Mathematical and Computational Applications, Vol. 15, No. 1, pp. 66-78, 2010.

© Association for Scientific Research

NEURAL NETWORK MODEL FOR MOMENT-CURVATURE

RELATIONSHIP OF REINFORCED CONCRETE SECTIONS

Muhiddin Bağcı
Civil Engineering Department, Celal Bayar University,
Manisa, Turkey. [Link]@[Link]

Abstract- The analysis of moment-curvature relationship of reinforced concrete

sections is complex due to large number of variables as well as non-linear material
behavior involved. Artificial Neural Networks (ANNs) are found to be a tool capable of
solving such problems. This has led to increasing use of ANN for analyzing the
behavior of reinforced concrete sections. This paper reports the details of a study
conducted using ANN for predicting moment-curvature relationship of a reinforced
concrete section. Using data generated based on the analytical solutions, the ANN
model was trained. The trained model was tested for a different set of input parameters
and the output values were compared with the values based on analytical results. The
agreement was found to be good.

Key Words- Moment-Curvature, Neutral Network

1. INTRODUCTION

The moment curvature for a cross-section envelope describes the changes in force
capacity with deformation during a nonlinear analysis. The relationship between
moment and curvature demonstrates the strength, ductility, energy dissipation capacity
and rigidity of the section under question. To obtain the moment-curvature relationship
of reinforced concrete section, various researchers have investigated using different
models. Parviz [1] used firstly the filament method. Ersoy and Özcebe [2] presented a
computer program to determine moment-curvature relationships of confined concrete
sections. Artificial neural networks (ANN) are one of the artificial intelligence (AI)
applications which have recently been used widely to model some of human interesting
activities in many areas of science and engineering. The generalized delta rule algorithm
of artificial neural networks is employed to predict the flexural behavior of Steel Fibre
Reinforced Concrete (SFRC) T-beams using a computer program developed using C++
by Patodi and Purani [3]. For some other examples of ANN applications in structural
analysis, the reader is referred to Jadid et al. [4] ; Berke et al. [5] ; Lee et al. [6];
Avdelas et al [7]; Abdalla and Stavroulakis [8]; Karlık et al. [9]. As far as structural
analysis and design are concerned, Hajela et al. [10] used BPNN to represent the force-
displacement relationship in static structural analysis. Jenkins considered the application
of neural nets to approximate structural analysis and especially to a comparatively
simple structure [11]. Mukherjee et al. [12] mapped the relationship between the
slenderness ration, the modulus of elasticity and the buckling load for columns. As the
input taken directly from the experimental results, factors affecting the buckling load of
columns are automatically incorporated in the model to a great extent. Adeli defined the
learning parameters as a function of iteration number of the training [13].
 M. Bağcı 67

In this study, the behavior values of reinforced concrete sections subjected to

flexure and axial load were obtained by using an analytical solution named the filament
model, and then the required data for the network training were prepared. To obtain the
behavior of confined concrete, several data points were used in training a multi-layer,
feed-forward and back propagation artificial neural network (ANN) algorithm. The
behavior values were calculated using the neural network and were compared with those
obtained from the analytical results. Finally, the reliability of the ANN solution was
validated by comparing experimental values with modeled values.

2. MATERIAL MODELS

Moment-curvature analysis for a reinforced concrete section, indicating the

available flexural strength and ductility can be carried out provided that the stress-strain
relationships for the concrete and steel reinforcements are known. Typical stress–strain
curves for concrete are shown in Fig. 1.
Stress, f c

f cc' Confined Concrete

f c'

Unconfined Concrete

εo kεo Strain , ε c
f ct'
Fig. 1. Typical concrete stress-strain curve

The material model for concrete used in this analysis is based on a model
suggested by Modified Kent–Park [14]. This model takes into account the different
stress-strain curves for unconfined and confined concrete as shown in Fig. 2. The
general shape of curve is modeled by a second degree parabola for the ascending branch
up to the maximum stress which corresponds to strain level of 0.002 and linear
horizontal part leading to the ultimate strain.
Stress, fc'
Confined concrete

kf c'

f c'

0.5kf c'
Unconfined
Concrete
0.2kf c'

εo kε o ε 50u + ε 50 h Strain , ε c
Fig. 2. Modified Kent-Park concrete stress-strain curve.
68 Neural Network Model for Reinforced Concrete Sections

In the ascending branch, the concrete compressive stress, f c , at a given strain, ε c is

given by:
  ε   ε 2 
f c = kf c'  2  c  −  c   ; ε c ≤ kε o (1)
  εok   εok  
 
where f c' is the concrete compressive strength and ε o is concrete strain at the
maximum stress assumed to be 0.002. It can be seen from Eq. (1) that concrete reaches
a maximum stress of kf c' at a strain of k ε o . k is a factor which accounts for the strength
increase due to the confinement. The value of the parameter k is obtained from;
ρ s f yh
k =1 + '
(2)
fc
where f yh is the yield strength of stirrups, and ρ s is the ratio of the volume of hoop
reinforcement to the volume of concrete core measured to the outside of stirrups. The
descending branch of the stress–strain curve is described as follows;

f c = kf c' [1 − Z m (ε c − ε o k )] ≥ 0.2 f c' ; ε c > kε o (3)

where
0.5
Zm = (4)
ε 50u + ε 50 h − ε o k
3 + 0.285 f c'
ε 50 u = (5)
1.42 f c' − 1000
ε 50 h = 0.75 ρ s b '' / s (6)

The term ε 50u defines the slope of the falling branch of the unconfined concrete,
which identifies the strain at which the stress has fallen to 0.5 kf c' . The ε 50 h is the
additional ductility in concrete which is provided by transverse reinforcement. The “b”
is the width of the confined core measured to the outside of stirrups, and s is the center
to center spacing of stirrups or hoop sets. At large strains, the value of compressive
stress is kept constant at 0.2 kf c' to account for the ability of concrete to support load at
large strains.
When a reinforced concrete member is subjected to tensile strains less than the
cracking strain of concrete, the stress-strain relationship is approximately linear. A bi-
linear model is used for concrete in tension. Rüsch [15] recommends the following
relationship:
f c tention = − E ct ε ct ; 0 ≤ εt ≤ ε ct (7)
f c tension = f ct − 5000 f ct (ε ct − 1000 ) ; ε t ≥ ε ct (8)
where E ct is the modulus of elasticity in tension, ε ct is the tensile strain. c1 is taken as
0.5, ε cto as 0.0001 and ε ctu as 0.0002. The relationships given in Eq. (7) and Eq.(8) are
shown in Fig. 3.
 M. Bağcı 69

A sample reinforcement stress–strain relationship is shown in Fig.4. The

constitutive model used for steel reinforcement is a simple elastic-plastic three linear
model. These are the linear segment, the yield plateau, and the strain hardening
segment. There are some other parameters of the reinforcement stress–strain
relationship, such as the reinforcement yield strength , f yk , the ultimate reinforcement
strength, f su , the reinforcement yield strain, ε sy , hardening strain , ε st , the ultimate
strain, ε su , and modulus of elasticity , E s .
f c tension
Stress, f s
f ct f su
ff
c1 f ct yk

εt Strain, ε s
ε ct ε ctu ε sy ε st ε su
Fig. 3. Tension model for RC Fig. 4. Idealized reinforcement stress-strain curve

3. METHOD OF ANALYSIS

The reinforced concrete section is modeled using filament method. As can be seen
from Fig. 5, the cross-section is divided into 40 filaments to determine a moment-
curvature relationship. For each filament, confined core and unconfined cover areas are
defined. For a given strain at the extreme fiber in compression, the depth of neutral axis
satisfying the force equilibrium is found by trial. For each filament, the average stresses
are calculated at the centroids of unconfined and confined portions of the filament. To
achieve this, first the strain at the centroids of the filament is calculated using the
compatibility requirements. This centroidal strain is later used along with the
appropriate concrete models to calculate stresses acting on the unconfined and confined
portions of the filament. Finite concrete forces for the confined and unconfined portions
of the filament (∆Fcc and ∆Fcu respectively) are given multiplying the stress with the
corresponding areas as follows:
∆Fcu = f cui . Acui (9)
and
∆Fcc = f cci . Acci (10)
where fcci and fcui are the concrete stresses for the confined and unconfined portions of
the layer i.
Stress in the reinforcement at a given level is found by entering the f-ε diagram of
steel with the strain value found from the compatibility requirements. Steel force at that
level is given by multiplying the stress found with the area of the reinforcement at that
level as:
Fsi = σ si Asi (11)
This algorithm is demonstrated in Fig. 5 where only some typical finite forces are
shown.
70 Neural Network Model for Reinforced Concrete Sections

The moment-curvature relationship for a given axial load is determined by step by

step incrementing concrete strain in the extreme compression fibre ε cm . For each value
of ε cm , the strain gradient, i.e. the curvature φ, is obtained by satisfying the force
equilibrium equation. The bending moment M corresponding to chosen value of ε cm
and axial load N are determined by taking moments of the internal forces about the
geometric centroids of the section.

εcmm
εsi Fsi = σsi Asi
Asi εci
Acui ∆Fci = σcui Acui +σcci Acci
Acui M
Acci
2 2
n.a
Actj N
εctj ∆Fctj = σctj Actj

Fig. 5. Strains and finite forces in the cross-section.

4. PARAMETRIC STUDY

In this section the effect of different variables on flexural behavior are

investigated using analytical solutions developed to predict the moment-curvature
relationship of reinforced concrete cross-sections shown in Fig. 6.
fck Characteristic strength of concrete ( Mpa)
400 fyk Yield strength of reinforcing steel ( Mpa)
fsh Yield strength of transverse steel ( Mpa)
50
εcu Extreme fiber strain of unconfined concrete in
compression .
400
εcto Strain of concrete in tension (0.0001)
εctu Extreme strain of concrete in tension (0.0002)
εsy Yield strain of reinforcing steel (0.0021)
50 εsh Hardening Strain of reinforcing steel (0.01)
εsu Extreme strain of reinforcing steel (0.1)
Fig. 6. The cross-section considered in analyses.
Table 1. Results according to different variables

Variable properties Curvature (rad/m)

M
εcm (kN-m)
fck N/No fsh s Ø fyk
No ρ TY TH CvC CoC
(Mpa) (N) (Mpa) (cm) (mm) (Mpa)
1 30 0 420 15 8 0.02 420 0.0085 0.0365 0.0321 0.0492 0.0125 241.0

2 20 0 420 15 8 0.02 420 0.0105 - 0.0245 0.0350 0.0125 228.2

3 16 0 420 15 8 0.02 420 0.0093 0.0363 0.0212 0.0323 0.0125 222.4

4 30 0.25 420 15 8 0.02 420 0.0112 - 0.0120 0.0141 0.0028 350.1

5 20 0.25 420 15 8 0.02 420 0.0171 - 0.0115 0.0125 0.0032 296.2

6 16 0.25 420 15 8 0.02 420 0.0167 - 0.0112 0.0118 0.0032 236.9

7 30 0.5 420 15 8 0.02 420 0.0251 - 0.0079 0.0089 0.0032 342.4

 M. Bağcı 71

8 20 0.5 420 15 8 0.02 420 0.0264 - 0.0077 0.0081 0.0040 279.2

9 16 0.5 420 15 8 0.02 420 0.0263 - 0.0076 0.0079 0.0040 253.1

10 30 0.75 420 15 8 0.02 420 - - 0.0078 0.0071 0.0028 246.1

11 20 0.75 420 15 8 0.02 420 - - 0.0054 0.0065 0.0036 217.2

12 16 0.75 420 15 8 0.02 420 - - 0.0048 0.0059 0.0036 173.8

13 30 0 420 15 8 0.011 420 0.0082 0.0338 0.0391 0.0648 0.0100 135.6

14 30 0 360 15 8 0.011 420 0.0082 0.0338 0.0391 0.0647 0.0100 135.5

15 30 0 300 15 8 0.011 420 0.0082 0.0337 0.0390 0.0646 0.0100 135.3

15 30 0 220 15 8 0.011 420 0.0082 0.0337 0.0390 0.0645 0.0100 135.0

17 30 0.25 420 15 8 0.011 420 0.0115 0.0520 0.0133 0.0151 0.0034 214.3

18 30 0.25 360 15 8 0.011 420 0.0115 0.0518 0.0133 0.0151 0.0032 213.7

19 30 0.25 300 15 8 0.011 420 0.0115 0.0515 0.0133 0.0150 0.0032 213.2

20 30 0.25 220 15 8 0.011 420 0.0114 0.0534 0.0132 0.0150 0.0032 212.6

21 30 0.50 420 15 8 0.011 420 0.0194 - 0.0082 - 0.0038 229.1

22 30 0.50 360 15 8 0.011 420 0.0193 - 0.0082 - 0.0038 227.2

23 30 0.50 300 15 8 0.011 420 0.0210 - 0.0081 - 0.0038 225.3

24 30 0.50 220 15 8 0.011 420 0.0208 - 0.0081 - 0.0038 222.7

25 30 0.75 420 15 8 0.011 420 - - 0.0057 - 0.0030 172.7

… ….. ….. ….. …. ….. ….. …. ….. ….. …... …... ……. …..

50 30 0.5 420 15 8 0.011 220 0.0104 - 0.0084 0.0094 0.0028 196.7

51 30 0.75 420 15 8 0.011 320 0.0151 - 0.0058 0.0065 0.0030 163.7

52 30 0.75 420 15 8 0.011 220 0.0124 - 0.0059 0.0067 0.0030 157.8

The results of the parametric study on reinforced concrete members presented

here allow the following conclusions to be drawn.
1- As the compressive concrete strength increases, the tendency toward a brittle, sudden
failure also increases. One of the disadvantages of a high-strength concrete is that it is
more brittle than a concrete of a lower strength. The increasing compressive strength
causes a decrease in ductility. The compressive strength, fck, does not have any effects
on the behavior in the case of pure bending. The compressive strength becomes
effective with increasing axial load. The maximum moment capacity changes ±25% due
to ±25% compressive strength variation.
2- The ductility decreases as level of the axial load increases. The variation of ductility
with the level of axial load is quite significant. It is interesting to note that, although the
sections considered are well confined, the behavior becomes very brittle under high
levels of axial load. The upper limits imposed on axial loads in seismic codes roots
from such considerations.
3- It is found that yield strength of transverse reinforcement, fsh, has no effect on the
behavior at all levels of the axial load.
4- The most important parameters for obtaining a ductile behavior are spacing of the
confinement and the reinforcement configuration. Generally, closer confinement
spacing and a denser reinforcement configuration does not contribute to a higher load
72 Neural Network Model for Reinforced Concrete Sections

capacity. The results presented in this study show that for a well-confined cross-section
it is an advantage to use a higher grade of steel, while for a lightly confined section it is
not. Table 1 shows that closer confinement spacing has little effect on maximum load.
However, by decreasing the reinforcement confinement spacing a less brittle behavior
can be achieved. From these tables it can be seen that the greatest effect of confinement
is gained in pure compression.
5- To achieve ductility, the transverse reinforcement volume ratio needs to be increased
and the reinforcement configuration should be designed to provide high confinement.
As can be seen from Table 1, the increase in ductility with transverse reinforcement
diameter has no significant effect on moment capacity. The crushing of core concrete
delays with an increase in the diameter of transverse reinforcement. The diameter of
transverse reinforcement becomes effective with the increasing axial load.
6- The reinforcement volumetric ratio, ρ has an important effect on the behavior of the
confined section. The reinforcement volumetric ratio has significant effect on the
behavior at low level axial load. The ultimate moment capacity increases from 10 % up
to 30 % with the reinforcement volumetric ratio. The moment capacity decreases with
the higher axial load. The reinforcement volumetric ratio is not effective on ductility.
7- The ductility increases remarkably when the reinforcement yield strength is increased
with reinforcement configuration. The reinforcement yield strength, fyk, , is an effective
parameter in case of pure bending. The ultimate moment capacity changes from ±10 %
up to ±30 % with the reinforcement yield strength.

5. ANN MODELING

Artificial Neural Networks (ANNs) approach is used to determine the behavior of

confined concrete sections in this study. ANNs do not require an explicit understanding
of the mechanism underlying the process, which is the main advantage. It has the
capacity to learn the relationship between input and output provided that sufficient data
are available for its training. The analytical results available for the confined sections
were used to prepare the training and testing data sets for the network.
The present study is concerned with the prediction of a confined section using
ANN. In this study a neural network program model developed by Karlık [9] in
PASCAL was used. The data for training and testing were formed using parametric
results. For generating the data analytically, filament method is used. The database
consists of 52 sets of results, of which 45 sets were used for training the network, and
the remaining 7 were used for testing in Table 1.
The training patterns should be normalized before they are applied to the neural
network so as to limit the input and output values within a specified range. This is due
to the large difference in the values of the data provided to the neural network. Besides,
the activation function used in the back propagation neural network is a sigmoid
function. The lower and upper limits of this function are 0 and 1, respectively. The
following formula is used to pre-process the input data sets whose values are between 0
and 1.
x − x min
v= [12]
x max − x min
 M. Bağcı 73

Since the output value of the sigmoid function is between 0 and 1, the following
function might be used. The combinations momentum rates {0, 0.3, 0.5, 0.7, and 0.9}
are used to investigate their effects on the behaviour of the neural network convergence.
The results are shown in Fig.7. The effects of all the given learning parameters and
momentum rates on the convergence epoch and generalization of the neural network are
shown in Table.2.
t − t min
o= [13]
t max − t min

Fig.7 . Effect of Momentum Rate on the Training of Neural Network

Table.2. Effect of Learning and Momentum Rate on the Behaviour of Neural

Network
Momentum Learning rate %
Behaviour of Neural Network
factor
Training of the neural network 7.82
0.9
Generalizations for test patterns 3.7

The parametric study was conducted to find out the optimum number of hidden
layers as well as the number of nodes for the present problem. The results of the
parametric study conducted were shown in Fig.8. Training for all these network
configurations was carried out initially for one thousand cycles with error tolerance
value of 0.025. When the number of hidden layers was made two, only the architecture
12-13 reached the smallest error tolerance in 1000 cycles. With one hidden layer, the
architecture was not able to attain the required error tolerance of 0.0065 within 1000
cycles. Hence, for the problem under consideration, the network with 2 hidden layers
having the 12-13 architecture was chosen since it reaches the required error tolerance
with the least number of cycles, which in turn would reduce the CPU time requirement.
74 Neural Network Model for Reinforced Concrete Sections

3,00

2,50

2,00
% error
1,50

1,00

0,50

0,00
10

11

12

13

14

15

11_11

12_12

13_13

14-14

12_13

13_12

11_13
Number of hidden layer

Fig. 8. The error changes due to the number of nodes in the hidden layer at 1000 iterations.

Using the 7-12-13-6 architecture in Fig.9, the network was trained and then tested.
For training the network, totally 45 data set were used which were listed under Table 3.
These input data sets were analytically generated using the filament model. The
network, after being trained, was tested with 7 data sets. These 7 input data sets were
formerly generated using the filament model. The remaining data sets used for testing
the network are shown in Table 4.
Hidden Layer Hidden Layer

Input
Output

fck
TY
N/No
TH
fsh
CvC
s
CoC
Ø
ε
ρ
M
fyk

Fig. 9. ANN architecture

Finally, the least required error convergence for 7-12-13-6 architecture was
reached within 5000 cycles. A numerical study of training and testing of the network
was conducted keeping the error tolerance values as 0.1, and 0.001. For an error
tolerance of 0.1, the number of cycles required is less; but the results are less accurate.
In the case of 0.001, even though the accuracy is high, the numbers of cycles required
are very high. Hence, keeping in mind the number of cycles required for convergence
together with the accuracy needed for training and testing, the minimum error tolerance
was chosen as 0.7% in Fig. 10
The training results predicted using ANN is compared with the parametric values
in Table 2. In these cases, results represent a one to one correspondence, that is, the
 M. Bağcı 75

predicted and the parametric values are identical. The average error between the
analytical − ANN / ANN
analytical and the ANN values ( ) produced is less than 0.2
number of solution
%. The maximum differences between the analytical and ANN for TY, TH, CvC, CoC,ε
and M are the outputs 0.965, 0.978, 1.039, 0.961, 0.962, and 0.976, respectively.
Technically speaking, these errors are regarded to be sufficiently low.
1,2
1,1
1
0,9
0,8
% error

0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Number of iterations

Fig. 10. The error change at optimum ANN architecture ([Link])

Table 3. Training process and results

Anal Anal Anal Anal Anal Anal
No Meth. TY / TH / CvC / CoC / ε / M /
ANN ANN ANN ANN ANN ANN
Analy. 0.0085 0.0365 0.0321 0.0492 0.0125 241.0
1 1,016 0,987 1,023 1,004 1,006 0,996
ANN 0.0083 0.0369 0.0313 0.0490 0.0124 241.99
Analy. 0.0105 - 0.0245 0.0350 0.0125 228.2
2 1,001 - 1,006 0,995 0,988 0,997
ANN 0.0104 - 0.0243 0.0351 0.0126 228.86
Analy. 0.0171 - 0.0115 0.0125 0.0032 296.2
5 1,004 - 1,027 0,977 1,005 1,003
ANN 0.0170 - 0.0111 0.0128 0.0031 295.29
Analy. 0.0167 - 0.0112 0.0118 0.0032 236.9
6 0,999 - 0,966 0,993 0,993 0,995
ANN 0.0167 - 0.0115 0.0118 0.0032 238.05
Analy. 0.0264 - 0.0077 0.0081 0.0040 279.2
8 0,999 - 1,029 0,983 1,002 0,995
ANN 0.0264 - 0.0074 0.0082 0.0039 280.52
Analy. 0.0263 - 0.0076 0.0079 0.0040 253.1
9 1,000 - 0,996 1,023 1,000 1,002
ANN 0.0263 - 0.0076 0.0077 0.0040 252.7
Analy. - - 0.0078 0.0071 0.0028 246.1
10 - - 0,992 0,969 0,994 1,000
ANN - - 0.0078 0.0073 0.0028 246.14
Analy. - - 0.0054 0.0065 0.0036 217.2
11 - - 0,996 1,009 1,004 1,001
ANN - - 0.0054 0.0064 0.0035 217.03
Analy. 0.0082 0.0338 0.0391 0.0648 0.0100 135.6
13 0,998 1,011 1,005 0,983 1,005 0,999
ANN 0.0082 0.0334 0.0388 0.0659 0.0099 135.74
Analy. 0.0082 0.0338 0.0391 0.0647 0.0100 135.5
14 0,991 1,008 1,010 0,987 1,000 1,003
ANN 0.0082 0.0335 0.0387 0.0655 0.0100 135.04
Analy. 0.0082 0.0337 0.0390 0.0645 0.0100 135.0
16 0,985 1,006 1,010 0,987 0,999 1,014
ANN 0.0083 0.0334 0.0386 0.0653 0.0100 133.11
Analy. 0.0115 0.0518 0.0133 0.0151 0.0032 213.7
18 1,008 1,008 0,984 0,981 0,988 1,007
ANN 0.0114 0.0513 0.0135 0.0153 0.0032 212.29
Analy. 0.0115 0.0515 0.0133 0.0150 0.0032 213.2
19 1,010 0,994 0,995 0,977 1,026 1,009
ANN 0.0113 0.0517 0.0133 0.0153 0.0031 211.2
Analy. 0.0114 0.0534 0.0132 0.0150 0.0032 212.6
20 0,985 1,011 1,013 0,988 1,028 1,002
ANN 0.0115 0.0518 0.0130 0.0151 0.0031 212.25
Analy. 0.0194 - 0.0082 - 0.0038 229.1
21 0,985 - 1,011 - 0,994 0,995
ANN 0.0196 - 0.0081 - 0.0038 230.16
Analy. 0.0193 - 0.0082 - 0.0038 227.2
22 0,968 - 1,012 - 0,991 0,996
ANN 0.0199 - 0.0081 - 0.0038 228.16
… ….. …… …… …… ….. …. ….. …… ……. …… …… …… ……
Analy. 0.0104 - 0.0084 0.0094 0.0028 196.7
50 0,995 - 1,022 1,018 0,984 1,004
ANN 0.0104 - 0.0082 0.0092 0.0028 195.85
Analy. 0.0151 - 0.0058 0.0065 0.0030 163.7
51 1,005 - 0,963 0,961 0,997 0,994
ANN 0.0150 - 0.0060 0.0067 0.0030 164.71
Analy. 0.0124 - 0.0059 0.0067 0.0030 157.8
52 0,992 - 0,995 0,972 1,006 1,005
ANN 0.0125 - 0.0059 0.0068 0.0029 157.09
76 Neural Network Model for Reinforced Concrete Sections

The trained model was tested for a different set of input parameters and the output
values were compared with the values based on analytical results. Seven different input
values were applied to the model for testing the training network and the results were
obtained in milliseconds. A comparison of the test and analytical values is given in
analytical − ANN / ANN
Table 4. The average error ( ) obtained is about 0.33 %. The
number of solution
maximum differences (analytical / ANN) for TY, TH, CvC, CoC, ε and M are about
0.967, 0.966, 0.972, 0.968, 0.991, and 0.992, respectively. Therefore, the results can be
said to indicate that the trained NN models have achieved good performance.

Table 4. Testing process and results

Anal Anal Anal Anal Anal Anal
No Method TY / TH / CvC / CoC / ε / M /
ANN ANN ANN ANN ANN ANN
Analy. 0.0093 0.0363 0.0212 0.0323 0.0125 222.4
3 0.987 0.966 1.025 1.031 0.999 0.992
ANN 0.0094 0.0375 0.0206 0.0313 0.0125 224.1
Analy. 0.0112 - 0.0120 0.0141 0.0028 350.1
4 1.014 - 0.976 0.968 1.004 1.005
ANN 0.0110 - 0.0122 0.0145 0.0027 348.1
Analy. 0.0251 - 0.0079 0.0089 0.0032 342.4
7 1.001 - 0.972 1.032 0.997 1.000
ANN 0..025 - 0.0081 0.0086 0.0032 342.3
Analy. - - 0.0048 0.0059 0.0036 173.8
12 - - 1.024 1.021 1.002 0.994
ANN - - 0.0046 0.0057 0.0035 174.0
Analy. 0.0082 0.0337 0.0390 0.0646 0.0100 135.3
15 0.986 1.004 1.009 0.988 0.999 1.008
ANN 0.0083 0.0335 0.0386 0.0654 0.0100 134.1
Analy. 0.0115 0.0520 0.0133 0.0151 0.0034 214.3
17 0.992 1.012 0.979 0.975 0.991 0.998
ANN 0.0115 0.0513 0.0135 0.0154 0.0034 214.7
Analy. 0.0208 - 0.0081 - 0.0038 222.7
24 0.967 - 0.978 - 0.991 0.992
ANN 0.0215 - 0.0082 - 0.0038 224.4

Compared to conventional digital computing techniques, neural networks are

advantageous because of their special features, such as the massively parallel
processing, distributed storing of information, low sensitivity to error, their very robust
operation after training, generalisation and adaptability to new information.

6. CONCLUDING REMARKS

In this study, a back-propagation neural network model was employed to predict

the influence of various parameters on the behavior of reinforced concrete sections. A
neural network model was applied to the data derived from the analytical solutions. The
analytical model is based on a filament modeling technique and capable of taking into
account the crushing of cover and core concrete, the strain hardening of steel and the
effect of confinement on core concrete.
To reduce the computing time of microprocessor of the system, a new computer
model, which replies in milliseconds, was developed based on ANN method. A multi-
layer, back propagation and feed-forward ANN algorithm was used to train the data.
The ANN algorithms are not able to replace the conventional analytical techniques
completely since they need some key values for training. However, in the determination
of reinforced cross-section behavior, they can be implemented as an efficient
supplementary tool to reduce the computational cost drastically. Modeling process in
neural network is more direct since there is no necessity to specify a mathematical
relationship between the input and the output variables. The trained ANN was able to
 M. Bağcı 77

produce quick results for the reinforced cross-section behavior with the same degree of
accuracy as the filament model analysis achieved under flexure and axial load.
Therefore, the trained ANN may be used in practice for determining the reinforced
cross-sections behavior as an alternative to the time consuming filament model analysis.
The ANN applications presented in this study have demonstrated the viability and
feasibility of using analytical results for the reinforced confined sections’ behavior. The
obtained results have shown that the neural network model is successful in modeling the
non-linear relationship between different input and output parameters even when it
involves a relatively smaller number of training patterns. It is envisaged that the model
developed may be used in practical structural engineering applications.

7. REFERENCES

1. S. Parviz, S. Jongsung and W. H . Jer, Axial / Flexural Behavior of Reinforced

Concrete Sections: Effects of Design Variables, ACI, 88, 17-21, 1991.
2. U. Ersoy and G. Özcebe, Moment-Curvature Relationship of Confined Concrete
Sections, First Japan-Turkey Workshop On Earthquake Engineering, Ankara,
Turkey, 10-21,1997.
3. S. C. Patodi and V. S. Purani, Modeling flexural behavior of steel fibre reinforced
concrete beams using neural networks, Journal of New Building Materials and
Construction World, 4, 28-35, 1998.
4. M. N. Jadid and D.R. Fairbairn, Neural-network applications in predicting moment-
curvature parameters from experimental data, Engineering Applications of Artificial
Intelligence, 9, 309-319,1996.
5. L. Berke and P. Hajela, Applications of artificial neutral nets in structural mechanics,
Structural Optimizations, 4, 90-9, 1992.
6. Y. Lee, S. H. Oh, H. K. Hong and M. W. Kim, Design rules of multi-layer
perception, Proc. SPIE, Science of Artificial Neutral Nets in Structural Mechanics.
Structural Optimization, 1710, 329-339, 1992.
7. A. V. Avdelas, P. D. Panagiotopoulos and S. Kortesis, Neutral networks for
computing in the elastoplastic analysis of structures, Meccanica, 30,1-15, 1995.
8. K. M. Abdalla and G. E. Stavroulakis, A Back propagation neutral network model
for semi-rigid steel connections, Microcomputers in Civil Engineering, 10, 77-87,
1995.
9. B. Karlık, E. Ozkaya, S. Aydın, S. and M. Pakdemirli, Vibration of beam -mass
systems using artificial neutral networks, Computers & Structures, 69, 339-347,
1998.
10. P. Hajela, and L. Berke, Neurobiological computational models in structural
analysis and Design, Computers & Structures, 41, 657-667, 1991.
11. W. M. Jenkins, An Introduction to Neural Computing for the Structural Engineer.
The Structural Engineer, 75, 38-4,. 1997.
12. A. Mukherjee, J. M. Deshpande and J. Anmala, Prediction of Buckling Load of
Columns Using Artificial Neural Networks. Journal of Structural Engineering, 122,
1385-1387, 1996.
78 Neural Network Model for Reinforced Concrete Sections

13. H. Adeli and H.S. Park, Counter propagation Neural Networks in Structural
Engineering. Journal of Structural Engineering, .121, 1205-1212, 1995.
14. R. Park, M. J. N. Priestly and W. D. Gill, Ductility of square confined concrete
columns, Journal of Structural Division, ASCE, 108, 929–950, 1992.
15. H. Rüsch and H. Hilsdrof, Verfomungei-genschaften von beton unter zentrishcen
zugspannangen, Materialprüfungsemt für das bauwesen der technischen hochshule
München, 1963.