Académique Documents
Professionnel Documents
Culture Documents
Project Team
May-2018
i
FACULTY OF ENGINEERING AND TECHNOLOGY
Certificate
This is to certify that the Project titled “Application of Artificial
Intelligence in Pile Foundation” is a bonafide work carried out in the
Department of Civil Engineering by Mr Amit Patel, Arjun Rathod, Arnab
Bhargav and Sachin Kulkarni bearing Reg. No. 14ETCE001003,
14ETCE001005, 14ETCE001006 and 14ETCE001018 respectively in partial
fulfilment of requirements for the award of B. Tech. Degree in Civil
Engineering of Ramaiah University of Applied Sciences.
August – 2018
i
M.S.Ramaiah University of Applied Sciences – Faculty of Engineering and Technology (FET)
Declaration
The project work is submitted in partial fulfilment of academic requirements for the
award of B. Tech. Degree in the Department of Civil Engineering of the Faculty of
Engineering and Technology of Ramaiah University of Applied Sciences. The project
report submitted herewith is a result of our own work and in conformance to the
guidelines on plagiarism as laid out in the University Student Handbook. All sections of
the text and results which have been obtained from other sources are fully referenced.
We understand that cheating and plagiarism constitute a breach of University
regulations, hence this project report has been passed through plagiarism check and the
report has been submitted to the supervisor.
1
M.S.Ramaiah University of Applied Sciences – Faculty of Engineering and Technology (FET)
Acknowledgements
We would like to sincerely thank our supervisors Dr. S. D. Anitha Kumari and Mr.
Abhishek Pulgur for their invaluable guidance, constant assistance, support, endurance
and constructive suggestions for the betterment of the project.
We also would like to convey my heartfelt thanks to Dr. H.M Rajashekhar Swamy, Head
of Civil Engineering Department, RUAS for giving us the opportunity to embark upon this
topic.
We wish to thank Ramaiah University of Applied Sciences for providing us a healthy
environment to proceed with the topic.
Finally, we are thankful to our parents and friends for their continued moral and
material support throughout the course.
Arjun. Rathod
Amit Patel
Arnab Bhargav
Sachin Kulkarni
Abstract
Table of Contents
Certificate ……………………………………………………………………………………………………………….……(i)
Declaration……………………………………………………………………………………………………………………1
Acknowledgements......................................................................................................................... 2
Abstract ............................................................................................ Error! Bookmark not defined.
Table of Contents ............................................................................................................................ 4
List of Tables ................................................................................................................................... 6
List of Figures .................................................................................................................................. 7
Nomenclature14…………………………………………………………………………………………………………………………….8
1. Introduction ................................................................................................................... 159
1.1 Artificial intelligence (AI)…………………………………………………………………………….10
1.1.1 Moments that Shaped AI……………………………10
1.1.2 Recent development in AI.............................11
1.1.2.1 Need of AI....................................................11
1.1.2.2 Backdrop of AI..............................................12
1.2 The Future of Artificial Intelligence……………………………………………………………..12
1.3 AI in Civil Engineering………………………………………………………………………………….13
Organisation of the report…………………………………………………………………….........................14
Appendix………………………………………………………………………………………………………………………14
2. Literature review.............................................................................................................. 15
3. Overview of ANN and Procedures..................................................................................... 17
3.1 Introduction .................................................................................................... 17
3.2 Neural Network:........................................................................................... 18
3.2.1 Comparison between Artificial Neural Network (ANN) with Biological
Neural Network (BNN): ......................................................................................... 18
3.2.2 Principle of Artificial Neural Network: ..................................................... 20
3.3 Artificial Neural Network Architecture: ....................................................... 22
List of Tables
List of Figures
________________________________________________________________________
Nomenclature
________________________________________________________________________
1. Introduction
Human mind is the most complex system in the world. The most complex networks
which the existing systems can’t imagine can be easily analysed by human brain. Making an
artificial mind to act in the same way as a human being tackles an objective, is the main aim of
the researchers and scientists all over the world. In 1956, John McCarthy introduced the term
artificial intelligence and defined it as ’the science and Engineering of making intelligent
machines’ (towardsdatascience.com). Artificial Intelligence (AI) is not based on making robots or
machines, it is about making an artificial mind that can solve problems like a human mind. Over
the years, many people have been successful in making robots that can interact and talk with
others like a normal human being. One such example is that of Hanson Robotics who has been
able to successfully develop a humanoid robot named Sophia as shown in figure 1.
Sophia is the most advanced robot to date and also is a cultural icon. Showing up in
various media outlets she has become a media darling, lighting up the interest of people towards
AI. Sophia has gained her name as a renowned speaker in business and showed her great
potential across several industries.
Today artificial intelligence has entered our lives and has become a part of it. From the
apple voice assistant SIRI to self-driving cars, all have been developed using artificial intelligence.
Science fiction movies portray artificial intelligence as robots with human like features but AI is
not only robots, it can be used in anything from Google’s search engine to autonomous
weapons.
In the 21st century artificial intelligence has only been able to perform small tasks like
only facial recognition or only driving a car. However, in future many researchers have the
motive to create a general AI that would be able to outperform human beings at nearly every
cognitive task.
A new field called deep learning was introduced which enhanced new corporate interests in AI
by revealing that their ideas could make speech and image recognition more precise.
2012
Neuro-fuzzy based prediction of the durability of self-consolidating concrete to various sodium
sulfate exposure regimes was developed by Bassuoni and Nehdi.
2016
The world champion of the board game named ‘Go’ was defeated by a google based computer
program called AlphaGo.
A few centuries ago, the thought of artificial intelligence was merely a science fiction.
However, in the past few years’ major milestones have been seen in the artificial intelligence
field. Experts viewed these milestones to be decades ahead, but the breakthrough has struck
everyone with surprise. But some experts still suggest that the human-level AI is centuries away.
1.1.2.1 Importance of AI
The first and foremost advantage of AI is that it will have a very lower error rate as
compared to humans. AI will have an outstanding precision and speed
AI will be able to work in a hostile environment, thus able to perform hazardous tasks,
like space exploration, mining and tackle life threatening problems
In case of laborious work AI can be a significant help for human beings, for example in
case of construction of a high rise building or other complicated structures all the
calculations can be done at the blink of an eye with the help of AI.
AI system doesn’t require to sleep, doesn’t get bored or is not required to get
entertained
In card-based systems AI can catch any frauds
The present practicing engineers are using AI to tackle a whole range of problems.
Future development of AI will be a boost to the construction industry which mainly will
include benefits in terms of optimisation, speed of processes and cost reduction
1.1.2.2 Backdrop of AI
Experts foretell that a super intelligent AI is unlikely to exhibit human emotions like love or
hate. While experts predict two most likely scenarios:
The AI is programmed to do something catastrophic: Autonomous weapons in the
hands of the wrong person, could easily cause mass destruction
The AI is programmed to benefit the society, but it takes a wrong path for attaining
the goal: this can be understood by considering the following example if a command is
given to an intelligent car to drive to the airport as fast as possible, the car might get
there in time but at the same time chased by the police, thus doing not what expected
but literally what it was asked for
1.2 The Future of Artificial Intelligence
At present, people can expect more AI powered features in their gadgets and
smartphones. The AI advancement will improve virtual assistance and smart speakers as
predicted by Google and Amazon in particular.
The present decade is the best time for an AI researcher. The AI investigating labs have
sprung up in number and also are better funded than ever in history. Despite the whirlwind of
late advances in AI and predictions about its future, there are as yet numerous things that
machines can't do, for example, understanding the variation of languages, common sense and
taking in knowledge from a couple of examples. AI programming should ace assignments like
these in the future if it is to draw near to the multifaceted, versatile and imaginative knowledge
of people.
Chapter 1: Introduction
Introduction to the topic of artificial intelligence and looking at the pros and cons of AI. Also
discussing about the future of AI and the application of AI in civil engineering.
Chapter 6: Conclusion
Inference from the results and future scope.
References
Appendix
2. Literature review
Proposed traditional theories for predicting the ultimate bearing capacity of foundations have a
similar format to the theory proposed by Terzaghi (1943), which was based on a similar
superposition formula. Eq. (1) illustrates the format for vertically loaded foundations on
cohesionless soils.
where B is the width of foundation (m), γ is unit weight of soil (kN/m3), q=γD, D is the depth of
footing embedment (m), Nq and Nγ are the bearing capacity factors, Fqs and Fγs are the shape
factors and Fqs and Fγs are the depth factors. All the mentioned factors in Eq. (1) are non-
dimensional. Terzaghi derived this equation from the theory of bearing capacity developed by
Prandtl (1920) that was obtained from the theory of plasticity for evaluating the punching of a
rigid base into a softer material (Bowles, 1982). There are several different equations to
calculate these factors, and various definitions have been proposed by numerous researchers,
including Terzaghi (1943), Meyerhof (1963), Hansen (1970), Vesic (1973), Lundgren and
Mortensten (1953), De Beer (1970), Kumbhojkar (1993), and Bolton and Lau (1993). In each of
these methods, the footing geometry and the shearing resistance angle of sand play key roles in
the bearing capacity of a foundation against external loads. They illustrate the effects of physical
conditions of the soil and foundation system. Among the foundation parameters, width (B) is the
smaller lateral dimension), length (L) is the larger lateral dimension), and shape; among the
environmental parameters, depth of the foundation embedment (D), unit weight of soil (γ) and
shearing resistance angle of soil (ϕ) have the most significant effects on the ultimate bearing
capacity value. Meyerhof (1950) believed that, of the physical characteristics of the footing,
foundation depth has the greatest effect on the bearing capacity. There are also some other
parameters that may affect the bearing capacity, such as the compressibility of the soil;
however, they have a low degree of importance in comparison with other parameters.
Considering the unit weight of soil (γ) and the internal friction angle of soil (ϕ) allows a formula
to be proposed without including the compressibility of soil. Additionally, the important effect of
the water table level should also be considered. A recent study by Foye et al. (2006) revealed
that the most important parameters to be considered are B, L, D, γ and ϕ, and the L/B ratio
interferes with the effect of the footing shape. All the above-mentioned ways to express shape,
depth and bearing capacity factors are based on these parameters.
By extensive computing methods (particularly for pattern recognition and data mining
problems), Padmini et al. (2008) used an Adaptive NeuroFuzzy Inference System (ANFIS), an
Artificial Neural Network (ANN) and a Fuzzy Inference System (FIS) to develop models that
establish correlations between the initial conditions of the soil-foundation system as the inputs
and ultimate bearing capacity as the output. Their study showed the high potential of soft
computing methods for modelling the ultimate bearing capacity of pile foundations.
The purpose of this study is to calculate the ultimate bearing capacity of the pile foundation
using artificial neural network both in cohesive as well as cohesionless soil samples. First, an
overview of the neural network methodology must be presented. This would be followed by
some practical guidelines for implementing back-propagation neural networks. Experimental
data sets for both cohesive and cohesion less soil sample will be provided. Finally analysis of the
results obtained from the back propagation neural network for both the soil sample will be
compared with the actual experimental data sets.
3.1 Introduction
The Artificial Neural Network (ANN) is a machine learning method developed from the
idea of simulating the human brain. The capability of artificial neural network can model
complex non-linear relationship. ANN also has precise fault tolerance and it is fast and highly
scalable with parallel processing.
Recent advancements in the field of artificial intelligence include voice recognition,
robotics and image recognition. Artificial neural networks is meant to derive meaning from the
imprecise and complex data. This Neural network is capable to extract patterns in the given data
as inputs and detect the trends that are way too complex to be noticed by human beings or
other techniques. A trained neural network can be considered as an expert, to whom input can
be given and output can be expected. This neural network is based on creating the projection
from given sets of data and capable to answer ‘what if’ situations. This computational model
which is inspired by the human brain (Figure 3.1) can be considered as the next major
advancement in the Computing Industry.
Figure 3.2 – Difference between computer and human brain (Ref: Xenonstack.com)
Human brains can recognise various patterns precisely, but the progress of machine pattern
recognition is commendable. For example, recent development in the cameras of the
smartphone or digital camera can able to recognise specific faces. But human brain can
recognise complex patterns and adapt to them.
3.2.1 Comparison between Artificial Neural Network (ANN) with Biological Neural
Network (BNN):
The artificial neural network functions similar to human body’s biological neural
network. A typical biological neuron is shown in Figure 3.3. The dendrites in Biological Neural
Network (BNN) are similar to the input (weighted inputs) based on their synaptic inter-
connection in the Artificial Neural Network (ANN).
The function of Biological Neural Network is presented in Figure 3.4. The task is commanded
from the left and through dendrites it goes to the cell body, where the summation of data and
threshold takes place. In the final step, the output comes from the axon (to the right) in
biological neuron network which can be compared to the output unit in artificial neuron
network. The artificial neuron unit in Artificial Neural Network (ANN) is similar to the cell body,
which also have basic task of summation and threshold unit.
The comparative analysis of Biological Neural Network (BNN) and Artificial Neural Network
(ANN) has been tabulated in the Table 3.1.
Table 3.1 Comparative analysis of BNN and ANN
Characteristics Artificial Neural Network Biological(Real) Neural
Network
Speed Faster in processing the Slower in processing the
information as compared to information as compared to
biological neural network. The ANN. The response time of
response time of ANN is in BNN is in milliseconds.
nanoseconds.
Processing Processing is of serial type Processing is of massively
parallel type
Size & Complexity Size and complexity is less Higher size and complexity as
compared to ANN.
Storage The information storage is The information is of
editable, that means new data adaptable type.
can be added by deleting the
older one.
Fault tolerance Once the information is This processing is of natural
corrupted, it cannot be type rather than computing
retrieved in the case of type.
system failure.
Control Mechanism control unit is present for No specific control
controlling computing mechanism.
activities
and b = bias. Bias is used to increase the flexibility of the network, it fits the data in the case of
zero as the output.
Each statement or input is multiplied with the respective weights. The weights are the crucial
part of the artificial neural network, it shows the strength of the interconnection between the
neuron inside the artificial neural network. These statements with assigned weights is
collectively fed to the individual unit of artificial neural network called neuron. If the weighted
sum value comes to zero, bias is added to make the output non-zero or set the system response.
Bias has the weights and inputs which always equal to ‘one’. It basically allows the shifting of the
activation function to the left or right. The sum of the inputs can be any numerical positive
value.
So to fix or limit the output, the threshold value is assigned. For this, the sum of the inputs has to
pass through the threshold value. This gate or activation function is set to transfer the sum to
get the desired output.
There are two types of activation function:
Linear activation function: the function is a line or linear in nature as shown in Figure
3.6. Therefore, the output function won’t be confined between any ranges.
In the typical artificial neural network, large number of artificial neurons (which is called
as units) are arranged in a series of layers as shown in the Figure 3.8.
These neural network layers can be divided into three main layers:
1. Input layer: Layer of neural network which receives input from the outside world. And
on this input, the neural network with progress and will work on
2. Hidden layer: Layer after input layer which is responsible for the transformation of the
inputs into something that output units can use.
3. Output layer: Last layer of neural network in which the units that respond to the
information processed are present.
Conjugate gradient Algorithm: In this algorithm, the search is performed along the conjugate
directions which results in the faster convergence than gradient descent.
Levenberg-Marquardt algorithm: This algorithm is also known as damped least squares method,
which has been designed to work with the loss functions which take the form of a sum of
squared errors. This algorithm makes use of gradient vector and the Jacobian Matrix instead of
Hessian Matrix.
3.5 Conclusion
The artificial neural networks have been applied successfully to many civil engineering
areas like hydrology, pile capacity prediction, tide level prediction and deflection of retaining
walls, etc. (referred article: Application of Artificial Neural Networks in Civil Engineering by Zode
Pramey Moreshwar) and it is evident that ANN perform even better than convention methods.
In many situations, a civil engineer encounters a complex issue and it consumes time. So, ANN
can be used to get the precise data in less amount of time. ANN have several significant benefits
that make it a powerful and practical tool for solving real life problems.
4. Experimental data
4.1 Introduction:
Foundation is the important part of any structure. While design of any foundation it is necessary
to know the type of soil, its behaviour and bearing capacity.
Bearing capacity of soil is the capacity of soil to bear the load applied by the structures
constructed
over it. It is perhaps the most important aspect of soil.
Various methods and efforts have been made to determine the ultimate bearing
capacity of foundation.
The data required for the study are collected from various sources. The collected data for
cohesionless soil as reported by : Dr. Khalid R. Mahmood Al-Janabi (Iraq) for the region is taken
for this study. Around 100 data is available of which a set is tabulated in Table 4.1. The data set
of 100 is provided.
Table 4.1 Data collected for cohesionless soil(Ref: Dr. Khalid R. Mahmood Al-Janabi : 2008)
Shearing Bearing
Footing Unit Weight
Width (B) Depth (D) Resistance φ Capacity qu
Geometry L/B Sand γd
(deg) (kPa)
0.6 0.3 2 9.85 34.9 270
0.6 0 2 10.2 37.7 200
0.6 0.3 2 10.2 37.7 570
0.6 0 2 10.85 44.8 860
0.6 0.3 2 10.85 44.8 1760
0.5 0 1 10.2 37.7 154
0.5 0 1 10.2 37.7 165
0.5 0 2 10.2 37.7 203
0.5 0 2 10.2 37.7 195
0.5 0 3 10.2 37.7 214
The data required for the study are collected from various sources. The collected data for
cohesive soil as reported by Harandizadeh, M. M. Toufigh and V. Toufigh (Iran, Mexico, India,
Alton-Illinois) for the region is taken for this study. Around 100 data is available of which a set is
tabulated in Table 4.2. The data set of 100 is provided.
Table 4.2 Data collected for cohesionless soil (Ref: Harandizadeh et al : 2017)
5.1 Analysis:
In most of the engineering streams, empirical relationships are often utilized to estimate design
parameters and engineering properties. Generally, this process involves several interacting
factors in which relationship between these factors is not precisely known. The procedure
utilized to calculate the output value follows a tedious process and demands human attention
along with the intuitions. These all errors can be minimized by utilizing artificial intelligence
techniques. Artificial Intelligence is “the science and engineering of making intelligent
machines”.
There are many techniques utilized in artificial intelligence as Artificial Neural Networks (ANNs),
Genetic Programming (GP), Evolutionary Polynomial Regression (EPR), Support Vector Machines
(SVM), M5 model trees, and k-nearest neighbors. ANN involves different techniques such as
Feedforward neural network, radial basis function, kohonen self-organizing, Recurrent,
conventional, back-propagation and modular neural network. In this study, the path followed
to arrive at the result is back-propagation neural networks. In back-propagation neural
networks, the mathematical relationships between various parameters are not specified.
Instead, they learn from the data fed to them. They can generalize responses that broadly
resemble the data in the learning phase.
Training of the neural network is essentially carried out through the presentation of a series of
example patterns of associated input and target (expected) output values. The hidden and the
intermediate output neurons process its inputs by multiplying each input by its weight, summing
the product and then passing the sum through a nonlinear transfer function to produce the
result. The neural network learns by modifying the weights of the neurons in response to the
errors between the actual output values and the target output values. The transfer function
translates the input signals to output signals. Some of the transfer functions are Unit step
(threshold), sigmoid, piecewise linear and gaussian. The S-shaped sigmoid curve is commonly
used as the transfer function. This is carried out through the gradient descent on the sum of
squares of the errors for all the training patterns. (Rumelhart et all, 1986) Training is carried out
by repeatedly presenting the entire set of training patterns (with the weights updated at the end
of each cycle) until the average sum squared error over all the training patterns are minimized
and within the tolerance. At the end of the training phase, the neural network should correctly
reproduce the target output values for the training data provided the errors are minimal, i.e.
convergence occurs. The associated trained weights of the neurons are then stored in the neural
network memory.
In the next phase, the trained neural network is fed a separate set of data. In this testing phase,
the neural network predictions (using the trained weights) are compared with the target output
values. This assesses the reliability of the neural network to generalize correct responses for the
testing patterns that broadly resemble the data in the training set. No additional learning or
weight adjustments occur during this phase. Once the training and testing phases are found to
be successful, the neural network can then be put to use in practical applications.
There are many paths such as C program, python, Java, R-language and MATLAB through which
back-propagation neural network can be triggered. The simplicity and the …… visual
representation of the output made MATLAB the best choice. MATLAB utilizes nntool to estimate
the output viz. ultimate bearing capacity of the pile foundation.
Wp
c= 2.5× W
(steam hammer on
WHL
λe =
AEs 2
Wp
Cd = 0.75 + 0.15
W
In table 1 W is the hammer weight, H is the hammer drop, sis the settlement, L is the pile length,
Wp is the pile weight, A is the pile cross-sectional area and E is the pile modulus of elasticity.
In back-propagation neural networks these formulae are neglected, only the data plays a critical
role. Thus, the data has been extracted from different experiments and case studies to estimate
the ultimate bearing capacity of the pile foundation. A summary of the source from which the
data has been extracted is presented in Table 2.
nntool in MATLAB.
Figure 5.1: Neural network tool in MATLAB
The data set i.e., the input values, output values (for training phase) and testing values must be
imported to the nntool. Then the neural network must be created using the imported values and
changing the training parameters. Finally, in the testing phase will occur after attaining the
required regression values.
Generally, the reliability of the model improved as the number of valid input parameters
increased. Experiments were carried out using several combinations of input parameters to
determine the most reliable neural network model. The neural network model with seven input
neurons representing E, A, Wp , H, W, s, and the hammer type (H type) and two hidden neurons
was found to be most reliable. The corresponding neural network model is represented in figure
Statistical methods are utilized in the development of standard relationships between various
factors. This is often complex and circuitous, particularly for nonlinear relationships. Also, to
formulate the statistical model, the important parameters must be known. By comparison, the
modelling process in back-propagation neural networks is more direct, as there is no necessity to
specify a mathematical relationship between the input and output variables. Neural networks
can be effective for analysing a system containing several variables, to establish patterns and
characteristics not previously known. In addition, it can generalize correct responses that only
broadly resemble the data in the training set. During training, irrelevant input variables are
assigned low connection weights. These variables can then be omitted from the model. In neural
networks, quantitative as well as qualitative information can be considered. As new data
become available, the neural network model can be readily updated by retraining with patterns
which include these new data.
From the analysis of the data it is observed that the back propagation neural network Adopted
in ANN is providing sufficiently good response of the system. The results can be improved by
adding more data of both cohesive and cohesionless soils from different parts of the world.
Since most of the data in cohesionless soils have been collected from Iran, the results are well
matching. However, if data from different soil types or regions are fed to the system, the
prediction will improve.
References
1. Abu-Farsakh, M.Y., Titi, H.H., 2004. Assessment of direct cone penetration test
methods for predicting the ultimate capacity of friction driven piles. Journal of
Geotechnical and Geoenvironmental Engineering 130 (9), 935e944.
2. Abu-Kiefa, M.A., 1998. General regression neural networks for driven piles in
cohesionless soils. Journal of Geotechnical & Geoenvironmental Engineering 124
(12), 1177e1185.
3. Alkroosh, I., Nikraz, H., 2011a. Correlation of pile axial capacity and CPT data
using gene expression programming. Geotechnical and Geological Engineering
29, 725e748.
4. Alkroosh, I., Nikraz, H., 2011b. Simulating pile load-settlement behavior from CPT
data using intelligent computing. Central European Journal of Engineering 1 (3),
295e305.
5. Alkroosh, I., Nikraz, H., 2012. Predicting axial capacity of driven piles in cohesive
soils using intelligent computing. Engineering Applications of Artificial
Intelligence 25 (3), 618e627.
6. Bowles, J.E., 1997. Foundation Analysis and Design. McGraw-Hill, New York.
7. Broms, B.B., Lim, P.C., 1988. A simple pile driving formula based on stress-wave
measurements. In: Proc., Proceedings of the 3rd International Conference on the
Application of Stress-wave Theory to Piles. BiTech Publishers, Vancouver, pp.
591e600.
8. Brown, M., Harris, C.J., 1995. A perspective and critique of adaptive neurofuzzy
systems used for modelling and control applications. International Journal of
Neural Systems 6 (2), 197e220.
9. Burland, J.B., 1973. Shaft friction of piles in clay. Ground Engineering 6 (3), 1e15.
10. Chan, W.T., Chow, Y.K., Liu, L.F., 1995. Neural network: an alternative to pile
driving formulas. Computers and Geotechnics 17, 135e156.
11. Chen, Y.J., Kulhawy, F.H., 1994. Case History Evaluation of the Behavior of Drilled
Shafts under Axial and Lateral Loading. Report No. TR-104601. Electric Power
Research Institute, Palo Alto, Calif.
12. Coyle, H.M., Castello, R.R., 1981. New design correlations for piles in sand.
Journal of Geotechnical Engineering 107 (7), 965e986.
13. Das, B.M., 1995. Principles of Foundation Engineering, third ed. PWS Publishing
Company, Boston, MA.
14. Das, S.K., Basudhar, P.K., 2006. Undrained lateral load capacity of piles in clay
using artificial neural network. Computers and Geotechnics 33 (8), 454e459.
15. Eberhart, R.C., Kennedy, J., 1995. A new optimizer using particle swarm theory.
In: Proc., Proceedings of the Sixth International Symposium on Micro-machine
and Human Science. IEEE, Nagoya, pp. 39e43.
16. Elshorbagy, A., Corzo, G., Srinivasulu, S., Solomatine, D.P., 2010. Experimental
investigation of the predictive capabilities of data driven modeling techniques in
hydrology-part 1: concepts and methodology. Hydrology and Earth System
Science 14, 1931e1941.
17. Flood, I., 2008. Towards the next generation of artificial neural networks for civil
engineering. Advanced Engineering Informatics 22 (1), 4e14.
18. Gandomi, A.H., Alavi, A.H., 2012. A new multi-gene genetic programming
approach to non-linear system modeling. Part II: geotechnical and earthquake
engineering problems. Neural Computing Applications 21, 189e201.
19. Gardner, M.W., Dorling, S.R., 1998. Artificial neural networks (the multilayer
perceptron) e a review of applications in the atmospheric sciences. Atmospheric
Environment 32 (14/15), 2627e2636.
20. Giustolisi, O., Savic, D.A., 2006. A symbolic data-driven technique based on
evolutionary polynomial regression. Journal of Hydroinformatics 8 (3), 207e222.
21. Goh, A.T., Kulhawy, F.H., Chua, C.G., 2005. Bayesian neural network analysis of
undrained side resistance of drilled shafts. Journal of Geotechnical and
Geoenvironmental Engineering 131 (1), 84e93.
22. Goh, A.T.C., 1994. Nonlinear modelling in geotechnical engineering using neural
networks. Australian Civil Engineering Transactions CE36 (4), 293e297.
23. Goh, A.T.C., 1995a. Back-propagation neural networks for modeling complex
systems. Artificial Intelligence in Engineering 9, 143e151.
24. Goh, A.T.C., 1995b. Empirical design in geotechnics using neural networks.
Geotechnique 45 (4), 709e714.
25. Goh, A.T.C., 1996. Pile driving records reanalyzed using neural networks. Journal
of Geotechnical Engineering 122 (6), 492e495.
26. Goldberg, D.E., 1989. Genetic Algorithms in Search Optimization and Machine
Learning. Addison e Wesley, Mass.
27. Hiley, A., 1922. The efficiency of the hammer blow, and its effects with reference
to piling. Engineering 2, 673.
28. Koza, J.R., 1992. Genetic Programming: on the Programming of Computers by
Natural Selection. MIT Press, Cambridge (MA).
29. Lee, I.M., Lee, J.H., 1996. Prediction of pile bearing capacity using artificial neural
networks. Computers and Geotechnics 18 (3), 189e200.
30. Maier, H.R., Dandy, G.C., 2000. Applications of artificial neural networks to
forecasting of surface water quality variables: issues, applications and challenges.
In: Govindaraju, R.S., Rao, A.R. (Eds.), Artificial Neural Networks in Hydrology.
Kluwer, Dordrecht, The Netherlands, pp. 287e309.
31. McCulloch, W.S., Pitts, W., 1943. A logical calculus of ideas imminent in nervous
activity. Bulletin and Mathematical Biophysics 5, 115e133.
32. Meyerhof, G.G., 1976. Bearing capacity and settlement of pile foundations.
Journal of Geotechnical Engineering 102 (3), 196e228.
33. Nawari, N.O., Liang, R., Nusairat, J., 1999. Artificial intelligence techniques for the
design and analysis of deep foundations. Electronic Journal of Geotechnical
Engineering 4. http://geotech.civeng.okstate.edu/ejge/ppr9909/index.html.
34. Rezania, M., Faramarzi, A., Javadi, A., 2011. An evolutionary based approach
for assessment of earthquake-induced soil liquefaction and lateral
displacement. Engineering Applications of Artificial Intelligence 24 (1),
142e153.
35. de Ruiter, J., Beringen, F.L., 1979. Pile foundation for large North Sea structures.
Marine Geotechnology 3 (3), 267e314.
36. Rumelhart, D.E., Hinton, G.E., Williams, R.J., 1986. Learning internal
representation by error propagation. In: Rumelhart, D.E., McClelland, J.L.
(Eds.), Parallel Distributed Processing. MIT Press, Cambridge, pp. 318e362.
37. Savic, D.A., Giutolisi, O., Berardi, L., Shepherd, W., Djordjevic, S., Saul, A., 2006.
Modelling sewer failure by evolutionary computing. Proceedings of the
Institution of Engineers, Water Management 159 (WM2), 111e118.
38. Schmertmann, J.H., 1978. Guidelines for Cone Penetration Test, Performance
and Design. FHWA-TS-78e209. U. S. Department of Transportation,
Washington, D. C., p. 145
39. Semple, R.M., Rigden, W.J., 1986. Shaft capacity of driven pipe piles in clay.
Ground Engineering 19 (1), 11e17.
40. Shahin, M.A., 2010. Intelligent computing for modelling axial capacity of pile
foundations. Canadian Geotechnical Journal 47 (2), 230e243.
41. Shahin, M.A., 2013. Artificial intelligence in geotechnical engineering:
applications, modeling aspects, and future directions. In: Yang, X., Gandomi,
A.H., Talatahari, S., Alavi, A.H. (Eds.), Metaheuristics in Water, Geotechnical
and Transport Engineering. Elsevier Inc., London, pp. 169e204.
42. Shahin, M.A., 2014a. Load-settlement modeling of axially loaded drilled shafts
using CPT-based recurrent neural networks. International Journal of
52. Vesic, A.S., 1977. Design of Pile Foundations. National Cooperative Highway
Research, Synthetic of Practice No. 42. Transportation Research Board,
Washington DC.
53. Wellington, A.M., 1892. The iron wharf at Fort Monroe, Va. ASCE Transactions
27,129e137.
54. Zurada, J.M., 1992. Introduction to Artificial Neural Systems. West Publishing
Company, St. Paul