factories where they earned
more money and experienced independence. This left
the middle-class kitchens without servants.
Middle-class women had no experience of kitchen
work, and therefore new tools and facilities were more
than welcome.
Christine Fredericks book became almost a bible
for architects at the Bauhaus working on developing a
modern kitchen. At the first exhibition at Bauhaus in
1923, the German architect Adolf Meyer (18811929)
presented a design of a modern kitchen. As Frederick
had prescribed, the kitchen was rationally organized,
and in line with Bauhaus' standardization efforts and
vision of mass production, it was even conceived as
units of elements with a continuous aesthetic
expression.
Another important effort of modernism also became
visible in Adolf Meyers kitchen. The hygienic
movement had enormous significance for the
development of modernistic architecture [1, 10], and

this topic was obvious in the field of kitchen design.

Thus, Meyers kitchen from 1923 had bright and shiny
surfaces that could be kept clinically clean, and the
traditional rows of plates were replaced by cupboard
space where cooking utensils and tableware could be
kept without gathering dust and without becoming
greasy from gas and frying.
The hygienic movement in architecture is also
evident in other kitchens, designed by contemporary
modernists. Le Corbusier (18871965) pointed out the
importance of using materials that are easy to clean,
and he adhered to this religiously in his housing design.
In the kitchen at Villa Savoye, not only the walls but
also the kitchen tables are covered with white tiles (see
Fig. 3). The kitchen, which previously had the
character of a workshop, transformed into a
laboratory clinically clean. In this context,
electricity was considered a great revelation. It was for
example reflected in the Weissenhof exhibition in 1927,
where Josef Frank thematized the electrical kitchen.
Electricity was highlighted as a clean source of energy,
without smoke and dirt like gas or charcoal.
Furthermore, electricity made the position of
appliances independent of chimneys etc. The electric

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

kitchen could therefore be arranged optimally on the

basis of the functional analyses.
The Frankfurt kitchen (Fig. 4) was the most famous
example, where the modernistic ideals of kitchen
design were realized and mass produced [14, 15]. It
was developed in 1926 by the Austrian architect
Margarete Schtte-Lihotzky (18972000). She was
part of a team around Ernst May, who in 1920 were
tasked with building social housing in Frankfurt.
Schtte-Lihotzky developed a number of standard
kitchens, which were used in many houses [15]. The
kitchens were simple and cheap, and they were
constructed on the basis of an analysis of workflow and
storage needs. Spatial dimensions were also
determined in order to optimize workflow. In the
Frankfurt kitchen the rooms were typically 1.90 m x
3.44 m and similarly narrow and deep kitchens became

461

common in contemporary housing also in Denmark

(see Fig. 5).

Fig. 3 In Villa Savoye of 1928, the kitchen is designed in

line with the contemporary ideals of clinical hygiene.
Everything is white and the tiles on the kitchen table are
easy to clean.

Fig. 4 The Frankfurt kitchen was designed for social housing in Frankfurt. It was the first mass-produced kitchen with a
comprehensive design. Note the adjustable chair.

Fig. 5 A kitchen in the Danish social housing estate Degnegaarden from 1936. The deep and narrow space goes back to the
Frankfurt kitchen. Degnegaarden was designed by Architects Cooperative - Alex Olsen, VagnKaastrup and Ole Buhl.

462

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

In pure modernism the individual space in housing

was optimized in relation to one specific function, and
rationalization efforts were to dispel anything that was
not related to traditional kitchen work, i.e., anything
other than cooking and dishwashing. However, in
smaller houses this requirement was often compromised.
Schtte-Lihotzky developed a range of multifunctional
solutions, such as a bathtub with a lid that could double
as workstation when not in use [15]. On the one hand,
inventive and innovative, on the other hand a sign that
Schtte-Lihotzky still had roots in the traditional use of
the kitchen for more than just cooking.
Modernist ideals of kitchens arrived rapidly in
Denmark. It was for instance expressed by Edward
Heiberg (18971958), one of the leading Danish
functionalists. Kitchen work was in focus when
Heiberg built a house for his wife and himself in 1925.
They had no servants, and as stated by Heiberg, it was
therefore important that it should be easy for his wife to
cope with the household chores. Heiberg continued
working with kitchens, as an architect, as well as a
researcher. His motto for kitchen design was: Not one
unnecessary step [17].
Despite being politically active and supporting
equality between the sexes, Heiberg had no doubt
about who was responsible for the household in his
own home. Similarly the location of kitchens in
modernistic white villas usually expresses a clear
continuation of the division between masters and
servants despite the modernists effort to break
down the hierarchical social structure. The rooms for
servants are located on the lower floor, together with
other secondary rooms such as storage rooms, laundry
room, etc.
Thus, wealthy families kitchens were still a
workplace for servants, but many middle-class families,
who could no longer afford a staff, felt embarrassed
about having to work in the kitchen themselves [3]. To
those women who reluctantly worked in their own
kitchen (as the servants disappeared), cooking was a
chore [4]. In the general population, kitchen work was

Fig. 6 The Danish housing estate, Classens Garden built

in 1925. The apartment is very well equipped compared with
other contemporary houses, for instance with a large
bathroom with bath and a WC. The kitchen is located in the
rear of all spaces with access from dining room through a
corridor.

still associated with low status, and the kitchen

continued to be regarded as a secondary space. It was
hidden away in relation to the more presentable spaces
of the dwelling, and there was often more than one door
between the kitchen and dining room.

4. Danish Kitchens Functionalism in the

1950s
After the Second World War, Europe had to be
rebuilt and in Scandinavia the ideals of the welfare
society were to be realized. In this context there was a
socio-economic interest in getting more women into
the labor market. By 1951 more than 20 percent of
married women in Denmark worked outside the home.
For comparison, this had only been the case for six
percent in 1930 [18]. A study from 1954 showed that
mens presence in the kitchen was still limited [14],
and consequently there was a strong need to facilitate
the housewifes domestic work.
However, only working class women had joined the
labor market in earnest [18]. The middle-class families
established themselves in the nuclear family with a
housewife. The housewife was in charge of all
domestic work, and as the families could no longer
afford servants, she was alone in her kitchen. The
kitchen, which had previously been the domain of the

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

staff, was now the womans room. This is reflected in

the kitchens location in the dwelling. In contrast to the
pre-war kitchens, reserved for servants and hidden
away, the kitchen now became a more integral part of
the dwelling.
However, the kitchen was still not a part of the
representative space. It was still a workplace for the
housewife, and when guests came, she took off her
apron, closed the door to the kitchen and welcomed the
guests in the living rooms. There was still not much
prestige in kitchen work and this was reflected in
housing architecture. Kitchens were relatively small,
they were designed for specific functions around
cooking and washing and you could close off the
kitchen with a door.
In continuation of attempts in the 1920s and 1930s to
develop functional and efficient kitchens, kitchens in
the 1950s were also regarded as practical spaces [4].
Several measurements were made relating to space
needs and room conditions in kitchens. This was partly
as a continuation of the standardization processes,
which enjoyed good conditions for growth in a
post-war period with extensive housing shortages, and
partly in an effort to facilitate domestic work and free
female labor, so that women could participate in the
labor market.
In Denmark Heiberg remained a leading figure in the
debate on kitchen design and organization. In 1947 he
developed a standardized kitchen for the Association of
Social Housing and demonstrated that money could be
saved by standardization: The kitchen was 20%
cheaper than similar kitchens erected on site [17]. At
the Danish Building Research Institute, which was
established in 1947, studies of conditions related to
work in the kitchen were substantial research tasks.
Kitchenwork was studied and divided into a number of
sub-functions such as washing, cooking,
preparing food etc, and the need for storage was
identified. The clear division between the different
functions had hygienic advantages and can be seen as a
direct extension of the modernist hygienic ideals from
the pre-war period, but more important was the

463

time-saving potential. Time studies showed the amount

of time that women spent in the kitchen, and that this
time could be reduced through rational design.
The same functional perspective determined the
location of the kitchen in the dwelling. In dwellings
from the 1950s, the kitchen is no longer hidden away in
a distant corner of the home. It is typically located next
to the entrance hall (meaning that heavy shopping bags
can quickly be put down on the kitchen table), and with
direct access to the dining room (so that food and plates
do not have to be carried too far).
Actually, reductions in time were not only achieved
through the physical layout, but also the handling of
shopping and cooking were rationalized and
industrialized. Where government information on food
and cooking today is typically about health,
contemporary information focused on reductions in
time and utilization of new technical aids. For instance,
the
Danish
Households
Council
[Statens
Husholdningsrd] conducted studies on how much
time a housewife could save by serving canned soup
and machine-peeled potatoes [11].

Fig. 7 Typical Danish detached housing from the 1950s.

The dwelling is relatively small, as was dictated by
governmental regulations, and the space in the kitchen is
similarly small compared with todays standard. The
kitchen is functionally placed in the house near the
entrance hall and close to the dining room.

464

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

In the 1950s a softening took place regarding the

modernists concept of the mono-functional space
designed solely for cooking and dishwashing. This was
partly due to a series of studies conducted at the Danish
Building Research Institute in the early 1950s.
Research surveyed residents use of their kitchens in
contemporary housing, and the residents were asked
whether there was anything they wanted to change.
They replied that the kitchens were too small and that
housewives wanted a small dining area in the kitchen.
The mono-functional space did not accommodate
the life led by a modern nuclear family. The housewife
served as the family hub, and it became clear that the
concept of the mono-functional room was not so
functional after all. The housewife had many functions
besides cooking to take care of, for example helping the
children with their homework, and the narrow spaces
of the 1930s were criticized. On this background,
Edvard Heiberg designed kitchens with small dining
tables in the Copenhagen housing estate, Bellahj
(19501954). In the housing estate Tingbjerg (1955),
another famous Danish architect, Steen Eiler
Rasmussen, did likewise (see Fig. 8).
Thus, the dining kitchens of the 1950s were initially
functionally justified. This is evident from a report

published by The Danish Social Housing Association

in 1949. The report states that a place for eating should
be designed into the kitchen, but should not be so
spacious as to detract from the functions of the living
room [17]. The dining area in the kitchen was not
meant to be a place for socializing, but as a practical
set-up that made it easier for the housewife to watch the
children and help them with their homework while
cooking. Nevertheless, the kitchen dining area brought
a lot of life into the kitchen and informal socializing
developed. This was to have a great impact on the
further development of the kitchen.

5. Kitchen in the 1960s and 1970s

In the early 1960s the nuclear family was still the
societal ideal, and although women had increasingly
entered the labor market, they still took care of the
housekeeping at home. The kitchen remained the
womens domain, but they were not alone in their
kitchen anymore. A dining area as a part of the kitchen
was becoming popular, and the dining table, which in
the 1950s usually consisted of a small table in a
cramped corner, had become larger.

Fig. 8 Left: Bellahj designed by Edward Heiberg et al. Right: Housing plan from Tingbjerg designed by Steen Eiler
Rasmussen.

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

Thus a new view on the kitchen developed. The

kitchen was no longer a functional workspace, and the
dining area in the kitchen was no longer seen only as a
practical measure. In many homes the dining area in the
kitchen developed into a cosy space for social activities.
There was a special atmosphere in the kitchen, when
the heat from hot plates and the smell from the oven
filled the room. Eating and talking unfolded side by
side with kitchen work, and an informal atmosphere
developed in the kitchen. Initially, the kitchen retained
its workspace atmosphere and was still not seen as a
representative place. Danish kitchens of the 1960s were
informal and practically furnished, and thus became a
place where the family was together in a friendly and
relaxed manner.
When guests were invited, they were still not invited
into the kitchen, but during the week the dining room
stood empty. The dining area in kitchens grew, and as it
was now often located near the dining room, it felt
strange to have two dining tables standing only a few
meters apart. Obviously the next step was to break
down the wall between the kitchen and the dining room
[16]. The kitchen-dining area was born.
Thus, the kitchen and cooking came out of hiding.
The kitchen became a central space in the house, and
cooking became a visible part of everyday life.
However, initially this did not lead to sophisticated
kitchen design or the conversion of the kitchen into a
new best parlor. Quite the contrary, cooking was
carried out with doughy hands and floury arms, and
along with cooking utensils and kitchen appliances a
relaxed atmosphere was created in the kitchen. This
was perfectly inline with the rebellion against
bourgeoisie correctness that took place in the 1970s.
Thus, the kitchen was still a practical workspace.
However this should not lead to the misconception that
kitchens had no representative value. This was very
much the case, but it was not the aesthetics of the room
that scored points. In contrast, the quality of the kitchen
was in its informal character and the influence of this
on the rest of the dwelling.

465

The youth rebellion took place in 1968, and many

people felt that essential qualities were lost as a
consequence of societal modernization. Economic
status had become too dominant and in the search for
values of more fundamental character, pre-modern
society came into focus. Urban planners looked back,
and in traditional rural settlements they found a social
life that seemed to have been lost in the modern city.
Also architects looked back and found inspiration in
traditional housing: In a typical book from 1977 an
American architect described, how the kitchen in
Saxon dwellings from 900 BC was a central space in
the houses and served as the true living room [4].
Similar qualities were found by Danish architects in
Nordic farm kitchens.
Looking back in time, the social significance of
cooking and sharing a meal was rediscovered. A shared
meal is one of the oldest exchange relationships between
people, and many traditional rituals involve eating and
drinking [13]. It was in this light that communal eating
became important in the Danish community houses of
the 1970s. They were set up in order to create social
qualities and experiences between people. It was also
time-saving when the residents in the communities took
turns at cooking the meals, but the goal was not to
increase societal production, but rather to allow time to
be together.
All contemporary Danish low-density settlements
had a common-house with a large kitchen. However,
these kitchens were not installed at the expense of the
kitchen in individual dwellings. The kitchen-dining
area was an important space in the development of the
familys sense of cohesion and it was a place to gather,
where all family members could be heard.
Finally, the more central location of the kitchen
within the dwellings of the 1970s is explained by
changing gender roles. In many families the man
increasingly began to take part in household work, and
it is hardly a coincidence that at the same time kitchen
work became visible in everyday life. The development
of the exhaust hood that exhausts the smell of cooking

466

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

and frying, facilitated this development. The delimiting

walls could disappear, and the open kitchen-dining area
was made possible.
This

development

of

collective

lifestyles,

community houses and changing gender roles did not

affect all families in the same way, but the ideals about
kitchen design were nonetheless spread widely in
Danish housing; in social housing as well as in
detached houses. For instance in Ishj one of the
largest contemporary social housing estates in
Denmark the individual dwellings were consistent
with the ideals described. The open kitchens in Ishj
(see Fig. 9) are large and the dining table in the area
between the kitchen and the sofa group is spacious,
with room for at least six people.

6. 1980s until Today Consumerism and

Lifestyle Kitchens
It is hard to point to one overarching trend in the
architecture of this period. The architectural
development took many directions, postmodernism
made a break with modernisms functional analysis,
and architecture was freed of its moral obligation. With
the liberation of architectural design, it became obvious
that personal taste was often crucial when consumers
had to choose from a plethora of products, all covering
the same functional needs [15]. Moreover it became
increasingly clear that the design of products reflected
values, and thus the products indirectly came to be

Fig. 9 Danish Social Housing in Ishj from 1973. The

spacious kitchen is part of the living area, and the dining
table gives room for at least six people.

expressive of those who bought the products. Choice of

design became a matter of personality and identity [15].
A dwelling could therefore no longer be regarded as
a machine to live in and correspondingly a kitchen
was not only a room for cooking. The kitchen was a
designed product, reflecting opinions and personality,
and not necessarily built to last as long as the dwelling
itself. Since the functional properties no longer reigned
supreme, a kitchen could be replaced even though it
still functioned well if it was not in line with the
residents preference of taste. This kind of renewal and
decoration in the dwelling may be called lifestyle
renovations [2], and todays kitchens have been given
a similar name: lifestyle kitchens.
The dwelling has developed into an important place
for residents self-realization as well as a reflection of
their personal identity, and kitchen design, details,
materials and aesthetic take precedence over the
functional [6]. The dwelling has turned into a mirror
into which we look to find ourselves. Who are we, and
what do we want to be like? That is what we seek when
we decorate our dwelling and when we convert our
kitchen. It is about finding your own style about
finding yourself [2].
Advertising agencies know this. An advertising
campaign for a large Danish kitchen manufacturer
(Invita) sells the idea of a kitchen as an individual and
unique product. It is not an Invita kitchen, its a
Mette and Lars kitchen, says the advertisement.
Similarly, kitchen dealers say that they never supply
two identical kitchens [6]. It is explained that the
spaces to be furnished are different, the buyers have
varying needs and financial situation, and that various
materials, details, shapes and colors can be combined
[6]. The question remains whether this differs from
older kitchens? Modern kitchens are more standardized
than ever before, and it can also be argued that older
kitchens made specifically for a particular room were
more individual than the new ones.
It is not the major technical innovations or
architectural changes that characterize the kitchens of

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

the two last decades. The kitchen-dining area is still the

preferred type of kitchen, its location in the dwelling is
equivalent to that which was developed in the 1970s,
and the kitchen is still considered the central space in
the house. Later literature on kitchen decor is not about
new functional requirements and structural changes.
Focus is rather on the kitchen style. This is often
apparent in categorized sections that present various
styles: High-tech kitchen, the minimalist kitchen, the
romantic country kitchen, etc. A consistent, general
style of the period cannot be identified, but it can be
pointed out that style is extremely important [7].
In general, the dwelling has rediscovered the
representative character of the past, and this has
become apparent in the design and shape of the kitchen.
Where the kitchen of the 1970s was thought of as a
workplace, able to tolerate the tough nature of kitchen
work, kitchens in recent decades have seen an entirely
different aesthetic approach. Where the informal nature
of the kitchen was reflected in the adjoining rooms in
the 1970s, today the representative qualities of the
living rooms have influenced the kitchen. Light woods,
brushed steel, and polished granite have become some
of the preferred materials for exclusive kitchens (the
use of the delicate materials was made possible because
goods from the supermarket arrive in purified and
processed form).
Todays lifestyle kitchens send out strong signals,
and they work as a framework for self promotion. Like
for instance, when guests are welcomed on a Friday
night: a roast is in the oven, red wine has been poured,
and a shiny kitchen bears witness that the host is in
control of his life and apparently manages to cook in a
trendy way. With the increased prosperity in society,
focus is no longer on food as necessary for survival,
and focus is less on the meal as a framework for social
interaction. In todays lifestyle kitchen food expresses
a family in control and with ability to appreciate and
prepare fine cooking. Work in the kitchen is no longer
hidden away behind closed doors.

467

However, repeatedly the question is being asked of

whether todays large and prestigious kitchens actually
form the setting for cooking. It has been argued, that
our kitchens have become more and more spacious as
we eat more and more fast food and thus spend less
and less time cooking. As early as the 1990s, the Italian
scientist and designer Mondadori discussed this
development and he pointed out that current kitchens
reflect two concurrent trends. On the one hand he sees
the emergence of the fast-food home, where
residents are increasingly buying ready-made meals,
heating them in the microwave oven and eating them in
front of the television. This can be seen as a further
development of the trends of the 1950s, when the use
of canned or other ready-made dishes was highlighted
for the potential time savings.
On the other hand, Mondadori sees signs of the
emergence of the convent home, where cooking and
meals are increasingly seen as a ritual part of a social
community. Here, delicious recipes and culinary
experiences are paramount, and slow food cooking is
considered as part of a lifestyle rather than an
abomination that should to be reduced and eliminated.
The trend can be seen as a continuation of the 1970s
organic movements and communities rediscovery of
cooking and meals as more than just putting food into
your mouth. In this light, the kitchen becomes a space
of the senses, where the experiences of exotic spices,
colorful vegetables, the texture of materials and
delicate tastes are some of the ingredients.
According to Mondadori, the two trends exist side
by side not only in the sense that different families
choose to live in different ways and their neighbors do
not necessarily share the same lifestyle, but also that
trends can co-exist within one and the same family. The
shift can happen between lunch and dinner, during the
week, on weekends, etc. With the same furniture and
the same equipment, you can returnto medieval times,
or you can move toward Star Wars, writes Mondadori,
seductively. The kitchen is open for dreams and these
dreams will continue in the 21st century.

468

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

Fig. 10 Typical Danish detached housing from 2011. The kitchen is as in the previous period considered the central
space in the house. However, today there is an increased emphasis on the kitchen style.

7. Conclusion
This paper has shown how societal and cultural
trends (for instance health, food culture, gender,
sustainability, consumerism class, wealth, technology
etc.) have influenced architectural design over the last
100 years. This has been done by focusing on the
layout of the kitchen and kitchen design in European
and Danish housing.
The following changes have been identified:
One hundred years ago, the kitchen of the
bourgeoisie and the middle class was only used by
servants and other employees the kitchen was not
designed for the residents. Therefore, the design of the
kitchen and work in the kitchen were not associated
with any prestige. Today we spend an enormous
amount of money on decorating and designing our
kitchens. One of the reasons for this is that the kitchen
has become an important showcase for lifestyle and
identity.
In the early 20th century, major efforts were made to
streamline the workflow in the kitchen in order to save
time. An important objective was to release the female
workforce for the labor market. These trends are still
detectable in todays kitchens in the form of
ready-made meals, microwaves and fast food. But at
the same time slow food has become a trend. Time
is a valuable asset today, and it is associated with high

status to have the time, personal ability to cope and

reserves of energy to prepare good food with the
emphasis on quality.
At the beginning of the century, the kitchen was a
secondary space. In villas the kitchen was situated side
by side with the coalstore and servants rooms in the
basement, and in the large residential blocks in the city
they were well hidden away in a far corner of the
apartment. During the 20th century, the kitchen moved
even further into the spotlight, and today it has
developed into a central room of the home with great
social qualities, and a lot of money is spent on the
aesthetic appearance of the kitchen.
So fundamental changes have taken place in the
kitchen over this relatively short period of time. Our
view of the kitchen has undergone drastic changes and
the physical location and layout of the kitchen have
also developed dramatically. This can serve as a
reminder that kitchens as well as housing will continue
to change. Although many people have a firm concept
of what a dwelling is, in fact the dwelling is changing
constantly, both physically and mentally.

References
[1]
[2]

Claus Bech-Danielsen, Moderne arkitektur hva er

meningen? Copenhagen, Systime, 2004.
Claus Bech-Danielsen and Kirsten Gram-Hanssen,
Home-building and identity The soul of a house and

The Kitchen: An Architectural Mirror of Everyday Life and Societal Development

the personal touch, in: Urban lifescape: Anthology,
Aalborg, Aalborg University Press, 2004.
[3] John Burnett, Social History of Housing: 1815-1970.
Cambridge, The University Press, 1978.
[4] Terrence Conran, The Kitchen Book, New York: Crown
Publishers Inc., 1977.
[5] Christine Frederick, The New Housekeeping: Efficiency
Studies in Home Management, available online at:
http://nationalhumanitiescenter.org/pds/gilded/progress/te
xt4/frederick.pdf, Originally publ. in: Ladies. Home
Journal, Sept.-Dec. 1912.
[6] Hans-Christian Jensen, Livsstilskkkenet Alle har ret
til et fedt kkken, in: Dybdahl & Engholm (ed.), Design:
Kkkenet, Kbenhavn, Gyldendal, 2008.
[7] Gilly Love, Making the Most of Kitchens, London,
Conran Octopus Ltd., 1997.
[8] Ellen Lupton and Abbott J. Miller, The Bathroom, The
Kitchen, and the Aesthetics of Waste: A Process of
Elimination, Cambridge, MIT List Visual Arts Center,
1992
[9] Arnoldo Mondadori (Ed.), Annual Cusina. Milano,
September 1991.
[10] Paul Overy, Light, Air and Openness: Modern
Architecture Between the Wars, London, Thames &
Hudson, 2008.

469

[11] Gerda Petri (Ed.), Kitchen.Denmarks largest site, lgod:

lgod Museum 1991.
[12] Janice Williams Rutherford, Selling Mrs. Consumer:
Christine Frederick and the Rise of Household Efficiency,
Athens, University of Georgia Press, 2003.
[13] Karen Lisa Salamon, Det universelle kkken
fremtidens rum? in: Dybdahl & Engholm (Ed.), Design:
Kkkenet, Kbenhavn, Gyldendal, 2008.
[14] Anne-Louise Sommer, Fra ildstedskkkenet til
samtalekkkenet, in: Dybdahl & Engholm (Ed.), Design:
Kkkenet, Kbenhavn, Gyldendal, 2008.
[15] Deyvan Sudjic, Home: The Twentieth-Century House,
Glasgow, Festival Company Ltd, 1999.
[16] Sunset, The Magazine of Western Living, Western
Kitchens, Menlo Park, Lane Publishing Co, 1955.
[17] Leif Leer Srensen, Edvard Heiberg og dansk
funktionalisme, En arkitektog hans samtid, Kbenhavn,
Arkitektens Forlag, 2000.
[18] Birgitte Vorre, Boligen i det 20. rhundrede. Indretning
og brug, Kbenhavn, Nyt Nordisk Forlag Arnold Busk,
2008.
[19] Finn Werne, Arkitekturens ismer, Stockholm, Arkitektur
Frlag AB, 1998.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 470478

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

A Refined Formula for the Allowable Soil Pressure Using

Shear Wave Velocities
Semih S. Tezcan and Zuhal Ozdemir
Bogazici University, Bebek, Istanbul, Turkey
Abstract: Based on a variety of case histories of site investigations, including extensive bore hole data, laboratory testing and
geophysical prospecting at more than 550 construction sites, an empirical formulation is proposed for the rapid determination of
allowable bearing pressure of shallow foundations in soils and rocks. The proposed expression corroborates consistently with the
results of the classical theory and is proven to be rapid, and reliable. Plate load tests have been also carried out at three different sites,
in order to further confirm the validity of the proposed method. It consists of only two soil parameters, namely, the in situ measured
shear wave velocity and the unit weight. The unit weight may be also determined with sufficient accuracy, by means of other
empirical expressions proposed, using P or S wave velocities. It is indicated that once the shear and P-wave velocities are
measured in situ by an appropriate geophysical survey, the allowable bearing pressure as well as the coefficient of subgrade reaction
and many other elasticity parameters may be determined rapidly and reliably.
Key words: Shear wave velocity, shallow foundations, allowable bearing pressure, dynamic technique, soils and rocks.

1. Introduction
Professor Schulze [1], a prominent historical figure
in soil mechanics and foundation engineering in
Germany, stated in 1943 that For the determination
of allowable bearing pressure, the geophysical
methods, utilising seismic wave velocity measuring
techniques with absolutely no disturbance of natural
site conditions, may yield relatively more realistic
results than those of the geotechnical methods, which
are based primarily on bore hole data and laboratory
testing of so-called undisturbed soil samples.
Since that time, various significant contributions
have been made to solving geotechnical problems by
means of geophysical prospecting. The P-wave
velocities, for instance, have been used to determine
the unconfined compressive strengths and modulus of
elasticity of soil samples by Coates [2]. Hardin and
Black [3], and also Hardin and Drnevich [4], based on
extensive experimental data, established indispensable
relations between the shear wave velocity, void ratio,
Corresponding author: Semih S. Tezcan, professor,
research field: civil engineering. E-mail: tezokan@gmail.com.

and shear rigidity of soils. Similarly, Ohkubo and

Terasaki [5] supplied various expressions relating the
seismic wave velocities to weight density,
permeability, water content, unconfined compressive
strength and modulus of elasticity.
The use of geophysical methods in soil mechanics
has been extensively studied for the purpose of
determining the properties of soils and rocks by Imai
and Yoshimura [6], Tatham [7], Willkens et al. [8],
Phillips et al. [9], Keceli [10, 11], Jongmans [12],
Sully and Campanella [13], and Pyrak-Nolte et al.
[14]. Imai and Yoshimura [6] proposed an empirical
expression for the determination of bearing capacity qf
and/or qa as
nqa = qf = Vs2.4 /(1590) (kPa)
(1)
which yields values unacceptably much higher than
the classical theory as will be evident in next section.
Campanella and Stewart [15], determined various soil
parameters by digital signal processing, while Butcher
and Powell [16], supplied practical geophysical
techniques to assess various soil parameters related to
ground stiffness. An empirical expression is also
proposed by Abd El-Rahman [17], for the ultimate

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

bearing capacity of soils, using the logarithm of shear

wave velocity.
A series of guidelines have been also prepared in
this respect by the Technical Committee TC 16 of
IRTP, ISSMGE [18], and also by Sieffert [19]. Keceli
[11], Turker [20], based on extensive case studies,
supplied explicit expressions for the allowable bearing
pressure, using shear wave velocity. In this paper, the
earlier formula presented by Tezcan et al. [21], has
been calibrated and improved with the soil data of 550
construction sites. Massarsch [22] determined
deformation properties of fine-grained soils from
seismic tests. As to the in situ measurement of P and S
wave velocities, various alternate techniques are
available as outlined in detail by Stokoe and Woods
[23], Tezcan et al. [24], Butcher et al. [25], Richart et
al. [26], Kramer [27], Santamarina et al. [28].

2. Theoretical Basis for the Empirical

Expression
In order to be able to arrive at a particular empirical
expression for the allowable soil pressure qa
underneath a shallow foundation, the systematic
boundary value approach used earlier by Keceli [11]
will be followed. The state of stress and the related
elastic parameters of a typical soil column is shown in
Fig. 1. Considering a foundation depth of Df with a
unit cross-sectional area of A = 1, the typical form of
the compressive ultimate bearing capacity at the base
of the foundation nothing but only as a format, may be
written approximately as;
qf = Df
(2)
qa = qf/n = Df/n
(3)
where qf = ultimate bearing capacity at failure, = unit
weight of soil above the base of the foundation, qa =
allowable bearing pressure, and n= factor of safety. In
order to be able to incorporate the shear wave velocity
Vs2 into the above expressions, the depth parameter Df
will be expressed as velocity multiplied by time as;
Df = Vs2 t
(4)
in which, the Vs2 is purposely selected to be the shear
wave velocity measured under the foundation, t = is

471

an unknown time parameter. The time parameter t is

introduced herein just as a dummy parameter in order
to keep consistency in appropriate units. Substituting
Eq. (4) into Eq. (3), yields
(5)
qa = Vs2 t / n
The unknown time parameter t, will be determined
on the basis of a calibration process. For this purpose,
a typical hard rock formation will be assumed to
exist under the foundation, with the following
parameters, as suggested earlier by Keceli [11];
qa = 10 000 kN/m2, Vs2 = 4 000 m/sec,
(6)
= 35kN/m3, n = 1.4
Substituting these numerical values into Eq. (5), we
obtain t = 0.10 sec, thus;
qa = 0.1 Vs2/n
(7)
This is the desired empirical expression to
determine the allowable bearing pressure qa, in soils
and rocks, once the average unit weight, , for the soil
layer above the foundation and the in situ measured
Vs2 wave velocity for the soil layer just below the
foundation base are available. The unit of Vs2 is in
m/sec, the unit of is in kN /m3 , then the resulting qa
value is in units of kPa. The unit weight values
may be estimated using the empirical expressions;
p = 0 + 0.002 Vp1
(8a)
0.25
and
s = 4.3 Vs1
(8b)
as proposed earlier by Tezcan et al. [21], and by
Keceli [29], respectively. The second expression is
especially recommended for granular soils, for which
the measured Vs1 values represent appropriately the
degree of water content and/or porosity. The wave
velocities must be in units of m/sec. The only
remaining unknown parameter is the factor of safety,
n, which is assumed to be, after a series of calibration
processes, as follows:
n = 1.4 (for Vs2 4 000 m/sec),
(9)
n = 4.0 (for Vs2750 m/sec)
The calibration process is based primarily on the
reference qa values determined by the conventional
Terzaghi method, for all the data sets corresponding to
the 550 construction sites considered. For Vs2
values greater than 750 m/sec and smaller than 4000

472

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

Pa = Aqa
P f = A qf
qa = Allowable stress (qa = qf / n)
qf = Ultimate bearing pressure
qf = Df qa = Df / n

Layer 1

Df

(Vp1 / Vs1)

Foundation

Fig. 1

qa = ks d

(Vp2 / Vs2)

for n = factor of safety. For soil types with Vs2 750

m/sec the factor of safety is n = 4, for rocks with Vs2
4000 m/sec, it is n= 1.4. For other intermediate values
of shear wave velocity, linear interpolation is
recommended. The validity of these values has been
extensively checked and calibrated by the soil data at
550 construction sites. The relatively higher value of
factor of safety assumed for soils is deemed to be
appropriate to compensate the inaccuracies and gaps
existing in the measured values of shear wave velocity.
In fact, Terzaghi and Peck [30] states that The factor
of safety of the foundation with respect to breaking
into the ground should not be less than about 3.

ks = qa /0.025 = 40qa (kN/m3)

It is determined by Terzaghi and Peck [30] also that

the width of footing, B, has a reducing influence on
the value of allowable bearing pressure for granular
soils. Therefore, a correction factor is introduced
into the formula, for sandy soils only, as shown in the
third line of Table 1. The proposed values of this
correction factor, for different foundation width B, are
as follows:
= 1.00
= 1.130.11 B
= 0.83 0.01 B

for (0 B 1.20 m)
for (1.2 B3.00 m)
for (3.0 B12.0 m) (10)

3. Coefficient of Subgrade Reaction

The shear wave velocity may be used successfully
to determine ks = coefficient of subgrade reaction of
the soil layer just beneath the foundation base by
making use of the expressions given in Fig. 1. The
coefficient of subgrade reaction ks, is defined, similar
to the definition of spring constant in engineering

Factors of safety, n, for soils and rocks(1)

Soil type
Hard rocks

(1)

Assuming d = 0.025 m

Soil column and related parameters.

m/sec a linear interpolation is recommended. The

engineering rock formations are assumed to start for
Vs2 > 750 m/sec. The factors of safety, as well as the
empirical allowable bearing pressure expressions, for
various soil (rock) types, are given in Table 1. It is
seen that three distinct ranges of values are assumed

Table 1

d = Pa H / AE = qa H/ E
d = q a / ks k s = q a / d

Layer 2

d =Settlement of layer H

Vs range (m/sec)
Vs 4000

qa (kN/m2)

qa = 0.071 Vs

n = 1.4
-4

Soft rocks

750 Vs 4000

n =4.68.10 Vs

qa = 0.1 Vs / n

Soils

750 Vs

n = 4.0

qa = 0.025 Vs

Linear interpolation is applied for 750 Vs 4 000 m/sec. correction factor is used for sands only (Eq. 10).

473

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

mechanics, to be the necessary vertical pressure to

produce a unit vertical displacement and expressed as:
ks = qa/d
(11)
For shallow foundations, the total vertical
displacement is restricted to 1 inch =0.025 m, as
prescribed by Terzaghi and Peck [30]. When, d =
0.025 m is substituted in Eq. (11), the coefficient of
subgrade reaction becomes in units of kN/m3 ;
ks = 40qa
(12)
or ks= 4Vs2/n

(13)

E = (3 4) G/( 1)
(21)
The Constraint modulus Ec , may be also obtained
in terms of as ;
Ec = (1)E/(3 4)
(22)
2
or, Ec = Vp /g
(23)
The Bulk modulus Ek, of the soil layer, may be
expressed, from the theory of elasticity, as
Ek = E/3 (1 2)

(24)

(25)

Ek = (1)E/3= V p 4 V s 3g

4. Elasticity Parameters

5. Case Studies

Once, Vp2 and Vs2 seismic wave velocities are

measured, by geophysical means, for the soil layer
No.2 just under the foundation, several parameters of
elasticity, such as G = Shear modulus, Ec =
Constraint modulus of elasticity, E = Modulus of
elasticity (Youngs modulus), Ek = Bulk modulus, and

The allowable bearing pressures have been also

determined at more than 550 construction sites in and
around the Kocaeli and Istanbul Provinces in Turkey,
between the years 20052010. At each construction
site, by virtue of City by-law, appropriate number of
bore holes were drilled, SPT counts conducted,
undisturbed soil samples were taken for laboratory
testing purposes, where shear strength -c, the internal

= Poissons ratio may be obtained easily. The Shear

modulus, G, and the Constraint modulus, Ec, are
related to the shear and P- wave velocities by the
following expressions, respectively;
2

G = V s (14)
and

Ec = V p (15)

where, mass density given by g. From the

Theory of Elasticity, it is known that, E = the Youngs
modulus of elasticity is related to Ec = the Constraint
modulus and also to G = the Shear modulus by the
following expressions:
E = Ec (1 + ) (1 2)/(1 )
(16)
E =2 (1+)G
(17)
Utilising Eqs. (14) and (15) and also substituting

as
= Ec/G = (Vp/Vs)2
into Eqs. (16) and (17), we obtain

(18)

= (2)/2(1)
(19)
or = (2 2)/(21)
(20)
The modulus of elasticity is directly obtained from
Eq. (17) as;

angle of friction -unconfined compression strength

-qu and unit weight -were determined. Subsequently,
following the classical procedure of Terzaghi and
Peck [30], the ultimate capacity and also the allowable
bearing pressures were determined, by assuming the
factor of safety as n = 3. For granular soils, immediate
settlement calculations were also conducted, in order
to determine whether the shear failure mechanism or
the maximum settlement criterion would control the
design.
The numerical values of the allowable bearing
pressures, qa, determined in accordance with the
conventional Terzaghi theory, are shown by a
triangular () symbol, in Fig. 2, where the three digit
numbers refer to the data base file numbers of specific
construction sites. Parallel to these classical soil
investigations, the P- and S- wave velocities have
been measured in situ, right at the foundation level for
the purpose of determining the allowable bearing
pressures, qa , which are shown by means of a circle
(o), in Fig. 2. Two separate linear regression lines

474

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

were also shown in Fig. 2, for the purpose of

indicating the average values of allowable bearing
pressures
determined
by
dynamic
and
conventional methods. In order to obtain an idea
about the relative conservatism of the two methods,

the ratios of allowable bearing pressures (r = qad/qac),

as determined by the dynamic and conventional
methods, have been plotted against the Vs values in
Fig. 3.

qa = allowable bearing pressure, kPa

350

300

250

200

150

100
250

300

350

400

450

500

550

Vs2 = shear wave velocity, m/sec

Fig. 2

Comparative results of Conventional and Dynamic methods.

2.00

St 502

1.50

= qa,d /

St 499

1.25

qa,c

1.75

1.00

Linear regression line

St 128

St 178

0.75

St 188

0.50

200

Fig. 3

300

400
500
Vs2 = shear wave velocity, m/sec

600

700

Ratios of allowable bearing pressures ( qa,d / qa,c ) as determined by the dynamic and the conventional methods.

475

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

It is seen that the linear regression line indicates for

Vs values smaller than 400 m/sec a narrow band of r
= 1.03 to r = 1.12, which should be regarded as quite
acceptable. The dynamic method proposed herein
yields allowable bearing pressures slightly (on the
order of 3 to 10 percent) greater than those of the
conventional method for Vs values smaller than
400 m/sec. In fact, the conventional method fails to
produce reliable and consistent results for relatively
strong soils and soft rocks, because it is difficult to
determine the appropriate soil parameters c, and for
use in the conventional method. At construction site
Nos: 133, 134, 138, 139, 206, 207, 214, 215, 219, 502,
507 and 544, where the soil conditions have been
mostly weathered andesite, granodiorite arena,
greywacke, limestone, etc did not allow for the
measurement of c and values. Therefore, the
use of dynamic method becomes inevitable for such
strong soils with Vs2 > 400 m/sec.
The list of soil parameters determined by in situ and
also by laboratory testing through geotechnical
prospecting, as well as the in situ measured Vp and Vs
velocities at each of the 550 construction sites, are
too voluminous to be included herein. Those
researchers interested to have access to these
particular data base, may inquire from internet
www.superonline.com, www.tezokan.com.

6. Seismic Wave Velocities

The seismic wave velocities have been measured
using P and S geophones by means of a 24

Channel Geometrics Abem Pasi seismic instrument,

capable of noise filtering. The P waves have been
generated by hitting 6 blows vertically, with a 0.15
kN hammer, onto a 250 250 16 mm size steel plate
placed horizontally on ground. For the purpose of
generating S waves however, an open ditch of size
1.4 1.4 1.4 m was excavated and then two steel
plates were placed on opposite vertical faces of this
ditch parallel to the conterline of the geophones.
Using the same 0.15 kN hammer, 6 heavy horizontal
blows were applied onto each of these vertical steel
plates. The necessary polarity of the S wawes was
achieved by hitting these vertical steel plates
horizontally in opposite directions, nonconcurrently.

7. Plate Load Testing

For purposes of correlating the allowable bearing
pressures determined by various methods, plate
loading tests have been carried out at three particular
construction Sites Nos: 335, 502 and 544. The soil
parameters c, qu , and as determined by laboratory
testing, as well as the P and S wave velocities
measured at site by geophysical prospecting are all
shown in Table 3. A thick steel bearing plate of 316.2
mm 316.2 mm = 0.10 square meter in size is used
under the test platform of size 1.50 m by 1.50 m. The
tests are carried out right at the bottom elevations of
foundations. One half of the bearing pressure 0,
which produced a settlement of s = 12.7 mm was
selected as the allowable pressure qa as shown in Fig.
4. It is seen clearly in Table 2 that the results of the

Table 2 Comparative evaluation of allowable pressures.

Site No Owner Lot Nos
Various soil parameters ( = 0)
(soil type)
(1)
Vp2
qu
Df
c
lab

335 Suleyman Turan 8

Paft./A/930 Pars. (silty
clay)
544 Ayhan Dede G22B/
574/11 (weathered diorite)
502 Ebru nar 30 L1C/
440/8 (clay stone)

kPa

kPa

kN/m3

172

1.50

86

0 = 16

190

1.50

95

0 = 16

147

1.00

140

0 = 20

18.9
18.0
22.7

qa = allowable pressure
Vs2

Tezcan, et.al.(3)
Eq. 7

Load test
Fig.4

kPa

kPa

kPa

Terzaghi
Eq. 26

(2)

m/sec

m/sec

896

390

157

173

180

1 020

453

172

204

208

1 210

489

248

274

280

(1) qu = unconfined compressive strength; (2) Terzaghi and Peck (1976); (3) qa = 0.025 p Vs (Eq.7), n = 4.

476

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

Table 3

Results of numerical example (H = 15 m, Vp2 = 700 m/sec, Vs2 = 200 m/sec, c = 52 kPa, =0), (Vp1 = 700 m/sec above

the base).
Formula
p 0 + 0.002Vp1
Laboratory
n=4
qf = c Nc + Df Nq
qf = 0.1 Vs2
qa = qf / n
ks = 40 qa= 4 Vs2 / n
2

G = Vs / g
= (Vp2 / Vs2)2
= (- 2) / 2(- 1)
E = 2 (1+) G
Ec = VP22 / g
Ek = E / 3 (1-2)
Ek = E ( -1)/3
d = displacement

Equation
eqn (8a)
Table 1
eqn (26)
eqn (7)
eqn (3)

Numerical calculations
p = 16 + 0.002 (700)
Vs2 750 m/sec
qf = 52 (5.14) +17.2 (2.9) 1
qf = 0.1 (17.4) 200
qa = 348 / 4

Result
17.4(1)
17.2
4
318
348
87

Unit
kN/m3
kN/m3
kN/m2
kN/m2
kN/m2

eqn (12)
eqn (14)

ks = 40 (87)
G = 17.4 (200)2 / 9.81

3 480
70 948

kN/m3
kN/m2

eqn (18)
eqn (19)
eqn (17)
eqn (15)
eqn (24)
eqn (25)

= (700 / 200)2
= (12.25 - 2) / 2(11.25)
E = 2 (1.456) 70 948
17.4 (700)2 / 9.81
206 537 / 3 (1-2)
206 537 (12.25-1) / 3
d = qa / ks = 87 / 3480

12.25
0.456
206 537
870 000
774 417
774 514

kN/m2
kN/m2
kN/m2
kN/m2

0.025

eqn (11)
3

(1) Result of Eq. (8a), = 17.4 kN/m is used in all subsequent expressions.

Fig. 4 Load test results at Sites No: 335, 502, and 544 ( = pressure under the test plate, kPa; 0 = 2, qa = pressure, which
produces s = 12.7 mm).

proposed dynamic method using P and S wave

velocities are in very close agreement with those of
the plate load testing. The allowable bearing pressures
qa, in accordance with the conventional theory
assuming Nc = 5.14 and Nq = 1, are also calculated
using
qa = (cNc+ Df Nq )/3.0

(26)

8. Numerical Example
For purposes of illustration, a soft clayey soil layer
of H = 15 m beneath a shallow strip footing of depth
Df = 2.90 m, width B = 1.30 m, is considered. The in
situ measured seismic wave velocities are determined
to be Vp2 = 700 m/sec and Vs2 = 200 m/sec, within the

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities

soil layer just below the foundation base. By

coincidence, the P wave velocity within the soil
layer above the foundation base is also measured to be
Vp1 = 700 m/sec. A comprehensive set of classical soil
investigations, including a number of bore hole data
and laboratory testing exist for this particular site,
together with the numerical values of various soil
parameters (c = 52 kPa, and = 0), including the
bearing pressure capacity determined to be qf = 322
kPa by the conventional method of Terzaghi and Peck
[30]. Therefore, the validity and the reliability of the
proposed empirical formulae have been rigorously
verified. Calculation of some elasticity parameters,
using the empirical expressions presented herein, is
summarised in Table 3.

9. Discussion on the Degrees of Accuracy

The degrees of accuracy of the proposed dynamic
method are quite satisfactory and consistent as attested
rigorously at more than 550 construction sites. The
conventional approach however, depends heavily on
the degrees of accuracy of in situ and laboratory
determined soil parameters. In fact, the allowable
bearing pressure calculations are very sensitive to the
values of c, and , determined in the laboratory using
so-called undisturbed soil samples, which may not
necessarily represent the true in situ conditions. This
may explain the reason why at a number of
construction sites, some inconsistent and erratic
results for qa are obtained using the classical theory, as
already depicted in Fig. 2, because the laboratory
measured c, and - values differed considerably from
one soil sample into the other. The Point Load tests
(Fig. 4) of rock samples have been carried out for Vs2 values greater than 400 m/sec as recommended by
Hunt [31].

10. Conclusions
The P and S wave velocities are most powerfull
soil properties representing a family of geotechnical
soil parameters, ranging from compressive and shear

477

strengths to void ratio, from subgrade coefficient to

cohesion etc,
Once the shear and P wave velocities are
measured, the allowable bearing pressure, the
coefficient of subgrade reaction, various other
elasticity parameters, as well as the approximate
values of the unit weight are rapidly and economically
determined, using relatively simple empirical
expressions. Bore hole drilling and laboratory testing
of soil samples including the point load method of
rock samples, may be beneficially utilised for
correlation purposes.

Acknowledgements
The writers gratefully acknowledge the supply of
voluminous site data by Mr. Mustafa Cevher, Head,
the Earthquake Research Department, Greater
Municipality of Kocaeli Province, and also the
scientific assistance by Dr. Osman Uyank, Assoc.
Professor of Geophysics at Suleyman Demirel
University, Turkey.

References
[1]

[2]
[3]

[4]

[5]

[6]

[7]

W. E. Schulze, Grundbau, in: B. G. Taubner, Deutsche

Forschungsgesellschaft fr Bodenmechanik (7th ed.),
Leipzig, Germany, also available: Technical University
of Istanbul, Issue 48, No DK624-15, Uler Printing
House, Istanbul, Turkiye, 1943.
D. F. Coates, Rock Mechanics Principles, Mines Branch
Monographs, No. 874, 1970.
B. O. Hardin, and W. L. Black, Vibration modulus of
normally consolidated clays, Journal of the Soil
Mechanics and Foundation Division 94 (SM2) (1968)
353369.
B. O. Hardin and V. P. Drnevich, Shear modulus and
damping in soils, Journal of the Soil Mechanics and
Foundation Division 98 (SM7) (1972) 667692.
T. Ohkubo and A. Terasaki, Physical Property and
Seismic Wave Velocity of Rocks, OYO Corporation,
Japan, 1976.
T. Imai and M. Yoshimura, The Relation of Mechanical
Properties of Soils to P and S- Wave Velocities for Soil
Ground in Japan, Urana Research Institue, OYO
Corporation, 1976.
R. H. Tatham, Vp/Vs and lithology, Geophysics 47 (1982)
336344.

478
[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

A Refined Formula for the Allowable Soil Pressure Using Shear Wave Velocities
R. Willkens, G. Simmons and L. Caruso, The Ration
Vp/Vs as a discriminant of composition for siliceous
limestones, Geophysics 49 (11) (1984) 18501860.
D. E. Phillips, D. H. Han and M. D. Zoback, Empirical
relationships among seismic velocity, effective pressure,
porosity, and clay content in sandstone, Geophysics, 54
(1) (1989) 8289.
A. D. Keceli, Determination of bearing capacity of soils
by means of seismic methods (in Turkish), Geophysical
Journal 4 (1990) 8392.
A. D. Keceli, Bearing capacity determination by seismic
methods, Geophysical Journal, 14, No. 12, 2000 (in
Turkish).
D. Jongmans, The application of seismic methods for
dynamic characterization of soils, Bulletin of
International Association of Engineering Geology. 46
(1992) 6369.
J. P. Sully and R. G. Campanella, Evaluation of in situ
anisotropy from crosshole and downhole shear wave
velocities measurements, Geotechnique 45 (2) (1995)
267282.
L. J. Pyrak-Nolte, S. Roy and B. l. Mullenbach, Interface
waves propagated along a fracture, Journal of Applied
Geophysics 35 (1996) 7987.
R. G. Campanella and W. P. Stewart, Seismic cone
analysis using digital signal processing for dynamic site
characterization, Canadian Geotechnical Journal 29 (3)
(1992) 477486.
A. P. Butcher and J. J. Powell, Practical considerations
for field geophysical techniques to used assess ground
stiffness, in: Proc. Int. Conf. on Advances in Site
Investigation Practice, ICE London, March 1995, Thomas
Telford, pp. 701714.
M. M. Abd El-Rahman, I. Setto and A. El-Werr, Inferring
mechanical properties of the foundation material, from
geophysical measurements, Egyptian Geophysical
Society, in: Proceedings of the 9th Annual Meeting, 1992,
pp. 206228.
IRTP, ISSMGE Technical Committee TC16, Ground
Property Characterization from In-situ Testing,
International Reference Test Procedure (IRTP) for the
Cone Penetration Test (CPT) and the Cone Penetration
Test with pore pressure (CPTU), in: Proc. XIIth
ECSMGE Amsterdam, Balkema, 1999, pp. 21952222.
J. G. Sieffert and Ch. Bay-Gress, Comparison of the
European bearing capacity calculation methods for
shallow foundations, Geotechnical Engineering 143
(2000) 6574.

[20] E. Turker, Computation of ground bearing capacity from

shear wave velocity, in: D. Bergman et al. (Eds.),
Continuum Models and Discrete Systems, Netherlands,
2004, pp. 173180.
[21] S. S. Tezcan, Z. Ozdemir and A. D. Keceli, Allowable
bearing capacity of shallow foundations based on shear
wave velocity, Journal of Geotechnical and Geological
Engineering 24 (2006) 203218.
[22] K. R. Massarsch, Deformation properties of fine-grained
soils from seismic tests, in: International Conference on
Site Characterization, ISC2, 19-22 Sept. Porto, 2004, pp.
133146.
[23] K. H. Stokoe and R. D. Woods, Insitu shear wave
velocity by cross-hole method, Journal of the Soil
Mechanics and Foundation Divison 98 (SM5) (1972)
443460.
[24] S. S. Tezcan, S. M. Erden and H. T. Durgunolu, Insitu
measurement of shear wave velocity at Boazii
University Campus, in: Proc. of the International
Conference on Soil Mechanics and Foundation
Engineering, Vol. 2, April 1975, pp. 157164.
[25] A. P. Butcher, R. G. Campanella, A. M. Kaynia and K. R.
Massarsch, Seismic cone downhole procedure to measure
shear wave velocity: A guideline prepared by ISSMGE
TC10: Geophysical Testing in Geotechnical Engineering,
in: Proc. of the XVIth International Conference on Soil
Mechanics and Geotechnical Engineering, 5p., Osaka,
Japan, May 2006.
[26] F. E. Richart, J. R. Hall and R. D. Woofs, Vibrations of
Soils and Foundations, Prentice Hall, Inc, Englewood
Cliffs, N.J., 1970.
[27] L. K. Kramer, Geotechnical Earthquake Engineering,
Prentice Hall international Series in Civil Engineering
Mechanics, Upper Saddle River, N.J., USA, 1996.
[28] J. C. Santamarina, A. K. Klein and M. A. Fam, Soils and
Waves, John Wiley & Sons, New York, N.Y., USA,
2001.
[29] A. D. Keceli, Applied Geophysics, UCTEA Chamber of
Geophysical Engineers, Ankara, Turkey, July 2009, http:
//jeofizik.org.tr (in Turkish).
[30] K. Terzaghi and R. B. Peck, Soil Mechanics in
Engineering Practice (2nd ed.), John Wiley & Sons,
London, 1976.
[31] R. E. Hunt, Geotechnical Engineering Investigation
Manual, Mc Graw Hill Book Co., New York, N.Y., 1984.
[32] Z. T. Bieniawski, Engineering Rock Mass Classification,
Wiley Interscience Publication, New York, NY, John
Wiley & Sons, USA, 1979.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 479488

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Structural Stress Identification Using Fuzzy Pattern

Recognition and Information Fusion Technique
Jun Teng, Ting Zhang and Wei Lu
Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, China
Abstract: In order to ensure the service security of space structures under wind load, the stress identification method based on the
combination of fuzzy pattern recognition and information fusion technique is proposed, in which the measurements of limited strain
sensors arranged on the structure are used. Firstly, the structure is divided into several regions according to the similarity and the most
unfavorable region is selected to be the key region for stress identification, while the different numbers of the strain sensors are
located on the key region and the normal regions; secondly, the different stress distributions of the key region are obtained based on
the measurements of the strain sensors located on the key region and the normal regions separately, in which the fuzzy pattern
recognition is used to identify the different stress distributions; thirdly, the stress distributions obtained by the measurements of
sensors in normal regions are selected to calculate the synthesized stress distribution of the key region by D-S evidence theory;
fourthly, the weighted fusion algorithm is used to assign the different fusion coefficients to the selected stress distributions obtained
by the measurements of the normal regions and the key region, while the synthesized stress distribution of the key region can be
obtained. Numerical study on a lattice shell model is carried out to validate the reliability of the proposed stress identification method.
The simulated results indicate that the method can improve identification accuracy and be effective by different noise disturbing.
Key words: Stress identification, Fuzzy pattern recognition, information fusion technique.

1. Introduction
Large space structures are usually the important
symbols in cities, which are wind-sensitive because of
their small damping ratio and low frequency in
vibration. Therefore, the safety and reliability of the
large space structures under wind load is one of the
most important research problems. Structural health
monitoring is a method which can obtain the structural
responses and give the estimation of the working
status of the structure, while different types of sensors
are arranged on the structure. Recently, the
researchers [15] put emphasis on realizing objectives
and functions of structural health monitoring system
using limited measurements of sensors. For example,
Ming Liu et al. [6] assessed the reliability of bridge
through the long-term monitoring measurements of
strain sensors under traffic loads and researched the

Corresponding author: Jun Teng, PhD, professor, research

field: structural health monitoring. E-mail: tengj@hit.edu.cn.

security limit using the actual traffic conditions and

measurements of strain sensor on Wisconsin Rive
Bridge in United States.
Key region is defined as the unfavorable region
with higher stress level than the normal regions of the
structure, which is related to the safety of the structure
directly. Generally, the stress distribution of the key
region is obtained by the measurements of limited
sensors located on the key region directly, in which
the stresses of elements without sensors locating on
can not be obtained. In addition, it is impossible to
measure the stresses of all the elements in the key
region due to the requirements in economic and the
construction of the structural health monitoring
system. Therefore, the stress identification method
based on measurements of limited strain sensors is
proposed.
In the structural health monitoring system of
Shenzhen Citizen Center in China, the stress fields of
brace steel brackets are identified by the limited

480

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

measurements of strain sensors located on the key

points [7]. However, the measurements of strain
sensors located on the normal regions should also be
used to obtain the stress distribution identification of
the key region, which can provide much more known
measurements and reduce the incompleteness of
information existing in structural health monitoring
system. Information fusion technique can fuse the
measurements of strain sensors located on key region
and the normal regions, and give synthesized
information for identification, eliminate redundancy
and contradictions that may exist. Information fusion
technique has been used in structural damage
identification, while it is demonstrated that the
identification results are better than that only using
one sensor information source by numerical examples
[8, 9].
The objective of the research is to develop a stress
identification method using fuzzy pattern recognition
and information fusion technique, while the
acquisition for the stress distribution of the key region
is discussed using the measurements of strain sensors
located on both the key region and the normal regions.
Correspondingly speaking, the acquisition for the
stress distribution of the key region using

Fig. 1

The framework of the stress identification method.

measurements of strain sensors located on the key

region is called the key region identification; and the
acquisition for the stress distribution of the key region
using measurements of strain sensors located on the
normal regions is called the normal region
identification. Firstly, the key region identification
and the normal region identification are obtained by
the fuzzy pattern recognition; secondly, information
fusion technique is used to give the synthesized stress
distribution of the key region; finally, the stress
distribution identification for a shell structure is
simulated and the method is proofed to be effective.

2. Basic Process
The proposed method is divided into early-stage
preparation part and real-time monitoring part, which
is shown in Fig. 1. In the early-stage preparation part,
firstly, the key region identification and the normal
region identification are obtained using fuzzy pattern
recognition; secondly, the selected normal region
identification results are decided using the D-S
evidence theory; thirdly, the different fusion
coefficients for the key region identification and the
selected normal identification are assigned using the

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

weighted fusion algorithm; finally, all the results are

saved in the fusion center for obtaining the
synthesized stress distribution of the key region in
real-time. In the real-time monitoring part, firstly, the
key region identification and the normal region
identification are acquired using fuzzy pattern
recognition; secondly, the synthesized stress
distribution of the key region can be obtained with
different fusion coefficients.

3. Fuzzy Pattern Recognition

The pattern recognition aims to classify the new
pattern into some pattern in the standard pattern
library based on a priori knowledge, in which the
feature vector is extracted to represent the
characteristics of the pattern and the fuzzy neartude is
used to measure the similarity between new pattern
and some pattern in the standard pattern library. To
represent the global characteristics of the feature
vectors, the fuzzy set constructed by the membership
function of each element in feature vector is used to
calculate the fuzzy neartude between patterns.
In the paper, the feature vectors of pattern
recognition can be obtained by the measurements of
strain sensors located on the key region and the
normal regions. The standard pattern library is formed
by calculating the different stress distributions of the
key regions under different load cases in the finite
element analysis, while the domain U of fuzzy pattern
recognition is formed by the feature vectors under
different load cases. In order to reflect the
characteristics of the stress distributions of the key
region better, the variation of stresses in measured
elements should be sensitive to the change of load
cases, so the elements with larger means and variation
coefficients in stress distributions are selected to
located sensors on.
3.1 Membership Function
The ith pattern in the standard pattern library, the
feature vector can be denoted as:

i 1i , 2i , , ni

481

(1)

The membership function of the stress in the jth

measured element in the feature vector can be
expressed as [7]:
s ij - s ij min
mij = i
j = 1, 2, , n
(2)
s j max - s ij min
where, j represents the stress of the jth measured
element of the ith pattern, jmax and jmin are the
maximum and minimum stresses of the jth measured
element in the domain U. In this way, corresponding
with domain U, the fuzzy subset Ai of the ith patter is:
Ai = [m1 , m2 , , mn ]
(3)
The standard pattern library is composed of m
patterns, whose fuzzy set is denoted as A,
A = { A1 , A2 , , Am }
(4)
And the fuzzy set B to be recognized can also be
calculated in the same way. If B is close to Ai, then B
is classified into Ai.
3.2 The Principle of Classification
The fuzzy neartude which can measure the nearness
of two fuzzy subsets is denoted as [10]:
n

D( Ai , B)

2 min( Ai ( x j ), B ( x j ))
j 1
n

j 1

Ai

( x j ) B ( x j )

(5)

j 1

where, B(j) represents the membership function of

the stress in the jth measured elements in the pattern to
be recognized. The fuzzy neartudes of fuzzy subsets
between the pattern to be recognized and the patterns
in the standard pattern library can be acquired by Eq.
(5).
In the original fuzzy pattern recognition method, the
pattern of stress distribution to be recognized is
classified into some pattern with the largest fuzzy
neartude in the standard pattern library, that is to say,
the stress distribution of the key region is the stress
distribution of the pattern with the largest fuzzy
neartude. However, the differences among the first
few patterns with larger fuzzy neartudes are very
small, that is to say, such patterns in the library can

482

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

also represent a certain feature of the pattern to be

recognized. Therefore, in the paper, the first few
patterns with larger fuzzy neartudes are considered to
decide the weighting coefficients for classifying the
new pattern, while an improved method is proposed
and the corresponding identified stress is known as:
m

si =

D s
j

ij

j =1
m

Dj

(6)

j =1

where, iis the stress identification result of the ith

measured element in the key region, Dj is the jth
largest fuzzy neartude, ij is the stress of the ith
measured element in the pattern with jth largest fuzzy
neartude, m is the number of patterns which
participate in the weighted average.

4. Information Fusion
4.1 D-S Evidence Theory
D-S evidence theory [11], a probability-based
information fusion classification algorithm, is useful
when the information sources contributing
information cannot associate a 100 percent probability
of certainty to their output decisions. The D-S
evidence theory is used to select the normal region
identification results which are applied to calculate the
synthesized stress distribution of the key region.
The hypothesis space is made of a finite set of n
mutually exclusive and exhaustive propositions,
which is called the frame of discernment and denoted
by . A power set of 2 is the set of all the subsets of
including itself and a null set . Each subset is
called a focal element. The basic probability
assignment (BPA) is critical variable of evidence
theory. The BPA of any subset A is represented as
m(A), which defines a mapping of 2 to the interval
between 0 and 1. The value 0 and 1 indicate no belief
and total belief in a proposition separately, and any
value between these two limits indicates different
partial beliefs. Formally, the description of BPA can
be represented as:

m : 2Q [0,1]
m(f ) = 0

m( A) = 1

(7)

A2Q

The assigned probability to is considered as the

amount of uncertainty within the BPA.
In the paper, the BPA is acquired by statistical
evidence method [12]. If a group of evidence is
acquired on the results of statistical test, then it is
called statistical evidence. Statistical evidence is the
usage of D-S evidence theory in statistical problems,
which is a new method Shafer used to handle
statistical problem, and is an attempt using
non-statistical methods to solve statistical problem.
The observation of statistical testing is determined
by a group of probability models {p |}, in which
is the frame of discernment, p is the probability
when the proposition is given.
The frame of discernment is defined as {accurate
identification A, inaccurate identification } in the
paper. When all the BPA cannot be assigned to m(A)
and m(), the rest of BPA is assigned to , that is
m()=m(A ). The probability of accurate
identification pA is defined as the probability of the
identification error less than 15%, which is acquired
by analyzing the errors of stress distribution in the key
region obtained by the measurements of sensors in the
key region and the normal regions. Then, the BPA of
is calculated by the statistical evidence method
listed as follows.
Assuming the consensus assumption is met, that is,
0
= {A, } is an ordered set of , which satisfies
pA>p, then the corresponding BPA is
m( A) =

pA - pA
pA

m( A) = 0
p
m(Q) = A
pA

(8)

If the consensus assumption can not be met, the

generalization method of acquiring BPA can be
achieved by weakening the consensus assumption.

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

That is, if 0 = {A, } is an ordered set of , which

satisfies pA<p, then the corresponding BPA is
m( A) = p A
(9)
m( A) = p
A

m(Q) = 0
After calculating the BPA, the Dempsters rule is
used to combine the information. The BPA of fusing
the identification results obtained by the
measurements of sensors in ith and jth regions are
given by
m ( A) = K ( m ( A) m ( A) + m (Q) m ( A) + m ( A) m (Q))
ij
i
j
i
j
i
j
m ( A ) = K ( m ( A ) m ( A ) + m (Q) m ( A ) + m ( A ) m (Q))
ij
i
j
i
j
i
j

(10)

m (Q) = Km (Q) m (Q)

ij
i
j

Where, mi and mj are BPA of the identification

results of the ith and jth regions.
K -1 = 1- (mi ( A)m j ( A) + mi ( A)m j ( A))

(11)
When there are three or more stress distributions
obtained by the measurements of sensors in the key
regions and the normal regions, the application of
Dempsters rule is repeated using the BPA calculated
from the first application of the rule and the other
BPA from another stress distribution. The Dempsters
rule is used to fuse all the possible combination of the
stress distributions obtained by measurements of the
key regions and the other regions. Finally the
combination which has the highest BPA for accurate
identification A will be selected to calculate the
synthesized stress distribution of the key region.

w =1.

where, the fusion coefficients satisfy

i =1

The fusion coefficients w is determined on the

purpose of minimizing the identification errors, which
is calculated by minimizing the fitness function with
genetic algorithm.
The fitness function is as follows:
m

(e

ki

f =

k =1

)2

[(s - s ) / s
i

]2

k =1

the ith elements in the key region with the training

samples; si is the true value of ith element stress;

si is the synthesized stress of ith element acquired
using information fusion, which is given by Eq. (12).

5. Numerical Simulations
5.1 The Finite Element Model
The model is a Schwedler-type single-layer sphere
lattice shell as shown in Fig. 2.
The section of the elements in radial direction is
133 mm in diameter and 4 mm in thickness and the
section of the elements in ring direction and in oblique
direction is 127 mm in diameter and 3mm in thickness.
The finite element model of the lattice shell consists

The weighted fusion algorithm is used to obtain

different fusion coefficients for the key region
identification and the normal region identification. If
the identified stress of one element in the key region
obtained by the measurements of sensors in n regions
are 1,2,n, which are independent of each other,
and the fusion coefficients corresponding to them are
w1,w2,,wn, then the synthesized results is defined as
n

i =1

(13)

m
m
where, m is the number of total identified stresses of

4.2 Weighted Fusion Algorithm

s = wi si

483

(12)
Fig. 2

The finite element model of the lattice shell.

484

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

of 145 nodes, 408 elements, and the supports are fixed

supports. In the analysis, the distributed mass is 200
kg/m2; Youngs Modulus is E = 2.061011 N/m2; the
Poisson ratio is = 0.3.

region 2

region 3

5.2 The Division of the Lattice Shell

The structure is divided into four regions according
to the similarity and the region 1 with 30 elements in
ring direction is selected to be the key region for stress
identification, which is shown in Fig. 3.
5.3 The Selection of the Measured Elements
The measured elements are selected based on the
larger means and variation coefficients in each ring of
the structure are selected the measured elements,
which are shown in Fig. 4 with light color. In addition,
the dark colors in Fig. 4 are the unmeasured elements
in the key region. Moreover, the measured elements in
normal regions have the same arrangements as the key
region.

region 1

region 4

Fig. 3

The division of lattice shell.

24
23
22
21
1920
43 44

In order to determine stress distribution patterns,

wind loads are simulated with mean wind speeds from
5 m/s to 60 m/s, in every 5 m/s, a total of 12 different
grades in wind loads. The simulation time of wind
loads is 10 minutes and the sampling frequency is 10
Hz, so there are 6000 load steps in each grade of wind
loads. There are 6000 patterns of stress distributions
corresponding to each grade of wind loads from the
finite element analysis.
The standard pattern library is formed of 2000
patterns of stress distributions under each grade of
wind loads. The training set is formed of 50 patterns
of stress distributions under each grade of wind load
are extracted to form the training set. The patterns in
standard pattern library and the training set are
different. The testing set is formed of 50 patterns of
stress distributions under the wind load with mean
wind speed 42.5 m/s.
The relative error is used to evaluate the accuracy
of the identification. Meanwhile, the elements with

72

96

46

95

45

120

71
70

69
67

5.4 The Establishment of Samples

48
47

119

94
118

68
93
92

117

91
116
115

Fig. 4 The arrangement of key regions and measured

elements.

large stress value are more important, so the effective

of the method is assessed by the identification relative
error of these elements. The identification results, in
which the stress of the element is more than 100 Mpa,
are considered to be used as assessment in the
effective of the method, while there are 286
identification results.
5.5 The Key Region Identification
The identification results of the original method and
the improved method using fuzzy neartude is
compared and shown in Fig. 5. It can be seen that the
number of cases with errors less than 5% using the
improved method is larger than that using the original
method and the improved method is much more
accurate than the original method.

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

When the measurements are disturbed by noise

level of 5%, 10%, 15% and 20%, which are defined as
the ratio of the root mean square (RMS) of the noise
to the RMS of stress time series, the identification
results of improved method are shown in Fig. 6, the
accuracy of identification results is not be affected
obviously in the situation of low noise level. However,
the accuracy under the disturbing of high noise level is
decreased.

selected by D-S evidence theory and are shown in

Table 1. The region in the Table 1 is the
corresponding regions selected to calculate the
synthesized stress value of the element in the key
region.
The identification results without using information
fusion is denoted as the identification results before
information fusion; the identification results using
information fusion is denoted as the identification
results after information fusion, which are analyzed
and shown in Fig. 7. The results show that the method
using information fusion increases the accuracy of
identification, but the effectiveness is not very
obvious.

5.6 The Synthesized Stress Distribution of the Key

Region
The corresponding regions for calculating the
synthesized stress distribution of the key region is

Identification results

250
200
150

The original method

196

The improved method

152
92

100

65

50

26 19

12 3

0
less than 5%

5%-10%

10%-15%

15%-20%

20%-25%

The distribution of errors

Fig. 5

The distribution of errors with fuzzy pattern recognition.

Identification results

250

Noiseless
5% noise

200

10% noise
15% noise

150

20% noise
100
50
0
less than 5% 5%-10%

10%-15%

15%-20%

20%-25%

The distribution of errors

Fig. 6

The distribution of errors with noise.

485

25%-30%

more than
30%

486

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

When the measurements of sensors are disturbed by

noise level of 5%, 10%, 15% and 20%, the
identification results using information fusion
technique are shown in Fig. 8. Comparing with the
method without using the information fusion, the
Table 1
element

identification results are more accuracy under noise

disturbing especially in the case of high noise level, as
shown in Table 2. The results illustrate that the
proposed method using information fusion technique
is effective.

The regions selected by D-S evidence theory.

region

element

region

element

region

element

region

element

region

19

14

43

124

67

91

124

115

124

20

124

44

1234 68

92

124

116

124

21

45

124

69

124

93

124

117

22

46

124

70

94

118

12

23

47

71

124

95

124

119

124

24

48

124

72

124

96

124

120

124

Identification results

250
200

196 197

Before information fusion

After information fusion

150
100

65 72

50

19 14

0
less than 5%

5%-10%

10%-15%

15%-20%

20%-25%

The distribution of errors

Fig. 7

The distribution of errors with information fusion.

250

Identification results

Noiseless
5% noise

200

10% noise
15% noise

150

20% noise
100
50
0
less than
5%

Fig. 8

5%-10%

The distribution of errors with noise.

10%-15% 15%-20% 20%-25% 25%-30% more than

30%
The distribution of errors

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

Table 2

487

The comparison of identification results with noise.

Noise level
0

Before information fusion

The ratio of errors less
The maximum error
than 15%
97.90%
23.20%

After information fusion

The ratio of errors less
The maximum error
than 15%
98.95%
23.20%

5% noise

96.85%

25.16%

97.90%

25.15%

10% noise

95.80%

23.20%

96.15%

23.27%

15% noise

88.11%

33.60%

90.56%

23.88%

20% noise

82.87%

36.25%

86.36%

32.99%

5. Conclusions
The stress identification method using the
measurements of sensors in the whole structure is
proposed. An improved fuzzy pattern recognition
method introducing the fuzzy neartudes as the
weighting
coefficients
is
discussed.
The
measurements of sensors in the normal regions are
used to improve the accuracy of stress identification in
the key region by information fusion technique. The
following conclusions can be obtained:
(1) The original fuzzy pattern recognition can be
used to identify stresses of elements with limited
measurements of sensors, but the identification
accuracy is not very good.
(2) The usage of improved fuzzy pattern
recognition and information fusion technology can
improve the reliability of stress identification.
(3) The D-S evidence theory is effective in selecting
the normal region identification, and the weighted
fusion algorithm can be used to acquire the fusion
coefficients of each region. But these two methods
could only be used in the early-stage preparation part
with data from the finite element model.
(4) When the measurements of sensors are
disturbed by noise, especially in the case of high noise
level, the use of information fusion technique will
bring in more reliable results.

National Science & Technology Pillar Program during

the Eleventh Five-year Plan Period of China (No.
2006BAJ13B03).

References
[1]

[2]

[3]

[4]

[5]

[6]

[7]

Acknowledgements
The work presented in this paper was partially
supported by Natural Science Foundation of China
(NSFC No. 50678052) and Key Projects in the

[8]

S. C. Liu and D. S. Jiang, The application of structural

health monitoring system technology FBG-based to No.2
Wuhan Bridge over Yangtze River, Photonics and
Optoelectronics Meetings, Fiber Optic Communication
and Sensors, Proceedings of SPIE 72780O, 2008.
F. N. Catbas, M. Susoy and D. M. Frangopol, Structural
health monitoring and reliability estimation: Long span
truss bridge application with environmental monitoring
data, Engineering Structures 30 (2008) 23472359.
T. H. T. Chan, L. Yu, H. Y. Tam, Y. Q. Ni, S. Y. Liu, W.
H. Chung and L. K. Cheng, Fiber Bragg Grating sensors
for structural health monitoring of Tsing Ma Bridge:
background and experimental observation, Engineering
Structures 28 (2006) 648659.
C. R. Farrar, H. Sohn, M. L. Fugate and J. J. Czarnecki,
Integrated structural health monitoring, Advanced
Nondestructive Evaluation for Structural and Biological
Health Monitoring, Proceedings of SPIE, Vol. 4335,
2001.
P. J. Barr, C. B. Woodward, B. Najera and M. N. Amin,
Long-term structural gealth monitoring of the San Ysidro
Bridge, Journal of Performance of Constructed Facilities
20 (1) (2006) 1420.
M. Liu, D. M. Frangopol and S. Y. Kim, Bridge system
performance assessment from structural health
monitoring: A case study, Journal of Structural
Engineering 35 (6) (2009) 733742.
J. Wang, W. L. Qu, J. Zhen and H. Liu, Fuzzy pattern
identification and experimental verification of stress field
of space steel structure, International Conference on
Health Monitoring of Structure, Material and
Environment, 2007, pp. 687793
H. Y. Guo, Structural damage detection using
information fusion technique, Mechanical Systems and
Signal Processing 20 (5) (2006) 11731188.

488
[9]

Structural Stress Identification Using Fuzzy Pattern Recognition and Information Fusion Technique

H. Li, Y. Q. Bao and J. P. Ou, Structural damage

identification based on integration of information fusion
and shannon entropy, Mechanical System and Signal
Processing 22 (6) (2008) 14271440
[10] Y. Xu, K. Y. Qin, J. Liu and Z. M. Song, Fuzzy pattern
recognition and its application, Xian Jiaotong University,
1998.

[11] L. A. Klein, Sensor and data fusion, The International

Society for Optical Engineering, 2004.
[12] Y. He, G. H. Wang, D. J. Lu and Y. N. Peng, Multisensor
information fusion with applications, Publishing House of
Electronics Industry, 2009.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 489495

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Static Requirement and Types Complexity in the

Diagnostics Phase
Antonella Guida and Antonello Pagliuca
Department of European and Mediterranean Culture: Architecture, Environment, Cultural Heritage (DICEM) - Matera, Italy
Abstract: The preservation of a historical building, whatever its architectural and/or artistic value, is more successful when undertaken
with a deep understanding of the buildings history, development, materials and construction techniques. The preliminary phase of any
restoration intervention must start with data acquisition regarding the characteristics and conditions of the building, including a survey
of significant alterations. In a great number of cases, restorations are not respectful of the buildings static efficiency, so its static
requirement is weakened. In fact, a very high percentage of instances in which a restoration effort results in building damage is
attributable to such modifications indifference to the structural balance of a structure, as consolidated over time. This study focuses on
the restoration intervention on two trilobate pillars that separate the central space from the presbytery in the Cathedral of Matera,
located in southern Italy. Through the use of sonic tests - carried out despite the complexity of the shape of the building and constructive
elements that characterize these pillars it was possible to show the effectiveness of the implemented intervention, highlighting
critical points and weaknesses. The research aims to show how despite the complexity of some structural elements of a building
in-depth knowledge of a structures structure and history is essential to for the success of restoration interventions, which are respectful
of a buildings type and material peculiarities.
Key words: Structural complexity, static requirements, refurbishment interventions, diagnosis phase, conservation.

1. Introduction
The preservation of a historical building, whatever
its architectural or artistic value, is more effective when
based on an in-depth understanding of the buildings
development, materials and constructive techniques.
In fact, the protection and conservation of heritage
buildings entails varied and sometimes even alarming
aspects; however, they help us to identify the
commonly called architectural emergency. This is
essentially related to the size and the great number of
cases to resolve and together to the limited
availability of usable resources and experiences. So,
its necessary to optimize in a better way the limited
finances and resources allocated to restoration efforts.
Faced with a degraded monument, the renovation
designer must answer three main questions: if to

Corresponding author: Antonella Guida, professor,

research fields: building refurbishment, recovery and
requalification. E-mail: antonella.guida@unibas.it.

perform a restoration, where to perform a restoration

and how to perform a restoration.
To these questions it is possible to add another one
that regarding the economic aspect: when to perform
a restoration intervention [10].
In order to respond adequately to these questions, it
is necessary to advance step by step, through easily
defined procedures: the identification of degradation,
the determination of its causes, the assessment of
residual safety and, finally, the appropariate
intervention and definition of its execution method [8].
So the preliminary action is the research of all
information about the monuments; this information is
required to describe the structure and all the
transformations that it has suffered; in this situation, it
becomes essential to know the history of the building,
from its construction until the last modification that it
has undergone. In the first phase of data acquisition, the
direct recognition of the building characteristics and
the survey of the checked alteration, should be

490

Static Requirement and Types Complexity in the Diagnostics Phase

complemented by research of design documentation

and of events that have affected the structure during all
its life cycle [9].

2. Shapes Complexity and Static Balance

The structural and static complexity of the historical
building represent the more important characteristics of
each monument. In fact, in the last years, the necessity
to operate interventions of masonry structural
reinforcement has become increasingly urgent,
stimulating both by the objective requirements related
to increased degradation of the monumental buildings
and by a greater interest towards conservation theory.
The restoration criteria, in fact, impose severe
restrictions on the possibility to realize a different
intervention types but at the same time stimulate
the choice of the suitable solutions.
Naturally, these solutions will seek in changes to
realize to the main parameters that constitute the
mechanical part of the structure (i.e., the geometry,
material and loads) [12].
In fact, changes, modifications, partial demolition
and reconstruction are quite frequent in a buildings
history, especially in a monumental building. A lot of
monuments, in fact, have modified its shape and its
structural and functional organization in its life cycle,
realizing for example of super elevation or
addition part of construction, modifying the static
distribution of the force, etc.; a very high percentage of
building damages are attributable to these
modifications are not always sufficiently respectful of
a structural balances consolidated over the time.
These changes led to new structural stresses - not
always proportionately allocated over the building
structure and very often they lead a new shape and
aesthetical characteristics.
Thus, the geometry changes are surely the most
effective (i.e., the increase of stiffness obtained with
the increase of resistant section or by adding a
constraint) [11].

The criterion for an intervention on the geometry

was almost abandoned, although structurally good; the
practice of complete replacement of degraded elements
with other more or less similar to the original is
realized very often, while frequently are offered in
parallel to existing structures structures that are able
to bear the load of the structure, in case of triggering
the collapse.
It is considered, therefore, that the elaboration of a
proposed consolidation intervention is essential to
achieve different objectives that ensure the
maximization of the monument respect.
These objectives are:
(1) to allow maximum persistence of the original
material, limiting at minimum the transformations
(demolition, replacement, etc.);
(2) to recognize the variable time as a positive
sign that are able to add value to the building, which
must be understood as a palimpsest on which are
layered different elements;
(3) to use objective knowledge that are closely
related to the building or specific pathological
situation;
(4) to make decisions on the basis of technical
assessments, supported by an in-depth knowledge and
never based on historical, critical or aesthetic
judgments;
(5) to realize recognizable and reversible
interventions, in order to make possible in the time
control, monitoring, additional maintenance (i.e., the
introduction of new technologies or systems more
effectively);
(6) to establish a correct maintenance program over
time, not aimed at any replacements intervention at
predetermined intervals, but based on observation of
the phenomena, in order to prevent possible dangerous
situations and slow down the aging process of the
building.

3. The Case Study

This research focus upon the Cathedral of Matera
located in a little city in the southern part of Italy (Fig. 1),

Static Requirement and Types Complexity in the Diagnostics Phase

491

a massive architectural structure, built at the end of the

XIII century. In Apulian Romanesque style, it keeps its
formal and architectonic connotations almost
unchanged outside, even though some interventions
have substantially changed the inside style. It does not
show any plain deterioration signs of its static structure,
but it is plain a pathology outside.
The study focuses on the analysis of the restoration
intervention on two trilobate pillars (Fig. 2) that
separate the central space from the presbytery.
Using sonic tests carried out despite the
complexity of the shape of building structural and
constructive elements that characterizes these pillars
it is possible to show the effectiveness of the
implemented intervention, highlighting its critical
points and its weaknesses [1].
In fact, with the sonic tests are measured and
analyzed the characteristics of propagation of elastic
waves inside the walls with the aim to understand the
homogeneity of the structure, the changes of material
properties caused by degradation, the structural defects
such as cavities or cracks and the magnitude of
the resistance of materials [2].
This test is carried out by measuring the time and
speed of the wave that passes through the wall, a
mechanical impulse generated by a transmitter and
picked up by a receiver [3].
The speed of an elastic waves in a homogeneous,
isotropic and perfectly elastic, is linked to the dynamic
elastic modulus Ed, by the following relation:

Ed v 2 d

Fig. 1 The Matera Cathedral.

1 v 1 2v
1 v

in which:
Ed = dynamic modulus in Pa
v = speed of the wave in ms-1
= Poisson modulus
d = density of the elements in kgm-3
The elastic wave loses energy in his propagate, and
this is due to a decrease in the intensity related to the
propagation law of the spherical type, while another
decrease is the interface between the gap, where the

Fig. 2 The trilobate pillars.

energy is partly reflected and partly refracted [4, 5].

High speed closely proportional to the elasticity
modulus and to the strength of the element and short
term indicate a compact structure (i.e., single body
masonry or multiple body well connected), while
discontinuities, fractures, porous materials are
considerably lower speed and consequently increases
time.

492

Static Requirement and Types Complexity in the Diagnostics Phase

The test was performed on the above said columns,

identifying different directions (Figs. 35) at
different heights in order to cover the largest part of
the surface; these directions would also consider the

Fig. 5 Damage visible on the surface of the pillars.

different materials connections, related to the particular

shape of the pillars. Certainly this complexity became
an important element to evaluate in order to operate a
correct and precise evaluation of the investigations
project [7].
Tables 12 show, for each chosen investigations
directions, the values of average speed; in fact, they
represent the average values of 12 speeds measured. In
fact, were made 12 different measurements that
returned 12 different values of propagation time of the
Fig. 3 Sonic directions of investigations carried out on the
left trilobate pillar.

wave and, consequently, 12 different values of

propagation speed.
The sonic test carried out on the left pillar (Table 1)
highlights the values between 1400 m/s and 1800 m/s
at different levels and in the range 1.502.73, which
usually characterize those walls not strongly degraded
and presence of voids.
From the observation of results, it is possible to note
that lower values are not recorded along the longest
paths such as those linking the two semi-columns or
the semi-column with the pilaster but along the
shortest paths within the elliptical shape. This leads us
to assume a more widespread damage, also visible on
the surface of the pillars [6].
While, the investigations on the right pillar (Table 2)
highlight of values between 1450 m/s and 1850 m/s at
different levels and in the range 1.502.73, which
usually characterize those walls not strongly degraded
and presence of voids.

Fig.4 Sonic directions of investigations carried out on the

right trilobate pillar.

Static Requirement and Types Complexity in the Diagnostics Phase

493

Table 1 Sonic tests results on the left trilobate pillar.

As
s(m) t(us)

1.57

Bs
V(m/s)

s(m) t(us)

Es
V(m/s)

s(m) t(us)

942

1315

1374

916

1321

1285

944

1317

1295

921

1352

1281

947

1331

1345

933
922

1686.902

2.37

1346
1343

1763.13

2.37

1340

909

1348

1376

922

1372

1345

941

1355

1363

983

1362

1361

919

1367
s(m) t(us)

1.87

Ms
s(m) t(us)

Hs
V(m/s)

s(m) t(us)

V(m/s)

s(m) t(us)

1133

613

1212

1002

634

1172

1023

628

1116

1017

691

1147

1020

639

1141

1036

1598.837

1.87

1025

1823.145

0.95

644
661

1147

1039

692

1176

1051

668

1161

1027

676

1192

1016

692

1188

1003

690

Ns
V(m/s)

Ls

1199

1173

s(m) t(us)

Os
V(m/s)

s(m) t(us)

Ps
V(m/s)

s(m) t(us)

V(m/s)

s(m) t(us)

448

341

391

553

407

408

336

393

531

400

415

349

389

573

396

429

350

382

563

412

405

348

381

575

394

1474.142

0.58

431
353

1494.075

0.96

343
351

2785.839

0.96

381
394

V(m/s)

1434.395

Qs

371

432

1768.393

1362

Gs

0.61

1332

V(m/s)

2442.748

0.84

633
657

384

354

342

399

630

472

354

350

405

636

406

348

344

405

656

435

354

336

401

648

482

379

343

395

651

V(m/s)

1372.998

Static Requirement and Types Complexity in the Diagnostics Phase

494

Table 2

Sonic tests results on the right trilobate pillar.

Ad

s(m) t(us)

1.49

Bd
V(m/s)

s(m) t(us)

Dd

Ed

V(m/s)

s(m) t(us)

886

1572

1260

871

1356

1287

855

1361

1257

879

1424

1300

899

1388

1240

848
894

1694.337

2.31

1434
1427

1644.713

2.31

#DIV/01

2.31

#DIV/01

2.31

1243
1019

863

1415

1251

883

1423

1249

870

1427

1271

894

1333

1284

922

1390

1292

Fd

Gd
s(m) t(us)

#DIV/01

1.81

Md
s(m) t(us)

0.67

Cd

Hd
V(m/s)

s(m) t(us)

s(m) t(us)

1037

620

1002

957

664

1043

970

650

1046

956

670

1011

937

1011

960

1005

1750.822

1.81

945

1883.259

#DIV/01

683
678

947

724

989

980

689

1037

963

701

1096

966

709

1116

967
s(m) t(us)

Pd
V(m/s)

s(m) t(us)

Qd
V(m/s)

s(m) t(us)

452

444

394

472

474

422

434

393

452

521

488

440

434

436

560

461

416

422

433

546

411

416

412

472

552

436
485

1459.059

0.60

408
402

1440.23

0.98

407
416

1477.541

670

Od
V(m/s)

V(m/s)

654
1

1005

s(m) t(us)

1828.4

Ld

V(m/s)

1082

Nd
V(m/s)

Id

V(m/s)

2401.372

0.98

436
407

2164.789

0.83

543
667

466

415

403

472

648

459

410

418

440

565

452

418

400

452

623

471

402

406

462

594

496

407

403

536

601

V(m/s)

1442.726

Static Requirement and Types Complexity in the Diagnostics Phase

Even the pillar right, by the observation of partial

results show that higher values are recorded along the
longer trajectories (as above said), the lowest (about
1400 m/s) along the shortest paths within the elliptical
shape. This leads us to do the same considerations of
the left pillar.
Sonic tests were carried out also on the stone base of
the pillars using the same investigation test. The
measured values are greater than 2000 m/s that
indicates a good state of preservation.
The research applied on the above described
pillars is on going also for the other structural
elements of the Matera Cathedral in order to provide a
complete knowledge of the monument and to develop a
suitable restoration intervention.

4. Conclusions
The research aims to show how despite the type
complexity of some structural elements of a building

becomes

essential

an

appropriate

building

knowledge to realize interventions that are able to

operate a correct restoration intervention of a building
and, at the same time, which it is respectful of its type
and material peculiarities.
The

study

aims

to

realize

an

applicative

methodology more respectful of the monumental

heritage, in order to avoid a simple critical
transposition of the calculation models in the recovery
interventions.
This research can also contribute to increment the
typology of the tests to carry out on the masonry; in fact,
in relation to all the problems of tests and
investigations (first of all the geometrical complexity),
it is not necessary to have only one kind of survey, but
an organic plan that consider in particular way the
shape and the morphology of the investigated element,
in order to have a more possible complete cognitive
frame. Only in this way it is possible to interpret the
real phenomena and give the necessary parameters to
realize a correct intervention.

495

References
[1]

A A.V V., Chiesa di S. Maria della Piet. Irsina (MT), in:

Mancosu (Ed.), Trattato sul consolidamento, Roma, 2003,
pp. C526C535,
[2] L. Binda, A. Saisia, C. Tiraboschi, S. Valle, C. Colla and
M. Forde, Application of sonic and radar tests on the
piers and walls of the Cathedral of Noto, Constr. Build
Mater 17 (8) (2003) 613627.
[3] L. Binda, Metodi statici di stima della capacit portante di
strutture murarie, in: Comportamento statico e sismico
delle strutture murarie, Roma, 1982,
[4] L. Binda, A. Saisi, S. Valle and C. Zanzi, Indagini
soniche applicate alle murature in mattoni: calibrazione
ed individuazione dei parametri significativi, V
Congresso Nazionale ASS.I.R.C.CO, Orvieto 1997, pp.
7781.
[5] L. Binda, C. Modena, G. Baronio and S. Abbaneo, Repair
and investigation techniques for stone masonry walls,
Construction and Building Materials 11 (3) (1997)
133142.
[6] P. Bonaldi, P. P. Rossi and L. Jurina, Osservazioni sulle
vicende statiche e geotecniche del Palazzo della Regione
di Milano, in: XIV Convegno Nazionale di Geotecnica,
Firenze, 1980.
[7] B. Calia and A. Mecchi, Microsismic tests in the analysis
and characterisation of high porosity stone materials, in:
International Conference on NDT, Roma, 1999.
[8] E. Carabelli, I metodi geofisici nelle indagini su vecchie
murature in Corso di restauro edilizio e monumentale,
Bergamo, 1980.
[9] S. Croce, Patologia edilizia: prevenzione e recupero, in
Manuale di progettazione edilizia, Milano, 1994.
[10] J. Kerisel, La citt e gli edifici antichi in Relazione
generale presentata al X Convegno Internazionale
SMFE, Stoccolma, 1981.
[11] A. Guida, F. Fatiguso, F. Guida and C. De Fino, Il
Castello di Cancellara, castrum seu fortilitium. Una
proposta metodologica per la tutela e la valorizzazione,
Ed.Ermes PZ, 2005.
[12] A. Pagliuca and A. Guida, La diagnostica comparata
come strumento per la durabilit degli interventi, in:
Proceedings of the CIB 11th International Conference
on Durability of Building Materials and Components,
Istanbul, Turkey, May 1114, 2008.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 496503

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Characteristics of User Redesign Process: A Study of

Changes Made by Users in Architect-Designed Housing
Nghai Ezekiel Suleman
Department of Architectural Technology, Federal Polytechnic, Kaura Namoda, Zamfara State, Nigeria
Abstract: Housing is dynamic. It changes with time. A sample of 60 houses in the Shagari Low Cost housing scheme in Kaura Namoda
was studied in order to find out the characteristics of changes made by the users. Physical observations and questionnaires were used as
the major instruments of data collection. The results show that 91% of occupants have made various changes to the original designs
without the help of professional designers, and some of the changes have completely transformed the outlook of the houses in locational
planning, materials and form. This implies that a post architectural design phase exists, suggesting that the design process can be
divided into two phases; the initial design by the architect and the subsequent redesign by the user. The paper presents preliminary
findings on the general characteristics of user redesign and concludes that a clear process can be established. The architectural design
process can, therefore, be expanded to include the user redesign process in order to reflect the entire lifespan of the building.
Key words: Architectural design process, changes, expanded design process, housing, user redesign.

1. Introduction
Built environments have great potentials for change.
Change induced by dwellers is necessary at all levels
for a healthy, vital steadily improving environment
Habraken [1]. Brand [2] discusses how various types of
buildings change, concluding that institutional
buildings are the most resistant to change, commercial
buildings change kaleidoscopically while domestic
buildings are the steadiest changers as they have to
respond daily to the interaction of man and the house.
Housing environments change constantly owing to a
number of factors which include changing family size
and structure, changing tastes, development of new
technology, fashion, and so on. Habraken [1] explains
that housing has great potentials for change and that
this fact is a major factor that architects should consider
in the design of housing. He calls on designers not to
see housing as a finished product like with the rest of
architectural artefacts but to consider the process of
change. The right attitude in design of housing should

Corresponding author: Nghai Ezekiel Suleman, MSc

(Architecture), research field: housing design process. E-mail:
nghaisule@yahoo.com.

be - one cannot finish things; one can only make them

possible. Similarly, Granath [3] calls on architects to
design architecture in such a way that it supports
participation in the use of architecture. Architecture,
especially housing, should be responsive to changes.
Users modify their housing environments
consistently. These changes often distort the original
intentions of the architect. Changes by users are
therefore significant in defining the housing design
process. Therefore, the architects inability to
eventually predict user needs that are not clearly stated
and the changing tastes of the user have often
culminated in changes in the design during occupation.
Boudon [4] studied Le Corbusiers Quartiers Modernes
Fruges, a rental housing scheme built in Pessac in 1926,
and discovered that the occupants changed the
character of the initial designs in order to express their
personality. This study points to the fact that no matter
how creative a designer may be, his designs could
undergo changes by the users if they are not involved in
the design process. The involvement of the user in
changing the original designs is a part played by the
user that has been overlooked but that is significant to
the design process.

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

The focus of this paper, therefore, is to study the

general characteristic of changes made by a sample of
users of the Shagari Low Cost housing scheme in
Kaura Namoda, Zamfara State, Nigeria and therefore
attempt to chart the processes that produced the
changes as having a distinct characteristic and to place
it on an expanded architectural design process
proposed by Suleman [5].

2. Background of Study Area

The study was conducted at the Shagari low cost
housing scheme in Kaura Namoda. The Shagari
Lowcost housing schemes were built all over the
country in the early1980s and were meant to house the
low income population. The design is basically a one
bedroom core house with a toilet and bath and a
kitchenette (see Fig. 1). The major concept was to
produce an initial minimal space that could be
expanded incrementally by the user.
The Shagari low cost at Kaura Namoda is located at
the north western part of the town, off the Kaura
Namoda - Shinkafi road. It consists of 100 housing
units. In the 1980s when the Shagari low-cost housing
was constructed, it was widely rejected throughout the
country based mainly on location, insufficiency of
space and lack of basic infrastructure. In Kaura
Namoda the houses were rejected and were finally
offered to the Federal Polytechnic which also rejected

Fig. 1 Original plan and view of house.

497

them. They were therefore sold to individuals. In the

1990s when increase in population of the polytechnic
brought acute shortage of staff housing, the polytechnic
had to acquire most of the housing units in the Shagari
estate at exorbitant rates and turned them into staff
quarters on affordable rents.

3. Methodology
For the study to be meaningful, various types of data
needed to be generated; data establishing that
significant changes occur in houses designed by
architects, data showing the general process of user
redesign, and data from which a general process of user
redesign can emerge. Therefore, different data
collection methods were used including physical
observations, questionnaires, interviews and protocol
analysis.
The initial plan was to show that changes do
significantly occur in housing designed by architects.
As a result physical observations were thought of as the
most appropriate way to study changes made by
occupants to architect-designed houses. Observed
changes were documented through notes, sketches and
photographs and measured drawings. An aspect of
physical observation recommended by Zeisel [6] and
which is central to this study is observing adaptive
traces. Adaptive traces are significant for designers
because they are direct manifestations of design by

498

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

users Zeisel [6]. Visual data were used mainly to

prove that users do effect changes in housing, and that
the changes are significant as they affect all aspects of
the building. Questionnaires were also administered to
occupants to find the condition of their housing at
occupation and the current state. The questionnaires
also focussed on the process of change. A random
sample of 60 houses, representing 60% of the houses,
was chosen for the study. Some few houses were
chosen purposefully to record the details of the changes
that had been effected by the users.
Interviews were also conducted with occupants
specifically on the process of change they adopted. The
unstructured method of retrospective protocol analysis
was adopted. Retrospective protocols of selected users
were taken using a tape recorder and these were later
transcribed and coded and finally compared to the three
stages of the architectural design process analysis,
synthesis and evaluation in order to establish the
process the users applied.

4. Results
Preliminary results from physical observations and
questionnaires are presented in this section.

4.1 Physical Observations

Observations showed that a lot of changes have
occurred in the estate. Despite this the estate has
maintained its general outlook except for few houses
where the complete outlook of the house has been
changed. Changes were dominated by changes of
functions of spaces, and increase in room sizes which
affected the structure of the building. Decoration was
also affected as people changed finishes applied to the
houses. Owing to the lack of pipe borne water in the
estate, toilets, bathrooms and kitchens were generally
relocated outside the houses. Figs. 28 show various
modifications of the original plans.
Fig. 2 shows the original plan converted to a
3-bedroom house with the living room and bedroom
retained while other facilities are relocated with
additions made. The conversion changed the entire
outlook of the house. Window hoods were added for
sun shading. Fig. 3 shows the house converted to 3
bedrooms with a shop. This has also affected the front
view. Fig. 4 shows addition of a high fence and gate,
while Figs. 58 show different configurations of
2-bedroom options as effected by various users.

Fig. 2 Plan and view of redesigned house converted to 3 bedrooms.

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

shop

Fig. 3 Plan and view of redesigned house with shop addition.

Fig. 4 Plan and view of redesigned house converted to 3 bedrooms with high fence wall and gate.

Fig. 5 Plan and view of house converted to 2 bedrooms.

499

500

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

Fig. 6 Plan and view of house converted to 2-bedrooms with fence and gatehouse.

development of the house. They cannot therefore be

ignored; they form part of the design process.
4.2 Questionnaires
Of the 60 questionnaires distributed, 48 representing
80% were returned. From the questionnaires, it was
discovered that 91% of respondents had made changes
to the original design and the 9% that were yet to make
changes indicated plans to make changes in the near
future. Detailed findings from the questionnaires are
presented in Tables 15 below.
Fig. 7 Plan of house converted to 2-bedrooms.
Table 1

Types of changes.

Type of Change

Percentage

Increase in number of rooms

55

Size of spaces

15

Change affecting form

Changes affecting materials/finishes

11

Table 2 Sources of redesign ideas.

Source

Percentage

Intuition

29

Experience from lived-in problems

49

Suggestions from laymen (family, friends, etc)

Assistance of professional

15

Table 3 Constraints encountered.

Type of Constraint
Fig. 8 Plan of house converted to 2-bedrooms.

These changes are significant as some of the houses

have completely been changed beyond their initial
outlook, showing that user redesigns are part of the

Percentage

Cost

77

Technical competence

11

Plot size

10

Availability of materials

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing
Table 4

Factors that caused changes.

Factor

Percentage

Seeking solution to problems

29

Increase in family size

49

Change of taste

20

Change in economic status

501

designers; rather than looking at houses as static

objects, designers should view them as dynamic. This
view is consistent with that expressed in Habraken [1],
calling on designers of housing not to see it as a final
product but consider the possibility of future change.
5.2 Changes affect All Aspects of Design

Table 5

Major Areas Affected by Changes.

Area

Percentage

Bedrooms

47

Kitchens

22

Store

14

Toilets

11

Living Room

Others (eg addition of shop)

5. Characteristics of User Redesigns

From the physical observations and questionnaire
the following deductions were made:
5.1 Change in Housing is Inevitable
Item 2 in Table 6 shows that a total of 91% of the
houses have undergone various changes that have
completely affected the original designs. Changes
affected the function of spaces away from what was
originally planned by the architect. It cannot therefore
be said to be the same building; it has been redesigned
by the occupant. This implies that changes are
inevitable in occupied houses and are a very significant
part of the design process. This is instructive to
Table 6 Some parameters that affected changes.
S/No Parameter

% Yes % No

Involvement in original design

100

Made changes to original design

91

Redesign was affected by initial design

68

32

96

89

11

Professional assistance in design

Affected by similar changes
neighbours
Produced alternative solutions

20

80

Used professional builders

32

68

Complete demolition
Communication of changes through
sketch
Need for training of users in basics of
design

100

12

88

85

15

9
10

by

Architectural design concentrates mainly on three

aspects; function, form (structure), and aesthetics. The
survey discovered that all these aspects of design were
affected. Though most changes were functional, there
were also changes in structure caused by addition of
rooms and changing of roof structure. Aesthetics is
affected by changes in finishes and painting and
creation of new forms. Figs. 28 show the various
changes in plan and form.
5.3 No Professional Assistance in Redesign
Most of the respondents indicated they did not
consult any professional in the redesign of their houses.
Item 3 in Table 6 shows that only 4% of respondents
got professional advice. This indicates that the
redesigns were made by the users themselves thus
using the architects initial design as a starting point for
their designs. Producing their designs without the help
of professionals is also significant. It brings out the
creativity of the user in solving the problem created
by the architect. The redesign by the user is part of the
evolution of the building and should form part of the
entire design process.
5.4 Hands-on (Trial-and-Error) Approach
Item 9, Table 6 shows that only 22% of respondents
communicated their ideas through sketches, while 88%
did not use sketches. This is a major difference between
the architects design and the redesign of the user.
Unlike the architect who uses drawings and sketches,
user redesigns are mental unaccompanied by sketches.
Most of the changes are made on site as the builders
work. Mistakes are corrected as observed
(trial-and-error). This is usually costly as it implies
undoing some of the work through demolition and

502

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

reconstruction. Table 3 shows that the major constraint

faced by occupants is cost (77%). This is likely because
of such trial and error approach. Lack of sketches also
affects evaluation. Ideas on paper can more likely be
easily understood and criticised than mental ideas.
Sketches are very useful in design as they help
reflective criticism Lawson [7]. One major
importance of sketching in the design process is that it
serves as external memory and reduces the load on
internal memory of the designer and thus helps the
designer to recall easily by looking through the sketch
Suwa et al. [8].
5.5 Problem-Driven Design Approach
Architectural designs are intuitive problem-solving
processes. Ideas for user redesign stem mainly from
experience of problems with the design. 49% of
respondents affirmed this while 29% got ideas by
intuition, 7% got ideas from other laymen and 15%
from professionals (Table 2). Item 6 in Table 6 shows
that only 20% of respondents produced alternative
solutions, while 80% did not produce alternative
solutions.
User
redesigns
are
therefore
problem-driven while that of the architect is regarded
as solution driven. Problem driven designs are seen as
design in which the designer focuses closely on the
problem at hand and only uses information and
knowledge that is strictly needed to solve the problem.
The emphasis lies on defining the problem, and finding
a solution as soon as possible, and solution-driven
design as designs in which the designer focuses on
generating solutions, and only gathers information that
is needed to further develop a solution. The emphasis
lies on generating solutions, and little time is spent on
defining the problem, which may be reframed to suit an
emerging solution. [9].
5.6 Users not Involved in Original Designs
Item 1 in Table 6 shows that none of the users was
involved in the design of the house occupied. In fact,
none had knowledge of the designer. Architects have

been criticised for designing without involving the

users and calls have been made for the involvement of
users in the design of their houses [6, 1014],.
5.7 Changes Caused by Needs
Increase in family size was the major reason
advanced for the need for changes by 49% of the
respondents, while 29% of changes were caused by
need to solve some problems. Other reasons are change
in taste (20%) and change in economic status (2%)
(Table 4). This is further confirmed by the dominance
of addition of rooms as the major type of change (47%)
as may be seen in Table 5. This further affirms that as
long as changes in taste occur due to changes in status
and family size for example, the house cannot remain
static; it must change with changing tastes.
5.8 Direction of Changes Affected by Original Designs
Item 3 in Table 6 shows that 68% of respondents
agree that the direction the changes took was detected
by the original designs. 32% felt the original designs
had nothing to do with changes made. This is
instructive to architects to realise they need to make
allowance in their designs for future changes. It also
calls for studies to identify likely direction of user
redesigns to help the architect make his initial design
flexible to accommodate the anticipated future
changes.

6. Inferences
Changes

from

Characteristics

of

From the study, the general process user redesigns

take is as follows:
(1)
Experience
lived-in
problems
from
architect-designed house
(2) Analyse problems
(3) Form mental picture of required changes
(5) Input from friends and family
(6) Effect changes
A clearly well-defined process of user redesign is
similar to that of the architect. It is necessary to note

Characteristics of User Redesign Process: A Study of Changes Made by

Users in Architect-Designed Housing

that because the user has lived long with the problem, it
is well-defined; he knows clearly what he wants.
Solving well-defined problems, such as mathematics,
tends to be linear. It is possible therefore that the user
redesign process, unlike the cyclical design process of
the architect, could be linear. Staying with the problem
for long means there has been some synthesis before
the solution could have been derived, however, lack of
procedural knowledge and absence of sketching have
limited the ability to synthesise. The synthesis stage
appears not well developed but forms part of the
problem analysis. Lack of alternative solutions
completely eliminates the evaluation stage. It is
possible to conclude tentatively therefore that user
redesign process is a linear process composed of a
strong analysis and a weak synthesis that are fused
together producing a final solution. Further analysis
through retrospective protocols will be helpful in
clearly defining the process of user redesign.

7. Conclusion
The study has shown that changes in buildings are
inevitable. Users make changes on designs initially
made by the architect to meet their changing tastes.
These changes often result in complete transformation
of the initial model created by the architect into
something completely different. Changes made by the
user are often independent of the architect. This
process of user redesign has its own distinct
characteristics from that of the architect. This shows
that the input of the user is very significant and should
be reflected in the design process to cover the entire
lifespan of the building. The study points to a possible
two-staged design process; the first stage by the
architect that has been found from studies to be cyclical,
leading to the final product, and the second stage of
user redesign which uses the architects final product
as its starting point to create a final solution. This calls
for further research, for example, using retrospective
protocol analysis, to plot an acceptable process of user
redesigns, and finally incorporating the resulting user

503

redesign process in the architectural design process to

produce a comprehensive design process that would
reflect the entire lifespan of the building.

References
[1]

[2]
[3]

[4]
[5]

[6]
[7]
[8]

[9]

[10]
[11]

[12]
[13]

[14]

N. J. Habraken, Design for adaptability, change and user

participation, in: Safran (Ed), Housing: Process and
Physical Form, Aga Khan Award for Architecture, 1980,
available online at: http://www.archnet.org, Dec. 2006.
S. Brand, How Buildings Learn: What Happens after
They Are Built, Viking, NY, 1994.
J. A. Granath, Architecture: Participation of users in
design activities, 2001, available online at:
http://granath.arch.chalmers.se/private.encyklopedia.htm,
Dec. 2006
P. Boudon, Lived in architecture: Le Corbusiers Pessac
revisited, Trans., M.I.T Press, 1972.
N. E. Suleman, Redefining the architectural design
process to reflect user redesigns, Journal of the
Association of Architectural Educators in Nigeria,
AARCHES J 6 (1) (2007) 4651.
J. Zeisel, Inquiry by design, Cambridge University Press,
1984.
B. Lawson, How Designers Think (4th ed.), Architectural
Press: London, 2006.
M. Suwa, J. S. Gero and T. A. Purcell, The roles of
sketching in early conceptual design process. Proceeding
of Twentieth Annual Meeting of Cognitive Science
Society, Lawrence Earlbaum, Hillsdale, New Jersey,
1998, pp. 10431048.
C. Kruger and N. Cross, Solution driven versus problem
driven design: Strategies and outcomes, Design Studies
37 (5) (2006) 527548.
N. J. Habraken, Supports: An Alternative to Mass
Housing, MIT Press, 1972.
R. G. Hershberger, Predicting the meaning of architecture,
in: Lang, Burnette, Moleski and Vachon (Eds.),
Designing for human behaviour, Dowden, Hutchinson &
Ross, Inc., Stroudsburg, Pennsylvania, 1974, pp.
147156.
A. Raporport, Development, culture change and
supportive design, Habitat Intl. 7 (5/6) (1983) 249269.
A. M. Duany, Architects are neglecting the bulk of
housing consumers, causing a dearth of good design and
rift in the profession. Architectural Record, April 22nd,
1999.
L. S. Popov, Architecture as social design: The social
nature of design objects and the implication for the
profession, The Journal of Design Research 1 (2) (2002).

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 504508

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Irradiation Damage of Nano-C2S Particles Studied by

In-Situ Transmission Electron Microscopy
Hongfang Sun1, Mahir Dham2, Eric A. Stach1 and Carol Handwerker1
1. School of Materials Engineering and Birck Nanotechnology Center, Purdue University, West Lafayette, IN 47907, USA
2. Mapei Corp., Deerfield Beach, FL 33442, USA
Abstract: Development of a reactive nanocement is a new approach to improve the physical and chemical properties of construction
materials. However, due to the decreased size of cement particles, beam damage during transmission electron microscope (TEM)
observation becomes more severe than in conventional cement. In this work, irradiation damage to nano-C2S (dicalcium silicate) is
observed and studied by in-situ evolution of diffraction patterns (DP), high resolution TEM (HRTEM), and electron energy-loss
spectroscopy (EELS). The results show that the damage to nano-C2S occurs through a decomposition reaction. Nano-C2S is first
amorphized, and then re-crystallized into CaO nano-crystals with average size of 7 nm surrounded by an amorphous matrix of Si and
SiO2. During this process, C2S particles exhibit volume shrinkage. The damage energy causing the reaction was analyzed and
electron-electron inelastic scattering produced radiolysis and heat, leading to the observed phenomena.
Key words: Nanocement, TEM, damage, inelastic scattering, decomposition.

1. Introduction
Cement is a mixture comprised of different
components, i.e., C3S, C2S, C3A, and C4AF (where C =
CaO, S = SiO2, A = Al2O3, F = Fe2O3), among which
dicalcium silicate (C2S) is one of the major components.
However, due to its slow hydrating nature, C2S is
usually used to provide long-term durability of
concrete, and its early strength is mostly dependent on
other minerals, such as C3S and C3A. A need to obtain a
reactive form of C2S mineral is based on this
consideration. By improving its reactivity, C2S may
participate in early strength evolution and has the
potential to improve medium term hydration as well.
The development of nanocement [2, 7, 8] brings focus
to this goal. An early age reactivity could be increased
tenfold through exposing much more reacting surface
area compared with conventional cements [8]. Based
on this idea, nano-sized C2S was therefore fabricated
Corresponding author: Carol Handwerker, professor,
research fields: develop and apply thermodynamic and kinetic
theory and experiments of phase transformations and interface
motion to solve important industrial and scientific problems.
E-mail: handwerker@purdue.edu.

through a chemical combustion method. To understand

the hydration properties of nano-C2S, the precursor
powder must first be well characterized. Transmission
electron microscopy (TEM) is a commonly used tool
for characterization of nanoparticles. However, with
TEM we have to assess the extent of irradiation
damage caused by the electron beam to
electrically-insulating cement specimens [6, 9, 10]. In
this work, beam damage of nano-C2S particles was
characterized during TEM observation and methods to
reduce the damage are proposed.

2. Experimental
Nano-C2S was synthesized from limestone and other
naturally occurring materials by NanoDynamics
(www.nanodynamics.com),
using
a
chemical
combustion reaction that is initiated at approximately
300C versus the 1500C used in the preparation of
conventional cement. The microstructure and phases of
the nano-C2S were characterized by scanning electron
microscopy (SEM, FEI Quanta 3D dual-beam system),
X-ray diffraction (XRD, Bruker D8 instrument), and
the MAUD Rietveld method [1]. To prepare TEM

Irradiation Damage of Nano-C2S Particles Studied by In-Situ Transmission Electron Microscopy

505

samples, particles were ultrasonically dispersed into

100% alcohol to form a suspension. Then a cement
sample was obtained by dipping a carbon-coated
copper grid support into the suspension for one second.
TEM characterization was performed using Tecnai 20
(FEI) operating at both 200 kV and 120 kV with a LaB6
filament and Titan 80-300 (FEI) at 300 kV with a Field
Emission Gun (FEG).

3. Results and discussion

The typical morphology of the nanocement in Fig.
1a consists of lightly agglomerated nanoscale particles
with an average size of 100 nm, in contrast with the
typical particles of ordinary Portland cement (OPC)
shown in Fig. 1b. The XRD pattern shown in Fig. 1c
indicates that the dominant phases of the nanomaterial
as C2S and CaCO3 (Fig. 1c) with the analysis shown in
Table 1.
Characterization by TEM is performed with
uniformly dispersed cement particles. Beam damage
induced phase transformation of small particles (less
than 100 nm) occurred essentially instantaneously on
exposure to the beam, much too fast for the reaction
path to be determined. Thus larger ones (no less than
100 nm) were chosen to study the chemistry before
radiation since the damage to them occurred more
slowly. Elements Ca, Si, O were observed from the
EELS (Electron Energy Loss Spectroscopy) spectra
(Figs. 2a2c), which indicates the composition of the
particle in Fig. 2b is C2S rather than CaCO3. After
electron irradiation, the resulting microstructure of the
particle is shown in Figs. 2d2g. Fig. 2e is a
high-resolution image taken from the region indicated
by the arrow in Fig. 2d, demonstrating that the
damaged particle consists of nano-sized subgrains with
an average size of around 7 nm. Fig. 2f is a fast Fourier
transformation (FFT) of the region in (b), which
matches the selected-area diffraction pattern (SADP) in
(g). The diffraction pattern can be indexed to the cubic
CaO phase, indicating that a decomposition reaction
transformed the C2S to CaO and a siliceous product
under electron beam irradiation.

Fig. 1 SEM image of a typical nanocement (NC) aggregate

(a) comparing with a particle of ordinary Portland cement
(OPC) in (b). XRD pattern in (c) showing the principal
phase of nanocement. A-C2S; B-CaCO3.
Table 1
Chemical analysis of nanocement by using
Rietveld method.
Phases
C2S
CaCO3
Others

Proportion in nanocement (Wt%)

82.65
13.79
3.56

Fig. 2 EELS spectra of the cement particle in (d) before

radiation damaged: (a) Ca, (b) Si, and (c) O. The existence of
Si demonstrated the composition of the particle in is C2S
other than CaCO3. (d) - (g) are the characterization of
damaged products of the particle in (d). (e): high resolution
image of the arrow indicated region in (d); (f): fast Fourier
transformation (FFT) of image (e); (g): selected-area
diffraction pattern. The index of rings in (f) and (g) indicates
the existence of CaO after radiation.

506

Irradiation Damage of Nano-C2S Particles Studied by In-Situ Transmission Electron Microscopy

In-situ observation of the phase transformation

process of nano-C2S was illustrated by a series of
diffraction patterns in Figs. 3a3f. Initially, a bright
diffraction pattern near the 100 zone axis were
observed (Fig. 3a). The corresponding particle
morphology is shown in Fig. 3g. With irradiation, the
diffraction spots became weak due to the partial
decomposition of C2S [see spots in (b)], until most of
the spots disappeared and an amorphous pattern was
obtained after 29 min of irradiation (Fig. 3c). The
particle corresponding to the pattern in Fig. 3c is shown
in Fig. 3h. Amorphization is considered to be one of the
signs of being radiation damaged. Similar phenomenon
was observed in C-S-H phase in Ref. [6]. Volume
shrinkage was also observed during the process
[regions indicated by the black arrows in (g)]. After 36
min, new diffraction spots started to form (black
arrows in Fig. 3d), suggesting the formation of new
crystals. More spots appeared after 50 min of
irradiation, indicating additional crystallization (Fig.
3e). Finally at 89 min, a well-shaped CaO
polycrystalline ring pattern was obtained with the
corresponding microstructure seen in Fig. 3i. The
particle showed additional volume shrinkage, as shown
in Fig. 3h (See regions marked with arrows in (h))
Fig. 4 shows the decrystallization process of C2S in
high-resolution mode using the Titan TEM. In order to
have the time resolution to observe the reaction, a
thicker region from a relatively large particle was
chosen. At 0 min, a perfect crystal lattice was observed
in the region circled by a dashed line (Fig. 4a). The
lattice spacing is measured to be 0.261 nm,
characteristic of C2S, not CaCO3. At 14 min of
irradiation, partial amorphization has occurred (Fig. 4b)
with close to complete amorphization after 16 min (Fig.
4c). During this process, CaO phase was not observed,
most probably because the CaO nano-crystals formed
in a different focal plane. The focal distance in
high-resolution imaging is quite short (of order several
nanometers), and thus it is not always possible to
imagine multiple regions of a three-dimensional
particle concomitantly with this approach.

Fig. 3 Series of selected area diffraction patterns (taken by

Technai 20) illustrating the radiation damage and
recrystallizing process with time. [(a)-(f)]. The radiation
time is marked at the right bottom of the diffraction. The
corresponding relationship between special diffraction and
particle morphology is marked by A1, A2, and A3 at the
bottom of the image, which means (a), (c), (e) correspond to
(g), (h), (i), respectively.

Fig. 4 Decrystallization process caused by radiation and

observed in high resolution mode on Titan. The radiation
time is marked at the right top of each image.

Decomposition of C2S results in the formation of

CaO in form of nanocrystals. Electron energy loss
spectroscopy (EELS) was performed using the silicon
L edges to investigate the form of the siliceous
decomposition product. Each spectrum presented in

Irradiation Damage of Nano-C2S Particles Studied by In-Situ Transmission Electron Microscopy

Fig. 5 was recorded for 10 s. Initially, distinct Si-L1, 2, 3

edges were observed, which suggests the existence of
Si4+. The broad peak observed at about 131 eV was
likely due to an inner-well resonance [11]. Five
minutes later, the intensities of the Si-L1, 2, 3 edges
started to decrease, and continue to decrease with
irradiation time. After 12 minutes of irradiation, an
identifiable shoulder near 100 eV appeared, indicating
the formation of Si0 [3]. The shoulder became more
distinct with time. After 66 min of irradiation, both Si0
and Si4+ were observed although the Si-L1, 2, 3 edges
became more flat. Combining these results with the
images of the microstructures in Figs. 2 and 3, we
conclude that Si0 and Si4+ are associated with the
formation of amorphous Si and SiO2 nanoscale phases,
a result that is consistent with the observations of the
development of diffuse scatter in the diffraction
patterns.
We investigate the effect of electron voltage in order
to further elucidate the mechanism of beam damage.
Electron irradiation can lead to sample damage through
several mechanisms, such as sample heading, knock-on
(displacement) damage and radiolysis (bond breaking)

Fig. 5 Series of EELS spectra of Si edge, illustrating the

effect of radiation with time. The Si-L2,3 edge indicates the
existing of Si4+. The edge at about 100 eV indicates the
formation of Si0, which becomes more distinct with time.

507

[4, 5]. The gradual amorphization observed in Fig. 3 is

consistent with a transformation from a crystalline
lattice to an amorphous state, and suggests that
electron-electron inelastic scattering is leading to
radiolysis during irradiation of C2S, while with further
irradiation, the atoms reconfigure to form CaO. In
general, radiolysis can be minimized by using the
highest available accelerating voltage to reduce the
inelastic scattering cross section [4, 5]. We found that
the reaction rate was slower with a voltage increase
from 120 kV to 200 kV using the Tecnai 20, a result
consistent with radiolysis being the primary damage
mechanism. Other ways to reduce the damage are using
a smaller condenser aperture or taking pictures at lower
magnification since both reduce the dose reaching the
sample. Another form of damage to C2S particle is
through direct heating by the beam [5] since C2S is a
ceramic with low thermal conductivity. Beam heating
may also lead to the observed volume shrinkage and
the mass loss of C2S particle shown in Fig. 3. To reduce
beam heating-induced damage, carbon coating both
sides of the TEM sample surface should help.

4. Conclusion
Beam damage to nano-C2S particles during TEM
observation was systematically investigated. C2S
particles are first amorphized, and then re-crystallized
into CaO nano-crystals with average size of 7 nm
surrounded by an amorphous matrix of Si and SiO2.
The reaction pathway suggested that the primary
damage was radiolysis induced by electron-electron
inelastic scattering. Beam-induced heating was
determined to be a secondary cause of damage, leading
to volume shrinkage and mass loss. According to the
forms of radiation damage, methods have been
proposed to reduce or even avoid the damage, such as
using smaller condenser apertures, increasing
accelerating voltage, and carbon coating the surface on
both sides. This work will help understand the data
collected from TEM observation of nanocement, and
provide useful suggestions for performing future

508

Irradiation Damage of Nano-C2S Particles Studied by In-Situ Transmission Electron Microscopy

in-situ TEM studies of the hydration reaction of

nanocement particles.

Acknowledgement
This work was supported by Army Corps of
Engineers. The authors also thank Bob R. Colby and
Jason Weiss for their useful discussions during
manuscript preparation.

References
[1]
[2]

[3]

[4]

[5]

http://www.ing.unitn.it/~maud/.
M. Aiu, C.P. Huang, 2006.
http://www.ce.udel.edu/cibe/reu/REU06/Aiu%20Report.p
df.
P. E. Batson, Simultaneous stem imaging and
electron-energy-loss spectroscopy with atomic-column
sensitivity, Nature 366 (6457) (1993) 727728.
W. B. David and C. C. Barry, Transmission electron
microscopy: A textbook for materials science, Springer,
2009.
R. F. Egerton, P. Li and M. Malac, Radiation damage in
the TEM and SEM, Micron 35 (6) (2004) 399409.

[6]

F. Frizon, J. Russias, C. Cau-Dit-Coumes, A. Malchere, T.

Douillard, C. Joussot-Dubien, Incorporation of aluminum
into C-S-H structures: From synthesis to nanostructural
characterization, Journal of the American Ceramic
Society 91 (7) (2008) 23372342.
[7] M. P. Ginebra, F. C. M. Driessens and J. A. Planell,
Effect of the particle size on the micro and nanostructural
features of a calcium phosphate cement: A kinectic
analysis, Biomaterials 25 (17) (2004) 34533462.
[8] S. C. Halim, T. J. Brunner, R. N. Grass, M. Bohner and
W. J. Stark, Preparation of an ultra fast binding cement
from calcium silicate-based mixed oxide nanoparticles,
Nanotechnology 18 (39) (2007).
[9] I. G. Richardson and G. W. Groves, Microstructure and
microanalysis of hardened ordinary portland-cement
pastes, Journal of Materials Science 28 (1) (1993)
265277.
[10] C. Rossler, J. Stark, F. Steiniger and W. Tichelaar,
Limited-dose electron microscopy
reveals the
crystallinity of fibrous C-S-H phases, Journal of the
American Ceramic Society 89 (2) (2006) 627632.
[11] D. Wang, S. Penner, D. S. Su, G. Rupprechter, K. Hayek,
R. Schlogl, Silicide formation on a Pt/SiO2 model
catalyst studied by TEM, EELS, and EDXS, Journal of
Catalysis 219 (2) (2003) 434441.

Apr. 2012, Volume 6, No. 4 (Serial No. 53), pp. 509513

Journal of Civil Engineering and Architecture, ISSN 1934-7359, USA

DAVID

PUBLISHING

Architecture, Faith and Culture Antagonism or

Harmony?
Hassib Rehailia
Department of Architecture, Badji Mokhtar University, Annaba 23000, Algeria
Abstract: The relationship between architecture and faith has always been of great wealth. Faith as a way of life and social memory
exercises a decisive influence on the shape of the construction environment. Yet this relationship is rarely analyzed. On a general level,
one reason for this deficiency is probably due to the spread of the ideology of modernity. Given that modernity is implicitly secular, it
does not prompt to understand faith in its relationship with other phenomena. It is the same for architecture. We believe on the contrary
that the construction environment and the cultural aspect must interact and complement each other. When this is not true or only
partially true, this results in user dissatisfaction, and refusal to engage and participate in the process of changing that environment.
Many authors have tried to show explicitly the influence of Islamic thought on the social organization and the housing as well as the
link between housing and identity, but its implications have not always been understood and transmitted concretely. One thing is certain:
the ideological -symbolic-metaphoric sphere is the essence of architecture, as are the program requirements: functionality, distributivity,
health of the environment and technology of construction. In this contribution we will address the thorny issue of cultural and religious
influences on the choice of the morphology of the habitat and its components, and we will try to highlight the interaction between
architecture, faith and culture through analysis of a Mediterranean type House with Patio as a particular architectural style, which has
been enriched through the centuries by different cultures. Although the course of this typology in each of these cultures and at different
times has not been the same, this has only strengthened the mission and spirit of the Patio until the appropriation of this model by the
Arab-Muslim peoples who were able to incorporate the Islamic perception in this space.
Key words: Islamic Architecture, faith and culture, the traditional urbanism, Mediterranean cities.

1. Domestic Spaces
The Arab-Islamic culture has developed a
representation of the very special place that has
allowed its original identity. This spatial representation
of the architectural framework is linked to living
experience. It is all daily actions which constitute the
appropriation of space by users. Thus, architectural or
urban design would be pointless if we did not
incorporate it in a study of living experiences, which
give the space its temporal and spiritual evolution. The
organization of the city in the Arab-Islamic world, is
based on common concepts which are expressed
differently depending on location, existing building
materials, and economic and cultural development of

Corresponding author: Hassib Rehailia, PhD, lecturer,

research fields: architecture and applied arts, health,
email:Hassib.Rehailia@gmail.com.

each site. The traditional and popular aspect of these

constructions manifests itself through the effect of
natural or cultural environment, as well as all
architectural features, structural and functional whose
hidden meanings we will try to dissect and decipher.
The traditional house is inhabited by memory. The
memory is built into the physical structures and
imaginary space. Home memory does not only rely on
images of specific acts, but also on pictures of places
and names of the lineage as a guarantee of social
excellence. The domestic memorial inscription is
socialized and based on the use of the space as it
evolves through the interactions between domestic
players. Home memory takes shape through fitting into
the domestic group structures and relationships that
drive the social universe which occupies space and
reinvents it each day [1].

510

Architecture, Faith and Culture Antagonism or Harmony

The traditional house was also, for travelers,

writers or artists, a mysterious place, enigmatic and
fascinating [2]. It has undergone multiple reflections,
particularly as regards the relationship between
physical structure and the religious and social
organization of Islam and has also resulted in
monographs of high quality [3]. There are many
stereotypes, as outlined by A. Raymond, concerning
the characteristics of the traditional house often
described as belonging to a single model, invariable in
space-time, isolated within the city, totally introverted,
with a rigorous window on the outside and an exclusive
opening on the inside with a patio. R. Ilbert [4]
highlights the dangers using terms such as Oriental,
Islamic or Arab city (fig. 1), as well as ideologies
that underpin it. Thus, whether the city or home, it is
more prudent to get rid of misconceptions and
uncontrolled designations in order to grasp the
specificity of each space in its context.

Fig. 1 A lemon tree in a Tunisian courtyard space

provides shade in summer, while its size high passes

the winter sun [5].

2. Cultural Influences and Theories about

the Shape of the House
During the last decade, the relationship between
domestic space and the society that uses it has been the
subject of much discussion and research. Among these
studies, the fundamental work of Pierre Bourdieus on
houses in Kabylia. In his research, he starts with a
central reflection of placing the relationship between
home and the ideological and symbolic parameters in
correlations. For Bourdieu, the house is not only
created by society, but should be considered as a
framework in which children learn the rules that
structure society, rules which themselves are created by
society.
Bourdieus study has caused other research dealing
with villages in the Aures, the M'zab, and to a lesser
degree the Atlas in southern Tunisia. This work has
demonstrated explicitly the link between ideology,
social organization and material homes. The attempt to
define the organizational characteristics of the houses
of the past and recent ethnography can help define
boundaries in domestic spaces. This gives us different
ways of looking at the effect of the cultural model on
the shape of the house.
Suka Ozkan had the courage to devote an article to
highlight the relationship between architecture, faith
and culture. She insists that most cultural analysis have
always sought to downplay the importance of religion,
if they have not neglected it altogether, because of
religions political nature and standing in society: The
relationship between architecture or, more generally,
between the building environment and faith has always
been of great wealth. Faith as a way of life and social
memory exercises a decisive influence on the shape of
the building environment. Yet this relationship is rarely
discussed. On a general level, one reason for this
deficiency is probably due to the spread of the ideology
of modernity. Given that modernity is implicitly
secular, it does not prompt to understand the faith in its
relationship with other phenomena. It is the same for
architecture [6]. It therefore seems important to

Architecture, Faith and Culture Antagonism or Harmony

511

explore the linkages between housing and social and

cultural aspects of traditional societies. This does not
mean that we neglect other influential factors on the
choice of housing types and forms.
2.1 Vocabulary Space
The spatial vocabulary assigns names, in Arabic, to
forms and spaces while emphasizing their content and
their actions. This linguistic practice, both written and
spoken, is strictly related to social and urban historical
developments. The meanings of these names are often
of great representative value of the image and the
description we want to give space. The spatial and
community vocabulary, linked to urban practices,
allows us to better understand some socio-urban
images and social behaviors.
We cite as an example in southern and eastern
Mediterranean countries (predominantly Arab), the
entire network from the alley until the dead-end
represented by: darb (maze), the zoukak (narrow
passage) The zankah (strangulation, alley which
narrows along its path), the Khanka (suffocation,
dead-end)...This illustration, which shows the
correspondence between the perceptual image and its
name, is a constant that we find both in urban and
architectural domains. The many names designating
the house and its components show the richness of this
phenomenon. This variety of language adds to the
space image as a temporal dimension that can not be
perceived without prior training of the observer (Fig. 2).

Fig. 2 Aerial view of the Medina of Tunis.

Fig. 3 Algiers: ground floor (staggered entry) [7].

The richness of the space-time vocabulary,

combined with the architecture, provides a typology of
spatial language that goes beyond the architectural
formalism. It then creates new dimensions
emphasizing understanding and images of space. This
illustration which shows the correspondence between
the perceptual image and its name is a constant that we
find both in urban and architectural domains (Fig. 3).
2.2 The Domestic Threshold, Symbolic of A Border
The traditional planning is characterized by a long
passage from private space to public space by
introducing a threshold boundary that we try to
describe t first, then the status of the liminal space [8]
and some of its symbolic and practical purposes, then
its interaction rituals that are, according to E. Goffman,
the rules of politeness and civility, how to behave, the
of protocol-oriented postures, and finally crossing
between modes of public space and private space,
emphasizing the basic texts and authors who have
addressed these issues.
The threshold (Alley, corridor, entrance hall,
vestibule), important structuring element, is primarily
an imaginary line of separation, as an internalized gap
between two entities constituting spaces, inside and out.

512

Architecture, Faith and Culture Antagonism or Harmony

It is because there are fractures that these spaces take

the quality of territories. The threshold is also a
symbolic space in which to experience what binds the
individual to the social body and in unties it. It marks
the transition [9] between two completely different
worlds. As an articulation point between home life and
social life, in and out, the threshold is the symbolic
place of the in-between. As an articulation point,
the threshold is also a non-place erected where the
transition takes place, encounter and separation, thus a
paradoxical element in that it allows and prohibits at
the same time communication with others.
The social actor always generates the border and in
return, the latter takes stage in the game of social
reciprocity: Says Simmel about the symbolic role of the
door: Man ceases to be home when he walks past the
door, and this means, of course, he breaks a part of
eternity of the uninterrupted natural existence. But, if it
is true that the shapeless limitation takes form, the
boundaries he sets find their meaning and dignity only
thanks to the symbol represented by the mobility of the
door, with the ability to exchange at any moment this
limitation in order to be free [10].
It is A. Van Gennep who probably has best
formalized the concept of rites of passage. He
showed that they all competed to mark a transition
from one social state to another, a transition that
introduces a time and a space to signal the difference
between the prior and posterior state. Mr. Boughali
defines rites of passage as both a manifestation of the
end of a spiritual or social state and integration into
another. These are manifestations which are truly
spatialized. It is in a space in its most material and
terrestrial aspects that the rites of passage which
rhythm existence integrate themselves [9]. For P.
Bourdieu, one must ask the theory of rite of passage
questions about the social function of the ritual and
social significance of the line, the boundary, whose
ritual makes lawful the passage, the transgression [11].
Marking the passage with a line that establishes such a
division of the social order, the ritual attracts the

Fig. 4 Entries in different places and ages [12].

attention of the observer on the passage, while the more

important, he says, is the dividing line that separates in
a visible manner the before and the after.
In traditional homes, the entrance hall is usually
set-up as a chicane and divided into two or three parts
separated by a door (fig.4). Past the entrance, one finds
himself in a stwan or sqifa, intermediate space where
men wait for permission to cross the border. A male
visit results in the withdrawal of women from visible
outer space. A space which generating first impressions
and subjected to specific coding, it is also protected and
part of a strategy of intimacy. The outer door remains
open in general, a symbol of generosity and hospitality,
referring to a social organization based on hospitality.

3. Conclusions
Despite their appearance as a basic courtyard house,
in daily practice, the use of these forms gives
Mediterranean houses a distinctly unique identity.
Typographical analysis and spatial composition of
homes have also shown that the design of the interior
space of these homes is different. Indeed, the study of
this model includes constant and variable features. We
assert that these architectural and social practices are
unchanging and reflect cultural and climatic influences

Architecture, Faith and Culture Antagonism or Harmony

of the region. They retain a considerable degree of

adaptation in individual cases, because a theoretical
model can not match the flexibility and spatial history
of architectural practice, which is a major feature of
Islamic tradition.
Muslim doctrine is at the origin of the design of
these houses where the outside artifice is of little
presence. This spirit is also reflected in the so-called
Islamic architecture which gives an image of
moderation, consistency and purity, with its maze of
multiple spaces that allow both grouping and isolation,
and finally with its judicious constructive systems and
optimal use of materials.
Under the influence of Islamic thought, we notice
that one of the characteristics of the urban Arab-Islamic
morphology lies in the virtual absence of external
marks of social distinction. This architecture features
modesty and a moral ideal of equality, even if the
interiors indicate clearly a search for distinction. This
style is the result of mixing of three dominant factors:
climate, religion and customs of its inhabitants, as well
as building materials available locally.
The living space is a space oriented towards ...
according to the reference culture. It is a constituent
part of the cultural system. It gives codes of conduct
and social representations. It is used to display what is
prescribed and what is prohibited, especially around
the categories of private and public. This opposition
and its corollaries such as the oppositions
inside/outside, familiar/foreign, near/far, male/female,
could be considered as universal categories of human
spirit that allow probing of the nature of the
relationship that the resident establishes with his
environment.
Arguably, the cultural unity of the Mediterranean
cities was formed by the mixing of successive
contributions influenced by social attitudes that
prevailed in the area. The Arab-Islamic architecture is
an architecture of forms, but it is not a formalist
architecture. The comparison between Mediterranean
cities reveals a number of basic elements common to

513

all these architectures. These elements are also found in

the Arab-Islamic architecture, such as the inlet chicane,
the spatial articulations, the introverted court, zone
separation, etc .... but each of these architectures has
assimilated and interpreted these elements differently.
A diversity of cultures is primarily determined in a
global manner before being fragmented or amplified by
the differentiated elements that have been joined
together.

References
[1]

Khalid Kajaj, La maison traditionnelle Tatouan,

Harmattan, Paris, 1999.
[2] J. C. Depaule, Figures de l'orientalisme en architecture,
Revue du monde musulman et de la Mditerrane, Paris,
1996.
[3] G. Marais, Les origines de la maison nord-africaine,
Cahiers des arts et technique d'Afrique du Nord, N7,
Paris.
[4] R Ilbert, La ville islamique: ralit et abstraction, Les
cahiers de la recherche architecturale, N10-11, Paris,
1982.
[5] John S. Reynolds, Courtyards, Aesthetic, Social, and
Thermal Delight, John Wiley & Sons, Inc., USA, 2002.
[6] Suka zkan, Foi, culture et architecture, dans Arch. &
Comport., Vol II, No. 34, pp. 187192.
[7] Andr Ravreau, La Casbah d'Alger, et le site cra la ville,
Sindbad, Paris, 1989.
[8] F.Navez-Bouchanine, L'espace limitrophe entre le priv
et le public, un no man's land? dans: Espaces et socits,
n 61-62, 1991.
[9] M Boughali. La reprsentation de lespace chez le
Marocain illettr, Anthropos, Paris, 1974.
[10] G. Simmel, Pont et porte, L'Herne, les symboles des lieux,
l'habitation de l'homme, N 44, 1983, p. 214. From Kh.
Kajaj, Ibid.
[11] P. Bourdieu, Le sens pratique, Minuit, Paris, 1980.
[12] Norbert Schoenauer, 6000 years of Housing, Garland
STPM, New York & London, 1981.