IRJET- Application of Emerging Artificial Intelligence Methods in Structural Engineering-A Review

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 11 | Nov 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 841
APPLICATION OF EMERGING ARTIFICIAL INTELLIGENCE METHODS IN
STRUCTURAL ENGINEERING-A REVIEW
Dr. Priyanka Singh1
1Assistant Professor (III), Civil Engineering Department, Amity University, UP, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The aspiration of the work presentedinthispaper
was to collect, organize, and write the knowledge and
experience about structural analysis-based design
improvements into a knowledge base for a consultative
advisory intelligent decision support system. The civil
engineering problems are not repetitive, as the problem
definition is always influenced by a number of factors like
financial modes, importance of structure and site conditions
and so on Therefore, although the use of computers in
structural analysis started almost four decades ago, the
profession has not been able to make use of computers fully,
especially, for structural design and planning. This is mainly
because of problem specific nature, need for logical
reasoning, feasibility constraints and use of past experience
required in actual design process and planning. Expert
systems have capabilities to incorporate some of these
requirements for programming a machine for solving a
design problem. Artificial Intelligence (AI) is a very versatile
and potential technology in the field of computer technology,
which enables computer users in various fields to solve
problems for which algorithmic approach cannot be
formulated and which normally requires human intelligence
and expertise. Expert Systems (ESs) and Artificial Neural
Networks (ANNs), the best known manifestations of AI, have
today gained immense credibility and acceptance in many
professional fields. Artificial neural networks arebiologically
inspired in the sense that neural network configurations and
algorithms are usually constructed with the natural
counterpart in mind.
Key Words: Artificial Intelligence (AI), Structural
Engineering, Civil Engineering, Pattern recognition (PR)
Expert Systems (ESs) Artificial Neural Networks (ANNs)
1. INTRODUCTION
Artificial intelligence (AI) is proving to be an efficient
alternative approach to classical modelling techniques. AI
refers to the branch of computer science that develops
machines and software with human-like intelligence.[1]
Compared to traditional methods, AI offers advantages to
deal with problems associated with uncertainties and is an
effective aid to solve such complex problems.[3] In addition,
AI-based solutions are good alternatives to determine
engineering design parameters when testing is not possible,
thus resulting in significant savings in terms of human time
and effort spent in experiments.[3] AI is also able to make
the process of decision making faster, decrease error rates,
and increase computational efficiency. Among the different
AI techniques, machine learning (ML), pattern recognition
(PR), and deep learning (DL) have recently acquired
considerable attention and are establishing themselves as a
new class of intelligent methods for use in structural
engineering. There are various advantages of artificial
intelligence approaches in different fields of civil
engineering. For example, the minimization of total weight
for a steel, concrete, and composite structure can be
obtained using genetic programming, estimating of energy
consumption based on the available data can be obtained
using artificial neural network, and optimal work schedules
for the activities of a construction management can bemade
by using met-heuristic optimization algorithm.Inparticular,
hybrid artificial intelligence studies inthefieldsofstructural
engineering,constructionmanagement,hydrology,hydraulic
engineering, geotechnical engineering, environmental
engineering, transportation engineering, coastal and ocean
engineering and materials of construction.
[4]Artificial Intelligence is a term thatdescribestheability of
a computational entity to perform activities in a fashion that
usually characterizes human thought. By deploying
appropriate models, algorithms and systems, the ultimate
goal of AI is to completely replicate intelligent human
behaviour. Thus, many scientists remain doubtful that true
AI with inherent, apparently intelligent behaviour, can ever
be developed because machines are not “mental” andcan, as
a result, neither incorporate intrinsic meaning nor a true
intelligence. However, with the rapidly increasing
advancements in modern sciences, the search for AI has
taken various directions comprising a multitude of AI-
related technologies and methods. Engineering design is a
very complex, iterative process. Physical and mathematical
modelling simulations and analyses are computationally
intensive but offer immense insight into a developing
product. Structural engineering analysis plays an important
role in this process, as the results of such analysis are often
used as basic optimizationparameterstoimprovethedesign
candidate being validated and analysed. The number of
iterations/cycles that are needed to reach the final design
solution depends directly on the quality of the initial design
and the appropriateness of the subsequent design changes.
Computer Aided Design (CAD) software is extensively
applied in performing various design activities, such as
modelling, kinematics, simulations, structural analysis or
just drawing technical documentation. Nowadays, the
software can be so complex, and offers such an extensive
assortment of different options, that one can easily be con-

fused.[5] This is the reason for the relatively low level of
control over these systems. Computer aided design (CAD)
has increased by orders of magnitude the power of
design tools available to the engineer. Advantages of CAD
include the reduction of computationtimeandthereforeits
cost, the elimination of the amount of tedious and error-
prone detailed calculations done by the engineer, and the
ability to develop and analyze much more complete
models of structures. All present applications of the
computer to structural design deal with later stages of
the design process, namely, analysis, proportioning and
drafting. With the advancements of a section of
computer science called artificial intelligence it is
now conceivable to create a knowledge-based system to
automate or assist in the early, preliminary stages of the
structural design process. The purpose of this report is
to explore the potentials of such a system.
2. STRUCTURAL ENGINEERING BACKGROUND
Design can be viewed as the general processinwhichanidea
is developed into detailed instructions for manufacturing a
physical product [7].The design process starts with a
definition of a need. The activitiesthatfollowcanbegrouped
into four phases:
1. Synthesis : the clarification of the input parameters
and their interaction to create a structure that will meet
design requirements.
2. Analysis: the modeling and solving of equations to
predict the response of a selected structure.
3. Evaluation: the activity of placing a worth on the
structure where worth may be cost, safety, or energy
consumption.
4. Optimization: the search over the range of possibilities to
improve the design as much as possible.
Today, CAD in structural engineering involves almost
exclusively analysis, proportioning of structural
components, and production of drawings and schedules.
There are very few applications to conceptual and
preliminary design. Conceptual and preliminary designs
are considered the creative aspects of design. Yet, generally
the preliminary design process is not new design but
redesign, where redesign involves the application of
existing structural ideas to a particular design. New
design implies the development of a new structural
configuration, e.g. new concepts in structural design.
Redesign actually is the application of a set of rules to
assign values to predefined variables. Thus, it appears that
preliminary structural design process may be placed in a
knowledge based system, where IF THEN rules are used
to instantiate values in a data structure.
A knowledge based program is developed using the
knowledge of experts. Once the program is developed
there should be close interaction between the designer
and the computer. The computer should be able to
respond to queries on the design process as well as
accept additional information. Since a design prepared by
the computer follows a limited number of rules, close
supervision by the designer is necessary. In this way
the designer will realize inadequacies in the existing set
of rules and make revisions or additions to the rules
when necessary.
3. ARTIFICIAL INTELLIGENCE BACKGROUND
Artificial intelligence is the study of ideas which enable
computers to do the things that make people seem
intelligent." [10] Ideas are being developed to facilitatethe
creation of knowledge-based systems using the experience
and knowledge of experts. The description of some of the
applications of artificial intelligenceisasfollows.Knowledge
based systems can do geometric analogy tests Knowledge
based systems acquire knowledge or learn ideas similar
to the way people learn. The study of how knowledge
based systems learn can provide insight into the way
people learn and vice versa. Learning new concepts may
be through sequences. Learning may also be done through
the acquisition of procedural knowledge. Knowledge based
systems can understand simpledrawings, simplelanguage,
and do expert problem solving.[8] There are programs
capable of doing integration problems (MACSYMA),
understanding mass spectrograms (DENDRAL), and
helping physicians diagnose and treat bacterial infections
(MYCIN). Knowledge based systems can also do
industrial work (i.e., robotics) and model
psychological processes. One approach to the development
of knowledge based systems is heuristic programming.
(Heuristic is defined as serving to guide or discover).A
model may be developed in the form of a goal tree. Some
existing applications of artificial intelligencetechniques are
AGE and SACON. AGE is a knowledge based program for
building knowledge based programs. AGE is an attempt
to formulate the knowledge used in constructing
knowledge based programs and put it at the disposal of
others. This is a developmentin‘knowledge engineering,the
process of writing application programs using primary
artificial intelligence methods. SACON is a knowledge based
consultant for structural analysis. SACON is an
"automated consultant, it advises engineers in the use of
a general purpose structural analysis program,
MARC.MARC offers a large choice of analysis methods.
SACON is an example of the use of AI techniques in
structural design in the analysis phase. The study and
development of knowledge based systems should enable
the structural designer to incorporate AI techniques into
the other phases of design. In the preliminary design of a
structure for example bridges, the first step is the
definition of a need, eg . A roadway between two points.
The preliminary design can be separated into two parts:
first, the horizontal and vertical alignment and second,
the selection of an alternative structure or structures. For
this example, we assume that the alignment is
determined before the designing engineer is involved.

Some of the constraints on the alignment are the cost of
property, the earthwork involved, and who will be
affected. The second part, the selection of alternatives,
will be discussed in terms of artificial intelligence
techniques. The selection of alternative structural types,
such as simply supported steel, pre-stressedconcrete, tied
arch, or truss, depends on span lengths and their ratios,
i.e., the comparison of the lengths of adjacent spans. The
selection of component types (e.g., deck vs. through truss,
plate girder vs. box beam, etc.) depends on span lengths
as well as available vertical clearance. Finally, material
type (eg. steel vs. Prestressed concrete beams) depends
largely on relative costs. The span lengths are
estimated in preliminary design by the application of
alignment and clearance constraints and cost analysis. It is
these constraints that will generate the rules to be executed
in a knowledge based system. The resulting facts, i.e.,
span lengths and types, and component types,
material and preliminary specificationscouldbeprocessed
and stored into frames. The frame representation may be
viewed as a network data base, where a frame containing
a level of information is linked to adjacent levels. The
highest level frame would contain general information
common to all alternatives. As a particular alternative
is developed, lower levels of frames are instantiated
with data. It may be possible to develop more than one
alternative within a frame structure by duplicating
frames and changing certain data within these frames.
3.1 Expert Systems for Analysis and Design
Developing an inference mechanism demands very high
programming skills particularly for developing a general
expert system shell, which can be used for diverse types of
applications. This is not an easy task particularly for those
who are not familiar with much programming. The
procedure outlined in IS: 10262 - 1982 known as Indian
Standard (IS) code method. This facility was found very
useful particularly for developing expert systems for
concrete mix design as most of the knowledge available for
mix design can be easily put in tabular form. Because of all
these features it was further used for development of expert
systems for the design of concrete mix for flexural strength
and also for selecting concrete constituents based on A. C. I.
Method. The developed expertsystemseliminatethetedious
procedure of referring to charts, graphs and tables of IS
codes and help the user to arrive at final quantities of
cement, water, sand and coarse aggregates per cubic meter
of concrete. Expert systems were developed to determine
safe bearing capacity of soil,toselectsuitablefoundationand
to design isolated footing subjectedtoaxial loadonlyortoan
axial load and moment or a combinedslabfootingasthecase
may be. Also many authors developed a knowledge based
expert system to determine the nature of loading on the
rectangular column and to calculateslendernessratioforthe
type of column, i.e. axial, uniaxial or biaxial and thus to fire
rules related to design of that particular type of column.
Further, expert system was developed to arrive at optimal
design of T-beam floors. For developing knowledge base, in
the rule form, available design charts for the cost of
materials and magnitudes of imposed loads for different
spans of slab were used and to obtainoptimal designsection.
Problems of design of singly reinforced section were chosen
for the optimal design. For simplicity, an optimal design
polynomial was considered forthedevelopmentofanexpert
system with the design constraints as equilibrium
constraints, bending moment constraint and beam
width/depth ratio constraint. The objective function
considered was the cost of beam, which included cost of
steel, concrete and shuttering. Some of the salient features
which offered a suitable base for development of rule based
expert system are menu driven navigation, simple English-
like rule syntax, the abilityto execute external DOSprograms
and good interface capabilities to external programs such
as spread sheets, databases and batch files. Particularly,
the facility of induce command that automatically creates a
knowledge base from a table contained in a data base was
found very much suitable in transforming directly the
tabular information available in design codes.
3.2 ANN in Non-destructive Testing
Two popular methods, namely, Schmidt test hammer
method and Ultrasonic pulse velocity method were
considered to study, for the first time, the feasibility of using
Artificial Neural Network (ANN) for correlation of Non
Destructive Testing (NDT) parameters to the strength ofthe
structure. As there is no direct relation between rebound
number and concrete strength or pulse velocity, the
development of an ANN simulator seems to be the natural
choice for such problems because predefined mathematical
relationship among the variables is not required in an
artificial neural network.Thefeedforwardback propagation
training algorithm was selected for the preparation of
program for its simplicity and good generalization
capabilities. A study of training andrecall resultsforboth the
concrete hammer and ultrasonic tests indicated that the
neural networks are able to learn examples of NDT and give
reasonable predictions of concrete strength for any new
value of rebound number or pulse velocity. To facilitate
rapid assessment of flexural behavior, multi-layer feed
forward ANNs were trained to learn the relationship
between input and output data generated from theavailable
experimental data. The error correcting back propagation
algorithm was used to map the relationship. The flexural
behaviour of two different types of steel fibre reinforced
concrete beam problems were modelled using neural
networks. The results obtained for both the problems were
found to be in excellent agreement with the actual
experimental values. The engineering importance of the
whole exercise was thus demonstrated by predicting the
behavior for new test values without performing any
expensive and time consuming experiments. A generalized
delta rule was used to train the networks based on the
existing experimental results for two different types of deep
beam problems, i.e., FRC deep beams with and without

reinforcement. In the case of FRC deep beam without
reinforcement, four inputs (length of beam, shear span,
span/depth ratio and percentage fibre content by weight)
were related to five outputs (first cracking load, failure
load, maximum average shear stress, maximum
experimental moment at failure and theoretical maximum
moment) using one hidden layer with 7 nodes.
3.3 ANN in Predicting Large Deflection Response
Recently, feasibility of using neural networks to evaluate
large deflection response of fixed immovable rectangular
plates subjected to patch loading has been investigated. The
error back propagation algorithm with sigmoidal function in
the range 0 to 1 was used to map the relationship between
the inputs-plate aspect ratio, the patch size and pressure
coefficient and the 8 outputs, namely, the central deflection,
bending and membrane stresses in the x and y directions at
key locations of the rectangular plate.
4. CONCLUSION
Developed expert systems for analysis-design, concrete
technology, design of R.C.C. and structural steel components
and use of artificial neural networks in non-destructive
testing, behavior modeling of fibre reinforced concrete
beams, and predicting large deflection response of
rectangular plates were discussed clearly the advantages of
using AI in these areas. Developed expertsystemsinthefield
of concrete technology arenotonlyusedby engineersduring
their laboratory work of concrete technology, but are also
used in commercial testing of material for arriving at proper
concrete mix for compressive and flexural strengths.
Developed artificial neural networks forthenon-destructive
testing are also used in the field for commercial testing work
for finding strength based on rebound number and pulse
velocity while using concrete hammer and ultrasonic
concrete tester respectively. The preliminary design
process relies heavily on the expert’s ability to identify
and analyze situations, and to evaluate alternatives. This
ability is developed through personal experience and
the passed on experience of other experts. Since it is
impossible for an expert to pass on all the knowledge he
has gained from experience, the departure of an expert
(from an office or the field of engineering) means the
loss of some of that experience. The development of
knowledge based systems will permit not only the
retention of expertise, but also its logical extension, as
well as access to the expertise by other structural
engineers.
REFERENCES
[1].Russell,S. and Norvig, P. (1995)―Artificial Intelligence:A
Modern Approach‖, Prentice Hall.
[2].Schalkoff, R.J. (1990), ―Artificial Intelligence: An
Engineering Approach‖, McGraw-Hill, New York.
[3]. Adeli, H. & Hung, S. L. ,‖Machine Learning—Neural
Networks, Genetic Algorithms, and Fuzzy Systems ―,John
Wiley &Sons, Inc., New York, 1995
[4].Flood, I.; Nabil, K. Neural networks in civil engineering
II: Systems and application. // Computing in Civil
Engineering. 8, 2(1994), pp. 149-162.
[5].Flood, I.; Paul, C. Modeling construction processes using
artificial neural networks. // Automation in Construction.
4,4(1996), pp. 307-320.
[6]. Jeng, D.S.; Cha, D. H.; Blumenstein, M. Application of
Neural Networks in Civil Engineering Problems. //
Proceedings of the International Conference on Advances in
the Internet, Processing, Systems and Interdisciplinary
Research, 2003.
[7].Boussabaine, A.H. and Kaka A.P.‖ A neural networks
approach for cost flow forecasting.‖ Construction
Management and Economics, 16, 1998, 471-479.
[8].Knezevic, M.; Zejak, R. Neural networks – application for
usage of prognostic model of the experimental research for
thin reinforced-concrete columns. // Materials and
constructions, (2008).
[9].Lazarevska, M.; Knezevic, M,; Cvetkovska, M.; Trombeva,
G. A.; Samardzioska, T. Neural network’s application for
predicting the fire resistance ofreinforcedconcretecolumns.
// Gradjevinar. 7, (2012), pp. 565-571.
[10].Lazarevska, Marijana ; Knezevic, Milos; Cvetkovska ,
Meri; Gavriloska, Ana Trombeva. Application of artificial
neural networks in Civil engineering.Issn1330-3651(print),
ISSN:1848-6339(online)udc/udk 032.26:[624.042:699.812]

IRJET- Application of Emerging Artificial Intelligence Methods in Structural Engineering-A Review

More Related Content

What's hot (20)

Similar to IRJET- Application of Emerging Artificial Intelligence Methods in Structural Engineering-A Review (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Application of Emerging Artificial Intelligence Methods in Structural Engineering-A Review