Vous êtes sur la page 1sur 8

Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP.

155-162

A Comparative Study on Nonlinear Static and


Dynamic Analysis of RC Frame Structures
Onur Merter*1, Taner Ucar2
Department of Civil Engineering, Dokuz Eylul University, Izmir, Turkey
Department of Architecture, Dokuz Eylul University, Izmir, Turkey
*1
onur.merter@deu.edu.tr; 2taner.ucar@deu.edu.tr

Abstract- Both static and dynamic analysis methods, which may be linear elastic or nonlinear, are used for structural analysis of
structures. In nonlinear analysis methods, geometric nonlinearity and behaviour of material beyond linear elastic limit are taken
into consideration. Artificial or recorded ground motions are used in nonlinear dynamic analysis, which are also known as nonlinear
time history analysis. In this study, both linear and nonlinear dynamic analyses of six and ten-story RC frame structures, which are
primarily designed according to Turkish Seismic Design Code, are performed by using seven ground motions recorded at different
soil sites of Turkey. However, nonlinear static analyses or so-called pushover analyses of frames are carried out. Base shear forces
and interstory drift ratios are obtained as analysis results and the comparison of above mentioned methods are made by using these
analysis results.
Keywords- Linear and Nonlinear Analysis Methods; Pushover Analysis; Time History Analysis; Base Shear Forces; Interstory Drift
Ratios

I. INTRODUCTION
Turkey, a country which experienced many destructive earthquakes in the past, is under significant seismic risk. In these
earthquakes, losses of many human lives, as well as economical losses, were occurred and many reinforced concrete buildings
were heavily damaged. In order to reduce these losses and damages, the design of structures in accordance with earthquake
codes and responsibility of structural engineers are quite imperative.
There are several methods for analysis of structures. These methods may be linear elastic or nonlinear as well as static or
dynamic. The Equivalent Earthquake Load Method, presented in Turkish Seismic Design Code ([1]), is a static and linear
elastic analysis method. Although it is linear elastic, in this method, nonlinear behaviour of structural systems is taken account
indirectly by using the earthquake load reduction factor. The most popular nonlinear and static analysis method is known as
pushover analysis ([2, 3, 4, 5, 6]). In pushover analysis, structures are pushed monotonically up to some target displacements
and lateral displacement versus base shear force curve of the top point (usually the top point) of the structure is obtained. This
force-displacement curve is known as the capacity curve of the structure and gives us some useful information about the
nonlinear behaviour of structures, which cannot be obtained from linearly elastic analysis of structures. Also it is possible to
observe the distribution of plastic hinges over the structure and so pushover analysis is a notably useful method in earthquake
performance evaluation, rehabilitation and retrofit of existing structures.
There exist some limitations in using of pushover analysis methods which consider only the fundamental vibration mode of
the structure. Both the lateral load distribution and target displacement are based on the assumption that the response is
controlled by the fundamental mode and that the mode shape remains unchanged after the structure yields. But after the
structure yields, both assumptions are approximate ([7]). So to overcome these limitations, modal pushover analysis methods
can be used.
As well as static analysis methods, dynamic methods can be used in the analysis of structures and these methods can also
be linear elastic or nonlinear. In dynamic analysis methods, or so-called time history analysis, both artificial and recorded
ground motions can be used.
From the studies related with earthquake-resistant design of structures ([8, 9, 10]), it is observed that some nonlinear
deformations may occur in the structures which are under the effect of strong ground motions. To estimate the distribution of
these kind inelastic deformations, it is quite imperative to use nonlinear methods in the analysis of structures. Much more
useful and effective results can be obtained by considering the nonlinear behaviour of structures both in static and dynamic
analyses. Nonlinear dynamic analysis methods are rather complex and time consuming. Also it is impossible to estimate the
strong ground motion that will affect the structure. All these reasons make nonlinear dynamic analysis methods impractical and
these methods are not so commonly used.
In this study, firstly, pushover analyses of six and ten-story RC frame structures, which are designed according to 2007
version of Turkish Seismic Design Code, are performed. Afterwards, linear and nonlinear dynamic analyses of the frames are
carried out. In dynamic analyses, acceleration time histories of seven earthquakes (1999 Kocaeli Earthquake, 1999 Duzce
Earthquake, 1995 Afyon-Dinar Earthquake, 2003 Bingol Earthquake, 2003 Denizli-Saraykoy Earthquake, 1992 Erzincan

- 155 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

Earthquake and 2005 Urla-Sigacık Earthquake) recorded in different soil sites of Turkey are used ([11]). Base shear forces and
interstory drift ratios of the selected frames are obtained as analysis results and these values, obtained from different analysis
methods, are compared with each other. It is found that each strong ground motion reflects its own characteristics to analysis
results. The best way to realize the real behaviour of structures under earthquake effects is to perform nonlinear time history
analysis.

II. ANALYSIS METHODS


In this study, pushover analyses, linear elastic and nonlinear dynamic analyses of six and ten-story RC frame structures,
designed according to Turkish Seismic Design Code, are performed. SAP2000, a structural analysis program, is utilized in all
analyses ([12, 13]). Total design vertical load, including live loads, is taken 40 kN/m at each beam span. The compression
strength of concrete is 25 MPa and the yield strength of reinforcing bars is 420 MPa. Frames are assumed to be on Seismic
Zone 1 and local site class is takes as Z3 for all frames. Dimensional properties of beams and columns are shown in Fig. 1.

Fig. 1 Geometric properties of frames

A. Nonlinear Modelling of Section and Nonlinear Static Analysis of Frames


Nonlinear properties of sections are determined by using the cross section program XTRACT ([14]). In nonlinear modeling
of RC sections, the effect of confinement is taken into consideration: stirrups with 8 mm diameter are used as transverse
reinforcement in all beams and columns. Stirrup spacing is taken as 10 cm at beam ends and 7.5 cm at column ends, which are
both potential plastic hinge regions. Reinforcing details of beams and columns are shown in Fig. 2.

Fig. 2 Reinforcing details of beams and columns

- 156 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

In nonlinear static analysis, so-called pushover analysis, the distribution of lateral load is taken proportional to the shape
fundamental vibration period of the frames. This lateral load distribution gives satisfactory results for regular low- and mid-rise
structures. Frames are pushed up to their target displacement, which is determined according to Turkish Seismic Design Code,
and distributions of plastic hinges that occur in structural members are observed. In pushover analyses, both geometric and
material nonlinearity is taken into consideration and lateral loads are increased monotonically while vertical loads are assumed to
remain constant. Pushover analyses are performed by SAP2000 structural analysis program and plastic hinges are assigned at both
ends of beams and columns by adopting a lumped plasticity model ([1, 15]). The capacity curves of frames are shown in Figs. 3
and 4. Target displacement of the top point and the corresponding base shear forces of the frames are given in Table 1.

Fig. 3 Capacity curve of six-story frame Fig. 4 Capacity curve of ten-story frame

TABLE 1 PUSHOVER ANALYSIS RESULTS

Target Displacement (δmax) Base Shear Force (V)


(m) (kN)
Six-story RC Frame 0.212 607.87
Ten-story RC Frame 0.326 592.92

B. Linear Dynamic Analysis of Frames


In linear elastic dynamic methods, analysis of structures under earthquake effects is performed for each time durations and
material assumed to remain linear elastic. Acceleration time histories used in the analyses are belong to 1999 Kocaeli
Earthquake, 1999 Duzce Earthquake, 1995 Afyon-Dinar Earthquake, 2003 Bingol Earthquake, 2003 Denizli-Saraykoy
Earthquake, 1992 Erzincan Earthquake and 2005 Urla-Sıgacık Earthquake ([11]). Acceleration time histories of earthquakes
are shown in Figs. 5-7 and detail of these strong ground motions are given in Table 2. In this table PGA is the peak ground
acceleration and PGV is the peak ground velocity.

Fig. 5 Acceleration time histories of Kocaeli and Duzce earthquakes

Fig. 6 Acceleration time histories of Afyon-Dinar and Bingol earthquakes

- 157 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

Fig. 7 Acceleration time histories of Denizli-Saraykoy, Erzincan and Izmir-Urla earthquakes

TABLE 2 STRONG GROUND MOTION DETAILS

Magnitude Recording Epicentral Distance PGA PGV


Earthquake Date
(Mw) Station (km) (cm/s2) (cm/s)
Kocaeli 17.08.1999 7.6 Kocaeli Meteorological Station 15.9 230.80 38.59
Duzce 12.11.1999 7.1 Duzce Meteorological Station 11 406.20 68.57
Afyon-Dinar 01.10.1995 6.4 Denizli-Cardak Health Center 5 47.79 3.39
Bingol 01.05.2003 6.3 Bingol Directorate of Public Works and Settlement 6 296.04 21.87
Denizli-Saraykoy 23.07.2003 5.3 Denizli-Saraykoy Geothermal Public Housing 11 122.16 4.79
Erzincan 13.03.1992 6.6 Erzincan Meteorological Station 23 478.77 78.22
Urla-Sigacik 20.10.2005 5.8 Ege University Agricultural Faculty 15.4 31.93 2.73

5% damping ratio is assumed for each vibration mode. After analyses, which are performed again by using SAP2000 ([11]),
total base shear force versus time graphs of six-story frame for Kocaeli and Duzce earthquakes are given in Figs. 8 and 9.
Similarly, total base shear force versus time graphs of ten-story frame for Afyon-Dinar and Bingol earthquakes are given in
Figs. 10 and 11. Base shear forces obtained from linear dynamic analyses of six- and ten-story frames for the considered
earthquakes are given in Table 3.

Fig. 8 Base shear force versus time graphs of six-story frame Fig. 9 Base shear force versus time graphs of six-story frame
(Kocaeli Earthquake) (Duzce Earthquake)

Fig. 10 Base shear force versus time graphs of ten-story frame Fig. 11 Base shear force versus time graphs of ten-story frame
(Afyon-Dinar Earthquake) (Bingol Earthquake)

- 158 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

TABLE 3 BASE SHEAR FORCES OBTAINED FROM LINEAR DYNAMIC ANALYSIS

Base Shear Forces of Six-Story Frame Base Shear Forces of Ten-Story Frame
Earthquake
(kN) (kN)
Kocaeli 1078.00 1037.00
Düzce 2973.00 2806.00
Afyon-Dinar 333.90 143.30
Bingol 865.20 733.80
Denizli-Saraykoy 121.70 310.90
Erzincan 2889.00 2771.00
Urla-Sigacik 125.70 141.80

C. Nonlinear Dynamic Analysis of Frames


In nonlinear dynamic analysis methods dynamic behaviour of structures under earthquake effects is investigated, as well as
in linear dynamic analysis. Nonlinear dynamic analysis method, so called nonlinear time history analysis, is accepted as the
best method to determine seismic behaviour of structures. In this method, analyses can be performed by direct integration
method, which gives the most reliable results. Dynamic earthquake loads are incrementally affected to the structure with time
intervals ∆t and equations of motion are solved by direct integration.
In nonlinear time history analyses, analysers may be faced with some problems such as determination of force-deformation
relation of structural member under cyclic loads and selection of strong ground motion that will used in analyses. Due to
complexity of nonlinear time history analysis, more practical analysis methods are needed and so in such cases nonlinear static
analysis methods are used.
In nonlinear time history analyses of this study, nonlinear force-deformation relationship of R/C members is defined by
Takeda hysteretic model, in which degradation of the stiffness due to increasing damage can be considered ([16]). Force-
displacement relationships of Takeda hysteretic model is shown in Fig. 12. In this figure, K0 is the initial stiffness, rK0 is the
post yield stiffness, δy is the yield displacement, δp is the peak displacement and Fy is the yield force. In order to define
nonlinear force-deformation relation of structural members under cyclic loads, any hysteretic model, theoretically and
experimentally validated, can be used ([17, 18]).

Fig. 12 Force-deformation relationship of Takeda hysteretic model

Nonlinear time history analyses of the frames are performed by using SAP2000 and strong ground motions are the same as
those used in linear dynamic analyses. Total base shear force versus time graphs of six-story frame for Kocaeli and Duzce
earthquakes are given in Figs. 13 and 14. The same graphs of ten-story frame for Afyon-Dinar and Bingol earthquakes are
given in Figs. 15 and 16. Base shear forces obtained from nonlinear dynamic analyses of six- and ten-story frames for the
considered earthquakes are given in Table 4.

Fig. 13 Base shear force versus time graphs of six-story frame Fig. 14 Base shear force versus time graphs of six-story frame
(Kocaeli Earthquake nonlinear time history result) (Duzce Earthquake nonlinear time history result)

- 159 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

Fig. 15 Base shear force versus time graphs of ten-story frame Fig. 16 Total shear force versus time graphs of ten-story frame
(Afyon-Dinar Earthquake nonlinear time history result) (Bingol Earthquake nonlinear time history result)

TABLE 4 BASE SHEAR FORCES OBTAINED FROM NONLINEAR TIME HISTORY ANALYSIS

Base Shear Forces of Six-Story Frame Base Shear Forces of Ten-Story Frame
Earthquake
(kN) (kN)
Kocaeli 261.50 486.80
Düzce 772.00 847.60
Afyon-Dinar 29.40 41.22
Bingol 189.50 239.90
Denizli-Saraykoy 61.00 78.94
Erzincan 708.80 846.60
Urla-Sigacik 41.36 53.40

D. Drift Ratios
After performing nonlinear static analyses, linear and nonlinear dynamic analyses, interstory drift ratios of the frames are
also determined. Interstory drift ratios obtained from linear and nonlinear analyses of six-story RC frame by using Duzce,
Erzincan and Bingol strong ground motion records are shown in Figs. 17, 18 and 19, respectively. In the same graphs,
interstory drift ratios obtained from nonlinear static analyses of six- and ten-story RC frames, which are pushed up to their
target displacements, are also shown.

Fig. 17 Interstory drift ratios of six-story RC frame (Duzce Earthquake)

Fig. 18 Interstory drift ratios of six-story RC frame (Erzincan Earthquake)

- 160 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

Fig. 19 Interstory drift ratios of six-story RC frame (Bingol Earthquake)

In nonlinear time history analyses, which are performed by using Duzce strong ground motion records, the largest
interstory drift ratios are obtained in upper stories. In bottom stories nonlinear static analysis gives larger interstory drift ratios
for six- and ten-story frames. The same evaluation can also be done for interstory drift ratios obtained from Erzincan
Earthquake (Fig. 18). As shown in Fig. 19, the smallest interstory drift ratios are obtained from linear dynamic analyses of six-
story frame by using Bingol records and the largest values are obtained from pushover analyses. Interstory drift ratios of six-
story RC frame obtained from other four earthquakes (Kocaeli, Afyon-Dinar, Denizli-Saraykoy and Urla-Sigacik) show
similarity with the results of Bingol Earthquake.
In Figs. 20 and 21, interstory drift ratios of ten-story RC frame obtained from linear and nonlinear dynamic analyses by
using Kocaeli and Duzce records are shown together with interstory drift ratios obtained from nonlinear static analyses of the
same frames. In nonlinear time history analyses, which are performed by using Kocaeli and Duzce strong ground motion
records, the largest interstory drift ratios are obtained in upper stories (Figs. 20 and 21). In bottom stories nonlinear static
analysis gives larger interstory drift ratios. In general, the largest intersory drift ratios of ten-story RC frame are obtained from
nonlinear static analysis while the smallest ones are obtained from linear dynamic analysis.

Fig. 20 Interstory drift ratios of ten-story RC frame (Kocaeli Earthquake) Fig. 21 Interstory drift ratios of ten-story RC frame (Duzce Earthquake)

III. CONCLUSIONS
In this study, analysis of six- and ten-story RC frames, which are primarily designed according to Turkish Seismic Design
Code, are performed by using different analysis methods. Total base shear forces and interstory drift ratios are obtained from
nonlinear static, linear and nonlinear dynamic analyses of the frames. The ratio of the base shear force obtained from linear
dynamic analysis to the base shear forces obtained from nonlinear dynamic analysis gives structural system behaviour factor
(R). In Turkish Seismic Design Code, the structural behaviour factor for nominal ductility buildings in which seismic loads are
fully resisted by frames is given as 4 (R=4). In this study, the structural behaviour factor of six- and ten-story frames is
obtained about or smaller than 4. But this result can change due to characteristic of earthquake records.
Interstory drift ratios obtained from nonlinear time history analyses are generally larger in upper stories. Particularly for
Duzce, Kocaeli and Erzincan earthquakes, which are relatively more severe earthquakes, the largest interstory drift ratios are
obtained from nonlinear time history analysis of frames. Interstory drift ratios obtained from nonlinear static analysis are larger
than those obtained from other four earthquakes. Interstory drift ratios obtained from linear dynamic analysis are generally
smaller.

- 161 -
Journal of Civil Engineering and Science Sept. 2013, Vol. 2 Iss. 3, PP. 155-162

Nonlinear static analysis methods are commonly used in structural and earthquake engineering. These kinds of analysis
methods provide an effective way to determine and evaluate the nonlinear behaviour of structures by civil engineers, who are
generally familiar with linear elastic analysis methods. Due to its complexity and uncertainties in selection of strong ground
motion records that will used in nonlinear time history analysis, nonlinear static analysis, so-called pushover analysis, becomes
much more popular. In this study, pushover analyses of six- and ten-story frames are performed and capacity curves of these
frames are obtained. For six-story frame, base shear forces obtained from nonlinear time history analyses are smaller than
those obtained from pushover analysis; expect Duzce and Erzincan earthquakes which give bigger base shear forces than base
shear forces obtained from pushover analysis. The same conclusions can be made for ten-story RC frame. While interstory
drift ratios obtained from different analysis methods are compared with each others, nonlinear time history analyses performed
by using Duzce, Kocaeli and Erzincan records give larger values. For other earthquakes, interstory drift ratios obtained from
pushover analyses are larger. Expect the nonlinear time history results of Duzce, Kocaeli and Erzincan earthquakes, pushover
analysis results are on the safe side. It may be concluded that, in case nonlinear time history analyses are not performed,
pushover analysis methods give valuable information about nonlinear behaviour of structures and they are more practical.
The analysis results show that each earthquake record reflects its characteristic to the analysis and different earthquake
records may lead to different results. It is not so realistic to make definite conclusion by using limited number of ground
motion in analyses. When numerous strong ground motion records are used in nonlinear time history analyses and analyses are
performed for different types of structures, more realistic and meaningful results can be obtained according to results obtained
from other analysis methods.

REFERENCES

[1] Turkish Seismic Design Code, (2007). Ministry of Public Works & Settlement, Ankara, Turkey [in Turkish].
[2] Krawinkler, H. and Seneviratna, G.D.P.K., “Pros and cons of a pushover analysis of seismic performance evaluation”, Engineering
Structures, vol. 20, No. 4-6, pp. 452-464, 1998.
[3] Chatpan, C. and Chopra, A.K., “Evaluation of modal pushover analysis using generic frames”, Earthquake Engineering and Structural
Dynamics, vol. 32, No.3, pp. 417-442, 2003.
[4] Mwafy, A.M. and elnashai, A.S., “Static pushover versus dynamic collapse analysis of RC buildings”, Engineering Structures, vol. 23,
No. 5, pp. 407-424, 2001.
[5] Fajfar, P., “A nonlinear analysis method for performance based seismic design”, Earthquake Spectra, vol. 16, No. 3, pp. 573-592, 2000.
[6] P. Yang and Y. Wang, “A study on improvement of pushover analysis”, Proceedings of 12 th World Conference on Earthquake
Engineering, 2000, Auckland, New Zealand.
[7] Chopra, A.K. and Goel, R.K., “A modal pushover analysis procedure for estimating seismic demands for buildings”, Earthquake
Engineering and Structural Dynamics, vol. 31, No. 3, pp. 561-582, 2002.
[8] P. Agorwal and M. Shrikhande, Earthquake Resistant Design of Structures, India: Prentice-Hall, 2006.
[9] F. Naeim, The Seismic Design Handbook, United States of America: Chapman & Hall, 1989.
[10] Miranda, E. and Bertero, V.V., “Evaluation of strength reduction factors for earthquake-resistant design”, Earthquake Spectra, vol. 10,
No. 2, pp. 357-379, 1994.
[11] (2012) Strong Ground Motion Database of Turkey, Available: http://www.deprem.gov.tr.
[12] SAP2000 Nonlinear, Version 14.2, Structural Analysis Program, (2010). Computers and Structures Inc., Berkeley CA.
[13] Computers and Structures Inc. (2002). Integrated Finite Element Analysis and Design of Structures Basic Analysis Reference Manual,
CSI SAP2000, Berkeley, California.
[14] Imbsen and Associates Inc. (2007). XTRACT-Cross Section Analysis Program for Structural Engineers-Step by Step Examples,
IMBSEN Software Systems v. 3.0.8, California.
[15] Korkmaz, A. and Duzgun M., “Evaluation of nonlinear analysis methods”, Engineering Journal of ITU, vol. 6, No. 3, pp. 11-23, 2007
[in Turkish].
[16] Takeda, T., Sozen, M.A. and Nielsen, N.N., “Reinforced concrete response to simulated earthquakes”, Journal of the Structural Division,
vol. 96, No. 12, pp. 2557-2573, 1970.
[17] A. Aviram, K.R. Mackie and B. Stojadinovic, Guidelines for Nonlinear Analysis of Bridge Structures in California, Pacific Earthquake
Engineering Research Center, Berkeley, 2008.
[18] Fahjan, Y.M., “Selection and scaling of real earthquake accelerograms to fit the Turkish design spectra”, Techanical Journal of Turkish
Chamber of Civil Engineers, vol. 19, No.3, pp. 4423-4444, 2008 [in Turkish].

- 162 -

Vous aimerez peut-être aussi