EXPERIMENTAL AND ANALYTICAL PERFORMANCE ASSESSMENT …conf.ncree.org.tw/Proceedings/i0951012/data/pdf/4ICEE-0154.pdf · EXPERIMENTAL AND ANALYTICAL PERFORMANCE ASSESSMENT OF IN-SITU

4th International Conference on Earthquake Engineering Taipei, Taiwan

October 12-13, 2006

Paper No. 154

EXPERIMENTAL AND ANALYTICAL PERFORMANCE ASSESSMENT OF IN-SITU PUSHOVER TESTS OF SCHOOL BUILDINGS IN TAIWAN

Yuan-Tao Weng 1 Ker-Chun Lin 2 Shyh-Jiann Hwang3

Abstract

The 1999 Chi-Chi earthquake revealed the poor performance of the RC school buildings in Taiwan. It also indicates an urgent need of seismic evaluation and retrofit for the remaining schools. In order to realize the behavior of a typical school building subjected to lateral load, a series of in-situ test for two existing 2-story RC school buildings was carried out. Three types of lateral loading patterns were subjected to three 2-story 2-classroom frame specimens respectively: monotonic static pushover, cyclic static pushover, and earthquake time history input. This present paper is devoted to the experimental program, test results and analytical assessment. Moreover, the responses to free vibration, forced vibration and mircotremor measurement were recorded in order to identify the dynamic behavior of frame specimens. These experiments provide reliable and efficient data of real interest for a clear understanding of the actual building behavior, especially the effects of the seismic performance of a school building subjected to cyclic lateral loading collocating with pseudo dynamic test (PDT). The advantage of integrating these data in the seismic evaluation is presented and discussed. Results of these tests are reported, analyzed and interpreted in this paper. Test results present that the effect of the decay of structural strength and stiffness induced by cyclic loading should be considered in seismic evaluation procedure properly.

Introduction

In Taiwan, many typical school buildings suffered severe damage by the Chi-Chi earthquake, 1999. Most of old RC school buildings were designed according to a standard plan that is functional for getting natural light and ventilation. The typical plan has all the openings and a corridor in the longitudinal direction and many partition walls in the transverse direction. Some common failure patterns were found because of the typical type of school buildings, such as failure in the longitudinal direction due to lack of walls, short-column effect due to constrain by windowsills, and strong-beam-weak-column effect due to non-ductile reinforcement and slabs that connect with the beams. For preventing possible damage in the future, it is urgent to develop the seismic evaluation and retrofit technology for the existing schools. Although there are already some assessment methods developed by international

1 Associate Research fellow, National Center for Research on Earthquake Engineering, Taiwan ROC. 2 Associate Research fellow, National Center for Research on Earthquake Engineering, Taiwan ROC. 3 Deputy Director, National Center for Research on Earthquake Engineering, Taiwan ROC.

researchers, usually they are verified by small-scale or partial structural assemblages subjected to monotonic lateral loading pattern, but not full-scale structure subjected to cyclic loading pattern. It is still questionable that if test results in the laboratory can represent the true behavior of actual buildings. Therefore, a series of in-situ structural test of an existing RC school building were carried out for realizing the real structural behavior.

The seismic analysis of the existing buildings requires a particular attention for several reasons among which:

1. a large amount of buildings were constructed without any consideration of seismic risk; in Taiwan for instance, the first regulations against seismic risks date from the 70th. Thus, a small percentage of structures have been built with a design taking into account the seismic risk;

2. the necessity of establishing reliable diagnosis not only for strategic and public buildings but also for ordinary or industrial structures in order to identify vulnerable structures and to justify decisions to improve the structure.

Several methodologies, based mainly on statistical approaches, have been developed to assess the structural seismic performance. As any statistical method, these approaches are reliable only if applied to a large number of buildings, i.e. at the urban scale, and only if sufficient amount of data is available Therefore, when only a few seismic events are available, and when a given building is to be evaluated, these global vulnerability approaches are no longer relevant.

Many common seismic evaluation methods are used to apply monotonic static pushover to gain the capacity curve, which represents the structural seismic resistant capacity simply, but monotonic loading pattern could not present the actual cyclic behavior induced by an earthquake. Therefore, in order to exceed this drawback through in-situ experiments, a 8-classroom building constructed in 1979 was cut in the staircase, One part of this building was set for monotonic and the other part was set for cyclic loading or earthquake loading alone the longer axis of building, respectively. These two pure frame specimens are arranged to be corresponsive each other. Moreover, on the other side, a 4-classroom building specimen retrofitted by adding composite columns to partition brick walls was cut in the middle where jacks were set for monotonic loading alone the longer axis. One half of this building was reinforced by steel bracings to provide reaction support, while the other half was pushed to failure.

Indebted to the Tao-Yuan County Government and Reui-Pu elementary school, the research team composed of crews of National Center for Research on Earthquake Engineering (NCREE), National Taiwan University of Science and Technology (NTUST), National Cheng Kung University (NCKU), Minhsin Institute of Technology (MHIT) and National Taiwan University (NTU) were allowed to use an old school building that is about to be demolished as the subject of push over test. Except for providing verification for seismic assessment and retrofit technology, this test also gives further understanding of seismic ability of existing school buildings.

In this context, a series of in-situ experimental programs using existing RC Reui-Pu elementary school buildings before and during their demolition has recently been finished. The aims of these in-situ experiments are threefold:

1. to identify the dynamic behavior of typical RC school buildings built according to the design rules of common practice in Taiwan;

2. to take advantage of their progressive demolition to obtain some qualitative and, where possible, quantitative answers of the actual influence on the dynamic characteristics including the ductility and the decay of structural lateral strength and stiffness induced by seismic cyclic loading.

3. to verify the feasibility of the innovative retrofitting technique using adding composite columns to partition brick walls on traditional RC school buildings.

The present paper, divided into three sections, focuses on the experiments. The first section is devoted to the experimental procedure, data processing and studied building. In the second section, the experimental results obtained on in-situ school building specimens are presented and discussed. The third section deals with the observations and measurements gathered during progressive demolitions and their interpretation.

Tested Structures and Test Results

Testing real structures offers the possibility to analyse phenomena that are difficult to reproduce in idealized laboratory experiments. In counter part, the tested structures are only partially known and the range of loading is limited. This section presents the used in-situ methods, the main structural characteristics of the tested buildings and the test results.

There are three primary frame specimen showed below: Specimen I: Pseudo dynamic test and the static cyclic pushover test of frame specimen,

Specimen II: Monotonic static pushover test of frame specimen,

Specimen III: Monotonic static pushover test of seismic retrofitted frame specimen.

The site plan of Reui-Pu elementary school is shown in Fig. 1. All frame specimens are located in Buildings C and D, respectively. In the North-South direction, the Building D with 2 floors contains 8 classrooms and a staircase in the middle. The Specimen I and Specimen II are two parts of the Building D parallel to each other. As the Fig. 2 shown, Specimen I is the south part of Building D including classrooms No.3 and No.4 was set for cyclic static pushover or PDT, and Specimen II is the north part including classrooms No.5 and No.6 was set for monotonic loading alone the longer axis of Building D, respectively. In the Classrooms No.1 and No.2, some members were removed and two pieces of 60 cm-thick RC reaction walls were constructed along the longitudinal direction to provide reaction support for actuators installed at Specimen I; Classrooms No.7 and No.8 at the 1st floor was reinforced by steel bracings to provide reaction support to Specimen II. These two pure frame specimens are arranged to be corresponsive each other. In the East-West direction, Building C with 2 floors contains 4 classrooms and a staircase in the east. As the Fig. 3 shown, Specimen III is the east part including classrooms No.11 and No.12 was also set for monotonic loading alone the longer axis of Building C, and classrooms No.9 and No.10 at the 1st floor was reinforced by steel bracings to provide reaction support to Specimen III. All building frame specimens were made of reinforced concrete, but the partition brick wall and windowsills are made of 1-brick-thick brick walls and 1.5-brick-thick brick walls, respectively. The building had no visible damage before the test, as shown in Fig. 4.

street

N

廁所 820

440

755

324

299

341

Specimen I: Pseudo dynamic test and the static cyclic pushover test of frame specimen

Specimen II: Monotonic static pushover test of frame specimen

Spec

imen

III:

Mon

oton

ic st

atic

pus

hove

r te

st o

f sei

smic

retro

fitte

d fr

ame

spec

imen

Figure 1. Site plan of Reui-Pu elementary school, Tao-Yuan, Taiwan

385

385

310

385

385

290

71

Figure 2. Structural plan of the Specimen I and Specimen II

Figure 3. Structural plan of the Specimen III

(a) (b)

Figure 4. Photo of the Specimen I & III

In order to identify the dynamic behavior of these three frame specimens, three types of excitation are used to identify modal frequencies and damping ratios before pushover tests or PDT: (i) Free vibration: Free vibration test were realized by exciting a structure harmonically at

its fundamental resonance frequency by an electromagnetic exciter installed on the roof floor. When steady state motion is obtained, the exciter is stopped and the free vibration response of the structure is observed.

(ii) forced vibration: Forced vibration tests were realized by impacting the upper part of the frame specimens in the two main directions by mean of a heavy mechanical shovel

(usually used for demolition). The damage (when it appears) is localized in the very vicinity of the impact and outside the zone the structure remains entirely intact. To our knowledge, this test method was not proposed before though the use of a shovel can be simply implemented and tests easily repeated. Compared to the mircotremor measurement and harmonic oscillations, the short impulsive record is of larger magnitude with a pick acceleration of about 10-2g even on the ground floor, i.e. thousand times greater than the ambient level.

(iii) mircotremor measurement: The method, initiated in the 70th (Stubbs and MacLamore, 1973; Trifunac, 1972), know actually large developments due to its simplicity (Ivanovic and Trifunac, 2000; Farsi, 1996). The level of horizontal acceleration is the order of 10-5g at the bottom and 10-4g at the building top. The density of probability of the random signals follows a Gaussian distribution so that the building responds to a white noise imposed motion. When measuring ambient or shake-excited building vibrations, one usually assumes that the structure can be approximated by a one-dimensional, linear, damped, discrete or continuous system. In some cases (Jennings and Kuroiwa, 1968) measurements indicated that floor diaphragms are sufficiently stiff so that the above assumption is acceptable.

To identify the dynamic behavior of these three frame specimens, three types of excitation are used to identify modal frequencies and damping ratios before pushover tests or PDT. It should notice that all the tested buildings are founded on footing foundations in transverse direction. The soil is red soil with soft mechanical condition. Table 2 shows the experimental modal frequencies and damping ratios of all frame specimens. In the general case, the 1st modal frequency measured by mircotremor is always higher than that measured by free vibration or forced vibration, because the mircotremor measurement is not easy to reveal the effect of soil-structure interaction. Secondly, it is quite close between the 1st modal frequencies measured by free vibration and forced vibration respectively.

Table 2. Modal characteristics of building specimens

Modal characteristics of building I

Longitudinal direction Transverse direction

Monitoring methd Frequency(Hz)

Damping(%)

Frequency(Hz)

Damping (%)

Free Vibration 2.88 2.63 * *

Forced Vibration 3.00 3.50 6.44 2.39

Microtremor Measurement 3.39 * 7.83 *

Modal characteristics of building II


Monitoring method Frequency(Hz)

Damping (%)

Frequency(Hz)

Damping (%)


Forced Vibration 2.65 2.82 5.85 2.86


Modal characteristics of building III


Monitoring methd Frequency(Hz)

Damping (%)

Frequency(Hz)

Damping (%)


Force Vibration * * * *


In addition, there are three primary tests: monotonic static pushover test at the Specimen I and Specimen III, cyclic pushover test and PDT at the Specimen I, prepared and executed from August 5 to 18th, 2006. Details of description of the three tests are as below.

Specimen I: Pseudo Dynamic Test and Cyclic Static Pushover Test

Fig. 5 shows the layout of the PDT and cyclic static pushover test at the Specimen I. Each 10 m wide classroom is consist of 3 spans, lies along the longitudinal direction. About half of the columns in Frame A and all the columns in Frame C have 80 cm high windowsills that usually cause the short-column effect besides them. Short-column effect happens because the column constrained by windowsills that were not considered in design, effective height of the column is then shortened and cause larger shear stress or even shear failure.

1500

Figure 5. Test layout of the Specimen I

The servo-control units used for these tests were MTS actuators with ±0.5 m stroke and load capacity of 1.0 MN. The external control displacement transducers were electromagnetic TEMPSONIC sensor with a stroke ±0.5 m stroke and a 5 μm resolution. All the controllers are connected a master unit by mean of MTS FlexTest digital (GT) controller. The corresponding software (based on C++ language) had to combine capabilities with high-level computation. However, for this occasion, due to the significant complexity added by the 2-DoF-per-floor model, this master unit conducts two tasks: 1) sending the targets to the controllers and receiving and display the associated measurements, 2) performing all computations for the pseudo dynamic integration. In this in-situ PDT and cyclic static pushover test, four hydraulic actuators were set at the cut beam and slab end of Frames A and C of 1F and 2F respectively, they were connected with post-tension steel tendons to steel clampings presented in Fig. 6. Behind the actuators, two pieces of RC reaction walls were constructed to provide reacting support. Two actuators per floor were designed to push the specimen in its weak axis and aim to the horizontal center line of each floor slab. Fig. 7 shows the locations of the actuators and reaction walls.

(a) (b)

Figure 6. Photo of actuators connected with post-tension steel tendons to steel clampings

(a) (b)

Figure 7. Locations of the actuators and reaction walls Initially, three types of vibration testing mentioned above were performed on per frame specimen to measure the fundamental period and damping ratio before PsD seismic test or pushover tests. Afterwards, the PsD seismic tests described below were performed. The earthquake scenario, including the earthquake intensities and sequence, for this experiment is shown in Table 3. To reveal the site effect and failure prevention was considered as an important criterion for selection of a record to obtain more controllable results and a stream of good response data in these tests, Tao-Yuan TCU006 station ground motion data recorded in 2000/6/10 was selected because no pronounced peak was observed. The earthquake intensity using in the PDT No.1 ~ No.4 were decided so that the seismic response of Specimen I is contained elastic. On the other hand, To simulate the Specimen I meet a destructive earthquake just likes the 1999 Chi-Chi Earthquake, Chia-Yi CHY041 station ground motion data recorded in 1999/9/21 was selected, thus the earthquake intensity of the PDT No.5 was decided so that the seismic response of Specimen I is contained that the lateral strength of Specimen I just degraded, thus containing the Specimen I has comparability with the Specimen II. Figure 8 shows three earthquake ground accelerations and their corresponding normalized acceleration spectra.

Figure 8. Response spectra of the ground accelerograms and original acceleration time history

The corresponding spectral curve representing the 10% probability of exceedance in 50 years (10% / 50yr) seismic level of Taiwan 2005 building seismic provision is also shown in Fig. 8. Experimental results show the fundamental period and damping ratios of the Specimen I is about 0.36 second (2.88 Hz) and 3.0 %. According to the Sa(T1) method proposed by Shome et al. (1998), the ground motions shall be scaled such that the spectral acceleration of the spectra is equal to the 5% damped smoothed design spectra at the fundamental period of the prototype building in the considered direction. The 5% damped Sa values for CHY041EW record is also shown on Fig. 8a. To scale the earthquake intensity by the Sa(T1) method mentioned above, the corresponding PGA value for the 10% / 50yr seismic hazard levels is 0.24g. It was used to design the earthquake intensity using in the PDT No.5 was decided to measure the seismic response of Specimen I in the seismic hazard level.

Table 3. Test schedule of Specimen I

Date Test No. Earthquake Intensity

2006/08/11, 9:30 A.M. PDT No.1 TCU006EW, PGA= 67.9 gal

2006/08/12, 12:00 A.M. PDT No.2 TCU006NS, PGA= 135.1 gal

2006/08/13, 9:30 A.M. PDT No.3 TCU006NS, PGA= 120.0 gal

2006/08/13, 12:00 A.M. PDT No.4 CHY041EW, PGA= 120.0 gal

2006/08/14, 9:30 A.M. PDT No.5 CHY041EW, PGA= 240.0 gal

2006/08/15, 10:00 A.M. Cyclic Pushover Test

2006/08/16, 9:30 A.M. Monotonic Pushover Collapse Test

Test Results of PDT No.1 and PDT No.2 The PDT No.1 and PDT No.2 tests were failed due to the water cooling system of servo-control unit was shut down due to its motor was damaged, or the fuse of the hydraulic power was damaged. Even though these accidents occur, the status of Specimen I still remains intact.

0 10 20 3Time (sec)

0-300

A

-200

-100

0

100

200

300

ccel

erat

ion

(gal

) 1999 Chi-Chi EarthquakeCHY041EW

PGA=297 gal

0 5 10 15 20 25Time (sec)

-20

-10

0

10

20

Acc

eler

atio

n (g

al) 2000/6/10

TCU006EWPGA=16 gal

0 5 10 15 20 25Time (sec)

-20

-10

0

10

20

Acc

eler

atio

n (g

al) 2000/6/10

TCU006NSPGA=17 gal

0 0.4 0.8 1.2 1.6 20

1

2

3

4

5S a

(g)

Specimen IIn-situ PDT

TCU006EW

Period (sec)

TCU006NSCHY041EWT1=0.36 sec

(b)

(c)

Code 2005

(a)

Test Result of PDT No.3

After repairing all the damage of the servo-control unit, the PDT No.3 was executed smoothly. Figure 9 shows the final scene after PDT No.3 executed, the roof drift ratio was actually nearly 0.15% radian, and the base shear versus roof drift ratio cycles obtained in the PDT No.3. The maximum base shear P is 654.6 kN when the roof displacement ∆2 reached 10.6 mm. Essentially the frame specimen was still remain elastic and on obvious damage after the PDT No.3 conducted.

-0.2 -0.1 0 0.1 0.2Roof Drift (%, radian)

-1000

-500

0

500

1000

Bas

e Sh

ear (

kN)

PDT No.3TCU006NS

PGA=120 galEXP(PDT No.3)

(a) (b)

Figure 9. Experimental base shear-roof drift ratio cycles and final scene of PDT No.3


To reveal the difference under different earthquake sources and sites, the PDT No.4 was designed to be inputted the 1999 Chi-Chi Earthquake, Chia-Yi CHY041 station E-W acceleration time-history, but PGA remained 120 gal. Figure 10 shows the final scene after PDT No.4 executed, the roof drift ratio was actually nearly 0.21% radian, and the base shear versus roof drift ratio cycles obtained in the PDT No.4. The maximum base shear is 825.0 kN when the roof displacement ∆2 reached 15.4 mm. Essentially the frame specimen was still remain elastic and on obvious damage after the PDT No.4 conducted.

-0.3 -0.2 -0.1 0 0.1 0.2 0.3Roof Drift (%, radian)

-1000

-500

0

500

1000

Bas

e Sh

ear (

kN)

PDT No.4CHY041EW


(a) (b)



Figure 11 shows the final scene after PDT No.5 executed, the roof drift ratio was actually nearly 0.64% radian, and the base shear versus roof drift ratio cycles obtained in the PDT No.3. The maximum base shear P is 1177.0 kN when the roof displacement ∆2 reached 28.2 mm. The base shear of Specimen I occurred the maximum value at a roof drift ratio of about 0.35% radian, and its lateral strength began to degrade.

-0.8 -0.4 0 0.4 0.8Roof Drift (%, radian)

-1500

-1000

-500

0

500

1000

1500

Bas

e Sh

ear (

kN)

PDT No.5CHY041EW


(a) (b)


Test Result of Cyclic Pushover Test

Figure 12. Loading pattern used for Specimen I

During the test, the loading was cyclic, but in every 0.5% drift, the actuators were hold for 15-20 minutes, so that the staff can mark cracks and record the damage condition. The specimen was loaded until the roof drift ratio reach 2% radian and its strength descended to 50% of the maximum strength.

Figure 13. Base shear-roof drift ratio cycles and final scene of cyclic pushover test

-3 -2 -1 0 1 2 3Roof Drift (%, radian)

-1500

-1000

-500

0

500

1000

1500

Bas

e Sh

ear (

kN)

Cyclic Pushover

(a) (b)

Specimen II: Monotonic Static Pushover Test

Fig. 2 shows the structural plan of the specimen. Each 10m wide classroom is consist of 3 spans, lies along the longitudinal direction. Over half of the columns in Frame A and all the columns in Frame C have 80cm high windowsills that usually cause the short-column effect besides them. Short-column effect happens because the column constrained by windowsills that were not considered in design, effective height of the column is then shortened and cause larger shear stress or even shear failure.

Two 150t and two 300t hydraulic actuators were set at the cut beam end of A, and C frames of 1F and 2F, respectively. They were designed to push the specimen in its weak axis. Behind the actuators, 2 spans of the classroom No.7 were reinforced by added steel bracings to provide reacting support. During the test, the actuators were controlled through their cylinder areas to keep the loading put on 1F and 2F being 1:2, which is the proportion of lateral load distributed by the fundamental mode. The loading was monotonic, but in every 0.05% drift, the actuators were hold for 15-20 minutes, so that the staff can mark cracks and record the damage condition. The specimen was loaded until its strength descended to 67% of the maximum strength. For preventing doing any harm to the neighbor in the north side and safety of the staff, two steel supports were set in the classrooms in 1F to prevent complete collapse of the specimen.

Test Result of Monotonic Static Pushover Test of Specimen II

Fig. 14 shows the push over curve of the specimen and the final scene of the specimen at a roof drift ratio of about 6%. However, most deformation happened at 1F, while the 2F seemed remain undamaged, so the drift ratio at 1F was actually nearly 12%. The maximum base shear P is 1190.0 kN when the roof displacement ∆2 reached 55.1 mm. Some indentations showed when the actuators were hold for recording damages. However, the structural strength decayed obviously, quickly and shows acceptable ductility.

As shown in Fig. 14a, the normal columns at A-frame obviously failed by flexural bending and concrete at the compressive side crushed. Other normal columns in C-frame showed the same failure pattern. Otherwise, the short columns mostly failed by both bending and shear. Both horizontal and diagonal cracks showed in these columns’ middle part or lower ends, as shown in Fig. 15a. The diagonal cracks caused by shear stress show that the short-column effect did happened. While all of the columns had failed, the beams and slab still remained almost undamaged. This phenomenon, so-called strong-beam-weak-column, was also found in those school buildings damaged by the Chi-Chi earthquake.

-4 -3 -2 -1 0 1 2 3 4Roof Drift (%, radian)

-2000

-1500

-1000

-500

0

500

1000

1500

2000

Bas

e Sh

ear (

kN)

Mono. Pushover Exp (Specimen II)

(b) (a)

Figure 14. Base shear-roof drift ratio and final scene of monotonic pushover test of Specimen II

Specimen III: Monotonic Static Pushover Test

Fig. 15 shows the design scheme of the retrofitting technique using adding composite columns to partition brick walls at Specimen III. Each 10m wide classroom is consist of 3 spans, lies along the longitudinal direction. Over half of the columns in Frame A and all the columns in Frame C have 80cm high windowsills that usually cause the short-column effect besides them. Short-column effect happens because the column constrained by windowsills that were not considered in design, effective height of the column is then shortened and cause larger shear stress or even shear failure.

1

6

5

4

3

2

65

hoops: #3 @ 15

bars: 8 -#6

Figure 15. Design scheme of retrofitting technique using adding composite columns to partition brick walls

Similarly, two 150t and two 300t hydraulic actuators were set at the cut beam end of Frames A and C of 1F and 2F, respectively. They were designed to push the specimen in its weak axis. Behind the actuators, 2 spans of the classroom No.7 were reinforced by added steel bracings to provide reacting support. During the test, the actuators were controlled through their cylinder areas to keep the loading put on 1F and 2F being 1:2, which is the proportion of lateral load distributed by the fundamental mode. The loading was monotonic, but in every 0.05% drift, the actuators were hold for 15-20 minutes, so that the staff can mark cracks and record the damage condition. The specimen was loaded until its strength descended to 67% of the maximum strength. For preventing doing any harm to the neighbor in the north side and safety of the staff, two steel supports were set in the classrooms in 1F to prevent complete collapse of the specimen.

Test Result of Monotonic Static Pushover Test of Specimen III

0 2 4 6Roof Drift Ratio (% radian)

0

400

800

1200

1600

2000

Bas

e Sh

ear (

kN)

(a) (b)

Figure 16. Base shear-roof drift ratio and final scene of monotonic pushover test of Specimen III

Figure 16 shows the pushover curve of the specimen and the final scene of the specimen at a roof drift ratio of about 5.7%, and the drift ratio at 1F was actually nearly 5.3%. The maximum base shear P is 1965.5 kN when the roof displacement ∆2 reached 86.9 mm. Some indentations showed when the actuators were hold for recording damages. However, the structural strength decayed obviously, quickly and shows acceptable ductility.

Comparison of Test Result

To connect every envelope of per cycle of experimental base shear-roof drift ratio curves obtained from the experimental results of the PsD seismic tests, cyclic pushover test of Specimen I in positive and negative loading direction respectively, and link up the post-part of the capacity curve obtained the monotonic pushover collapse test, the entire capacity curves of Specimen I could be presented as the Specimen I(+) and the Specimen I(-) curves shown in Figure 17a. Here the Specimen I(+) curve is used to be representative capacity curve of Specimen I. Figure 17b shows the capacity curves of Specimen I, II and III. Comparing with the capacity curves of Specimen I and II, it shows the effects of the seismic performance of a school building subjected to cyclic lateral loading collocating with the PsD seismic tests. Table 4 shows the structural lateral strength, stiffness and peak roof displacement of Specimen I, II and III.

-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7Roof Drift (%, radian)

-2000

-1500

-1000

-500

0

500

1000

1500

Bas

e Sh

ea

2000

r (kN

)

Mono. Pushover Exp (Specimen II)EXP(Specimen I, PDT No.5)Cyclic(Specimen I)Collapse(Specimen I)Specimen I (+)Specimen I (-)

0 100 200 300 400 500

RFL Displacement (mm)

0

400

800

1200

1600

2000B

ase

She

ar (k

N)

Specimen ISpecimen IISpecimen III

(a) (b)

Figure 17. Comparison with base shear-roof displacement results of Specimen I, II and III

Table 4. Comparison of stiffness, ultimate base shear, and peak roof displacement Frame Specimen

No. Stiffness (kN/mm) Ultimate base shear (kN)

Peak roof displacement (mm)

Specimen I 52.3 1177 28.2 Specimen II 52.9 1190 55.1 Specimen III 82.1 1966 86.9

Experimental results show the stiffness and shear strength of Specimen I and II are very proximate. After the ultimate base shear occurred, the maximum difference of the shear strength was about 33% between Specimen I and Specimen II. In other words, the monotonic pushover procedure obviously overestimates the shear strength and stiffness after the structural strength begins to decay.

After all the tests were finished, some members were sampled to check their actual

configurations. A large number of column cover thickness are larger than 10 cm as Fig. 18 shown. This construction error is easy to reduce the capacity of RC columns.

Figure 18. Damage on RC column members

Conclusion

In-situ tests provides a precious chance to realize the behavior of a real building and to verify the analytical methods. Especially the experimental results of cyclic pushover test collocating with the PsD seismic tests confirmed the damaging behavior of school buildings observed in the Chi-Chi earthquake, and reveal the effects of cyclic loading patterns. Typical failing characteristics of school buildings, such as short-column effect and strong-beam-weak-column behavior, did happen to the specimens.

The monotonic pushover procedure obviously overestimates the shear strength and stiffness after the structural strength begins to decay.

Further research subjects would include study on out-of-plane behavior of brick walls, torsional effects, retrofit measures for resisting horizontal and vertical loads, and improving the analytical method.

Acknowledgement

The authors appreciate the National Science Council and Ministry of Education for sponsor the test, the Tao-Yuan County Government and Reui-Pu elementary school for providing so much help, the staff and students from NCREE, NTUST, MHIT, NCKU and NTU for their contribution.

References

Farsi, M.N.. (1996), “Identification des Structures de Genie Civil a Partir de Leur Reponse Vibratoire,” Vulnerabilite du bati existant. These Universite Joseph Fournier, Grenoble, December.

Ivanovic, S.S., Trifunac, M.D., Todorovska, M.I. (2000), “Ambient Vibration Tests of Structures—A Review,” Bulletin of Indian Society of Earthquake Technology: Special Issue on Experimental Methods, December.

Jennings, P.C. and Kuroiwa, J.H. (1968), “Vibration and Soil-Structure Interaction Tests of a Nine-story Reinforced Concrete Building,” Bull. Seism. Soc. Am. 58, 891-916.

Shome, N., Cornell, C. A., Bazzurro, P. and Carballo, J. E. (1998), “Earthquake, records, and nonlinear responses.” Earthquake Spectra, Vol.14, No.3, 469-497.

Stubbs, I.R. and MacLamore, V.R. (1973), “The Ambient Vibration Survey,” Proceedings of Fifth World Conference on Earthquake Engineering, Rome, Italy.

Trifunac, M.D. (1972), “Comparsion between Ambient and Forced Vibration Experiments,’ Earthquake Engineering and Structural Dynamics”, 1:133-150.

EXPERIMENTAL AND ANALYTICAL PERFORMANCE ASSESSMENT …conf.ncree.org.tw/Proceedings/i0951012/data/pdf/4ICEE-0154.pdf · EXPERIMENTAL AND ANALYTICAL PERFORMANCE ASSESSMENT OF IN-SITU

Documents