Eurasian Journal

Of
Civil Engineering and Architecture

Volume 1, Issue 2

1 December 2017

www.cessciencegroup.com
 EDITORIAL BOARD

Editor-in-Chief

Assoc. Prof. Dr. Serkan SAHINKAYA

Editors

Prof. Dr. Arzu KOCABAS DIREN (Mimar Sinan Fine Arts University, Turkey)

Prof. Dr. Aynur KAZAZ (Akdeniz University, Turkey)

Prof. Dr. Bahadır YUKSEL (Selcuk University, Turkey)

Prof. Dr. Kabir SADEGHI (Near East University, TRNC)

Prof. Dr. Mehmet Faik SEVIMLI (Karatay Technical University, Turkey)

Prof. Dr. Necati CATBAS (University of Central Florida, USA)

Assist. Prof. Dr. Abdulkerim ILGUN (Karatay Technical University, Turkey)

 EURASIAN JOURNAL OF
CIVIL ENGINEERING AND ARCHITECTURE

CONTENTS

Title of Articles Page Numbers

Interpreting Load-Settlement Curves of Pile Foundations by Graphical 1 – 10
Methods
Performative Approaches in Tall Buildings: Pearl River Tower 11 – 20
Reducing Sediment Deposition in a Clarification Tank Using Numerical 21 – 29
Modeling
Structural Deformation Measurement Using Matlab Image Processing 30 – 37
Toolbox
 Interpreting Load-Settlement Curves of Pile Foundations by Graphical Methods

M. Olgun, Y. Yenginar and A. Hanati

Department of Civil Engineering, Engineering Faculty, Selçuk University, Konya, TURKEY

(Corresponding Author’s E-mail: yavuzyenginar@selcuk.edu.tr)

ABSTRACT

In Geotechnical Engineering, deep foundation systems, especially pile foundations, are

typically used when shallow foundations are inadequate due to design criteria in terms of
bearing capacity, settlement, liquefaction and stability. The load settlement behavior of the
pile foundations can be determined realistically by full-scale loading tests conducted on the
piles after the pile construction. In this study, it was tried to determine the ultimate bearing
capacity of the diriven and bored piles manufactured in different soil conditions using the
load-settlement data. For this, 9 different graphical methods such as Tangent, Fuller-Hoy,
Butler-Hoy, De Beer-Wallays, Chin Kondner, Decourt, Brinch Hansen 80%, Mazurkiewicz
and the Corps of Engineers have been used. Among these methods, there are considerable
differences between the predicted ultimate bearing capacities of the piles which decreases to
35% for the piles loaded up/over to the collapse load, and increases up to 120% for the piles
loaded below the failure load. In generally, methods of Brinch Hansen, Mazurkiewicz, Chin
Kondner and Decourt predict the pile capacity greater than maximum test load and the other
ones predict smaller than it. The closest average collapse load was obtained from methods of
Corps of Engineering and Butller-Hoy.

Keywords: Bearing capacity, field pile loading tests, graphical methods.

INTRODUCTION

Pile loading experiments stand out as the most reliable approach that can be applied to
determine load-bearing capacities and load-settlement behavior of piles. Because this
experiment is a full scale model experiment. The application purposes of pile load tests are
determining the pile capacity, prediction pile service load after pile construction, determining
load-settlement relation of pile foundation.

Static loading experiments are a type of pile loading experiments. Axial pressure and axial
tensile tests are the types of static loading experiments. The standards used for these
experiments are ASTM D 1143-81 (1994), ASTM D-3689 (1995), ASTM D 3966-90 (1995),
ASTM D 3966-07 (2013), ASTM D-1143/D1143M (2013). The principles of the experiments
and the points to be noted are detailed in these standards.

The most reliable way of estimating the ultimate bearing capacity of the piles under vertical
loads is to apply the static axial loading test. These experiments are carried out by measuring
the settlements of the piles against these loads by applying predetermined loads to the pile. It
is estimated that from the data obtained as a result of these experiments, the settlement of the
pile will take place on the service load and ultimate load. As a result of these obtained results,
it is reached that the load bearing capacity of other piles to be produced in the project area
will be sufficient.

Several criteria have been proposed in the literature depending on settlement restriction to

1
determine the ultimate bearing capacities of the piles under axial compression and tension
loads. However, by analyzing the load-settlement curve obtained from the pile loading tests,
many graphical methods developed by Hansen (1963), Mazurkiewicz (1972), Chin-Kondner
(1970), Decourt (1999), Corps of Engineers (1991), Fuller and Hoy (1970), Butller and Hoy
(1977) and De Beer and Wallays (1989) have been proposed to find the pile bearing capacity.

Lastiasih and Sidi (2014) concluded that, using 130 pile loading test results, many graphical
methods predict the failure load of pile if the pile is loaded up to ultimate pile capacity.
Decourt (2008) improved some correlations using test result of standard penetration, cone
penetration and menard pressuremeter to predict pile capacity when the pile loading test
results did not achieved. Petek et al. (2012) evaluate the full scale loading test results of
driven pile and they concluded that there was %20 difference of pile capacity among the
graphical methods.

In this study, the ultimate capacities of the piles were determined by using 9 different methods
and the results were interpreted by using the pile load test data on 3 piles constructed in
different areas.

GRAPHICAL METHODS USED TO DETERMINE PILE CAPACITY

In the literature and in various standards, it is possible to find many criteria and methods for
interpreting pile loading test results to determining the ultimate load capacity of the test pile.
These methods can be summarized as methods for interpreting the distribution of load-
settlement data with various criteria that limit total settlement, plastic settlement or
settlement/load ratios.

Method of Brinch Hansen %80

Hansen (1963) proposed a definition for ultimate pile capacity as the load that gives four
times the settlement of the pile head as obtained for 80% of that load. This ‘80%- criterion’
can be estimated directly from the load-settlement curve, but it is more accurately determined
in a plot of the square root of each settlement value divided by its load value and plotted
against the settlement. This graph continues linearly after a certain point. Normally, the 80%-
criterion agrees well with the intuitively perceived “plunging failure” of the pile. The
following simple relations can be derived for computing the ultimate capacity, Qu;

1
Qu = (1)
2 C1 ⋅ C 2

where, C1: slope of the straight line, C2: intersection of load axis of the straight line.

Mazurkiewicz Method

Mazurkiewicz (1972) suggested a method of extrapolating the load-settlement curve,

assuming it is same as a parable. In this method, vertical parallel lines with equal distance to
settlement axis are drawn and the curve is intersected, then horizontal lines are drawn, starting
from each point of intersection to the load axis. Straight line segments make an angle of 45
degrees with the load axis are plotted, each with ends at the point of intersection of the load
axis and the next horizontal parallel line. Finally, the line passing through the intersections of
the segments with the horizontal lines and then this line extended to load axis. This
intersection point gave the ultimate bearing capacity of pile.
2
Chin-Kondner Method

The method proposed by Chin (1970), in a study based on work by Kondner, allows the
extrapolation of the failure load in the static load tests. In this method, a settlement/load-
settlement graph is drawn by dividing each load value by the settlement value corresponding
to this load value. The points on the graph show a linear trend after a certain value (Chin,
1970). The inverse of the slope of this line (1/C1) gives the ultimate bearing capacity of pile.

Decourt Method

In the Decourt (1999) method, a load/settlement vs load graph is drawn. The curve in the
resulting graph becomes linear when it approaches axis of abscissas and intersects the
abscissas when it is extended. Linear regression analysis is applied to these points with linear
trend. The ultimate bearing capacity of pile can be obtained by dividing the value at
intersection of vertical axis and the regression line by the slope of the regression line.

C2
Qu = (2)
C1

where; C2=value at intersection of vertical axis and the regression line, C1=slope of the
regression line

Tangent Method

In this method, the ultimate bearing capacity of piles can be determined by drawing first
tangent lines to the starting and ending portions of the load-settlement curves; the intersection
point of these two tangents was assumed to represent the ultimate bearing capacity of pile.

Corps of Engineers Method

This method is mostly used by the U.S. Army Corps of Engineers (1991). In this method,
load-settlement curve is drawn first, and then three different loads are detected. The firs load
(Q1) is the load corresponding to the 6.4 mm settlement level. The second one (Q2) is the load
corresponding to the point obtained by the tangential method. The third one (Q3) is the load
corresponding to the point at where the line make an angle of 0.025 mm/kN intersects with
load-settlement curve. The average of these three load values is considered as the ultimate
bearing capacity of pile.

Fuller and Hoy Method

In the Fuller and Hoy (1970) method, the ultimate load is determined by finding the point at
where the line make an angle of 0.127 mm/kN intersects with load-settlement curve.

Butller and Hoy Method

In this method, a line make angle of 0.127 mm/kN which tangent to the load-settlement curve,
as in the case with Fuller-Hoy method, is drawn. In addition to this, a new line tangent to the
initial part of the load-settlement curve is drawn. Finally, the load at which corresponds to
intersection of these two lines is the ultimate bearing capacity of pile (Butller and Hoy, 1977).

De Beer and Wallays Method

3
In the De Beer and Wallays (1989) method, the load-settlement graph is plotted on a
logarithmic scale for both axes. If the load applied to test pile passes the ultimate load, it is
observed that the points on the graph are located around the straight lines on the different
slopes. There is no result to be obtained from the slope of these straight lines, but the point
where the straight lines intersect is the point at which the reaction of the pile to the applied
load changes, and the load corresponding to this point is the ultimate load.

EVALUATION OF FIELD PILE LOADING TEST RESULTS

Design Features of Pile Foundations

In this study, authors evaluate the full scale pile loading test results performed in the field for
three different piles. Load-settlement data of pile 1 and pile 2 were taken an existing study
(Dinç, 2010) and load-settlement data of pile 3 obtained from loading test results achieved
from pile load test performed in Konya Industrial Zone, Turkey. Load-settlement curves and
design features of test piles are given in Fig. 1 and Table 1, respectively. Pile 1 and pile 2 can
be classified as mini piles because of their dimensions and they are prefabricated driven piles.
Pile 3 can be classified as cast-in-situ reinforced bored piles. It is understood that, pile 1 and 2
were loaded up to the ultimate load (Qu). Because there is no significant change in the amount
of load carried by the pile with further increasing deformations. Pile 3, however, was loaded
not to ultimate capacity but it was loaded up to the 2.25 times of design load, Qd=1600 kN
(Fig. 1).

Table 1. Design features of test piles

Pile 1 Pile 2 Pile 3

Construction process Prefabricated Prefabricated In-situ
Pile type Driven pile Driven pile Bored pile
Pile diameter, D (m) 0.3 0.3 1.0
Pile length, L (m) 7.6 8.0 25.0
Soil profile Sandy loam Sandy loam Soft clay
Design load (kN) 200 350 1600
Maximum test load (kN) 520 800 3600
Total settlement, δT (mm) 36.01 16.09 4.02
Plastic settlement, δP (mm) 29.74 11.00 3.05

4
 560 900
480 750
400
600
Q (kN)

Q (kN)
320
450
240
300
160
80 a) Pile 1 150 b) Pile 2
0 0
0 5 10 15 20 25 30 35 40 0 3 6 9 12 15 18
δ (mm) δ (mm)

3750

3000
c) Pile 3

2250
Q (kN)

1500

750

0
0 1 2 3 4 5
δ (mm)

Figure 1. Load-settlement curves obtained from field pile loading tests

Ultimate Bearing Capacities of Piles

Firstly, ultimate bearing capacities of test piles (Qu) determined by using above mentioned
graphical methods and these results compared with each other. For example, determination of
the ultimate capacity of pile 2 using 9 different graphical methods is given in Fig. 2. Then, the
safety factors (FS) were obtained by dividing the Qu values found in each method by the
project loads for all piles;

Qu
FS = (3)
Qd

where; Qu=ultimate bearing capacity of pile, Qd=design load of pile

Ultimate bearing capacities of test piles, according to different methods, changes from 408.3
kN to 555 kN for pile 1, 661.7 kN to 974.6 kN for pile 2 and 3100 kN to 6993 kN for pile 3.
The differences between minimum and maximum predicted ultimate pile capacities are
35.9%, 47.3% and 125.6% for pile 1, pile 2 and pile 3, respectively. Ultimate bearing capacity
value of a pile may change in a wide range according to the methods used to determine pile
capacity. Therefore, the most logical way to determine pile capacity is using average failure
load (Qu,avg)obtained from several methods. Then, ultimate bearing capacity of test piles may
be accepted as 495.8 kN, 784.3 kN and 4409.9 kN for pile 1, pile 2 and pile 3, respectively
(Table 2).

5
 0.014 900
Qu=810kN
0.012 C1=0.00012 750
sqrt(δ)/Q (mm/kN) C2=0.003
0.010 Qu=833.3 kN 600

Q (kN)
0.008 450
0.006 y = 0.00012x + 0.003 300
0.004 150
0.002 0
0 3 6 9 12 15 18 0 3 6 9 12 15 18
δ (mm) δ (mm)

a) Method of Brinch Hansen %80 b) Mazurkiewicz method

0.021 150

0.017
C1=0.0011
Qu=909.1kN 120
δ/Q (mm/kN)

Q/δ (kN/mm)
0.013 C2=295.59
90 C1=0.3033
0.009 Qu=974.6 kN
y = 0.0011x + 0.0025 60
0.005
y = -0.3033x + 295.59
0.001 30
0 3 6 9 12 15 18 0 150 300 450 600 750 900
δ (mm) Q (kN)
c) Chin-Kondner method d) Decourt method
900 900

750 Qu=700 kN 750 Q2

Q3
600 600 Q1
Q (kN)

Q (kN)

450 450
Q1=635 kN
Q2=700 kN
300 300 Q3=650 kN
Qu=661.7 kN
150 150

0 0
0 3 6 9 12 15 18 0 3 6 9 12 15 18
δ (mm) δ (mm)
e) Tangent method f) Corps of Engineers method
900 900
Qu=760kN
750 750 Qu=710kN
600 600
Q (kN)

Q (kN)

450 450

300 slope: 0.127 mm/kN 300 slope: 0.127 mm/kN

150 150

0 0
0 3 6 9 12 15 18 0 3 6 9 12 15 18
δ (mm) δ (mm)
g) Fuller and Hoy method h) Butller and Hoy method

Figure 2. Graphical methods used to determine ultimate capacity of pile 2

6
 1000

Qu=700kN

Q (kN)
100

10
0.1 1 10 100
δ (mm)
ı) De Beer and Wallays method

Figure 2 (continued). Graphical methods used to determine ultimate capacity of pile 2

Table 2. Failure loads of piles found by different methods and the safety factors (FS)

Pile 1 Pile 2 Pile 3

Methods
Qu (kN) FS Qu (kN) FS Qu (kN) FS
Brinch Hansen %80 527.0 2.64 833.3 2.38 6993.0 4.37
Mazurkiewicz 520.0 2.60 810.0 2.31 3700.0 2.31
Chin Kondner 555.0 2.78 909.1 2.60 5882.0 3.68
Decourt 543.0 2.72 974.6 2.78 5984.0 3.74
Tangent 502.0 2.51 700.0 2.00 3100.0 1.94
Corps of Engineers 408.3 2.04 661.7 1.89 - -
Fuller-Hoy 485.0 2.43 760.0 2.17 3320.0 2.08
Butter-Hoy 420.0 2.10 710.0 2.03 3200.0 2.00
De Beer and Wallays 502.0 2.51 700.0 2.00 3100.0 1.94
Average (Qu,avg or FSavg) 495.8 2.48 784.3 2.24 4409.9 2.76

Table 2 gives security numbers obtained by different methods for test piles. The average of
these security numbers is FSavg=2.48 for pile 1. The maximum load is 520 kN and the safety
factor according to this load is FS =520/200=2.60 for pile 1 since design load is 200 kN. This
value is very close but greater than FSavg=2.48. In this case, it is understood that pile 1 is
loaded over the ultimate bearing capacity. The average of security numbers is FSavg=2.24 for
pile 2. The maximum load is 800 kN and the safety factor according to this load is
FS=800/350=2.29. This value is very close but greater than FSavg=2.24. Pile 2 is also loaded
over the ultimate bearing capacity. In Table 2, the average of the security numbers obtained
for the different methods is FSavg=2.76 for pile 3. The maximum load is 3600 kN and the
safety factor according to this load is FS=3600/1600=2.25. This value is smaller than
FSavg=2.76. In this case, it is understood that pile 3 is loaded smaller than its ultimate bearing
capacity. The greatest ultimate bearing capacity and factor of safety are obtained from the
methods of Chin Kondner and Decourt for Pile 1 and pile 2. Brinch Hansen method, for pile
3, gives the greatest values in addition to these methods. This situation may be occurred if a
pile is not loaded up to the ultimate capacity. The smallest pile capacity and safety factor are
obtained for all piles by Corps of Engineers method. In addition to this, Butter-Hoy and De
Beer-Wallays methods also give small values.

In generally, methods of Brinch Hansen, Mazurkiewicz, Chin Kondner and Decourt predict
the pile capacity greater than maximum test load and the other ones predict smaller than it
(Table 3). The closest values to the maximum test load were obtained except from Corps of
7
Engineers and Butter-Hoy methods for pile 1. Brinch Hansen, Mazurkiewicz and Fuller-Hoy
methods give the loads closest to maximum test load for pile 2. The closest values to the
maximum test load, for pile 3, were obtained from Mazurkiewicz and Fuller-Hoy methods. If
the settlement value is smaller than 6.4 mm such as in the pile 3, the Corps of Engineers
method cannot be used. After that explanations, it can be said that all methods can predict the
ultimate pile capacity correctly if the test pile is loaded up to the ultimate capacity, if not
many methods such as Tangent, Corps of Engineers, Fuller-Hoy, Butter-Hoy, De Beer and
Wallays predict the pile capacity smaller than the real value.

Table 3 shows the mean ratio (Qu/Qavg) of the ultimate load (Qu) found by a certain method to
the average ultimate load (Qavg) values found by all methods for the same pile. Thus, it has
been determined which method gives a closer result to the mean capacity (Qave). It can be
misleading to using any method to determine the ultimate bearing capacity of a pile. Finding
individual results with different methods and getting their average values helps to stay on the
safer side. The following result can be concluded that the failure load values obtained from
the Brinch Hansen, Chin Kondner and Decourt methods predict pile capacity greater than
average capacity for all piles according to the other methods.

Table 3. Comparison of ultimate pile capacity with average failure load and maximum test
load for each method.

Pile 1 Pile 2 Pile 3

Methods
Qu/Qu,avg Qu/Qmax Qu/Qu,avg Qu/Qmax Qu/Qu,avg Qu/Qmax
Brinch Hansen %80 1.06 1.01 1.06 1.04 1.59 1.94
Mazurkiewicz 1.05 1.00 1.03 1.01 0.84 1.03
Chin Kondner 1.12 1.07 1.16 1.14 1.33 1.63
Decourt 1.10 1.04 1.24 1.22 1.36 1.66
Tangent 1.01 0.97 0.89 0.88 0.70 0.86
Corps of Engineers 0.82 0.79 0.84 0.83 - -
Fuller-Hoy 0.98 0.93 0.97 0.95 0.75 0.92
Butter-Hoy 0.85 0.81 0.91 0.89 0.73 0.89
De Beer and Wallays 1.01 0.97 0.89 0.88 0.70 0.86
Average 1.00 0.95 1.00 0.98 1.00 1.22

CONCLUSIONS

In many projects, some of the manufactured piles are loaded to determine the pile bearing
capacity. This is the most reliable way to determine the pile capacity. However, it is not easy
to determine the point where the pile has reached its ultimate capacity on the load-settlement
curve. In this study, it was tried to determine the ultimate bearing capacity of the diriven and
bored piles manufactured in different soil conditions using the load-settlement data. For this, 9
different graphical methods such as Tangent, Fuller-Hoy, Butler-Hoy, De Beer-Wallays, Chin
Kondner, Decourt, Brinch Hansen 80%, Mazurkiewicz and the Corps of Engineers have been
used.

According to the values of the ultimate loads obtained by different methods within the scope
of the study;
1. There can be considerable differences between the predicted ultimate bearing capacity
values of the piles according to the method used. This difference decreases to 35% for
the piles loaded up/over to the collapse load, and increases up to 120% for the piles
loaded below the failure load.
8
 2. In generally, methods of Brinch Hansen, Mazurkiewicz, Chin Kondner and Decourt
predict the pile capacity greater than maximum test load and the other ones predict
smaller than it.
3. If the pile is loaded up to the ultimate pile capacity, the results obtained by the
methods give a result close to the maximum test load, if not, greater values than
maximum test load is obtained.
4. The failure load values obtained from the Brinch Hansen, Chin Kondner and Decourt
methods predict pile capacity greater than average capacity for all piles according to
the other methods.
5. It has been found that if the average of the safety factors obtained for different
methods is close to the calculated safety coefficient for the maximum test load, the
determination of the bearing capacity by the different methods of the loaded piles will
be within acceptable limits.
6. It has been determined that the values closest to ultimate loads are obtained by
Mazurkiewicz and De Beer method.
7. The Corps of Engineering and Butller-Hoy method gave the smallest ultimate values.

Although the Mazurkiewicz method has the closest value to the ultimate load, some
operations are performed on the load-settlement graph to determine the ultimate load. These
operations can increase the error margin. In the Chin Kondner and Decourt methods, it is
stated that the ultimate load test data is predefined functions and that these functions are
expressed by the asymptote, that is, the defined ultimate load is analytically determined. The
advantage of analytical determination of ultimate load is that the result of the assessment is
not dependent on the scale of the applicant and the graph drawn when compared to graphical
methods. Therefore, a single method of evaluating the results of the pile loading test may lead
to a misleading situation.

In addition, no assessment was made of the criteria for the scope of this study. However, for
many batches, acceptance prerequisite is largely a settlement criterion from bearing capacity.
In this context, it will be the most realistic approach to evaluate the results by taking into
consideration the settlement criteria as well as determining the bearing capacities of the piles.

As a continuation of this study, it is planned to evaluate the pile loading test results with more
literature methods and to make a statistical study on this subject.

REFERENCES

ASTM D1143/D1143M-07. 2013. Standard Test Methods for Deep Foundations Under Static
Axial Compressive Load. ASTM International, West Conshohocken, PA, 2013,
https://doi.org/10.1520/D1143_D1143M
ASTM D1143-81. 1994. Standard Test Method for Piles Under Static Axial Compressive
Load. ASTM International, West Conshohocken, PA, 0, http://www.astm.org/cgi-
bin/resolver.cgi?UOP43
ASTM D 3689-90. 1995. Standard Test Method for Individual Piles Under Static Axial
Tensile Load. ASTM International, West Conshohocken, PA, DOI: 10.1520/D3689-
90R95
ASTM D 3966-90. 1995. Standard Test Method for Piles Under Lateral Loads. ASTM
International, West Conshohocken, PA, DOI: 10.1520/D3966-90R95
ASTM D3966/D3966M-07. 2013. Standard Test Methods for Deep Foundations Under
Lateral Load, ASTM International, West Conshohocken, PA,
https://doi.org/10.1520/D3966_D3966M
9
Brinch Hansen, J., 1963, Discussion, Hyperbolic Stress-Strain Response of Cohesive Soils,
Journal of Soil Mechanics, Foundation Division ASCE, V. 89 No.SM4, 241-242.
Butler, H.D. & Hoy, H.E. 1977. User’s Manual for the Texas Quick Load Method for
Foundation Load Testing, FHWA, Office of Development, Washington, D.C., FHWA-
IP-77-8, 59.
Chin, F.K. 1970. Estimation of Pile Not Carried to Failure, Proceedings 2nd Southeast Asian
Conference on Soil Engineering, Singapore, 81-90.
De Beer, E.E. & Wallays, M. 1989. Franki Piles with Overexpanded Bases, La Technique des
Travaux, No:333, 48.
Décourt, L. 1999. Behaviour of Foundations Under Working Load Conditions, Proc. of the
11th Pan-American Conf. on Soil Mechanics and Geotechnical Engineering, Dolguassu,
Brasil, V. 4, 453-488.
Décourt, L. 2008. Loading Tests: Interpretation and Prediction of their Results. Congress on
From Research To Practıce In Geotechnıcal Engıneerıng, 452-470.
Dinc, H. 2010. Determination of pile bearing capacity with Osteberg cells. MS thesis, İTÜ,
İstanbul.
Fuller, F.M. & Hoy, H.E. 1970. Pile Load Tests Including Quick Load Test Method
Conventional Methods and Interpretations, Highway Research, Record No.333, pp. 78-
89, Transportation Research Board, USA.
Lastiasih, Y. & Sidi, I.D. 2014. Reliability of Estimation Pile Load Capacity Methods. J. Eng.
Technol. Sci., Vol. 46, No. 1, 1-16.
Mazurkiewicz, B.K. 1972. Test Loading of Piles According to Pollish Regulations, Royal
Sweedish Academy of Engineering Sciences, Commission on Pile Research, Report
No.35, pp20, Stocholm.
Petek, K.A., Mitchell, R.A., Buechel, G.J. & Goodyear, D. 2012. Full Scale Instrumented Pile
Load Test for the Port Mann Bridge, Surrey, British Columbia, Canada. Full-Scale
Testing & Foundation Design, 362-375.
U.S. Army Corps of Engineers. 1991. Design of Pile Foundations. Engineering Manual 1110-
2-2906, Washington, DC 20314-1000.

10
 Performative Approaches in Tall Buildings: Pearl River Tower

S. Arslan Selçuk1, H.E. Ilgın2

1
Department of Architecture, Gazi University, TURKEY.
2
Department of Architecture, METU, TURKEY.
(Corresponding Author’s E-mail: hemreilgin@gmail.com)

ABSTRACT

21st century has become prominent with two main concepts in architecture; the first one is
“sustainability in architecture” which has been seeking for a less environmental footprint
in the ecosystem and the second one is digital technologies that drive a novel approach in
all kinds of man-made products including architecture. Potentials of digital solutions
(CAD/CAM/BIM) in the design of "energy efficient" buildings can be considered as
innovative way of thinking/designing/simulating/optimizing for the actors working in all
levels of discipline of architecture. In other words, integration of these tools with the
design process offers designers many alternatives through the analysis of both climatic
and physical data usage during the optimization process of architectural forms and
structures, especially, designing unique/extreme buildings like, tall buildings, wide span
structures and buildings having complex forms. From this context, this paper discusses
and exemplifies the term "performative architecture" as a melting point of the concepts
mentioned above. It is aimed to present and discuss how cutting edge technologies help
designers to design not only the building but also the “design of design process” in a
sustainable way. For this purpose, tall buildings are examined within the light of
performative approach. After presenting the theoretical framework of the concept of
performative design; Pearl River Tower, (Guangzhou, China) an energy efficient tall
building, in which digital design and optimization tools are integrated with the design
process, is exemplified and the criteria of performance based design used in the case study
is discussed.

Keywords: Energy efficiency, Pearl River Tower, performative architecture, tall

buildings.

TALL BUILDINGS

“Tall building”, “high-rise building”, and “skyscraper” are difficult to define and distinguish
solely from a dimensional perspective because height is a relative matter that changes
according to time and place. The term “high-rise building” has been recognized as a building
type since the late 19th century, while the history of the term “tall building” is very much older
than that of the term “high-rise building”. As for the use of the term “skyscraper” for some
tall/high-rise buildings reflecting social amazement and exaggeration, it first began in
connection with the 12-storey Home Insurance Building, built in Chicago towards the end of
the 19th century (Harbert, 2002, Peet, 2011)

11
Definition

There is no general consensus on the height or number of storeys above which buildings
should be classified as tall buildings or skyscrapers. The architectural/structural height of a
building is measured from the open-air pedestrian entrance to the top of the building, ignoring
antennae and flagpoles (Gunel and Ilgın, 2014a) According to the CTBUH (Council on Tall
Buildings and Urban Habitat), buildings of 14 storeys or 50 metres’ height and above could
be considered as “tall buildings”; buildings of 300 metres’ and 600 metres’ height and above,
are classified as “supertall buildings” and “megatall buildings” respectively. According to the
Emporis Standards, buildings of 12 storeys or 35 metres’ height and above, and multi-storey
buildings of more than 100 metres’ height, are classified as “high-rise buildings” and
“skyscrapers” respectively (Emporis Data Standards ESN 18727, ESN 24419).

Tall buildings are defined: by structural designers as buildings that require an unusual
structural system and where wind loads are prominent in analysis and design; by architectural
designers as buildings requiring interdisciplinary work in particular with structural designers,
and with experts in the fields of aerodynamics, mechanics and urban planning that affect
design and use; and by civil engineers as buildings needing unusual and sophisticated
construction techniques. César Pelli (1982) defines a skyscraper as a supertall building and
highlights the word “super” within this definition as changing according to time and place.
Structures such as The Eiffel Tower (Paris, 1889) cannot be classified as skyscrapers because
of the lack of a habitable interior space.

In the view of the authors of this paper, “tall building or high-rise building” is a local concept
and “skyscraper or supertall building” is a global concept. To be able to define a tall building
as a skyscraper or supertall building, it is not sufficient for it only to be tall in its own region;
it is necessary for it to be recognized in the world as a skyscraper or supertall building (Gunel
and Ilgın, 2014b ).

Emergence and Historical Development

Architects contribute to the social and economic changes of the age, reflecting the
environment they live in with their designs and creating a development/evolution by
developing new styles or building types. In addition, underlying the first appearances of
skyscrapers in Chicago was a social transformation triggered by the economic boom of that
era and by the increase in value of urban building plots. The advance in construction
technology has played a much more important role in the development of tall buildings than
in the case of other types of structure. At the end of the 19th century, beginning with the
discovery of the elevator for the vertical transportation system, and structural metal beam-
column framing system, the construction of tall buildings commenced as an American
building type owing to innovations and developments in new structural systems, high-strength
concrete, foundation systems, and mechanical systems; this continues to drive the race for
height in skyscrapers that is spreading across the world.

The Home Insurance Building (Chicago, 1885) (Figure 1), designed by engineer William Le
Baron Jenney with 12 stories, is recognized as being the first skyscraper. The use of a
structural frame in the building won it the title of the first skyscraper, marking a new epoch in
the construction of tall buildings, and it became a model for later tall building designs. When
800m was passed at the beginning of the 2000s, heights have been reached that could not have
even been dreamed of in engineer William Le Baron Jenney’s time. In other words, while 10-
storey buildings were classified as skyscrapers in the 1890s, about 40 years later the Empire
State Building (New York, 1931) (Figure 2) exceeded 100 stories, and about 100 years later
the Burj Khalifa (Dubai, 2010) (Figure 3) exceeded 150 stories.
12
Figure 1. Home Insurance Building
Figure 2. Empire State Building (photo courtesy of Antony Wood)
Figure 3. Burj Khalifa (photo courtesy of Adrian Peret, adrian.peret@gmail.com)

PERFORMATIVE ARCHITECTURE

Today, the use of digital technologies in architecture has been shifted from being a tool of
representation to media of design, optimization and manufacturing. There is no doubt that,
CAD/CAM/BIM applications have been rapidly changing the conventional architectural
design and construction process since the end of last century. In this process researchers and
practitioners have been seeking for new tectonics, materials, systems producing its own
energy, more efficient, more comfortable etc. revealing the beauty of using cutting edge
technology in a sustainable viewpoint.

From this respect, thanks to these technologies, some parameters like being recyclable, being
convertible, security, interaction and circulatory systems have been taken into consideration
and more holistic approaches have been used under the umbrella of sustainability. It possible
to argue that the increasing number of design parameters can only be turned out into an
economical, aesthetical and performance based/optimized “design” by the help of
computational technologies. In this context, recent studies reveal that these two main concepts
can be melted in a pot and called performative architecture. It is possible to claim that
performative design in architecture represents a combination of two critical characteristics of
computational design. First, generation of a solid model that enables all kind of analytical
evaluation of environmental performance based upon simulating physical conditions such as
solar or structural loadings etc. Second, helping designers to create "architecture” performing
as an art, with the surroundings, acting as the stage on that the building can perform and be on
show via computational model that enables from file to fabrication.

Followingly, in the discipline of architecture “parametric design and performative

architecture" has been discussed and published by Kolarevic and Malkawi to describe the idea
of performance as a guiding principle in architectural design (Kolarevic, 2003, Kolarevic and
Malkawi, 2005). Menges and Hensel (2008) highlight that being performative is usually
associated with sustainability and complex digital models analyzing the structural and
environmental behavior of buildings. This limits performance to a merely technical

13
interpretation. Hagan (2008) argues that performative architecture must also consider other
aspects, because architecture has always performed socially, semantically, ideologically, and
in a basic manner as a shelter. Therefore, the question “what is architectural performance in
the digital age” gains importance: "Is this performance comparable to the performance of a
machine or a theatrical performance?"

Oxman (2009) argues that, there is no single answer for this question because of the
multiplicity of the meanings and connotations of the word performance have. Albayrak (2011)
concludes that, determining different performative aspects in a particular project and
reconciling conflicting performance goals in a creative and effective way are key challenges
in performative architecture. Accordingly, performative architecture can be considered as a
paradigm that defines complex and ill-defined design problems, identifies constraints/criteria
and goals of design, evolves the designing process to an optimizing process. This definition
reflects the “architectural design process for sustainability” accurately. More lightweight
structural system, less energy, more day lighting, more natural material, interactive building
systems... all these “more” constitutes the themes of performative architecture and provides
them being realized through digital designing and manufacturing techniques. Even, as
transforming to a meta discipliner identity, architecture is only possible to construct
smart/intelligent buildings via smartness/intelligence of the design process. Therefore, it is
possible to claim that architecture is already an action for sustainability.

PERFORMATIVE TALL BUILDINGS: PEARL RIVER TOWER

China has very little in a way of oil and gas reserves and coal represents the solution to
providing 80% of electricity demand. The countries’ developing economy has increased the
energy consumption and resulted with rapid growth in carbon emissions (Tomlinson et.al
2014). Therefore, Chinese government set a goal of reducing carbon emissions by 10% by
2010 and Guangzhou was the focus of this policy. Furthermore, China sets 15% non-fossil
energy consumption target by 2020. According to Frechette and Gillchrist (2008)
performative approaches in the design of built environment has become a necessity in this
way.

The Pearl River Tower designed by Skidmore, Owings & Merrill LLP (SOM) (completed in
2013) is one of many sustainable or green tall buildings that have recently developed (Figure
4). The project’s site area: 10,635m2 project area: 214,000m2 number of stories: 71 building
height: 309m) Karlatornet Tower in Sweden, Pertamina Energy Tower in Jakarta, Al-Bahar
Tower in Abu Dhabi, Agile Corporation Headquarters Tower in Guangzhou, Shanghai Tower,
Guohua Financial Tower and Nanjing Keyne Centre in China can be mentioned as
performative tall buildings recently completed with the help of computational design
optimization and manufacturing tools and technologies.

14
Figure 4. Images from The Pearl River Tower [URL1] [URL2]

Performative Quests in the Design Process of Pearl River Tower

Selected through a competition in 2005, Pearl River Tower with its 214,000m2 usage area and
309m building height (71-story) is an important case to demonstrate what is possible in
performative design by integrating the latest green technologies and engineering
advancements. The tower's sculpted form directs wind to a pair of openings at its mechanical
floors, where passing winds trigger turbines that generate energy for the building. Other
integrated sustainable elements including double-skin curtain wall, a chilled ceiling system,
solar panels, under-floor ventilation, and daylight harvesting, all of which contribute to the
building’s energy efficiency. As completed the building uses approximately 30% less energy
than would be used by a similar structure built to China stringent energy codes (Tomlinson
et.al 2014). These quests are examined in detail starting with the “performative structural
design” in the following sections.

Performative Structural Design

The lateral load resisting structural system of Pearl River Tower utilizes both reinforced
concrete and structural steel to resist gravity and lateral loads, seismic and wind loads. The
Tower’s structural system is outriggered frame system 1 which is composed of reinforced
concrete core (supercore), steel perimeter columns and composite corner mega columns,
outrigger and belt trusses and steel end bracing. The rectangle-shaped central core, namely
“supercore”, consists of reinforced concrete shear walls with thickness varying between about
150cm at the bottom and 70cm at the top through height of the building (Tomlinson et.al 2014,
Smith, 2012, Gonchar, 2015) Closed form of shear wall provides torsional stiffness.

The “supercore” shear wall system, which is linked to the steel perimeter columns by a
system of outrigger and belt trusses, at two mechanical equipment floors at levels 23-27 and
levels 49-53, and composite corner mega columns linked by steel end bracing. Figure 5

1
An “outrigger” consists of a horizontal truss or shear wall. It is a lateral extension of the core shear truss / shear
wall to the perimeter columns in the form of a knee. An outriggered frame system is formed by the addition of an
outrigger to a shear-frame system with a structural core. To make them effective, outriggers are at least one
storey deep, and have a high flexural and shear rigidity. They are generally located at the mechanical equipment
floors in order to not to hinder the use of normal floors. At the levels of the outriggers, connecting the perimeter
columns to each other with belts consisting of a horizontal truss or shear wall. By means of the belt, the column
connected to the core by the outrigger, distributes the axial load effect of the outrigger to other columns.
15
represents how this structural elements work together to achieve structural performance in a
wise and economical approach (Tomlinson et.al 2014, Smith, 2012, Gonchar, 2015

Figure 5. Structural system of the tower 2

More Green Energy

Integration of wind power into the design of the building is one of the tower’s distinct
features. The towers location is shifted from the city’s predominant grid to help capture the
wind for energy generation. In the building design it can be easily seen that the east and the
west elevations are straight while the south façade is concave and the north façade is convex.
The building has four large openings in the concave face, where mechanical floors and four -
2m wide 5m tall- wind turbines are placed (Figure 6). As Epstein (2008) highlights wind
speed in Guangzhou is usually a mild 9 mph at that elevation, but the air is literally pulled
through the holes in the envelope by the negative pressure on the protected south side,
accelerating winds to speeds of about 18 mph (Epstein, 2008). (Figure 7). The turbines are
estimated to provide 1% of the building’s energy needs (Sharpe, 2010) It can be concluded
that together with the design of building mass and its facade the geometry significantly
enhances turbine performance.

2
Drawings for structural system of Pearl River Tower are made by Havva Nur Tümbaş in the MSc
course of BS 536: Studies on Tall Buildings: Design Considerations, spring 2014-2015, Department
of Architecture, METU
16
Figure 6. Principles of building design in terms of wind energy use (Choroba, Siemen,
2012.)

Figure 7. Wind tribunes[URL3] [URL4]

More Effective Building Skin

The tower's smooth, aerodynamic form was developed through analysis of solar and wind
patterns of the city. The achieved design optimizes the solar path and utilizes the sun to the
building's advantage. Similarly, façade design is performed according to energy saving
principles. For example, the façade has an insulated exterior layer and an inside layer - with
air space sandwiched in between the two layers (Epstein, 2008)

Figure 7. The principles of double façade design and PV panels on sunshades[URL5]

Furthermore, a photovoltaic system is integrated into Pearl River Tower's external solar
shading system and glass outer skin. Motorized sunshade devices are provided within cavity
for solar shading and glare control, and are controlled automatically in response to photocells
that track the sun position relative to the elevation. Building Management Systems controls
the angle of the sunshades automatically in response to solar intensity, solar altitude angle and
solar azimuth angle to minimize solar heat (Tomlinson et.al 2014 ). Similarly, a low energy

17
high-efficient lighting system uses radiant panel geometry to assist in the distribution of light.
The radiant cooling system provides improved human thermal comfort, efficient heat
exchange and improved office acoustics. Additionally, daylight harvesting uses daylight
responsive controls integrated with the automated blinds. These measures have yielded a large
decrease in the building's MEP and lighting electrical consumption (45% reduction) as
compared to the National Chinese Energy Code [URL6]

More Natural Ventilation

Pearl River Tower integrates innovative technologies to significantly reduce the amount of
energy required to operate the building, and to promote the highest levels of human comfort
and indoor air quality. Of all the elements in Tower’s energy-efficient design, the "radiant
cooling" system is the biggest energy saver. In this system, water flows through ceiling panels
and provides cooling from there. Cooling is delivered by water instead of air (and no heating
is necessary in the tropical climate), the system does not need to move air to cool the building.
The ventilation system is delivered via a raised access floor, providing improved indoor air
quality and air change effectiveness. There is also a reduced cost of tenant fit-out and future
retrofits due to the absence of fan coils, VAV boxes, filters, ductwork, insulation, and other
items typically requiring tenant-specific alterations.

Figure 8. Typical section of tower showing "radiant cooling" system

“We have a building that takes outside air, filters it, puts it into the building and then vents it
from the building,” said Frechette. “Conventional buildings use recirculation to cool or heat
air because the air has already been cooled, and it saves energy. We aren’t using air for
cooling, so the small quantity of air we bring in from the outside, for breathing, is not that
expensive or difficult to cool.” (Frechette and Gillchrist, 2008). To conclude, the facade of the
tower hide an eight-inch air gap inside them. When the sun beats through the double-glazed
outer skin, the heat -instead of entering the interior- is trapped in this cavity and then rises to
heat exchangers on the equipment floors. This, with other features, lets the building use an
AC system that is 80% smaller than those in conventional skyscrapers [URL6].
 

18
CONCLUSIONS
 
Tall buildings are creating the architectural language/identity of cities in this century. They
are, as symbol of power and prestigious, the centre of business and economics, hosting
thousands of people at the same time. Hence, they are considered as a great consumer of
energy which utilized huge amount of resources and materials; produce massive volumes of
waste discharge into the environment and more often called as unsustainable buildings.
However, the architecture of 21st century is changing. Thanks to the computational design,
optimization, manufacturing and operating systems, architecture is not only performing as an
"art" but also enables all kind of analytical evaluation and optimization of environmental
performance.

The Pearl River, through its 309-meter high sculpted faces, is designed to be one of the
world’s highest performing tall buildings. It weaves together highly innovative, yet proven,
technologies that work together to significantly reduce the amount of energy required to
operate the building. The design of the tower is intended to minimise harm to the environment
and it will extract energy from the natural and passive forces surrounding the building. Major
accomplishments are the technological integration of form and function in a holistic approach
to engineering and architectural design. Its performative features include; performative
façade, wind turbines, integrated photovoltaics (PVs), radiant cooling coupled with underfloor
ait ventilation, day-light responsive control, daylight reflectors, efficient lighting and chiller
system. It is important to highlight that to create a performative building it is critical to
assemble all of the primary stakeholders -the design team, consultants, contractors, city
officials, local utilities, as well as the owner of the project- at the very beginning of the design
process. High performance can sometimes mean “highly complicated” yet BIM technologies
provide a unique environment for them to share all detailed information from the very
beginning of design process to the end of the construction. The last but not the least the
completed structure still consume energy, but 58% less than a traditional building of similar
size.

REFERENCES

L. Harbert, 2002. Home Insurance Building-The First Skyscraper, Journal of American

Society of Civil Engineers (ASCE), Vol. 43, No. 2, p. 1-2,
G. Peet, 2011.The Origin of the Skyscraper, CTBUH Journal, Issue I, 18-23.
M.H. Gunel and H.E. Ilgın, 2014a. Tall Buildings: Structural System and Aerodynamic Form,
Routledge.
M.H. Gunel and H.E. Ilgın, 2014b.Yüksek Binalar: Taşıyıcı Sistem & Aerodinamik Form,
Middle East Technical University, Publication of Faculty of Architecture, Ankara.

B. Kolarevic, 2003. Architecture in the Digital Age: Design and Manufacturing, Spon Press,
New York, NY.
B. Kolarevic and A.M. Malkawi, 2005. Performative Architecture: Beyond Instrumentality,
Spon Press.
A. Menges and Hensel, M. 2008. Inclusive Performance: Efficiency Versus Effectiveness
Towards a Morpho-Ecological Approach for Design, Architectural Design. 78(2), 54-
63.
S. Hagan, 200. Digitalia- Architecture and the Digital, the Environmental and the Avant-
Garde, Routledge.
R. Oxman, 2009. Performative design: a performance-based model of digital architectural
design, Environment and Planning B: Planning and Design, 36, pp.1026-1037.

19
C. Albayrak, 2011. Performative architecture as a guide line for transformation of the defence
line of Amsterdam, Doctoral dissertation, TU Delft, Delft University of Technology.
R. Tomlinson, W. Baker, L. Leung, S. Chien, and Y. Zhu, 2014. Case Study: Pearl River
Tower, Guangzhou, China, CTBUH Journal, 14(2).
R. Frechette and R. Gillchrist, 2008. Case Study: Pearl River Tower, Guangzhou, China.
CTBUH 2008 8th World Congress, Dubai.
Smith, A., From Jin Mao to Kingdom: Searching for an Asian Supertall Vernacular, CTBUH
2012 9th World Congress, Shanghai, 2012.
Gonchar, J. (2015, June 2). Towering Ambition: The architects and engineers behind an office
building in rapidly expanding Guangzhou put super green before supertall. Retrieved
from http://archrecord.construction.com/projects/portfolio/2014/03/1403-Pearl-River-
Tower-Skidmore-Owings-and-Merrill.asp
K. Epstein, 2008. How Far Can You Go? Case Study: Pearl River Tower, High Performing
Buildings http://www.hpbmagazine.org/attachments/article/12002/08W-Pearl-River-
Tower-Guangzhou-China.pdf
T. Sharpe, 2010. The Role of Aesthetics, Visual and Physical Integration in Building Mounted
Wind Turbines-an Alternative Approach. INTECH Open Access Publisher.
Z. Choroba, P. Siemen, 2012. Pear River Tower http://www.adip.tu-berlin.de/wp-
content/uploads/2011/11/PearlRiverTower.pdf
URL1 http://www.som.com/FILE/19407/pearlriver_1575x900_timgriffth_08.jpg
URL2 http://megaconstrucciones.net/images/rascacielos/foto/pearl-river-tower-3.jpg
URL3 http://www.circlelaw.net/images/SOM3.jpg
URL4 http://www.som.com/FILE/17902/pearlriver_sustainable_700x800_som_07.jpg
URL5 http://www.som.com/FILE/21822/pearlriver_diagram_830x630_lsi-ye_zhangrsom.jpg
URL6 http://archive.fortune.com/2010/02/22/technology/zeb_pearl_river.fortune/index.htm

20
 Reducing Sediment Deposition in a Clarification Tank Using
Numerical Modeling
C. Yilmazer*1, A. O. Celik2, V. Kiricci3
1
Department of Civil Engineering, Anadolu University, 26555, Eskisehir, Turkey.
(Corresponding Author’s E-mail: cemyilmazer@anadolu.edu.tr)
2
Department of Civil Engineering, Anadolu University, 26555, Eskisehir, Turkey.
3
Department of Civil Engineering, Anadolu University, 26555, Eskisehir, Turkey.

ABSTRACT

This study aims to resolve a hydraulic engineering problem using Computational Fluid
Dynamics (CFD) method. The problem is related to a sedimentation basin in a water
treatment plant of ESKI (Eskisehir Water and Sewerage Administration) which is one of the
municipal facilities in Turkey. Accumulation of sediment in distribution channels for the
clarification tanks were reported to be generating operation-wise problems. The plant is
entirely gravity driven and the flow conditions at various channels are difficult to control. The
manual cleaning process of deposited sediment is required periodically during operation due
to accumulated sediment in the distribution channels. This work puts an effort for detecting
the problem and stopping the sediment deposition purely by simple geometrical improvement
using numerical modeling. In this study, the main goal is to offer a reasonable solution based
on basic hydraulic principles.

Sediment accumulation (as a result of deposition) is characterized by low stream velocity and
also low turbulence kinetic energy. Based on the most recent suspended sediment theory, the
developed local flow conditions under which the suspended particles start gravitating was
identified and used as a criterion for controlling the flow conditions. The objective then was
to hydraulically redesign the feeding channels for the clarification tank in an effort to increase
the stream velocity and stop the early occurring sediment deposition. Low-velocity flow
regions (at the downstream of the channel) were identified using CFD method. Also, initial
conditions (water height) were identified and the CFD model was validated by a 1:10 scale
physical model of the clarification tank. Consequently, by altering the geometry of the
channel, these low power regions were activated in terms of suspension of sediment using
contractions in the channel. The results are believed to be leading a low cost but effective
solution to the problem which eliminates manual intervention during the treatment plant’s
operation.

Keywords: Clarification Tank, CFD, physical modelling, sedimentation, water treatment.

INTRODUCTION

This study aims to identify and resolve a hydraulic engineering problem with CFD method.
The problem is about a sedimentation basin in a physical water treatment plant of ESKI
(Eskisehir Water and Sewerage Administration) in Turkey.

The physical water treatment process has two integral parts; settling basin and clarification

21
basin. The clarification tanks are widely used for water treatment plants. They are based on a
physical process which suspends solid particles and not precipitate them as in a settling basin.
Settling basins are ponds constructed for removing undesirable entrained solids by
sedimentation. Clarifiers are tanks built with mechanical means for continuous removal of
finer size solids being suspended by the water as shown in the upper left corner in Fig. 1. In
this treatment plant, the problem is identified as the accumulation of sediment in distribution
channels seen at the Fig. 1. The clarification tanks were generating operation-wise problems
in the plant as the facility is entirely gravity driven and the flow conditions at various
channels are difficult to control. Costly processes have been employed to remove the
accumulated particles manually.

Figure 1. The clarification tank.

As a first step, the present system was examined to detect the source of the problem. In this
stage, CFD method was performed for the clarification tank in the water treatment plant.
ANSYS v.14 CFX was used as a solver. Initially, a computer aided drawing (CAD) model
was generated representing the actual size of the clarification tank. Then, meshing process
was applied to the flow domain for which a considerable time has been devoted in an effort to
generate an accurate mesh in the entire domain. Following the meshing, proper boundary
conditions were applied.

In this study, the main problem is low-velocity values at the downstream of the clarification
tanks distribution channel. Low-velocity profiles cause sediment particles to settle. To solve
this problem, velocity, as well as turbulence level, needs to be increased. In this model, the
particle size is accepted as spherical with 0.1 mm diameter. The weight of solid particles is
balanced by lift forces also directly related to stream velocity and particles settling velocity.
Lift force must be equal to or greater than the weight of the solid particles which is supposed
to be clarified completely from distribution channel to the clarifier. Lift force will be
elaborated below.

To give background information, relevant studies in the literature will be reviewed here.
Larsen (1977) is one of the researchers who employed a CFD model in studies involving
clarification tanks [11]. Larsen (1977) identified the phenomenon which is also known as
“density waterfall” that causes the fluid to fall down at bottom of the clarification tanks.

22
Shamber and Larock (1981) [13] used Finite Volume Method (VOF) with Navier-Stokes
equations and they also used k-ɛ turbulence closure with the addition of solid concentration
equation using a settling velocity. McCorquodale et al. (1991) [12] built a model using Finite
Difference Methods (FDM) for the boundaries and Finite Element Methods (FEM) for flow
functions. Krebs et al. (1995) [10] also investigated different inlet arrangements and assessed
them as inlet baffle locations by Phoenics code. Deininger et al. (1998) [5] created the
software, Champion3D, which is a numerical flow solver that could obtain velocity and solids
distribution of a circular secondary clarifier system. Imam et al. (1983) [7] were able to
determine vertical velocity profiles and an accurate model of vorticity transport flow
functions with a constant eddy viscosity in a numerical simulation. Goula et al. (2008)
investigated sedimentation tanks successfully using CFD method in potable water treatment
[9]. A. Gkesouli et al. (2013) [6] formulated a validated CFD model to observe flow fields at
tank inlet. They identified recirculation zones where the concentration of solid particles is
uniformly distributed and concluded that as the flow rate increases, the concentrated regions
shift to the outlet which causes a reduction in the efficiency of the tank.

CFD is a useful tool in design or rehabilitation processes for physical water treatment plant
facilities. For a typical distribution channel, upstream has a higher velocity than downstream.
Therefore, the improvement on the channel geometry will be performed for the only
downstream. For fast, reliable and fluid mechanics based decisions on the modifications in the
geometry, CFD method is employed in this study.

The clarification tank is connected with seven outlet pipes at the bottom. A plane was fixed to
obtain area averaged velocity (stream-wise direction) in z-direction near these outlets. The
plane allows determining average local velocities. These velocity values are compared against
a critical velocity value associated with lift forces sufficient to keep the particles suspended.
This approach can be used for various geometry changes until a satisfactory result is obtained.

MATERIALS AND METHODS

CFD method consists of four steps. Firstly, known geometry is drawn on CAD model by
considering interacting solid parts with fluids. Secondly, the numerical network was created
by separating the fluid domain into the small cells. The other step is to define the initial and
boundary conditions. Finally, the numerical model is run by means of a solver. These steps
will be elaborated below.

Geometry of The Clarification Tank

The clarification tank is a rectangular box with a 66 cm width (-x direction), 112 cm height (-
y direction) and 1150 cm depth (z-direction) as shown in Fig. 2. The channel has seven
identical outlet pipes with 10 cm diameter at the bottom of clarification tank, each 100 cm
apart. The first outlet pipe is located 100 cm away from the inlet. Top of the clarification tank
is open to the free entrance and exit of the air to the system.

23
 INLET

Figure 2. The CAD model of the clarification tank.

Meshing Process

One of the most critical points in CFD method is to obtain a proper mesh. The shape of the
cells directly affects the accuracy of the solution. The mesh quality which is the key factor to
reach to desired aim accurately as the important parameter. Related to this term, there are
useful benchmark works such as the one performed by W.G.Habashi (2000) [8] regarding
how the nodes in the mesh affect the result.

There are several factors that affect the mesh quality. “Skewness” is the major factor among
the others. This study uses the Skewness parameter assessing the mesh quality. It is a rule of
thumb in the literature now that the maximum Skewness shouldn’t exceed 0.85 and also
average value shouldn’t exceed 0.25. The model meshed as shown in Fig.3 has 0.824
maximum and 0.204 average in Skewness respectively. In this study, the total number of
elements and nodes are 4383872 and 1350712 respectively.

Figure 3. The view of the numerical grid (mesh) of the clarification tank.

24
Boundary Conditions

In the present work, as an initial condition, water height and operating discharge were
identified by experiments on a 1:10 scale physical model of the clarification tank as shown in
Fig. 4. With this approach, CFD model was also validated in terms of gross flow parameters.
“Froude Model” assumes that the acceleration due to gravity is identical in both the model
and the prototype [4]. 0.1, 0.2, 0.3, 0.4, 0.5, 0.6 lt/s of discharges were tried in the physical
model. At 0.5 lt/s flow depths near outlet pipes became stable at 85 mm. At 0.6 lt/s of
discharge, the model overflows. After the experiment, results were transformed to prototype
values by appropriate scale factors which are listed as 157.92 lt/s of discharge and 85 cm of
flow depth to be used in the CFD model.

Figure 4. The 1:10 scale physical model of the clarification tank.

Top of the clarification tank is open to the atmosphere (1 atm pressure). Namely an “opening”
type boundary is considered here. For observing the free surface, surface tension model is
necessary to define fluid pair (water/air) interaction. In this numerical model, surface tension
model was selected as continuum surface model. The continuum method eliminates the need
for interface reconstruction, simplifies the calculation of surface tension, enables accurate
modeling of two- and three- dimensional fluid flows driven by surface forces and does not
impose any modeling restrictions on the number, complexity, or dynamic evolution of fluid
interfaces having surface tension [3]. The surface tension coefficient is defined as 0.072 N/m
to represent the interaction between water and air.

K-epsilon turbulence model was used as a closure in this study. “k” is the turbulence kinetic
energy which means the variance of the fluctuations in velocity and epsilon (ԑ) is the
turbulence eddy dissipation which is defined as the rate at which the velocity fluctuations
dissipate. As described by J.E Bardina et al. (1997) [1] this closure is generally useful for
free-shear layer streams which have small pressure gradients. Similarly, for wall-bounded and
internal streams, if mean pressure gradients are small, k-epsilon gives reasonably well results.

To complete the boundary conditions on the model, outlet pipes were identified as an outlet
with 1 atm pressure. This is also necessary to prevent a redundant solution because the inlet is
defined as normal speed (value of which was obtained from the discharge and cross-sectional
area from the physical model). Rest of the flow domain was assigned as a wall with sand
grain roughness of 0.1 mm to reflect the concrete walls in the actual channel.

25
RESULTS AND DISCUSSION

At each outlet pipe in -xy plane, an intersecting plane was assigned to determine the water
velocity profile. As mentioned above, velocity values must be equal or higher than threshold
velocity values which ensure lift forces sufficient for sedimentation of particles as shown in
equation 1 below.

FLIFT = 1/ 2 ρW CL U 2 A (1)

In equation 1, “ρw” is the density of the water which equals to 1000 kg/m3 “A” is the
projected area of the solid particles and “CL” is the lift coefficient (assumed 0.35 in this
work). “U” is the superficial water velocity in z-direction acting on the projected area, A.
Weight of those particles opposing the lift force can be computed by equation 2 below. In
equation 2 below, “g” indicates the acceleration due to gravity, “r” is the radius of the
spherical particle. According to situ investigations, the particles were reported to be silt whose
density (ρs) equals to 1100 kg/m3 [2].

W = 4 / 3 π r 3 ρS g (2)

The approach was to model the channel as it is (Fig. 5 left image) and to verify the
problematic zones. That is, before the modifications, the zones near the exit holes were
investigated to determine flow velocities below the critical value. This way, the modification
regions would also be determined. Subsequently, the modification was implemented on the
geometry to increase the flow velocity in the zones where the deposition of sediment would
potentially occur.

INLET

Figure 5. a) Planes in the x-y direction at each outlet pipes for unmodified geometry, b)
Planes in the x-y direction at each outlet pipes for modified geometry.

Table 1 below summarizes the results for all 7 sections seen in Fig. 5 for both the original and
the modified channel geometries. It is not surprising to see that the low-velocity regions are
near planes 4-7. This is where the deposition actually occurs at the plant. In an effort to offer
an effective and affordable solution, geometry near planes 4-7 was modified. Contractions as
shown in Fig. 5 (right image) were applied at the channel walls to reduce the cross-sectional
area and manipulate the flow. Table 1 summarizes the modified channel results as well and

26
suggests even the most critical section, plane 7 has an average flow velocity above the critical.
The effect of the modification on the flow is also given in Fig. 6. The contraction and the
resulting increase in the flow velocity near the downstream of the channel is seen on the
velocity contour figure.

Table 1. Distances from the inlet and superficial velocity at each plane for unmodified and
modified geometry.

Superficial Superficial
Distance Velocity Velocity
from Z direction Z direction
The Inlet For For modified
unmodified model
model
(cm) (m/s) (m/s)
Plane 1 100 0.196 0.271
Plane 2 250 0.159 0.243
Plane 3 400 0.124 0.212
Plane 4 550 0.091 0.172
Plane 5 700 0.061 0.148
Plane 6 850 0.034 0.112
Plane 7 1000 0.011 0.102

Figure 6. a) Superficial Velocity in Z direction at the x-z direction for existing geometry, b)
Superficial Velocity in Z direction at the x-z direction for modified geometry.

27
CONCLUSIONS

This paper studied an engineering problem observed at a water treatment plant and offered a
sound solution using CFD method. The existing tank was observed to have very low
velocities near downstream end. A 1:10 scale physical model was set up to help determine the
model initial and steady conditions. Boundary conditions were also determined from the
experimental data and in situ investigation of the water treatment plant. The numerical model
was also validated using the flow depths and discharge values from the physical tests. For
each plane near outlet pipes, velocity profiles were investigated. Necessary geometry changes
were done and their effect was assessed using again the validated CFD model. The results
indicate that the suggested simple changes in the geometry create flow regions with higher
average velocities which prevent deposition in the tank. If implemented, the plant operation
will not be affected by the sediment deposition as the particles will remain suspended and
discharged into the settling basin.

ACKNOWLEDGEMENT

The authors would also like to thank Senay Subasi and ESKI (Eskisehir Water and Sewerage
Administration) for their assistance and guidance in acquiring the plant data.

REFERENCES

[1] Bardina, J.E., Huang P.G., Coakley T.J. 1997. Turbulence Modeling Validation, Testing,
And Development, NASA Technical Memorandum 110446.

[2] Bayazit M., Avci I. 2010. Akarsularda Akım ve Sediment Taşınımı (River Flow and
Sediment Transportation), Birsen, Istanbul, Turkey.

[3] Brackbill J.U., Kothe D.B. and Zemach C. 1992. A Continuum Method for Modeling
Surface Tension, Journal of Computational Physics, 100, 335-354. DOI: 10.1016/0021-
9991(92)90240-Y.

[4] Chanson H. 1999. The Hydraulics of Open Channel Flow by Hubert Chanson, Published
by Arnold, 338 Euston Road, London NW1 3BH, UK.

[5] Deininger A., Holthausen E., Wilderer P.A., 1998. Velocity and Solids Distribution In
Circular Secondary Clarifiers: Full-Scale Measurements And Numerical Modeling, Water
Research, 32, 2951–2958, DOI: 10.1016/S0043-1354(98)00072-4.

[6] Gkesouli A., Stamou A. I., Xanthaki M. and Georgiadis S. 2013. Validation of a CFD
Model in Rectangular Settling Tanks, Proceedings of the 13th International Conference
on Environmental Science and Technology, Athens, Greece.

[7] Goula A. M., Kostoglou M., Karapantsios T. D., Zouboulis A. I. 2008. A CFD
Methodology for The Design of Sedimentation Tanks in Potable Water Treatment Case
Study: The Influence of a Feed Flow Control Baffle, Chemical Engineering Journal, 140,
110–121, DOI: 10.1016/j.cej.2007.09.022.

28
[8] Habashia W. G., Dompierrea J., Bourgaulta Y., Ait-Ali-Yahiaa D., Fortinband M. and
Vallet M. 2000. Anisotropic Mesh Adaptation: Towards User-Independent, Mesh-
Independent and Solver-Independent CFD Part I: General Principles, International
Journal For Numerical Methods In Fluids, 32, 725-744, DOI: 10.1002/(SICI)1097-
0363(20000330)32:6<725::AID-FLD935>3.0.CO;2-4.

[9] Imam E., McCorquodale J.A., Bewtra J.K. 1983. Numerical Modeling of Sedimentation
Tanks, Journal of the Hydraulics Division, ASCE 109, 1740–1754,
DOI:10.1061/(ASCE)0733-9429(1983)109:12(1740).

[10] Krebs P., Vischer D., Gujer W. 1995. Inlet-Structure Design for Final Clarifiers, Journal
of Environmental Engineering, ASCE 121, 558–564, DOI: 10.1061/(ASCE)0733-
9372(1995)121:8(558).

[11] Larsen P. 1977. On The Hydraulics of Rectangular Settling Basins, Report No. 1001,
Department of Water Research Engineering, Lund Institute of Technology, Lund,
Sweden.

[12] McCorquodale J.A., Yuen E.M., Vitasovic Z., Samstag R. 1991. Numerical Simulation
of Unsteady Conditions in Clarifiers, Water Pollution Research Journal of Canada, 26,
201–222.

[13] Shamber D.R., Larock B.E. 1981. Numerical Analysis of Flow in Sedimentation Basins,
Journal of the Hydraulics Division, ASCE 107, 575–591.

29
 Structural Deformation Measurement Using Matlab Image Processing
Toolbox

Z.F. Alemdar1 and S. Bilici2

1,2
Yıldız Technical University, Department of Civil Engineering, Istanbul, TURKEY.
(E-mail: zalemdar@yildiz.edu.tr, snmbilici@gmail.com)

ABSTRACT

This paper describes the use of digital photogrammetry for measurement of deflections in
concrete beams during the bending test. This new method to measure displacements in the
experimental mechanics without applying digital sensors to the specimens is described.
The optical technique uses a camera, a computer with frame grabber board and image
analysis techniques within the MATLAB Image Processing Toolbox. The displacements
at the surface of the test specimens are obtained by the analysis of the movements of dots
painted on the specimen. The image analysis algorithm makes it possible to automatically
track the motions of the dots and compute their centroid coordinates in a sequence of
images. The aim of the study is to examine the surface deformations of the reinforced
concrete under uniaxial loading and to compare with the results obtained from traditional
instruments.

Keywords: Bending test, deflection measurement, digital image processing.

INTRODUCTION

Conventional deformation gauges produce recordings only for one or a few points and
they are very difficult to use when it comes to small specimens or a small area of a
specimen. In contrast to conventional instrumentation, photogrammetry is non-contact,
requires no manual reading of dials, yields three-dimensional measurements and provides
visual records of the testing. It is ideally suited for destructive testing since only a set of
inexpensive targets is lost or damaged rather than expensive LVDTs or dial gauges. The
new method to measure displacements in mechanical test pieces without applying sensors
to the specimens is comprised of a camera and digital image analysis techniques within
the MATLAB Image Processing Toolbox. Therefore, they are used to investigate relative
displacements by the use of distinct markers on the specimen.

A literature survey shows that digital image processing has been widely used in a range of
engineering topics in recent years. Austrell et al. (1995) investigated measuring
displacements and strains in mechanical test pieces without applying sensors to the
specimens. The image analysis algorithm made it possible to automatically track the
motions of the dots and computed their centroid coordinates in a sequence of images. A
method to interpolate displacements and compute surface strains have also been devised,
by the use of finite element shape functions. An automatic procedure for detection of the
dots on the first image would be time saving. The application of the dots on the test
specimen also can be made more effective.

Yue et al. (2003) focused that a digital image processing based finite element method for
the two-dimensional mechanical analysis of geomaterials and their materials had

30
inhomogeneities and microstructures. Digital image techniques were used to acquire the
inhomogeneous distributions of geo-materials including soils, rocks, asphalt concrete and
cement concrete in digital format. The numerical results show that this new digital image
process based finite element method can take into account the material inhomogeneities in
the geomechanical analysis which can have significant effects on the tensile stress
distribution along the loading axis of the Brazilian indirect tensile tests.

Whiteman et al. (2002) used digital photogrammetry for measurement of deflections in

concrete beams during destructive testing. Results were presented from several tests of
different types of beams. Some of advantages over contact methods including three-
dimensional measurement of deformation components, unrestricted measurement range
and immunity to nonlinear systematic errors were demonstrated. Comparison was made
of photogrammetric and linear variable differential deflection measurements.

Jauregui et al. (2002) studied vertical deflection measurement of bridges using digital
close-range terrestrial photogrammetry. First of all, the initial camber and dead load
deflection of 31.1 m (102 ft) prestressed concrete bridge girders were measured
photogrammetrically and compared with level rod and total station readings. Secondly,
the vertical deflection of a 14.9 m (49 ft) noncomposite steel girder bridge loaded with
two dump trucks was measured. Finally, Photogrammetric results are compared with
deflections estimated using elastic finite-element analysis, level rod readings, and
curvature-based deflection measurements.

Comak et al. (2011) investigated the potential for the utilization of image processing
techniques in the area of civil engineering, specifically in concrete technology. The
evaluation of the investigated parameters yielded that the utilization of image processing
techniques in construction technologies, specifically in studies regarding the
determination of the properties of concrete, has been progressively increasing.

Fırat Alemdar et al. (2011) studied the determination of the location of nonlinear response
in structural systems which were under different loading conditions. Data was collected
during the NEESR investigation of the seismic performance of four-span large-scale
bridge systems that details deformations in column hinging regions during response to
strong shaking events. A photogrammetry method was applied using a reference grid on
the top and bottom column surfaces to record and analyze deformations in the plastic
hinging regions. The surface deformations and rotations of a reinforced concrete bridge
column under dynamic loading has been examined and compared with the results
obtained from traditional instruments. The photogrammetry method performed very well
to track the lateral and vertical displacements at the points on the grid surface as well as
the deformed shape of the hinging regions, but the results of secondary calculations, such
as rotations of the column, had limited success.

MATERIALS AND METHODS

The experiments were done as a graduate thesis at construction laboratory of Yıldız

Technical University (Öztürk, 2016). The purpose of our study is that recorded video data
during the experiments are converted to images and processed by using MATLAB
program.

31
In general, the implementation of the DIC method comprises the following three
consecutive steps, namely (1) specimen and experimental preparations; (2) recording the
videos of the beam surface during loading; (3) processing the acquired images by
converting the recorded videos using a computer program to obtain the desired
displacement information. In this section, issues on specimen preparation and image
capture are introduced first. Then, the basic principles and concepts of DIC are described.

Geometry of beams

In this study; the polypropylene fiber reinforced concrete beams without stirrups are
subjected to the point load in the middle of the beams and the displacements have been
examined. In the experiment, so as to distribute 1%, 2% and 3% volumetric content of
polypropylene fiber in the concrete homogeneously, polypropylenes were added in the 3
different mixers and stirred for 10 minutes. Then, fiber reinforced concrete evacuated in
the molds. In the experimental specimens, the ratio of the shear distance to the effective
height of the beam was taken respectively 2.5, 3.5 and 4.5. The series of the beams are
called as B25, B35 and B45. Due to consisting of 1% and 2% volumetric contents of
polypropylene fiber, the beams are named as B35P10 and B35P20. The mechanical
behaviors of these beams were investigated (Öztürk, 2016; Arslan et al., 2017).

The cross section of the beams used in the experiment, boundary conditions and loading
conditions are given in Figure 1. Width of beam (bw) is taken 150mm, height of beam (h)
is taken 240mm. In the produced beams; tension reinforcements is used as 2Ф16 (Öztürk,
2016).

Figure 1. Beam sections, reinforcement settlements and locations of the linear variable
displacement transducers (LVDTs)

The point at which vertical displacements on the load cell are measured by PDT
(potentiometric displacement transducer) is defined as "1", and the point at which
displacements of in the middle of the beam is defined as "2". For B35 series, the defined
points “3” and “4” are 15 cm far away from the right and left side of the middle of the
beam, the defined points “5” and “6” are 30 cm far away from the the right and left side
of the middle of the beam as shown in Figure 1 (Öztürk, 2016).

Using Digital Image Processing

The video recordings were unaware that the image processing method could be used, for

32
this reason records were taken by a normal camera and the effect of daylight was not
prevented. These camera records have been converted to images by using VirtualDub-
1.10.4 program. Front surfaces of the beams were divided into grids before the
experiment (Figure 2). By means of the image processing method, displacements are
captured from these grids which are closest to the midpoint of the beam, and the measured
values using the LVDTs at the middle point are compared.

Figure 2. Front surface of the B35P10 beam

RESULTS AND DISCUSSION

While the image processing program was running, the colors in the grid were getting
lighter as displacements increase. Therefore, these color differences caused deviations
from the actual displacement due to the poor quality images. First of all, displacements of
the beam called as B35P10 during the test was calculated and compared with the result of
the LVDT numbered as 2. There is a different trend between the plots (Figure 4). The
maximum measured value by the LVDT was 7.54 mm, the captured value using Digital
Image Correlation was 6.69 mm. The error rate is equal to 11 %. The reason is considered
that the video recording was in poor quality.

33
Figure 3. Reference point of recorded displacement of the B35P10 beam

Figure 4. Comparison of the results of LVDT and DIC methods

Secondly, displacements of the B35P20 beam were calculated and compared to the result
of LVDT. Almost the same trends and nearly the same maximum displacements were
obtained as shown in Figure 6. The measured value of LVDT was 8.17 mm, the obtained
value via Digital Image Correlation was 8.16 mm. The error rate is around 0.1 %. It can
be seen that as the color difference decreases and the quality of the picture increases, the
error rate decreases and the method gives a closer result comparing to the measured
values.

34
Figure 5: Reference point of recorded displacement of the beam B35P20

Figure 6. Comparison of the results of LVDT and DIC methods

Digital image correlation method generally suffers some disadvantages: (1) surface of the
tested beam section may have a random gray intensity distribution; (2) the measurements
depend heavily on the quality of the imaging system; (3) During recording video of the
specimen surface, various noises (e.g. shot noise, thermal noise, cut off noise) are
unavoidably presented in the digital images (Pan et al., 2009).

Compared with other interferometric techniques for deformation measurement, one

significant advantage of the DIC is that it has fewer requirements in experimental
environment, and can easily be implemented with a simple experimental setup. However,
this does not mean that the measurement accuracy of DIC is not or less affected by the
measuring system.

35
CONCLUSIONS

A simple photogrammetry analysis method was applied to evaluate the deformations of

several beam specimens. The experiments were done as a post graduate thesis at
construction laboratory of Yıldız Technical University and the video recordings were
unaware that the image processing method could be used. Natural light prevented us from
getting homogenous images of the beam specimens and video quality was also
insufficient for the using image processing method. Such external factors have prevented
us from getting close results to the image processing method. However, a near-realistic
trend was obtained from the second sample thanks to good image quality and the error
rate is relatively low compared to the other sample. In this study, only two beam
specimens were used and other samples can also be used to improve this method. In
addition, increasing the resolution of the camera and improving the capacity of the image
processing machine gives better precision. Though having limitations, this method is cost-
effective and accurate. Besides, it is of great importance when it comes to measurement of
distances and deflections in inaccessible areas.

REFERENCES

Arslan G., Keskin R.S.O. & Öztürk M. 2017 Shear Behavior of Polypropylene Fiber-
Reinforced Concrete Beams without Stirrups, Structures & Buildings, 170 (3), 190-
198.

Austrell P.E., Enquist B., Heyden A. & Spanne S. 1995 Contact Free Strain Measurement
Using MATLAB Image Processing Toolbox, Nordic MATLAB Conference,
Stockholm.

Çomak B., Beycioğlu A., Başyiğit C. & Kılınçarslan Ş. 2011 The Use of Image
Processing Techniques in Concrete Technology, International Advanced
Technologies Symposium, Elazığ.

Fırat Alemdar Z., Browning J. & Olafsen J. 2011 Photogrammetric Measurements of RC

Bridge Column Deformations, Journal of Engineering Structures, 33 (8), 2407-2415.

Jauregui D.V., Leitch K.R., White K.R. & Woodward C. 2002 Vertical Deflection
Measurement of Bridge Structures with Digital Close-Range Terrestrial
Photogrammetry, First International Conference on Bridge Maintenance, Safety and
Management, Barcelona, Spain.

Öztürk M. 2016 Shear Strength Of Polypropylene Fiber Reinforced Concrete Beams

without Stirrups. MSc. thesis, Department of Civil Engineering, Graduate School Of
Natural And Applied Sciences, Yıldız Technical University, Turkey.

Pan B., Qian K., Xie H. & Asundi A. 2009 Two-Dimensional Digital Image Correlation
for In-Plane Displacement and Strain Measurement, Measurement Science And
Technology Conference, UK.

Whiteman T., Lichti D.D. & Chandler I. 2002 Measurement of Deflections in Concrete
Beams By Close-Range Digital Photogrammetry, Symposium on Geospatial Theory,

36
 Processing and Applications, Ottawa.

Yue Z.Q., Chen S. & Tham L.G. 2003 Finite Element Modeling of Geomaterials Using
Digital Image Processing. Computers and Geotechnics, 30, 375–397.

37