confinement of circular rc columns with fine mesh · 2013. 7. 27. · international conference on...

International Conference on Earthquake Engineering and Disaster Mitigation, Jakarta, April 14-15, 2008

CONFINEMENT OF CIRCULAR RC COLUMNS WITH FINE MESH

Tavio1, R. Purwono1 and M.L. Ashari1

1 Department of Civil Engineering, Sepuluh Nopember Institute of Technology (ITS), Surabaya, Indonesia

Email: [email protected]

ABSTRACT: The behavior of circular RC columns confined with fine mesh was investigated. The experimental investigation comprised strength and ductility tests using small-scale circular concrete column specimens with different grid spacing of fine mesh and diameter of fine mesh as lateral reinforcement. The column specimens were tested under concentric loading. The results indicate that fine mesh can be effective in confining the core concrete, resulting in significant improvements in strength and ductility of columns. These improvements were achieved even though the column specimens contained a relatively small percentage of fine mesh. Although some practical problems remain, fine mesh can potentially be used in earthquake-resistant structures as confinement reinforcement, particularly for retrofitting purposes.

1. INTRODUCTION

Tests of reinforced concrete columns have indicated that strength and ductility of concrete in compression are improved very significantly when confined by reinforcement. Concrete under high axial compression develops transverse strains due to internal cracking, but in the presence of reinforcement, the core concrete applies pressure on the steel, which in turn applies reactive pressure on the concrete.

Experimental and analytical research has been conducted in the past to investigate confinement of concrete by rectilinear ties (Sheikh, 1978). The main variables considered in that research were the size, amount, and spacing of lateral reinforcement. Prior to 1975, researchers ignored the effect of longitudinal reinforcement on concrete confinement. The effect of longitudinal reinforcement was discussed by Park and Paulay in 1975 and Vallenas et al. in 1977. However, it was not until 1978 that Sheikh and Uzumeri demonstrated the substantial improvement achieved in column strength and ductility by distributing the longitudinal steel around the core perimeter and providing a support for each bar by means of cross ties and/or hoops (Sheikh, 1978). This observation was later confirmed by large-scale column tests by Scott et al. in 1982 and Ozcebe and Saatcioglu in 1987.

It has become clear that both transverse and longitudinal bar spacings play important roles in confining the core concrete; therefore, it is reasonable to expect that concrete confinement will be increased if the concrete is placed in a cage that consists of closely spaced reinforcement in both longitudinal and transverse directions. Fine mesh (FM) appears to satisfy this requirement; however, no attempt has been made in the past to investigate the effect of FM on concrete confinement.

The use of FM as confining reinforcement in the structural members such as columns has not been investigated so far. Up to present, the implementation of FM is only limited for use in the non-structural elements, e.g., fences, animal cages, steel basket, etc. The idea to use FM as confining reinforcement in the structural components was raised from the following considerations: its economical price, higher precision (fabricated), lighter in weight, and easy to implement for existing buildings (retrofitting purposes).

The current research at the Sepuluh Nopember Institute of Technology (ITS) includes investigation of FM and WWF (Kusuma and Tavio, 2007a, b and c; Tavio et al., 2008) as confinement reinforcement. Twelve small-scale circular RC column specimens have been tested as part of this investigation. Various combinations of FM have been used as confinement steel. It is the objective of this paper to present the results of the experimental program.

Back to Table of Contents 502


2. RESEARCH SIGNIFICANCE

The importance of ductility in earthquake-resistant structures has long been recognized. However, due to the brittle nature of plain concrete, the required ductility is difficult to achieve, especially in members subjected to high compressive stresses. The research project reported in this paper deals with a new application of fine mesh, i.e., to improve ductility of concrete in reinforced concrete members. Therefore, it has a potential application to earthquake-resistant structures (Purwono and Tavio, 2007).

3. EXPERIMENTAL PROGRAM

3.1 Test Specimens

Figure 1 illustrates the geometry of a typical column specimen. A total of twelve small-scale columns were prepared in one set, each cast from the same batch of concrete, then tested under concentric axial compression. Two identical specimens were prepared for each reinforcement configuration.

150

180 mm22,92 134,16

180 mm

22,92

450 mm

CROSS-SECTION ELEVATION180 mm

grid 50 x 50 mmØ 4 mmfine mesh

C 450

Figure 1 Specimen geometry.

The length of each FM piece used in a column was between 532 and 574 mm in the transverse direction. A summary of all test specimens and their properties is provided in Table 1.

Table 1 Summary of specimen properties. Fine mesh (FM)

Column pair

cf ′ , MPa

Spacing, mm Gage

yfmf , MPa

fmρ , percent

Reinforcement configuration

C 000 26 –– N/A –– –– 1 C 225 26 27.1 × 27.1 4 678 0.44 2 C 250 26 52.1 × 52.1 4 657 0.23 3 C 325 26 28.2 × 28.2 4 690 0.95 4 C 350 26 53.2 × 53.2 4 691 0.50 5 C 450 26 54.2 × 54.2 4 697 0.83 6

Reinforcement configurations used in the specimens are illustrated in Figure 2. Configuration 1 did not include FM. Configuration 2 consisted of FM placed around the perimeter of the column with concrete cover.

The material properties, as determined from standard concrete cylinder tests and reinforcement coupon tests, are shown in Figures 3 and 4. The stress-strain relationships for FM also were obtained experimentally. The yield strength for the FM was taken as the stress corresponding to a strain of 0.35 percent as stipulated in ACI 318-08 and SNI 03-2847-2002 (Purwono et al., 2007) since no clear yield point was observed in the test results.



(1) (2) (3)

C 000 C 225 C 250

(4) (5) (6)

C 325 C 350 C 450

Figure 2 Reinforcement configurations.

0

5

10

15

20

25

30

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40STRAIN (%)

STR

ESS

(MPa

)

Figure 3 Concrete stress-strain relationship.

0

100

200

300

400

500

600

700

800

900

1000

0 1 2 3 4 5 6 7 8STRAIN (%)

STR

ESS

(MPa

)

FM 250FM 350FM 450

Figure 4 Stress-strain relationships for fine mesh.

3.2 Instrumentation

Axial deformations of columns were measured by linear variable differential transformers (LVDTs). One LVDT with a gage length of 150 mm was placed on each column face. Strains in ties were measured using electric strain gages. The data were recorded using a computerized data acquisition system.

3.3 Test Setup and Procedure

The columns were tested using a compression testing machine with a 2000-kN load capacity. Upon fixing the LVDTs, each column was placed in the center of the testing machine. An initial



load of 50 to 100 kN was applied, and the LVDTs were monitored to insure concentric loading. Shims were used when necessary to minimize accidental eccentricity.

The specimens were loaded slowly and the data were recorded at selected load and/or strain increments. The loading continued until a significant drop in load capacity was observed.

0

5

10

15

20

25

30

35

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00STRAIN (%)

STR

ESS

(MPa

)

(a)

(b)

Figure 5 Response of column pair C 250: a) stress-strain relationship and b) Column C 250 after testing.

4. OBSERVED BEHAVIOR AND TEST RESULTS

The columns showed similar response up to their peak loads. The peak load and corresponding axial strain varied somewhat depending on the confinement characteristics of the core concrete. The first set of cracks appeared on column faces at a strain of approximately 0.2 percent. These cracks propagated vertically and increased in width before the peak load was reached.

At ultimate load, concrete cover was spalled off in most of the specimens. It was noted that columns reinforced with closely spaced FM continued resisting the peak load even after the cover concrete had completely spalled off. Well-confined columns developed significant inelastic deformations at approximately the peak load level. The load resistance started dropping when bending and buckling of longitudinal bar of FM was observed. At this load stage, LVDT readings started deviating substantially from each other, indicating redistribution of the load resulting from eccentricity in columns.

The eccentricity in columns could be attributed to uneven spalling and buckling of longitudinal bars at different times. Once most of the longitudinal bars had buckled, the load was observed to be nearly concentric. The reduction in load resistance continued at different rates, depending on the confinement characteristics of specimens.



0

5

10

15

20

25

30

35

40

45

50

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00STRAIN (%)

STR

ESS

(MPa

)

(a)

(b)

Figure 6 Response of column pair C 325: a) stress-strain relationship and b) Column C 325 after testing.

Typical results of column tests are shown in Figures 5 and 6 for lightly confined and well-confined columns. All figures contain curves of average values since the test results of each individual in column pairs were in very close agreement. A summary of test results is presented in Table 2. The table contains average values for the two columns with identical test parameters within each pair. The table includes computed and measured values determined as:

Table 2 Summary of test results.

Column Po, kN

Poconc, kN

Pocore, kN

Ptest, kN

Pcmax, kN Ptest/Po Pcmax/Pocore

ε1, percent

ε85, percent ε85/ε1

C 000 664.9 664.9 –– 670 670 1.008 –– 0.195 0.281 1.439 C 225 712.3 663.0 328.7 900 851 1.263 2.588 0.278 1.306 4.691 C 250 689.0 663.9 306.0 760 735 1.103 2.402 0.270 0.595 2.205 C 325 777.0 660.5 372.6 1150 1033 1.480 2.774 0.453 1.098 2.425 C 350 723.7 662.6 334.9 950 889 1.313 2.654 0.311 1.055 3.393 C 450 767.2 660.9 365.4 1010 904 1.317 2.473 0.334 2.579 7.722

( ) yfmfmfmgco fAAAfP +−′= α (1)

( )fmgcoconc AAfP −′= α (2)

( )fmcorecocore AAfP −′= α (3)

testP = maximum column load applied in test (4)



yfmfmtestc fAPP −=max (5)

where α is the ratio of unconfined concrete strength to cylinder strength. The value of α was taken as 1.0 in computing the values given in Table 2 because of the similarities in size and shape of the column specimens and the standard cylinder. However, α varies between 0.85 and 0.90 in large-size members. Strength enhancement of core concrete due to confinement is indicated in the table by the ratio Ptest/Po or Pcmax/Pocore. Ductility of concrete is indicated in the same table by the ratio ε85/ε1. The significance of test variables and related test results are presented and discussed in the following section.

5. ANALYSIS OF TEST DATA

The test data were analyzed to investigate the significance of the variables considered. The main variables studied in the research program included spacing/grid and diameter of FM. Responses of each column pair is compared with that of different spacing/grid and diameter. The comparisons are shown in Figures 7 and 8. The results indicate that FM can be used as lateral reinforcement if the ductility is to be expected from the column. Figure 7 indicates a ductile response of column pair C 225 after reaching the peak load. It was observed during the tests that FM could provide sufficient lateral support for the longitudinal bars. At ultimate load, the pressure applied by bending and buckling longitudinal bars caused FM to rupture suddenly. The closer the spacing, the more ductile the column is.

0

5

10

15

20

25

30

35

40

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00STRAIN (%)

STR

ESS

(MPa

)

C 225C 250

Figure 7 Effect of FM spacing/grid as confinement reinforcement.

0

5

10

15

20

25

30

35

40

45

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00STRAIN (%)

STR

ESS

(MPa

)

C 250C 350C 450

Figure 8 Effect of FM size as confinement reinforcement.

Different sizes of FM were used in column pairs C 250, C 350, and C 450. The comparison of results, shown in Figure 8, indicates that while all sizes of FM produce ductile response the one with less area results in a higher rate of strength drop after the peak load.



6. CONCLUSIONS

The following conclusions can be made based on the experimental investigation reported in this paper:

1. The use of FM as confinement reinforcement improves concrete strength and ductility very significantly. Concrete strength confined with FM was observed to increase by as much as 48 percent.

2. FM is effective in improving concrete ductility since buckling of longitudinal wires is prevented by lateral wires.

3. Columns confined with FM show ductile response with a slow strength drop after the peak load. FM is potential and promising of providing the necessary lateral support to longitudinal reinforcement.

4. FM can be used as confinement reinforcement either between the longitudinal and lateral tie reinforcement, outside the reinforcement cage, or outside the existing column for retrofitting purposes.

5. For approximately the same area of steel, finer mesh produces better confinement than coarser mesh.

7. ACKNOWLEDGMENTS

The fine mesh used in the research program reported in this paper was supplied by PT. Partiwa Unggul Abadi, Surabaya, Indonesia. The experimental program was conducted at the Laboratory of Concrete and Building Materials and Structures Laboratory, Department of Civil Engineering, Sepuluh Nopember Institute of Technology (ITS), Surabaya, Indonesia.

8. NOTATION

coreA = area of core concrete enclosed by center-to-center of exterior fine mesh

gA = gross area of column cross section

fmA = area of fine mesh (FM)

d = effective depth of column section measured from extreme compression fiber to the centroid of tension steel

cf ′ = concrete cylinder strength

yfmf = yield strength of fine mesh (FM)

maxcP = maximum axial load carried by concrete during a column test

oP = computed capacity of column under concentric loading

oconcP = computed concrete contribution to column strength under pure concentric loading

ocoreP = computed core concrete contribution to column strength under pure concentric loading

testP = maximum axial load recorded during a column test

s = spacing of fine mesh (FM)

α = ratio of plain (unconfined) concrete strength in a member to concrete cylinder strength

1ε = minimum axial strain corresponding to the maximum load resistance

85ε = axial strain corresponding to 85 percent of the maximum load resistance on the falling branch of the load-strain relationship



fmρ = ratio of volume of lateral wires in FM to volume of concrete core measured center-to-center of outer fine mesh

9. REFERENCES

ACI Committee 318 (2008). “Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary (ACI 318R-08)”, American Concrete Institute, Farmington Hills, Michigan, USA, 456.

Kusuma, B. and Tavio (2007a). “Welded wire reinforcement sebagai tulangan pengekang pada kolom beton mutu tinggi”, Proceeding of the Seminar Regional Material, Desain dan Rekayasa Konstruksi pada Bangunan Tahan Gempa, Universitas Merdeka, Malang, Indonesia, 7 June, 44–49.

Kusuma, B. and Tavio (2007b). “Studi perbandingan model-model pengekangan pada kolom beton mutu tinggi terkekang welded wire reinforcement”, Proceeding of the Seminar Nasional Pascasarjana VII-2007, Institut Teknologi Sepuluh Nopember (ITS), Surabaya, Indonesia, 2 Aug., Paper No. SKH-14, 1–6.

Kusuma, B. and Tavio (2007c). “Usulan kurva tegangan-regangan beton mutu tinggi terkekang welded wire reinforcement”, Proceeding of the Seminar dan Pameran Teknik HAKI: Konstruksi Tahan Gempa di Indonesia, Hotel Borobudur, Jakarta, Indonesia, 21-22 Aug., Paper No. SPB-2, 1–13.

Kusuma, B. and Tavio (2008). “Kurva tegangan-regangan eksperimental baja kawat-las mutu-tinggi”, Proceeding of the Seminar Nasional Teknik Sipil IV-2008, Institut Teknologi Sepuluh Nopember (ITS), Surabaya, Indonesia, 13 Feb., B-102–B-110.

Ozcebe, G. and Saatcioglu, M. (1987). “Confinement of concrete columns for seismic loading”, ACI Structural Journal, Vol. 84, No. 4, July-Aug., 308–315.

Park, R. and Paulay, T. (1975). “Reinforced Concrete Structures,” John Wiley & Sons, New York, USA, 769.

Purwono, R., Tavio, Imran, I. and Raka, IG.P. (2007). “Tata Cara Perhitungan Struktur Beton untuk Bangunan Gedung (SNI 03-2847-2002) Dilengkapi Penjelasan (S-2002)”, ITS Press, Surabaya, Indonesia, Mar., 408.

Purwono, R. and Tavio (2007). “Evaluasi Cepat Sistem Rangka Pemikul Momen Tahan Gempa”, ITS Press, Surabaya, Indonesia, Sept., 51.

Scott, B.D., Park, R. and Priestley, M.J.N. (1982). “Stress-strain behavior of concrete confined by overlapping hoops at low and high strain rates,” ACI Journal, Proceedings, Vol. 79, No. 1, Jan.-Feb., 13–27.

Sheikh, S.A. (1978). “Effectiveness of Rectangular Ties as Confinement Steel in Reinforced Concrete Columns”, Ph.D. Dissertation, Department of Civil Engineering, University of Toronto, 256.

Tavio, Suprobo, P. and Kusuma, B. (2007). “Effects of grid configuration on the strength and ductility of hsc columns confined with welded wire fabric under axial loading”, Proceeding of the 1st International Conference on Modern, Construction and Maintenance of Structures, Vol. 1, 10-11 Dec., Hanoi, Vietnam, 178–185.

Vallenas, J., Bertero, V.V. and Popov, E.P. (1977). “Concrete confined by rectangular hoops and subjected to axial loads,” Report No. UCB/EERC-77/13, Earthquake Engineering Research Center, University of California, Berkeley, USA, 114.



UNIFIED STRESS-STRAIN MODEL FOR CONFINED COLUMNS OF ANY CONCRETE AND STEEL STRENGTHS

B. Kusuma1 and Tavio1

1Department of Civil Engineering, Sepuluh Nopember Institute of Technology (ITS), Surabaya, Indonesia


ABSTRACT: In this paper, a unified stress-strain model of confined concrete columns is developed and presented. The model is based on the extensively obtained data from tests of column specimens subjected to concentric compression loading. The model covers a wide range of varieties including both normal- and high-strength concretes and steels. The model is sensitive to the influencing parameters of confinement, such as concrete strength, yield strength of confining reinforcement, volumetric ratio of confining reinforcement to concrete core, spacing between confining reinforcement, cross section of confined core, configuration of lateral confining reinforcement, and distribution of longitudinal bars. The model can also be used for various concrete columns confined by spirals, crossties, and even combinations of these reinforcements. Comparison with extensive experimental data available in the literatures illustrates the validity of the proposed model in predicting the actual stress-strain curves of any confined columns of various concrete and steel strengths.

1. INTRODUCTION

The stress-strain relationships of confined concrete columns, including their empirical formulas, have long been developed and proposed by many researchers in a way to describe the actual stress-strain relationships of confined concrete columns. To date, several available analytical models for predicting the actual stress-strain relationship do not provide similar predictions on the descending branch of the curves beyond the peak stress.

The behavior of confined concrete column is normally characterized by its strength and ductility enhancements. The increase of the strength and ductility of confined concrete column highly depends on the lateral confining stress in the concrete core produced by the transverse steel. The value considerably depends on the configuration, yield strength, size, and spacing of lateral and longitudinal reinforcements.

The objective of this study is to present a comprehensive and relatively simple analytical stress-strain model for confined concrete column. The model was developed from a large database of NSC and HSC column confined by normal- and high-strength steel ties with or without crossties as well as spiral with square and circular column cross sections. The compiled database involves a total of 231 square and circular column specimens from extensive experimental tests involving monotonic and concentric axial compression loading. The concrete strength considered ranges from approximately 20 to 124 MPa, whereas the yield strength of lateral steel (fyh) ranges from about 260 to 1390 MPa. The model is shown to be applicable for a wide range of quantity and configuration of lateral reinforcement with volumetric ratio to concrete ranges from 0.2 to 5.6 percent. In addition, the peak stress and strain of the relationship are presented in a comparative study for different confinement situations.

2. AVAILABLE EXPERIMENTAL DATA FROM LITERATURE

The reviewed stress-strain models for confined NSC and HSC were applied to predict the results of experimental tests on small and large-scale specimens carried out and reported by Sheikh and Uzumeri (1980), Yong et al. (1988), Nagashima et al. (1992), Sheikh and Toklucu (1993), Cusson and Paultre (1994), Pessiki and Pieroni (1997), Saatcioglu and Razvi (1998), Razvi and Saatcioglu (1999), Li et al. (2001), Assa et al. (2001), Lin et al. (2004), and Sharma (2005). Table 1 presents the experimental work used in the analysis including 231 concrete column specimens with



different cross sections, heights, compressive strengths, and transverse reinforcements. The range of material strength is shown in Figure 1, while the range of volumetric ratio of transverse reinforcement in Figure 2.

Table 1 Experimental work included in the analysis.

Sheikh (1980) 24 Square 305 1,960 31.3 - 40.8 0.76 - 2.40 269 - 767Yong (1988) 6 Square 152 457 83.6 - 93.5 0.55 - 1.64 496Nagashima (1992) 20 Square 225 716 57.3 - 112.1 1.66 - 3.98 807 - 1,387Sheikh (1993) 27 Circular 203 - 356 812 - 1,424 35 0.58 - 2.30 452 - 629Cusson (1994) 27 Square 235 1,400 52.6 - 115.9 1.40 - 4.80 392 - 770Pessiki (1997) 8 Circular 559 2,235 37.9 - 84.7 1.32 - 2.61 476 - 537Saatcioglu (1998) 24 Square 250 1,500 60 - 124 0.99 - 4.59 400 - 1,000Razvi (1999) 16 Circular 250 1,500 60 - 124 0.41 - 3.05 400 - 1,000Li (2001) 13 Circular 240 720 52.0 - 82.5 0.82 - 2.94 445 - 1,318Assa (2001) 24 Circular 145 300 20 - 90 1.13 - 4.15 909 - 1,296Lin (2004) 24 Square 300 1,400 27.6 - 41.3 0.86 - 2.16 365 - 554

9 Square 150 600 61.85 - 83.15 2.2 - 5.62 412 - 5209 Circular 150 600 61.85 - 83.15 2.2 - 5.5 412 - 520

Sharma (2005)

Author Height (mm)

Cross section (mm)

ShapeNumber of specimens

Transverse reinforcement

Volumetric ratio (%)

Yield strength (MPa)

Compressive strength (MPa)

0200400600800

1000120014001600

0 20 40 60 80 100 120 140

f' c (MPa)

f yh (M

Pa)

Figure 1 Range of material strength in database.

0

1

2

3

4

5

6

0 20 40 60 80 100 120 140

f' c (MPa)

ρ s (%

)

Figure 2 Range of volumetric ratio of transverse reinforcement in database.

3. ANALYTICAL STRESS-STRAIN MODEL

3.1 Stress-Strain Relation

Figure 3 shows the proposed complete stress-strain curve of confined NSC and HSC column. In most of the previously proposed stress-strain models, the ascending branch was formulated by the modified Sargin’s curve (Sargin, 1971). This is because a fractional equation is a simple mathematical expression and it represents well the stress-strain relation. The equation is given by:

( ) bb

bbbccc K

Kffε

εε21

2

−+−

= for ccc εε < (1)

where; cc

cccb f

EK ε= ,

cc

cb ε

εε =

cf is the confined concrete stress at strain of cε , and and ccf ccε are the peak stress and the corresponding strain. is calculated as per ACI 318-08 equation as follows: cE



ccc fwE ′= 5.1043.0 (2)

where is the unit weight of concrete in kg/m3, and cw cf ′ is in MPa.

The descending branch of the stress-strain curve consists of a linear segment originating from the peak, as indicated by the test results from literature shown in Figure 3. The slope of this segment is defined as . The stress-strain relation of the descending branch can be determined by: desE

( cccdesccc Eff )εε −−= (3)

where, = deterioration rate, which is developed from regression analysis of test data in the range of

desE

ccε to cuε .

Figure 3 Proposed stress-strain curve and definition of ultimate strain.

3.2 Formulation of Confinement Effect

In the proposed model, expressed by Eqs. (1) and (3), factors for controlling the stress-strain relation of confined concrete are the peak stress, the strain at the peak stress and the slope of the falling branch. The effect of confinement on these three parameters was determined based on the test results from literature, as described next.

The main parameters that are likely to influence the confinement effect are the strength of concrete, yield strength of the confining reinforcement, volumetric ratio of the confining reinforcement to the concrete core as well as spacing between confining reinforcement, cross section of confined core, configuration of lateral confining reinforcement, and distribution of longitudinal bars. All these parameters are considered in the presented model.

In the study, the effective confinement index was defined as the effective lateral pressure ( ) calculated from Eq. (4).

lef

yhsele fkf ρ5.0= (4)

where, sρ is the volumetric ratio of the confining reinforcement to the concrete core; is the yield strength of the confining reinforcement; and is the modified Sheikh and Uzumeri (1982) factor for calculating the effectiveness of confinement given by the following formula.

yhf

ek

For square hoops/ties, 22

16

1 ⎟⎟⎠

⎞⎜⎜⎝

⎛−⎟

⎟⎠

⎞⎜⎜⎝

⎛−= ∑

ccc

ie b

sdbb

k (5)

For circular hoops/spirals, 5.0

1 ⎟⎟⎠

⎞⎜⎜⎝

⎛−=

ce b

sk (6)



where, is the center-to-center distance between consecutive restrained longitudinal bars, ib s the center-to-center spacing of transverse reinforcement, and the center-to-center width and height of the outer tie, respectively.

cb cd

A regression analysis was performed using the experimental results to formulate the peak strength , the strain at peak strength ( ccf ) ( )ccε , and the slope of the descending branch ( in terms of

(see Figures 4, 6, and 8). The results of regression analyses are presented in Eqs. (7) to (9), respectively.

)desE

lef

1

2

3

4

5

6

0.0 0.2 0.4 0.6 0.8 1.0

f le /f' c

f cc/f'

c

f cc /f' c = 1.0+3.7 f le /f' c

for Square and Circular; R2 = 0.888

Figure 4 Relation between the effective lateral pressure and peak stress.

Figure 4 shows the relation between an effective lateral pressure, cle ff ′ , and the ratio of peak stress to the strength of concrete, ccc ff ′ , for the circular and square experimental specimens. The following relationships are deduced from regression analysis to predict the compressive strength of confined concrete columns;

⎥⎦

⎤⎢⎣

⎡′

+′=c

leccc f

fff 7.31 (7)

Figure 5 shows the relationship between the experimentally recorded results for the confined concrete strength versus the values derived analytically from the proposed model. The correlation between the two sets of values is terrific for most of the specimens. The calculated, , value is 0.9299 that displays the fine matching between experimental and analytical results.

2R

R2 = 0.9299

020406080

100120140160180

0 20 40 60 80 100 120 140 160 180f cc Analytical - MPa

f cc E

xper

imen

tal -

MP

a

Figure 5 Correlation between experimental and analytical concrete confined strengths.

Figure 6 shows general relationship between the confining pressure ratio to the concrete strength, cle ff ′ , and the peak strain, ccε , of the experimented specimens. Clearly the cle ff ′ versus ccε

relation may be approximated by a linear function. The following relations are obtained from regression analyses:

c

lecc f

f′

+= 055.00029.0ε (8)



0.00

0.01

0.02

0.03

0.04

0.05

0.0 0.2 0.4 0.6 0.

f le /f' c

ε cc

8

ε cc = 0.0029+0.055 f le /f' c


Figure 6 Relation between the effective lateral pressure and strain at peak stress.

R2 = 0.8245

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0 0.01 0.02 0.03 0.04 0.05 0.06

ε cc Analytical

ε cc E

xper

imen

tal

Figure 7 Correlation between experimental and analytical strains at peak point.

Figure 7 shows the correlation between the prediction of ccε calculated by Eq. (8) and the

experimental results. The calculated, , value is 0.8245 that displays good performance between experimental and analytical results.

2R

Figure 8 shows a confinement index factor ( )2cyhs ff ′ρ versus the deterioration rate, . is defined as the slope of the straight line connecting the point of the peak strength and the point at which the stress drops to 50 percent of the peak strength. The following expression can approximate the test data for both circular and square sections, and approximate relation is written as

desE desE

( )22.12

cyhsdes ff

E′

=ρ

(9)

The experimental results reported by Li et al. (2001) and Assa et al. (2001), in which the value of strain along the descending branch of the confined concrete curve when stress drops to

ccf50.0 ( 50cc )ε is not available. The strain corresponding to 50 percent of the peak stress is assumed as the ultimate strain

ccf

cuε because the strain at is usually close to the point of failure due to hoop fracture and/or shear failure of the confined core (Cusson and Paultre, 1994). The definition of ultimate strain,

ccf50.0

cuε , is important.

05

10152025303540

0 3 6 9 12 15

ρ s f yh /f' c2 (x10-3)

E d

es (x

103 )

E des = 12.2f' c2 / ρ s f yh


Figure 8 Relation between confinement index factor and deterioration rate.

By substituting into Eq. (3), the ultimate strain ccc ff 50.0= cuε is obtained as



des

cccccu E

f2

+= εε (10)

Figure 9 compares the prediction of cuε calculated by Eq. (10) to the experimental results. The

calculated, , value is 0.6955 represent well between experimental and analytical results. 2RR2 = 0.6955

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.00 0.01 0.02 0.03 0.04 0.05 0.06

ε cu Analytical

ε cu

Expe

rimen

tal

Figure 9 Correlation between experimental and analytical ultimate strains.

4. COMPARISON OF PROPOSED MODEL

To show the effectiveness of the proposed model, the stress-strain relations predicted by the previous models were computed for all the test specimens, and compared against the experimental results. Curves obtained from the models of Hoshikuma et al. (1997), Kappos and Konstantinidis (1999), Legeron and Paultre (2003), and El-Dash and El-Mahdy (2006) are also shown in Figures 10 and 11 for four selected specimens, besides those from the proposed model. It is shown that significant scatter exists in the post-peak range.

0

10

20

30

40

50

60

0.000 0.005 0.010 0.015 0.020

Strain

Stre

ss (M

Pa)

SU (1980) : 2C1-16ProposedHKNT modelKK modelLP modelDM model

(a) 2C1-16 (Sheikh and Uzumeri, 1980)

020406080

100120140160180

0.000 0.005 0.010 0.015 0.020 0.025 0.030

Strain

Stre

ss (M

Pa)

NSKI (1992) : HL08LAProposedHKNT modelKK modelLP modelDM model

0

20

40

60

80

100

120

140

160

0.000 0.005 0.010 0.015 0.020 0.025

Strain

Stre

ss (M

Pa)

CP (1994) : 1DProposedHKNT modelKK modelLP modelDM model

(b) HL08LA (Nagashima et al., 1992) (c) 1D (Cusson and Paultre, 1994)

Figure 10 Comparison of stress-strain curves for square specimens.

From the figures, it can also be seen that the model derived in this study is reasonably accurate in predicting the actual response of both normal- and high-strength concrete and steel for either circular or rectangular column cross-section with various confinement types. The proposed model can simulate very well the actual response of confined column specimens, prior to and beyond the peak stress. By those plots, it is also observed that the proposed curve predicts the peak stress, the strain at the peak stress and the deterioration rate reasonably better than the previous models.



0

20

40

60

80

100

120

0.000 0.005 0.010 0.015 0.020 0.025

Strain

Stre

ss (M

Pa)

LPT (2001) : 6HBProposedHKNT modelLP modelDM model

0

20

40

60

80

100

120

140

0.000 0.005 0.010 0.015 0.020

Strain

Stre

ss (M

Pa)

SBK (2005) : CFProposedHKNT modelLP modelDM model

(a) 6HB (Li et al., 2001) (b) CF (Sharma et al., 2005)

Figure 11 Comparison of stress-strain curves for circular specimens.

From the previous comparison, it may be said that the proposed model generally provides better agreement with the stress-strain relation of confined concrete over a wider range of concrete strength, yield strength of confining reinforcement, volumetric ratio of confining reinforcement to concrete core, spacing between confining reinforcement, cross-section of confined core, configuration of lateral confining reinforcement, and distribution of longitudinal bars than the previous models.

The model proposed by Kappos and Konstantinidis (1999) is not capable to predict the confined column specimens with circular cross-section. This is because the proposed model, which was derived from the regression analysis, was mainly based on a few test data of rectangular column specimen only. All the predictions from the other models indicate that the actual stress-strain relationships, except those from Legeron and Paultre (2003) model, overestimate the strength gain of confined concrete columns.

5. CONCLUSIONS

A unified model is presented to predict the stress-strain relationship for normal- and high-strength concrete columns of square and circular cross-sections confined with spirals, ties, and/or crossties. The model is based on the experimental results of 231 concrete column specimens subjected to different types and amounts of transverse reinforcement and tested under concentric loading. Comparisons are made between the predictions of the model and the available experimental results. It can be concluded from the study that:

1. The model demonstrates good predictive capability and is applicable for a wide range of variables that include range of concrete compressive strength from 20 to 124 MPa and yield strength of confining reinforcement from 270 to 1390 MPa.

2. All models predict the ascending branch of the stress-strain curve fairly well, whereas the predicted descending branch is not consistent, except for the proposed model.

3. The statistical model proposed in the present study, which is based on a large experimental database, resulted in lower uncertainties than other models for most parameters.

4. The maximum strength of confined concrete is well predicted by the proposed than the other models as well as the strain along the descending branch of the confined concrete curve when stress drops to ccf50.0 .

5. The proposed model provides more reasonable values than the other relations.

6. REFERENCES

ACI Committee 318 (2008). “Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary (ACI 318R-08)”, American Concrete Institute, Farmington Hills, Michigan, USA, 456.

Assa, B., Nishiyama, M. and Watanabe, F. (2001). “New approach for modeling confined concrete i: circular columns”, Journal of Structural Engineering, ASCE, 127(7), 743-750.

Cusson, D. and Paultre, P. (1994). “High strength concrete columns confined by rectangular ties”, Journal of Structural Engineering, ASCE, 120(3), 783-804.



International Conference on Earthquake Engineering and Disaster Mitigation 2008

El-Dash, K.M. and El-Mahdy, O.O. (2006). “Modeling the stress-strain behavior of confined concrete columns”, International Symposium on Confined Concrete, ACI, SP-238, China, 177-192.

Hoshikuma, J., Kawashima, K., Nagaya, K. and Taylor, A.W. (1997). “Stress-strain model for confined reinforced concrete in bridge”, Journal of Structural Engineering, ASCE, 123(5), 624-633.

Kappos, A.J. and Konstantinidis, D. (1999). “Statistical analysis of confined high strength concrete”, Materials and Structures, 32, 734-748.

Li, B., Park, R. and Tanaka, H. (2001). “Stress-strain behavior of high-strength concrete confined by ultra-high- and normal-strength transverse reinforcements”, ACI Structural Journal, 98(3), 395-406.

Lin, C.H., Lin, S.P. and Tseng, C.H. (2004). “High-workability concrete columns under concentric compression”, ACI Structural Journal, 101(1), 85-93.

Legeron, F. and Paultre, P. (2003), “Uniaxial confinement model for normal-and high-strength concrete columns”, Journal Structural Engineering, 129(2), 241-252.

Nagashima, T., Sugano, S., Kimura, H. and Ichidawa, A. (1992). “Monotonic axial compression test on ultra-high strength concrete tied columns”, Proceedings of 10th World Conference on Earthquake Engineering, Madrid, 5, 2983-2988.

Pessiki, S. and Pieroni, A. (1997). “Axial load behavior of large-scale spirally-reinforced high-strength concrete columns”, ACI Structural Journal, 94(3), 304-314.

Razvi, S.R. and Saatcioglu, M. (1999). “Circular high-strength concrete columns under concentric compression”, ACI Structural Journal, 96(5), 817-825.

Sargin, M., Ghosh, S.K. and Handa, V.K. (1971). “Effects of lateral reinforcement upon the strength and deformation properties of concrete”, Magazine of Concrete Research, 23(75-76), 99-110.

Sheikh, S.A. and Uzumeri, S.M. (1980). “Strength and ductility of tied concrete columns”, Journal of Structural Division, ASCE, 106(ST5), 1079-1101.

Sheikh, S.A. and Uzumeri, S.M. (1982). “Analytical model for concrete confinement in tied columns”, Journal of Structural Division, ASCE, 108(ST12), 2703-2722.

Sheikh, S.A. and Toklucu, M.T. (1993). “Reinforced concrete columns confined by circular spirals and hoops”, ACI Structural Journal, 90(5), 542-553.

Saatcioglu, M. and Razvi, S.R. (1998). “High-strength concrete columns with square section under concentric compression”, Journal of Structural Engineering, ASCE, 124(12), 1438-1447.

Sharma, U.K., Bhargava, P. and Kaushik, S.K. (2005). “Behavior of confined high strength concrete columns under axial compression”, Journal of Advanced Concrete Technology, 3(2), 267-281.

Yong, Y.K., Nour, M.G. and Nawy, E.G. (1988). “Behavior of laterally confined high strength concrete under axial load”, Journal of Structural Engineering, ASCE, 114(2), 332-351.



STUDY OF CONFINEMENT MODELS FOR HIGH-STRENGTH CONCRETE COLUMNS CONFINED BY HIGH-STRENGTH STEEL

Zulfikar Djauhari1 and Iswandi Imran2

1PhD Student of Civil Engineering Study Program at Institut Teknologi Bandung, Indonesia 2Faculty of Civil and Environmental Engineering, Institut Teknologi Bandung, Indonesia


ABSTRACT: The analysis of structural members requires an analytical model for the full stress-strain relationship of concrete in compression, both in confined and unconfined states. Many empirical confinement models for normal strength materials have been developed and documented in the literature during the last three decade. However, the models developed for normal strength material are not applicable to high-strength material. This paper reviews the various proposed stress-strain model for high-strength concrete materials confined by high-strength steel.

The main objective of the study is to examine the capabilities of the various models available in literature to predict the actual experimental behavior of high-strength concrete columns confined by high-strength steel. The experimental data used are the results of the tests conducted by the author, involving testing of 18 short column specimens with 110 mm diameter circular section confined by 6 mm spirals or hoops made of wire rod reinforcement. The test variables include yield strength, type (spiral or hoops), spacing and volumetric ratio of confining steel. The resulting stress-strain curves from the tests are then compared with the various models available in the literature. It is shown from this study that there is no models which can accurately predict the complete behavior of high-strength concrete columns confined with high-strength steel. 1. INTRODUCTION

The most fundamental requirement in predicting the behavior of reinforced concrete structures is the knowledge of stress-strain behavior of the constituent materials. As concrete is basically used to resist compression, the knowledge of its behavior in compression is very important. If the behavior of unconfined and confined concrete in uniaxial compression is known, its flexural behavior can be predicted. A considerable volume of research has been directed towards generating the stress-strain relationship for compressed concrete both in confined and unconfined state. As a result, many empirical stress-strain models have been proposed for confined concrete, both for normal and high-strength concretes.

Inelastic deformability of reinforced concrete columns is essential for overall stability of structures in order to sustain strong earthquakes. This can be achieved through proper confinement of the core concrete. The increase in the strength and ductility of normal and high-strength concrete confined by normal strength steel has been well documented (Assa et al., 2001; Li et al., 2001; Legeron and Paultre, 2003; Sharma et al., 2005; Hong et al., 2006). Question, however, have been asked as to whether a similar amount of confinement is suitable for high-strength concrete columns confined by high-strength steel. Existing code provisions for minimum amount of confining reinforcement are based on experiences with normal strength materials (ACI 318, 2005). However, the influence of the increase in material strength used needs to be taken into account. This initiates research studies on the effect of high strength materials on the behavior of confined concrete in various countries. Most of these studies have unanimously concluded that high-strength concrete columns need a considerably higher amount of confinement to attain the level of ductility that is commonly achieved by a nominal amount of confining reinforcement in normal-strength concrete columns (Sharma et al., 2005). It means that the confining pressure required for high-strength concrete columns is significantly higher than that for normal-strength concrete columns. The research findings have also indicated that for well confined columns, increasing the yield strength of confining steel results in an increase in strength and ductility, and can alleviate


mailto:[email protected]


the congestion of reinforcement (Li et al., 2001; Legeron and Paultre, 2003; Sharma et al., 2005). So, the higher confinement requirement of high-strength concrete columns can be satisfied by either increasing the volumetric ratio of lateral steel or by using higher grade of lateral steel. Li et al. (2001) have recommended the use of lateral steel grades above 1000 MPa for high-strength concrete columns.

2. SUMMARY OF CONFINEMENT MODELS

Many confinement models had been proposed for high-strength concrete columns confined by high strength steel (Muguruma et al., 1993; Cusson and Paultre, 1995; Azizinamini et al., 1994; Nakatsuka et al., 1995; Razvi and Saatcioglu, 1999; Li et al., 2001; Legeron and Paultre, 2003; and Hong et al., 2005). However, the models reported had various limitations. For example, Li’s model is only applicable to certain range of concrete compressive strength. Moreover, Azizinamini and Hong’s concluded in their studies that there is no significance effect on the use of high-strength steel for confining reinforcement. All the models, except Li’s model, do not take into account a reinforcement type as a parameter. Among the models available in the literature, Legeron’s model is the most unified, widely accepted and with a wider scope than those of other models.

Cusson and Paultre (1995) developed a confinement model for high-strength concrete on the basis of test results of 50 large-scale high-strength concrete tied columns tested under concentric loading. Out of them, 30 high-strength concrete tied columns (225 x 225 mm) were tested by authors themselves and 20 high-strength concrete tied columns (225 x 225 mm) were tested by Nagashima et al. (1992). The concrete compressive strengths of the specimens ranged from 60 to 120 MPa. The ties with yield strength from 400 to 800 MPa were used. The proposed model takes into account tie yield strength, tie configuration, and longitudinal reinforcement ratio. Two-part of stress-strain relationship with separate expressions for ascending and descending parts were formulated in their study (Figure 1).

Azizinamini et al. (1994) modified the model developed by Yong et al. (1988) based on their test data, as well as test data obtained by Yong et al. The model consists of a linear ascending branch, followed by a linear descending branch with a constant residual at 30% of peak stress (Figure 1). The model was developed on the basis of test results of nine 2/3-scale column specimens with square cross section. The main test parameters were in place concrete strength (54 to 101 MPa), spacing of confining reinforcement (41.3 to 66.7 mm), axial load, and yield strength of confining reinforcement (414 to 828 MPa).

Razvi and Saatcioglu (1999) proposed a model for confined normal and high strength concrete columns using extensive test data from the authors own test results as well as experimental results from other research studies. This included the test results of nearly full size specimens of different shapes, sizes, reinforcement configuration, tie yield strength (ranging from 400 to 1387 MPa) and concrete strength (ranging from 30 to 130 MPa). The parameters incorporated in the model were type, volumetric ratio, spacing, yield strength, and arrangement of transverse reinforcement, distribution and amount of longitudinal steel as well as concrete strength and geometry. The two-part stress-strain model proposed by the authors is in the form of the ascending parabolic branch up to peak and a linear descending branch up to 20% of the peak stress (Figure 1).

Silva (2000) proposed a model for circular confined high strength concrete columns. He tested 15 circular columns that were made of normal and high strength concrete with compressive strength ranging from 35.5 to 125.4 MPa, confined with spiral reinforcement with yield strength of 440 to 560 MPa. The variable considered in the tests were diameter, spacing, volumetric ratio and yield strength of confining reinforcement. A two-part stress-strain relation was proposed to predict the constitutive behavior of confined high-strength concrete. He developed the confinement model based on the Popovic’s model (1973) for ascending branch and Fafitis and Shah’s model (1985) for descending branch.

Assa et al. (2001) proposed the steel concrete interaction model to predict the strength and stress-strain curves of confined concrete. This model is based on the confining stiffness of transverse reinforcement. A total of thirty-two 150 x 300 mm concrete cylinders were tested under monotonic



concentric compression. No concrete cover was provided in all the specimens. The target strength of concrete ranged from 20 MPa to 90 MPa. The confining reinforcement of 6.25 mm helical spirals and welded circular hoops were used. The nominal yield strengths of the confining reinforcement were 1300 MPa and 800 MPa, respectively. The spacing of spirals or circular hoops was varied from 19 mm to 75 mm. The test results can be used to model the stress-strain behavior of concrete confined by several types of reinforcement configurations. If the lateral pressure for any lateral strain is theoretically obtained for any transverse reinforcement configuration as indicated by lateral pressure-lateral strain line, the lateral pressure at peak load can be easily obtained at the intersection point between the peak load condition line and lateral pressure-lateral strain line. Then, stress-strain coordinate at peak load can be obtained.

Li et al. (2001) proposed a three-part stress-strain model for high-strength concrete confined by high-strength steel based on their experimental results (Figure 1). Fourty reinforced concrete short columns of both cylindrical (240 mm diameter) and square (240 x 240 mm) cross sectional shapes were tested. The main parameters were in place concrete strength (35.2 to 82.5 MPa) and lateral steel grade (445 and 1318 MPa). Other parameters like spacing and volumetric ratio of lateral steel, were also varied in the tests. From test results, a confinement model was developed based on Muguruma’s model (1993).

Legeron and Paultre (2003) proposed a stress-strain confinement model for normal and high-strength concrete columns based on the large number of test results of circular, square, and rectangular columns tested by themselves and by a number of other researchers. The concrete compressive strength ranged from 20 to 140 MPa and tie yield strength ranged from 300 to 1400 MPa. The model incorporates almost all the parameters of confinement. The stress-strain relationship is basically the same as that proposed by Cusson and Paultre (1995), but the parameters of the model were recalibrated on the basis of large number of test data collected by the authors.

Most of those models have limited validity in terms of concrete strengths, column geometry, transverse reinforcement yield strength and loading conditions. With exception of the models proposed by Razvi and Saatcioglu (1999), Li et al. (2001), Legeron and Paultre (2003), which cover both circular and rectilinear sections, all other models are applicable to only square or rectilinear shapes.

It has now been proven experimentally by many researchers that for high-strength concrete columns confined with high yield strength ties, lateral confining ties may not yield when the peak of confined concrete stress-strain is reached (Li et al., 2001; Cusson and Paultre, 2003). Even for lower yield strength of lateral confining steel, if enough degree of confinement is not provided, the transverse steel may not yield at peak confined strength. However, most of the models use yield strength of lateral steel to calculate lateral confining pressure at peak confined strength. Only Razvi and Saatcioglu (1999) and Legeron and Paultre (2003) have incorporated this fact into their respective models by proposing procedures to calculate actual confining stress at peak of confined stress-strain response. Li et al. (2001) has also accounted for this indirectly by suggesting a different expression for confined strength when higher grades of lateral steel are to be used, but no explicit expression for finding the confining stress at peak is proposed. In addition, none of the models take into account all the loading conditions, namely monotonic, cyclic, strain rate, and eccentric loading.

A critical review of the expression for ascending and descending portions proposed in the various confinement models indicates that the basic forms of these expressions are mostly the same and only the evaluation of the parameters differed from model to model. The expression of ascending branches as proposed by Li et al. (2001) is basically the same with those proposed by Muguruma et al. (1993). The ascending curves proposed by Cusson and Paultre (1995), Razvi and Saatcioglu (1999), Legeron and Paultre (2003) are also quite similar. The expression of descending curves proposed by Muguruma et al.(1993), Razvi and Saatcioglu (1999), Li et al. (2001) are basically the same and those given by Cusson and Paultre (1995) and Legeron and Paultre (2003) are also similar.



Figure 1 Example of stress-strain relationship of confined high strength concrete.

3. COMPARISON OF ANALYTICAL AND EXPERIMENTAL STRES-STRAIN BEHAVIOR

In this section, an attempt has been made to investigate the relative performance of the various proposed analytical models with regard to their capabilities of producing the experimentally observed stress-strain profiles on different test specimens. For this purpose, eighteen short column specimens (height of 600 mm) with 110 mm diameter circular section confined by 6 mm spirals or hoops made of wire rod reinforcement tested by Zulfikar et al. (2006) are used. The key parameters in the tests include yield strength, type (spiral or hoops), spacing and volumetric ratio of confining steel. The test parameters of the test specimens are shown in Table 1.

It should be noted that all the six confinement models of the study are applicable to square sections whereas only three, namely Razvi and Saatcioglu’s model (1999), Li et al.’s model (2001), Legeron and Paultre’s model (2003), can also be applied to circular sections. Figure 2 to 9 illustrate the uniaxial stress-strain curves obtained using the various models for circular columns.

Table 1 The test parameters.

Pitch Volumetric ConfinementConcrete Confining Reinforcement (mm) ratio Index

1 2 3 4 5 6 7 81 S400P30 30 0.0419 0.242 S400P60 60 0.0209 0.123 S800P30 30 0.0419 0.484 S800P60 60 0.0209 0.245 S800P120 120 0.0105 0.126 S960P30 30 0.0419 0.577 S960P60 60 0.0209 0.298 S960P70 70 0.0180 0.259 S960P120 120 0.0090 0.12

10 H400P30 30 0.0419 0.2411 H400P60 60 0.0209 0.1212 H800P30 30 0.0419 0.4813 H800P60 60 0.0209 0.2414 H800P120 120 0.0105 0.1215 H960P30 30 0.0419 0.5716 H960P60 60 0.0209 0.2917 H960P70 70 0.0180 0.2518 H960P120 120 0.0090 0.12

960

Spiral

70

400

800

960

Hoop

400

800

No Confining Type Strength (MPa)Specimen



In order to judge the predictive capability of various analytical models, the computations were carried out for the peak confined stress fcc and the corresponding strain εcc for all the specimens. The ratios of the predicted peak confined strengths to respective experimental peak values for all the specimens are shown in Table 2. The corresponding ratios for peak strains are shown in Table 3. It should be emphasized that the post-peak description of the stress-strain curve is of greater significance in order to ascertain the ductility and failure related aspects of concrete columns. Keeping this fact in mind, the study also aimed to determine the strain values at one specified stress levels in descending portion; i.e. 85% of the peak stress fcc. It is apparent from Figure 1 that linear or exponential descending branch belonged to most of the models employs either the definition of strain at 85% of peak stress, εc85, or strain at 50% of peak stress, εc50, or ultimate failure strain εu as an alternate means to characterize the post-peak curve. These expressions for softening branches enabled us to compute εc85 and εc50 strain for various models. Most of the reported experimental stress-strain curves of the test specimens do not reach down to a strain value of 50% of the peak stress. So, for these cases, no comparison is possible at strain of 50% of peak stress. The comparison at post-peak is only made at strain of 85% of peak stress. Table 4 show the ratios of these predicted to experimental post-peak strains.

Most of the confinement models assume that the lateral reinforcement stress to attain its yield value at the confined strength of concrete columns is irrespective of whether a high yield strength or normal yield strength reinforcement are being used. However, some of the recently developed confined models propose a method to work out the actual stress level in the lateral confining reinforcement. This issue has also been addressed in the present work by computing the ratios of predicted confining reinforcement stress at peak by various models to real stress of confining reinforcement obtained from the test, as shown in Table 5.

Table 2 Comparisons for peak confined stress fcc.

Assa Azizinamini Legeron Li Razvi SilvaS400P30 88.61 1.11 0.92 1.05 0.95 1.04 0.98S400P60 73.03 1.15 1.05 1.16 0.99 1.13 0.95S800P30 100.06 1.26 0.92 1.08 1.18 1.09 1.21S800P60 80.67 1.22 1.03 1.16 0.99 1.14 1.07S800P120 70.76 1.19 1.11 1.19 1.00 1.16 0.98S960P30 125.94 1.09 0.77 0.90 1.09 0.92 1.07S960P60 83.30 1.25 1.03 1.16 1.01 1.15 1.12S960P70 78.24 1.26 1.08 1.20 1.01 1.18 1.12S960P120 71.80 1.21 1.12 1.20 0.99 1.18 1.01H400P30 86.23 1.11 0.92 1.05 0.93 1.04 0.98H400P60 71.65 1.15 1.05 1.16 0.99 1.13 0.95H800P30 99.55 1.26 0.92 1.08 1.06 1.09 1.21H800P60 73.91 1.22 1.03 1.16 1.01 1.14 1.07H800P120 68.65 1.19 1.11 1.19 1.02 1.16 0.98H960P30 102.50 1.09 0.77 0.90 1.18 0.92 1.07H960P60 75.50 1.25 1.03 1.16 1.01 1.15 1.12H960P70 72.46 1.26 1.08 1.20 1.01 1.18 1.12H960P120 70.38 1.21 1.12 1.20 1.00 1.18 1.01

5.20 10.99 8.52 6.53 7.44 7.75

Specimen fcc exp (MPa) fcc predicted/fcc experiment

COV (%)



Table 3 Comparisons for peak confined strain εcc.

Assa Azizinamini Legeron Li Razvi SilvaS400P30 0.0062 1.74 0.76 1.07 2.41 0.72 1.89S400P60 0.0055 1.25 0.63 0.83 0.89 0.64 0.94S800P30 0.0178 1.04 0.33 0.63 1.12 0.34 1.84S800P60 0.0072 1.50 0.55 0.92 3.74 0.62 1.63S800P120 0.0047 1.45 0.54 0.96 3.19 0.75 1.09S960P30 0.0346 0.62 0.18 0.38 1.29 0.19 1.27S960P60 0.0035 3.48 1.17 2.12 8.37 1.36 4.28S960P70 0.0208 0.53 0.18 0.32 1.27 0.22 0.58S960P120 0.0059 1.30 0.44 0.84 2.77 0.63 1.06H400P30 0.0081 1.74 0.76 1.07 1.39 0.72 1.89H400P60 0.0036 1.25 0.63 0.83 1.06 0.64 0.94H800P30 0.0099 1.04 0.33 0.63 1.72 0.34 1.84H800P60 0.0070 1.50 0.55 0.92 3.19 0.62 1.63

H800P120 0.0046 1.45 0.54 0.96 2.18 0.75 1.09H960P30 0.0229 0.62 0.18 0.38 1.80 0.19 1.27H960P60 0.0211 3.48 1.17 2.12 1.17 1.36 4.28H960P70 0.0020 0.53 0.18 0.32 10.70 0.22 0.58

H960P120 0.0047 1.30 0.44 0.84 2.27 0.63 1.0658.55 56.54 56.66 93.13 56.27 65.08

εcc predicted/εcc experimentSpecimen εcc exp

COV (%)

S400P30

0

20

40

60

80

100

120

0 0.02 0.04 0.06 0.08

Strain

Stre

ss (M

Pa) Assa (2001)

Azizinamini (1994)Legeron (2003)Li (2004)Razvi (1999)Silva (2000)Eksperimen

S400P60

0

20

40

60

80

100

0 0.01 0.02 0.03 0.04

Strain

Stre

ss (M

Pa) Assa (2001)

Azizinamini (1994)Legeron (2003)Li (2004)Razvi (1999)Silva (2000)Experiment

Figure 2 Stress-strain curve for specimens S400P30. Figure 3 Stress-strain curve for specimens S400P60.

S800P30

020406080

100120140

0 0.03 0.06 0.09 0.12 0.15

Strain

Stre

ss (M

Pa) Assa (2001)


S800P60

0

20

40

60

80

100

120

0 0.02 0.04 0.06 0.08

Strain

Stre

ss (M

Pa) Assa (2001)



S960P30

020406080

100120140160

0 0.05 0.1 0.15 0.2

Strain

Stre

ss (M

Pa) Assa (2001)


S960P60

0

20

40

60

80

100

120

0 0.02 0.04 0.06 0.08

Strain

Stre

ss (M

Pa) Assa (2001)





H400P60

0

20

40

60

80

100

0 0.01 0.02 0.03 0.04

Strain

Stre

ss (M

Pa) Assa (2001)


H960P60

0

20

40

60

80

100

120

0 0.02 0.04 0.06 0.08

Strain

Stre

ss (M

Pa) Assa (2001)


Figure 8 Stress-strain curve for specimens H400P60. Figure 9 Stress-strain curve for specimens H960P60.

4. DISCUSSION

The relative performance of the various confinement models of high-strength concrete columns confined by high-strength steel in terms of their capabilities to predict the actual confined compressive test behavior has been evaluated. The behavior parameters considered are experimental peak confined strengths fcc, corresponding peak strain εcc, post-peak strains εc85, stress level in confining reinforcement fs at peak and stress-strain curves. In the absence of any conclusive pattern with respect to these parameters for the selected specimens, the coefficient of variation concept has been employed to quantify the performance index of each model.

The analysis using Azizinamini et al.’s model (1994) indicates that this model, in general, underestimate the peak confined strength, fcc, and the associated peak strain, εcc, and post-peak strain, εc85, of the test results. The coefficient of variation in estimating peak confined stress, fcc, for this model are the largest among all models considered (i.e. 10.99 % (Table 2)). This model proposes that the confining reinforcement always yield as the peak confined strength is reached. Therefore, the model is not able to predict the actual level of spiral or hoop stress at the peak confined stress for those specimens having higher spiral content or hoop yield strengths.

The Razvi and Saatcioglu model (1999) has the advantages of being applicable to all the cross-sectional shapes, covering a wider concrete strength range (60-124 MPa) and considers the implications of using high-strength steel, although it suffers from the limitations of being unable to produce an ascending and descending branch that is in agreement with the experimental results. However, the peak stress fcc parameter obtained from this model shows close agreement with the test values (Table 2). Nevertheless, the predictions of post-peak strain, εc85, and descending branches of predicted stress-strain curves indicate that the model consistently overestimate with considerable magnitude the actual post-peak behavior. But, strain at peak stress, εcc, is underestimated by the model. The procedure adopted in the model to compute the actual spiral or hoop stress at peak, fs appears to be more rational as its predictions for the specimens are better if not more accurate than those given by the other models considered in this study. The coefficient of variation in estimating the actual spiral or hoop stress at peak from this model is the minimum of all models (i.e. 43.12% (Table 5)).

The Li et al. model (2001) can predict stress-strain curves both for circular and square confined specimens. Its prediction of the peak confined strength, fcc is close enough to the test values for almost all test specimens (Table 2). It is evident from the study that the model is quite erroneous in predicting peak strains, εcc. The model produces the largest COV in this case (Table 3). For the specimens confined by normal and high strength steel in particular, the errors corresponding to the estimation of post-peak strain parameters εc85 are also significant (Table 4). A separate expression for calculating peak strength of column confined by high-strength steel is included in this model to account for delayed confining effect in such cases. So, the fact that high-strength steel spiral or hoop in high-strength concrete columns may not yield at peak is taken into account, but still the model give more erroneous results for specimens having high-strength confining steel. As no specific expression is given to calculate the level of stress in the lateral confining reinforcement at peak confined strength, then the ratio of actual lateral spiral or hoop stress at peak to yield strength can not be obtained for this model. The present study shows that this model is capable of



predicting reasonably the peak strength of high-strength concrete columns confined by high-strength steel only.

The Legeron and Paultre model (2003) can also be used to predict the stress-strain behavior of columns of all cross sectional shapes. This model predicted both peak strength, fcc and peak strain, εcc of the test value remarkably well. However, the stress-strain curves predicted by the model overestimate the experimental stress-strain curves for almost all specimens (Figure 2 to 9). The expression proposed in the model to calculate the actual level of confining reinforcement stress at peak confined strength produces higher estimate than the experimentally reported values as indicated in Table 5.

Table 4 Comparisons for strain at 85% of peak stress εc85.

Assa Azizinamini Legeron Li Razvi SilvaS400P30 0.0155 1.82 1.33 2.37 1.68 1.73 1.11S400P60 0.0036 8.12 4.16 4.01 2.41 3.49 2.03S800P30 0.0205 1.88 1.39 4.78 1.70 2.53 2.08S800P60 0.0090 3.13 1.99 4.09 5.20 2.31 1.92S800P120 -S960P30 0.0368 1.17 0.86 3.35 2.14 1.68 1.45S960P60 0.0570 0.53 0.34 0.85 0.90 0.42 0.39S960P70 0.0200 1.42 0.86 1.92 2.31 0.99 0.90S960P120 0.0100 2.72 1.16 1.78 2.84 1.11 0.89H400P30 0.0137 2.06 1.50 2.69 1.44 1.96 1.26H400P60 0.0088 3.29 1.69 1.62 0.77 1.41 0.82H800P30 0.0205 1.88 1.39 4.78 1.46 2.53 2.08H800P60 0.0064 4.41 2.80 5.75 6.13 3.25 2.69H800P120 0.0100 2.88 1.14 1.42 1.75 1.00 0.72H960P30 0.0714 0.60 0.44 1.73 1.01 0.86 0.75H960P60 0.0159 1.88 1.20 3.03 2.70 1.50 1.38H960P70 0.0350 0.81 0.49 1.10 1.07 0.57 0.51H960P120 -

76.73 67.13 52.58 67.28 53.44 51.31COV (%)

Specimen εc85 expεc85 predicted/εc85 experiment

Table 5 Comparisons for stress in confining reinforcement at peak stress.

fs exp (MPa) Cusson Hong Legeron RazviS400P30 570.67 0.94 0.70 0.70 1.79S400P60 211.60 2.40 1.89 1.89 4.11S800P30 483.69 2.68 1.41 1.41 2.43S800P60 843.81 0.78 0.81 0.81 1.19S800P120 735.12 0.62 0.93 0.93 1.23S960P30 525.17 4.18 1.78 1.78 1.69S960P60 293.14 1.15 3.20 3.20 2.57S960P70 261.05 6.43 3.59 3.59 2.80S960P120 385.94 1.47 2.43 0.63 1.68H400P30 273.20 2.47 1.46 1.46 3.75H400P60 221.00 1.55 1.81 1.81 3.94H800P30 444.75 1.85 1.54 1.54 2.65H800P60 444.75 1.44 1.54 1.54 2.25H800P120 347.24 1.27 1.97 1.97 2.61H960P30 179.99 8.68 5.20 5.20 4.93H960P60 319.97 5.10 2.93 2.93 2.36H960P70 228.10 0.85 4.11 4.11 3.21H960P120 764.95 0.59 1.22 0.32 0.85COV (%) 90.71 57.09 66.33 43.12

Specimenfs predicted/ fs experiment

5. CONCLUSIONS

The use of high strength steel is now being advocated for confining reinforcement in concrete columns to compensate for the lower ductility of high-strength concrete columns. Given the fact



that confining reinforcement may not yield when the peak confined strength of columns is reached, there are only few models available in the literature which include procedures to calculate the actual stress of confining reinforcement. Only the models proposed by Razvi and Saatcioglu (1999) and Legeron and Paultre (2003) cover a wide range of concrete strength, is applicable to all the cross-sectional shapes and can be used for both normal and high-strength confining steel.

A comparative study was undertaken to evaluate the capabilities of the various confinement models of high-strength concrete columns available in literature in predicting the actual experimental behavior of the specimens tested by the author. The study indicated that almost all the models are able to correctly estimate the ascending part of the experimental stress-strain curve. However, there are wide variation in the prediction of peak strength, peak strain, post peak strain and the descending part of stress-strain curves, with a few models underestimating and a few overestimating the test values.

There is no model which correctly predicts the stress in the confining reinforcement at peak confined stress. The resulting coefficient of variation in predicting experimental values for all specimens ranges from 43.12 to 90.71. The large values of COV show that the models considered are not accurate enough in predicting the stress in confining reinforcement at peak.

6. REFERENCES

ACI Committee 318 (2005). “Building Code Requirements for Structural Concrete ACI 318 and Commentary, American Concrete Institute”, Farmington Hills, Mich.

Assa, B., Nishiyama, M. and Watanabe, F. (2001). “New approach for modeling confined concrete I : circular columns”, Journal of Structural Engineering, V.127, No.7, July, pp 743-750.

Azizinamini, A., Baum Kuska, S.S., Brungardt, P. and Hatfield, E. (1994). “Seismic behavior of square high-strength concrete columns”, ACI Structural Journal, V.91, No.3, May-June, pp. 336-345.

Cusson, D. and Paultre, P. (1995). “Stress-strain model for confined high-strength concrete”, Journal of Structural Engineering, V.121, No.3, March, pp. 468-477.

Fafitis, A. and Shah, S.P. (1985). “Lateral reinforcement for high-strength concrete columns”, SP-87-12, American Concrete Institute, Detroit, pp. 213-232.

Hong, K.N., Akiyama, M., Yi, S.T. and Suzuki, M. (2006). ” Stress-strain behavior of high-strength concrete columns confined by low-volumetric ratio rectangular ties”, Magazine of Concrete Research, V.58, No.2, March, pp. 101-115.

Legeron, F. and Paultre, P. (2003). “Uniaxial confinement model for normal- and high-strength concrete columns”, Journal of Structural Engineering, V.129, No.2, February, pp. 241-252.

Li, B., Park, R. and Tanaka, H. (2001). ”Stress-strain behavior of high-strength concrete confined by ultra-high- and normal-strength transverse reinforcement”, ACI Structural Journal, V. 98, No. 3, May-June, pp. 395-406.

Mugurama, H., Nishiyama, M. and Watanabe, F. (1993). “Stress-strain curve model for concrete with a wide-range of compressive strength”, Proceeding High-Strength Concrete, Lillehammer, Norway, pp. 314-321.

Nakatsuka, T., Nakagawa, H. and Suzuki, K. (1995). “Strength – deformation characteristic of confined concrete and spiral reinforcement: of high strength”, Concrete 95 toward Better Concrete Studies, pp. 427-469.

Popovics, S. (1973). “Analytical approach to complete stress-strain curves”, Cement and Concrete Research, 3 (5), pp. 583-599.

Razvi, S. and Saatcioglu, M. (1999). ”Confinement model for high-strength concrete”, Journal of Structural Engineering, V. 125, No. 3, March, pp. 281-289.

Sharma, U., Bhargava, P. and Kaushik, S.K. (2005). “Comparative study of confinement models for high-strength concrete columns”, Magazine of Concrete Research, V. 57, No. 4, May, pp. 185-197.




Silva, P. (2000). “Effect of Concrete Strength on Axial Load Response of Circular Columns”, Department of Civil Engineering and Applied Mechanics, Mc Gill University, Montreal, Canada.

Yong, Y.K., Nour, M.G. and Nawy, E.G. (1988). “Behavior of laterally confined high-strength concrete under axial loads”, Journal of Structural Engineering, V. 114, No. 2, February, pp. 332-350.

Zulfikar, D., Imran, I. and Setio, H.D. (2007). ”Strength and ductility behaviour of high-strength concrete confined by high-strength steel (in Indonesia)”, Proceeding Earthquake Resistant Construction in Indonesia, Jakarta, pp. 23-34.



STUDY ON THE SEISMIC PERFORMANCE CURVES OF REINFORCED CONCRETE SHORT COLUMNS FAILED IN SHEAR

Shyh-Jiann Hwang 1, Yi-An Li 2 and Pu-Wen Weng 3

1Department of Civil Engineering, National Taiwan University, Taipei, 106, Taiwan 2Department of Civil Engineering, National Taiwan University, Taipei, 106, Taiwan

3National Center for Research on Earthquake Engineering, Taipei, 106, Taiwan [email protected], [email protected], [email protected]

ABSTRACT: Reinforced concrete (RC) columns failed in shear were commonly observed during Chi-Chi earthquake. Especially, the shear failure of RC short columns was one of the major failure modes which caused building collapse. Therefore, this study focuses on the behavior of shear failure till collapse. Eight specimens were constructed and tested to study the seismic behavior of short columns. These specimens with different height-to-depth ratio, shear reinforcement detailing and axial load ratios were subjected to double curvature bending with constant axial forces to observe the behavior of the shear and axial failure. Test results show that the different shear reinforcement detailing results in different behavior and the axial failure takes place much early with high axial load ratio. Finally, this study recognizes that the RC short column can possess the gravity-load carrying capacity even its lateral-load strength is lost completely.

1. INTRODUCTION

During Chi-Chi Earthquake in Taiwan (1999), we have found the failure of the columns was the major damage in the reinforced concrete (RC) buildings, especially low-rise RC school buildings. In order to understand the characteristic of the columns in low-rise RC school buildings, we observed that the typical columns become the captive columns because of the windowsill which commonly exist in low-rise RC school buildings. The shear failure of RC captive columns was one of the major failure modes which caused building collapse. However, it is noted that the columns still possess its gravity-load carrying capacity even the columns have been failed by shear. Therefore, this study focuses on understanding the post-strength behavior of the short columns after shear failure.

According to the experience surveyed during Chi-Chi earthquake, we found the low-rise RC school buildings collapsed due to the failure of the columns on the first story. The collapse of low-rise RC school buildings results from the lost of the gravity-load carrying capacity of columns. In order to understand the seismic behavior of low-rise RC school buildings, we should know the failure modes of short columns subjected to the horizontal and vertical loads. In this research, we varied three different parameters of short columns, such as height-to-width ratio, the amount of transverse steel and axial load ratio. We want to study the failure modes and the collapse behavior of the captive columns failed by shear.

2. EXPERIMENTAL PROGRAM

The experimental program consists of eight tests under cyclic lateral and vertical load. We totally have three parameters of the specimens such as height-to-width ratio, ductile or nonductile detailing and axial load ratio. The specimens were divided into four groups and every group has two specimens. These two specimens are subjected to low axial load ( , where gc Af ′1.0 cf ′ = designed concrete compressive strength and = gross cross-sectional area) and high axial load ( ) separately. Table 1 shows the parameters of the specimens. 3 and 4 represent the height-to-width ratio. D and N represent ductile detailing and nonductile detailing, respectively. L and H represent low axial load (

gA

gc Af ′3.0

gc Af ′1.0 ) and high axial load ( gc Af ′3.0 ), respectively. Figure 1 shows the reinforcement details of the specimens. First group of specimens is the nonductile



shorter columns shown in figure 1(a). The second group is the ductile shorter columns shown in figure 1(b). The third group is the nonductile longer columns shown in figure 1 (c). The fourth group is the ductile long columns shown in figure 4 (d). The height of the columns are 150 cm and 200 cm and the columns have a cross section of 50×50 cm.

Table 1 Specimens layout. Parameters

Specimens Aspect ratio Detailing Axial load

4DL gc Af ′1.0 (L)

4DH Ductile (D)

gc Af ′3.0 (H)

4NL gc Af ′1.0 (L)

4NH

4

Nonductile (N) gc Af ′3.0 (H)

3DL gc Af ′1.0 (L)

3DH Ductile (D)

gc Af ′3.0 (H)

3NL gc Af ′1.0 (L)

3NH

3

Nonductile (N) gc Af ′3.0 (H)

5

7015

070

50

(a) 3NL & 3NH (b) 3DL & 3DH

Unit : cm

10

8585

150

200

50

#8 longitudinal bars

#3 ties @ 30 cm

90-degree hooks

#3 ties @ 30 cm

16#8 longitudinal bars

50

(c) 4NL & 4NH (d) 4DL & 4DH

Figure 1 Specimen details.



The specified concrete compressive strength was 210 kgf/cm2 at 28 days. Mean concrete strengths obtained from standard compression tests on 152×305 mm cylinders on the day of column testing ranged from 303 to 352 kgf/cm2 (Table 2). Mean yield and ultimate strengths of the 25.4-mm diameter (No.8) deformed longitudinal bars were 4812 kgf/cm2 and 7004 kgf/cm2, respectively. Mean yield and ultimate strengths of the 9.5 mm-diameter (No.3) deformed longitudinal bars were 4570 kgf/cm2 and 6707 kgf/cm2, respectively. The axial loads of the specimens L and H are 63.75 tf and 191.25 tf.

Table 2 Material strength. reinforcement Specimens cf ′ ( kgf/cm2)

No. fy(kgf/cm2) fu(kgf/cm2) 4DL 310 4DH 303 4NL 347 4NH 314

#3 4570 6707

3DL 352 3DH 344 3NL 331 3NH 341

#8 4812 7004

The lateral support consists of four steel columns and two steel braces. It is designed to restrain the testing columns against lateral movement. Figure 2 shows the test setup in the laboratory. The loading systems are composed of two horizontal actuators and two vertical actuators. The force of horizontal actuators goes through the L shaped steel frame to the specimens. The vertical actuators apply axial load on the specimen. Figure 3 shows the loading history, which consists of the following lateral drift cycles: three cycles each at 0.25%, 0.5%, 0.75%, 1%, 1.5%, 2%, 3%, 4%, 5%, 6%. The control of actuator loading system is mixed with displacement and force controls. The loading systems which consist of four actuators are shown in Figure 4. The horizontal actuator 1 is the displacement control. It followed the loading history (shown in Figure 3). The horizontal actuator 2 provided the same load of the actuator 1. The resultant of the actuators 1and 2 went through the center of the column. The displacement of the vertical actuator 3 is equal to actuator 4. The resultant of the vertical actuators 3 and 4 was maintained as the targeted value throughout the test. The reason that we choose this kind of loading systems is to confirm the column deformed in double curvature.

Figure 2 Test setup.



drift

rat

io (%

)

Figure 3 Loading history.

Figure 4 Actuator control.

3. EXPERIMENTAL RESULTS AND DISSCUSION

3.1 Test Results

According to the crack patterns shown in Figure 5, we observed not only a lot of diagonal cracks after the specimens reach the maximum strength but also the vertical cracks, especially for ductile detailing columns. From Figure 5, it is noted that the inclined angle of the principal cracks with respect to the horizontal axis is larger when the axial load applied on the specimens is higher. Table 3 shows the test results contains the peak point of the test data and the observed failure modes. We think the failure mode of the short columns is the shear failure because of the diagonal crack patterns of the specimens. However, we note that there is another failure mode of the short columns, which is the vertical bond splitting along the column longitudinal bars.

Figure 6 shows the load-displacement hysteretic relationship of the specimens. We can see that the maximum strength of the specimens can not reach the flexural strength. The nominal flexural strength (Mn) of the test columns in Figure 6 were calculated according to ACI Committee 318 (2005). Therefore, we confirmed all the tested short columns were failed by shear. According to the hysteretic loop of testing (Figure 6), we concluded that the displacement of the nondcutile detailing specimens at the ultimate strength is lower than ductile detailing specimens and that the displacement of the specimens with high axial load is also higher than those of low axial load.

More details of the test results can be found in the reference of Weng (2007).



(a) 4DL (b) 4DH (c) 4NL (d) 4NH

(e) 3DL (f) 3DH (g) 3NL (h) 3NH

Figure 5 Failure patterns of the specimens.



Table 3 Test results. Peak Point

Specimens Strength (tonf) Displacement (cm) Description of Failure

4DL 73.7 3.5 shear failure with bond splitting

4DH 80.1 2.5 shear failure with bond splitting

4NL 47.9 1.7 shear failure

4NH 68.5 1.6 shear failure

3DL 78.5 1.7 shear failure with bond splitting

3DH 87.4 1.9 shear failure with bond splitting

3NL 48.2 0.8 shear failure

3NH 71.9 0.7 shear failure

Mn-8 -4 0 4 8 12

-18 -12 -6 0 6 12 18

-120

-80

0

60

120

Late

ralf

orce

(ton

f)

-12 -8 -4 0 4 8 12

-12 -6 0 6 12 18

-8 -4 0 4 8 12

-12 -6 0 6 12 18

-8 -4 0 4 8 12

-120

-60

0

60

120

-12 -6 0 6 12 18

-8 -4 0 4 8 12

3DL 3NL 3DH 3NHDisplacement (cm)

Drift Ratio (%)

Late

ralF

orce

(ton

f)

-16 -8 0 8 16 24

-8 -4 0 4 8 12

-24 -16 -8 0 8 16 24

-100

-50

0

50

100

-12 -8 -4 0 4 8 12

-16 -8 0 8 16 24

-8 -4 0 4 8 12

-100

-50

0

50

100

-16 -8 0 8 16 24Displacement (cm)

Drift Ratio (%)

4DL 4NL 4DH 4NH

Figure 6 Load-displacement relationship of the specimens.



3.2 Compare with ASCE 41

Some empirical equations to predict the drift ratio of the flexural shear failure (Elwood and Moehle, 2005a) and the drift ratio of the axial failure (Elwood and Moehle, 2005b) are available. Some of these findings were incorporated into the ASCE 41 (2007). Figure 7 shows the prediction of strength and displacement by the ASCE 41 (2007) and the envelope of the test results. According to the ASCE 41 (2007) the 4DH specimen is flexural-shear failure and the others are shear failure. That is different from the test observation because all specimens are shear failure. Also, the post-strength behavior predicted by the ASCE 41 (2007) is too conservative. The predicted displacement at the ultimate strength and at failure is much smaller than test data. Therefore, the prediction of the ASCE 41 (2007) is too stiff prior to strength and too conservative to post strength.

TestASCE 41

-24 -16 -8 0 8 16 24

-100

-50

0

50

100

Late

ralF

orce

(ton

f)

-12 -8 -4 0 4 8 12

-16 -8 0 8 16 24

-8 -4 0 4 8 12

-16 -8 0 8 16 24

-8 -4 0 4 8 12

-16 -8 0 8 16 24

-8 -4 0 4 8 12

4DL 4NL 4DH 4NH

Displacement (cm)

Drift Ratio (%)

Drift Ratio (%)

Displacement (cm)

-12 -8 -4 0 4 8 12 -8 -4 0 4 8 12 -8 -4 0 4 8 12 -8 -4 0 4 8 12

-18 -12 -6 0 6 12 18

-120

-60

0

60

120

-12 -6 0 6 12 18

3DL 3NL 3DH 3NH

-12 -6 0 6 12 18 -16 -8 0 8 16 24

Figure 7 Test results compared with ASCE 41.

4. CONCLUSION

According to the test result, we observed the failure modes of the tested short columns is shear failure. Besides, we have found the bond splitting is another failure mechanism involved with the short columns. As the test parameters were varied, different behavior was observed. The inclined angle of the principal cracks with respect to the horizontal axis is larger when the axial load applied on the specimens is higher. The displacement of the nondcutile detailing specimens at the ultimate strength is lower than that of ductile detailing; and the displacement of the specimens with high axial load is also higher than that of low axial load.

The prediction of the ASCE 41 is too stiff prior to strength and too conservative to post strength. More research works are needed.




5. REFERENCES

ACI Committee 318 (2005). “Building Code Requirements for Structural Concrete (ACI 318-05) and Commentary (ACI 318R-05)”, American Concrete Institute, Farmington Hills.

Elwood, K.J. and Moehle, J.P. (2005a). “Drift capacity of reinforced concrete column with light transverse reinforcement”, Earthquake Spectra, Vol. 21, No. 1, pp. 71-89.

Elwood, K.J. and Moehle, J.P. (2005b). “Axial capacity model for shear-damaged columns”, ACI Structural Journal, Vol. 102, No. 4, pp. 578-587.

Weng, P.W. (2007). “Study on the Seismic Performance Curves of Reinforced Concrete Short Columns Failed in Shear”, Master Thesis, Department of Construction Engineering, National Taiwan University of Science and Technology, Taiwan. (in Chinese).

ASCE/SEI 41 (2007). “Seismic Rehabilitation of Existing Building”, American Society of Civil Engineering, Reston, Virginia.



EXPERIMENTAL STUDY ON SEISMIC PERFORMANCE OF R/C FRAMES WITH CONCRETE BLOCK INFILL PANELS

Yasushi Sanada1, Botirjon Yorkinov2 and Taizo Hirose2

1 Faculty of Engineering, Toyohashi University of Technology, Toyohashi, Japan

2 Graduate School of Engineering, Toyohashi University of Technology, Toyohashi, Japan Email: [email protected]

ABSTRACT: This paper discusses seismic performance of reinforced concrete (R/C) frames with concrete block panels, focusing on internal shear transfer. Two R/C one-bay frame specimens were prepared, and concrete block panels were installed. Although one of the specimens was a conventional type, the other was specially designed to investigate its shear transfer on the critical section. In the latter case, shear forces carried by the columns were independently measured using a new force measuring system developed in this study. The measuring system was verified through comparing between the test results of both specimens. As a result of the force measurement, it was found that the installed block panel significantly affected the seismic performance of this kind of structure. In particular, shear force on the critical section concentrated around the compressive column bottom due to an interaction between the panel and its surrounding frame, which indicates that the column bottom is not only a structural weak point but also a key part for retrofitting.

1. INTRODUCTION

Accurate seismic performance evaluation of structures and/or structural elements plays important roles to mitigate earthquake disasters. A huge number of laboratory tests on structural members have been carried out for the seismic performance evaluation. Actually, however, experimental data on stress or internal force are extremely limited because of technical difficulties when measuring them.

Therefore, the authors have collected experimental data on internal forces of reinforced concrete (R/C) shear walls through several laboratory tests (e.g. Sanada and Kabeyasawa, 2006). In this study, a series of structural tests on concrete block infilled frames was conducted to obtain such experimental data. A new force measuring system was also developed to accomplish this objective. Focusing on internal shear transfer, the seismic behavior and performance of this kind of structure were investigated experimentally.

2. SPECIMENS

Two 3/10 scale R/C one-bay frame specimens were prepared, and concrete block infill panels were installed, as shown in Figure 1. Although one of the specimens was a conventional type (Specimen-C), as Figure 1 (a), the other was a specially designed one (Specimen-S) to investigate its shear transfer on the critical section, as Figure 1 (b). The details of both specimens were the same except for the bases. In particular, the base of Specimen-S was split into three pieces along the span, to measure local internal forces on the critical section. They were obtained using specially developed force measuring systems fixed under the specimen. Figure 2 shows the close-up around the bottom of Specimen-S placed on the force measuring systems. As shown in this figure, the internal forces were measured on the exterior bases. The details of the systems are mentioned in the following section. The material properties of concrete, reinforcements, and joint mortar used for the specimens are given in Tables 1, 2, and 3, respectively. The joint mortar was produced with 150 % water/cement and 300 % sand/cement ratios (by volume), as used for conventional masonry constructions in Japan.



a

a`

b

a-a`

D4@120D4

b-b`

Unit: mm

180

180

150

132

D10@50

8-D19

12-D19

D10@50

665 880 665

2250

20 20

700

51.5143 311 14351.5

350

900

350

b`

D4@120

260

180

260

180360 1170 180 360

a

b

a-a`

D4@120D4

b-b`

180

180

150

132

D10@50

8-D19

12-D19

D10@50

2250 51.5143 311 14351.5

350

900

500

b`

D4@120

260

180

260

180360 1170 180 360

400

a`

8-D10

700

700

8-D10

8-D10

Upper beam

Lower beam

Upper beam

Lower beam

D4

400

D4

8-D10

(a) Specimen-C (b) Specimen-S (a) Specimen-C (b) Specimen-S

Figure 1 Details of specimens. Figure 1 Details of specimens.

350

300

200

430 680 680 430

765 765

120

66520440

Force measuring system

Unit: mm

Figure 2 Close-up around the bottom of Specimen-S. Figure 2 Close-up around the bottom of Specimen-S.

Table 1 Mechanical properties of concrete. Table 1 Mechanical properties of concrete.

Specimen Specimen Age (day) Age (day) fc (MPa) fc (MPa) fcr (MPa) fcr (MPa)

Specimen-C 54 20.0 4.0

Specimen-S 69 21.9 4.3

where, fc: peak compressive strength of concrete cylinder, fcr: cracking stress of concrete in tension.



Table 2 Mechanical properties of reinforcements.

Bar no. Type Es (GPa) fy (MPa) εy (μ) ft (MPa)

D10 Deformed 184 352 1913 492

D4 Deformed 164 383 2332 537

where, Es: initial modulus of reinforcement, fy: yield stress of reinforcement, εy: yield strain of reinforcement, ft: peak strength of reinforcement.

Table 3 Mechanical properties of joint mortar for masonry construction.

Specimen Age (day) jmfc (MPa) jmfcr (MPa)

Specimen-C

Specimen-S 27 22.1 2.9

where, jmfc: peak compressive strength of joint mortar, jmfcr: cracking stress of joint mortar in tension.

3. TEST METHODS

3.1 Test System and Loading Program

The tests were carried out at a testing facility in Toyohashi University of Technology, as illustrated in Figure 3. The specimens were subjected to cyclic lateral loading under a constant axial load of 200 kN (≈ 0.15 x (2 Ac fc), where Ac: cross-sectional area of each column). However, the shear span to depth ratio of 0.8 was maintained by controlling both vertical jacks in Figure 3. The applied loading history in the lateral direction is shown in Figure 4.

Figure 5 shows the set-up of transducers to measure the horizontal, vertical, and diagonal displacements of the specimens. In Figure 5 (b), however, only the additional transducers for the Specimen-S are indicated. The single and double arrows in the figure mean the absolute and relative displacements, respectively, and the alphabets indicate the displacements used to control the horizontal jack in Figure 3. Lateral drifts of the specimens were defined as “a–b” in the tests. Moreover, crack widths were also observed using crack gauges at the peak and residual drifts in Figure 4.

Figure 3 Loading system.



-1/800

1/400

1/200

1/100

1/50

0.00

Rot

atio

n an

gle

(rad

.)

-1/200

-1/100

-1/50

1/800

-1/400

Figure 4 Loading history.

Unit: mm

(a) Specimen-C (b) Specimen-S

Figure 5 Transducers set-up.

3.2 Force Measuring System

A new force measuring system was developed to measure internal forces on the critical section of Specimen-S. Figure 6 shows the details of the system, in which two steel load cells are placed side by side, and connected with steel devices at the top and bottom. As shown in the figure, this system can be planned to increase its bending stiffness in proportion to the square of the span length between both load cells, provided that the connecting devices are rigid. The details of the load cells are also shown in Figure7.

Figure 6 Force measuring system.



200

100

0

-100

-200

210-1

-200

-100

0

100

200

Late

ral f

orce

( kN

)

-2

Drift ratio ( % )

Late

ral f

orce

( kN

)

Initial yielding of the column longitudinal reinforcementInitial yielding of the column transverse reinforcement

Initial shear cracking in the column and panel Initial flexural cracking in the column Initial separation cracking between the column and panel

Shear failing of the column

210-1-2

Drift ratio ( % )

Unit: mm

410

50

200

100

50

200

Figure 7 Details of load cell.

4. TEST RESULTS

Firstly, effects of splitting the base of Specimen-S on its behavior are discussed through comparing to the test results of Specimen-C, as a verification of the force measuring method proposed in this study. Subsequently, internal shear forces carried on the column bottom of Specimen-S are investigated experimentally.

4.1 Verification of Force Measuring System

Figure 8 compares the relationships between lateral force and top drift ratio for both specimens. In this figure, however, their failure processes are also indicated, which are similar to each other as follows. Separation cracks were initially observed along the boundaries between the columns and panel, and then flexural cracks of the columns and diagonal shear cracks occurred and propagated. Subsequently, the longitudinal and transverse column reinforcements yielded. After the maximum strengths of –223.5 kN and 234 kN for Specimen-C and -S were observed during the cycles to ±1/100 rad., the strengths of specimens began to degrade due to shear failure of the columns during the cycles to –1/100 rad. As a result, the similar final crack patterns, as shown in Figure 9, as well as the similar hysteresis loops in Figure 8 were observed in both specimens. Moreover, Figure 10 compares the relationships between maximum crack widths, observed at the peak and residual drifts, and top drift ratio. Good correlations were also obtained for both the panel and columns in each specimen. These results mean that the Specimen-S could substantially reproduce the behavior of Specimen-C, in spite of splitting the base.


Figure 8 Lateral force-top drift ratio relationships.

4082.5 82.5 82.5 82.540

410

200

100

105

105

50 50 105105 50 100 50



W E

Negative PositiveE

Negative PositiveW


Figure 9 Final crack patterns.

1

3

4

5

2

1/8001/400

1/2001/100

1

3

4

1/50

5

2

Cra

ck w

idth

(mm

)

-1/800-1/400

-1/200-1/100

-1/500

Column

Rotation ratio (rad.)

Specimen-SSpecimen-C

1/8001/400

1/2001/1001/50-1/800

-1/400-1/200

-1/100-1/50

0Rotation ratio (rad.)

Specimen-SSpecimen-CWall panel

Cra

ck w

idth

(mm

)

(a) At the peak drifts

1

3

4

5

2

1/800

1

3

4

1/50

5

2

Cra

ck w

idth

(mm

)

1/8001/400

1/2001/100

1/50-1/800-1/400

-1/200-1/100

-1/500


1/800 1/501/8001/400

1/2001/100

1/50-1/800-1/400

-1/200-1/100

-1/500


Specimen-SSpecimen-CColumn Specimen-S

Specimen-CWall panel

Cra

ck w

idth

(mm

)

(b) At the residual drifts

Figure 10 Maximum crack widths-top drift ratio relationships.

4.2 Detection of a Structural Weak Point

Local internal shear forces on the critical section of Specimen-S are investigated herein, based on experimental data obtained by the force measuring system. Figure 11 gives the ratio of shear force, acting on the compressive column bottom, to the total shear force. In particular, this figure focuses on the test results of the west column at the peak drifts during the negative loading (refer to Figures 3 and 4). As a result, more than half of the total shear force concentrated on the compressive column bottom from the relatively small drift level. The ratio was increased up to about 80 % before the maximum strength of the specimen was observed during the cycle to 1/100 rad. According to our past study (Choi et al., 2005), this shear stress concentration seems to be caused by an inclined compression strut forming in the infill panel when the infill is subjected to



shear deformation by the surrounding frame, as shown in Figure 12. Therefore, the resultant punching shear acted on the compressive column bottom, and increased its damages. Then, after the column failed in shear at the bottom during the cycle to –1/100 rad., the ratio began to decrease significantly. These results indicate that the compressive column bottom is a weak point in this kind of structure, and that retrofitting of the weak point possibly induces rational improvements of the seismic performance.

100

80

60

40

20

0

Drift ratio (rad.)

Rat

io o

f she

ar fo

rce

(%)

1/400(1)0 1/400(2) 1/200(1) 1/200(2) 1/100(2)1/100(1) 1/50(1)1/800

Figure 11 Ratio of the shear force, acting on the east column bottom, to the total shear force.

Punching shear Compression

strut

Figure 12 Assumed lateral force resisting mechanism of infilled frames.

5. CONCLUDING REMARKS

Seismic performance of concrete block infilled R/C frames were experimentally investigated focusing on internal shear transfer. Major findings from a series of structural tests are summarized as below.

1. A new force measuring system was developed and applied to the structural tests conducted in this study. This system can be planned to increase its bending stiffness by controlling the span length between neighboring load cells.

2. The specially designed specimen, for measuring its local internal forces on the critical section, could approximately reproduce the behavior of the conventional type specimen. This result means that the proposed force measuring method is effective to obtain internal forces in this kind of structure.

3. The force measurement could locate a structural weak point experimentally. This was around the column bottom on the compressive side, where the high punching shear acted due to a compression strut forming in the infill panel.




6. ACKNOWLEDGMENTS

The authors acknowledge financial supports by Grant-in-Aid for Young Scientist (A) (No. 19686033), Japan Ministry of Education, Culture, Sport, Science, and Technology (MEXT), and the Tatematsu Foundation.

7. REFERENCES

Choi, H., Nakano, Y. and Sanada, Y. (2005). “Seismic performance and crack pattern of concrete block infilled RC frames,” Bulletin of ERS, No. 38, pp. 119-134.

Sanada, Y. and Kabeyasawa, T. (2006). “Local force characteristics of reinforced concrete shear wall,” Proceedings of the 8th U.S. National Conference on Earthquake Engineering, San Francisco, Paper No. 324.



SEISMIC PERFORMANCE OF SUNGAI MERANG BRIDGE IN TERENGGANU UNDER LOW EARTHQUAKE GROUND MOTION

Azlan Adnan1, Ismail Mohd Taib2 and Meldi Suhatril1

1Faculty of Civil Engineering, Universiti Teknologi Malaysia, Johor Bahru, Malaysia

2Public Work Department, Kuala Lumpur, Malaysia Email: [email protected]

ABSTRACT Recently the effects of earthquake have been often felt in Peninsular Malaysia. The bridges as a pass or connecter from one town to another are essential structures that should undergo seismic vulnerability analysis. The seismic structure vulnerability for bridges was conducted under linear and nonlinear analysis. In this analysis, Sungai Merang Bridge in Terengganu is modeled under 2D and 3D-concept using SAP2000 and IDARC. In this study, the site-specific analysis has been done to determine the earthquake loading (Time history at the surface and design response spectrum) using two borehole data (BH7 and BH8) at the bridge location. Design response spectrum was constructed using IBC2000 modification. The seismic analyses conducted in this research were namely, free vibration analysis, time history analysis, response spectrum analysis and damage inelastic analysis. Free vibration analysis represents the period and mode shape of the structure while time history and response spectrum analysis represents the applied forces on the deck and piers. For damage inelastic analysis, the results showed where the critical part of the bridge structure failure under several peak ground acceleration (PGA).

1. INTRODUCTION

In 2004 and 2005, repeated tremors from the Sumatran earthquakes have brought concerns to the public, government authorities, engineers and researchers on safety especially when no earthquake design practice had been taken into consideration for structures in Malaysia. The assumptions saying that Malaysia is free from earthquake and its effects have changed since most of the tremors were felt. Thus the structure’s safety and adequacy in resisting earthquake effects have been questioned. Detailed research should be conducted in the future to determine the exact performance of the structure before and after earthquake. The effects of earthquake for critical structures like bridges should be considered. In this paper, the performance of Sungai Merang Bridge in Terengganu was investigated.

Sungai Merang Highway Bridge is a beam slab deck structure that uses T – beams as prestressed girder. The prestressed concrete bridge consists of five spans supported by four intermediate piers and two abutments. The bridge is approximately 85 Meters long and 11 Meters wide. The elevation of Sungai Merang Bridge components can be seen in Figure 1. Table 1 shows the properties of elastomeric laminated bearing for Sungai Merang Bridge.

Figure 1 Elevation of Sungai Merang Bridge.



Table 1 Properties of elastomeric laminated bearing.

2. RESEARCH OBJECTIVES

The overall objective of this phase of the study was to evaluate the seismic response of Sungai Merang Bridge with emphasis on two and three-dimensional effect of ground excitation;

(i) To perform 2D and 3D modeling analysis to investigate the seismic response of the actual Sungai Merang Bridge in the longitudinal direction reported to be the most vulnerable direction for multi span simply supported bridges under earthquake ground motion.

(ii) To determine the time history and design response spectrum at the location surface for Sungai Merang Bridge in Kuala Terengganu, and

(iii) To determine the seismic response of bridge under earthquake ground motion from initial failure up to collapse. In this stage, the results show the critical parts of the bridge for every earthquake ground motion (PGA).

3. LOCAL SITE EFFECT ANALYSIS

Two (2) soil data in Kuala Terengganu (BH-7 and BH-8) were collected from existing soil investigation (SI) of Sungai Merang Bridge. The shear wave velocity (VS) were obtained by converting the N-SPT value from Standard Penetration Test to shear wave velocity using empirical formula proposed by Ohsaki and Iwasaki (1973), Imai and Tonouchi (1982) and by averaging those two formulas. Based on analysis, the VS-30 of BH-7 and BH-8 are 229 m/s and 231 m/s, respectively. Based on these results, generally the site can be classified as stiff soil or site class D (SD) in accordance with 1997 UBC. Based on macrozonation study (Figure 2), the peak ground acceleration (PGA) at the bedrock of Kuala Terengganu for 500-year return period of earthquake is 0.039g (39 gal). The results of analysis can be seen in Figures 3. Generally, the accelerations at the bedrock were amplified on the surface in the range of 1.5-1.8. The predominant periods of the spectra generally occur in the range of 0.5 - 1.0 second. Time history at the surface can be seen at the Figure 4.

Figure 2 Macrozonation map for 500 years return period at T=1.0 sec.



0

5

10

15

20

25

30

35

40

45

0 0.05Maximum Acceleration (g)

Dep

th (m

)

BH-8

0.10

5

10

15

20

25

30

35

40

45

0 0.05 0.1Maximum Acceleration (g)

Dep

th (m

)

BH-7

Figure 3 Result of ground response analysis.

Figure 4 Time history at surface (0.072g) for Sungai Merang Bridge.

Response spectrum analysis used IBC2000 to construct the design response spectrum. The design response spectrum can be seen in Figure 5.

0.00

0.20

0.00 1.00 2.00 3.00 4.00Periode (Second)

Spec

tral

Acc

eler

atio

n (g

)

mean + 1 standar deviasi

IBC 2000 (modified)

BH7BH8

Figure 5 Design response spectrum using IBC2000 (modified).

4. SEISMIC ANALYSIS

A number of computer models of the Sungai Merang Bridge were created, analyzed and compared to evaluate the structural response of the bridge under earthquake loading. All models were linear elastic simulation in SAP2000 and nonlinear inelastic in IDARC.

4.1 Two Dimensional Analysis of Sungai Merang Bridge

The seismic analysis of Sungai Merang Bridge used SAP2000 for two-dimensional modeling (Figure 6). The deck and pier of the bridge was modeled by using beam elements. There were three types of seismic analysis that were implemented in this study. Namely: Free vibration, Time History and Response Spectrum analysis.

4.1.1 Free vibration analysis

The free vibration analysis considered five modes. The periods of structure are shown in Table 2. Figure 7 shows the mode shape 1 of the bridge structure.



Figure 6: SAP2000 2-D frame element model

Figure 7 Mode shape 1 of bridge structure.

Table 2 5 natural periods of Sungai Merang Bridge.

No Period (Sec)

1 0.50887

2 0.50034

3 0.47537

4 0.42613

5 0.34392

4.1.2 Time history analysis

In this analysis, the effect of time history at surface with PGA 0f 0.072g (Figure 4) was applied to two dimensional analysis. The maximum axial, shear and bending moment forces and maximum displacement at bridge pier and deck can be seen in table 3 and 4.The maximum displacements are presented by U1 and U3 as horizontal and vertical displacement.

Table 3 The results of 2 dimensional time history analysis.

Component Value

Maximum axial force pier 9086 KN

Maximum shear force pier 458KN

Maximum BM pier 6172 KNM

Maximum axial force deck 417 KN

Maximum shear force deck 1980 KN

Maximum BM deck 7355 KNM

Table 4 The maximum vertical and horizontal displacement for TH-2D analysis.

U1 10.2 Deck

U3 69

U1 2.22

Maximum Displacement

(mm) Pier U3 0.26

4.1.3 Response spectrum analysis

Response spectrum analysis used IBC2000 to construct the design response spectrum. The applied design acceleration response spectrum is shown in Figure 5.The maximum forces result of response spectrum analysis can be seen in Table 5. Table 6 shows the deck and piers maximum displacement.

Table 5 The result of 2 dimensional response spectrum analysis.

Component Value

Maximum axial force pier 9252 KN



Maximum shear force pier 1559KN

Maximum BM pier 7906KNM

Maximum axial deck 786 KN

Maximum shear force deck 2029 KN

Maximum BM deck 7708 KNM

Table 6 The maximum vertical and horizontal displacement for RS-2D analysis.

U1 15.4 Deck

U3 74

U1 4


(mm) Pier U3 0.3

4.2 THREE DIMENSIONAL ANALYSIS OF SUNGAI MERANG BRIDGE

For three dimensional analysis, the maximum applied forces are less than two dimensional analysis. Three dimensional decks are represented by a number of beam elements per span while the two dimensional model is represented by one beam element per span. It is same as well for pier model. Three dimensional model of Sungai Merang Bridge can be seen in Figure 8.

Figure 8 The three dimensional modeling of Sungai Merang Bridge.

4.2.1 Free vibration analysis

The free vibration analysis considered five modes. The period of structure are shown in Table 7. The first mode shape of bridge structure can be seen in Figure 9.

Table 7 5 Natural periods of Sungai Merang Bridge.

No Period (Sec)

1 0.4030

2 0.36286

3 0.35574

4 0.34491

5 0.33223

Figure 9 Mode shape 1 of three-dimensional analysis.



4.2.2 Time history analysis

Time history analysis of Sungai Merang Bridge model was performed using three-dimensional models. It used the time history with PGA of 0.072g (Figure 4), The Maximum applied forces presented in Table 8, while Table 9 shows the maximum horizontal (U1) and vertical (U3) displacement for deck and piers.

Table 8 Maximum applied force of Sungai Merang Bridge for 3 dimensional time history analysis.

Pier Deck

Max axial force 1699 KN Max axial force 210 KN

Max shear force 397KN Max shear force 297 KN

Max BM 1242KNM Max BM 1391KNM

Table 9 The maximum vertical and horizontal displacement for TH-3D analysis.

U1 4 Deck

U3 10.3

U1 2.2


(mm) Pier U3 0.22

4.2.3 Response spectrum analysis

Response spectrum analysis used IBC2000 to construct the design response spectrum. The applied design acceleration response spectrum is shown in Figure 5. From the analysis, the maximum applied forces can be seen in Table 10. Maximum displacement of three dimensional response spectrum analysis (Table 11) is almost same with three dimensional time history analysis.

Table 10 Maximum applied force of Sungai Merang Bridge for 3 dimensional response spectrum analysis.

Response spectrum analysis

Pier Deck

Max axial force 1670KN Max axial force 210KN

Max shear force 398KN Max shear force 297KN

Max BM 1242KNM Max BM 1392KNM

Table 11 The maximum vertical and horizontal displacement for RS-3D analysis.

U1 3.94 Deck

U3 10.3

U1 2.2


(mm) Pier U3 0.22

4.3 IDARC NONLINEAR SEISMIC DAMAGE INELASTIC ANLYSIS

The seismic analysis of Sungai Merang Bridge in Terengganu used IDARC for 2 dimensional modeling. From the analysis, the bridge started to crack at 0.17g and collapsed at 0.26g.



Figure 10 The sequence of IDARC analysis under earthquake ground motion for initial failure up to collapse.

= Initial Cracking

= Plastic Hinge Develop

= Local Failure

The first column yielding at T = 4.52 second and the first beam cracked at T = 4.55 second. The sequence of segment cracking or yielding can be seen at Figure 10.

5. DECK AND PIER CAPACITY OF SUNGAI MERANG BRIDGE

To know the ability of the Sungai Merang Bridge to resist the force from external load, the external force should be not greater than structure capacity of the bridge. In this study the structure capacity will be divided into 2 parts; deck and pier section. For the deck capacity, the forces in consideration are bending moment and shear stress capacity. While for the pier capacity, the forces in consideration are bending moment and axial force. It should be noted that due to lack of field strength testing of the concrete, no increase in concrete strength due to aging is considered and on the other hand, no strength reduction factors were applied for capacity calculations.

By using strain compatibility method, we found that for two dimensional analysis, the ultimate moment resistance of prestressed T-beam of Sungai Merang Bridge is 1667 KNM and the ultimate shear force is 828 KN, while for three dimensional analysis ultimate moment resistance is 15003 KNM and ultimate shear resistance is 7452 KN. The moment and axial force interaction diagram can be seen in Figure 11.




Figure 11 M-P interaction diagram for Sungai Merang pier.

6. RESULT AND DISCUSSION

Based on the comparison between the maximum applied force and capacity of Sungai Merang Bridge in Terengganu, it has shown that all applied forces for deck and pier is not more than capacity of the bridge (Table 12 and 13).

Table 12 The comparison of maximum applied and capacity force for bridge deck.

TH-2D RS-2D CAPACITY TH-3D RS-3D CAPACITY

Max Shear (KN) 1980 2029 7452 297 297 828 DECK

Max BM (KNM) 7355 7708 15003 1391 1392 1667

Table 13 The comparison of maximum applied and capacity force for the bridge pier.

TH-2D RS-2D TH-3D RS-3D 3D

CAPACITY

Max Axial (KN) 9086 9252 1699 1670 Refer to PIER

Max BM (KNM) 6172 7906 1242 1242 Figure 11

7. CONCLUSIONS

From the analysis, it can be concluded that the Sungai Merang Bridge is safe under earthquake analysis which subject to local site effect (PSA = 0.072g). The bridge will begin the initial cracking at 0.17 g and collapse at 0.26g.

8. REFERENCES

Chopra, A.K. (1995). ”Dynamics of Structures: Theory and Application to Earthquake Engineering”, New Jersey, Prentice Hall, Inc.

Adnan, A. (1998). ”Low Intensity Earthquake Effects on Steel Girder Bridges”, PhD. Theses, Universiti Technology Malaysia.

Yazdani-Motlagh, A. (2002). ”Inelastic Seismic Behaviour of Stiffening Systems Multi Span Simply Supported (MSSS) Bridges”, PhD. Theses, New Jersey Institute of Technology.



SEISMIC ASSESSMENT AND RETROFITTING OF AN IRREGULAR REINFORCED CONCRETE STRUCTURE

Marco Valente1

1Dipartimento di Ingegneria Strutturale, Politecnico di Milano, Milano, Italia


ABSTRACT: The efficiency of a seismic retrofit intervention using RC jacketing on an existing RC building structure designed according to pre-1970s standards was presented, based on analytical investigation supported by experimental tests carried out at the JRC Ispra. The aim of this study was to improve the understanding of the influence of plan-irregularity on the seismic response and to assess the opportunity of using RC jacketing as an effective technique for the seismic rehabilitation of RC building. The structure was analyzed in the “as-built” and retrofitted configurations. The study presented the design criteria and the procedures followed for the choice of the column to be retrofitted, the comparison with experimental tests and with the “as-built” structure. On the basis of test results, numerical models of the test structure were developed. The experimental tests and numerical results of the analyses performed on the structure in the “as-built” and retrofitted configurations highlighted the effectiveness of the design criteria and the RC jacketing to improve the global performance of an under-designed irregular RC building in terms of strength and stiffness of the structure.

1. INTRODUCTION

Based on results of experimental tests carried out on a plan-wise irregular RC structure at the JRC ELSA Laboratory, numerical models were developed to evaluate the effectiveness of the RC jacketing technique in improving the seismic performance of existing RC structures. The retrofit design criteria to enhance the global response of the full-scale under-designed irregular RC structure to the seismic excitation were presented. The design strategy was based on the basic concept of reducing the torsional components of the seismic response of the structure, which were highlighted in the response of the “as-built” structure, by means of a reduction of the eccentricity of the centre of strength (CP) with respect to the centre of mass (CM). The strength relocation was achieved using the traditional technique of RC jacketing, limited to selected elements. Non-linear time-history simulations of the pseudo-dynamic tests of the structure were performed. The experimental tests and numerical results of the analyses performed on the structure in the “as-built” and retrofitted configurations highlighted the effectiveness of the strength centre relocation strategy based on RC jacketing in reducing the torsional response and improving the global seismic performance of an under-designed irregular RC building.

2. TEST STRUCTURE AND RETROFIT INTERVENTION

Within the SPEAR Research Project, a series of experimental tests on a torsionally unbalanced three-story reinforced concrete framed structure were carried out at the ELSA Laboratory of the JRC at Ispra, Negro et al. (2004). An overview of the test building and the plan of a typical floor are presented in Figure 1. The structure represents a simplification of a typical older construction in Southern Europe. The main deficiencies of the SPEAR test structure are represented by irregular plan layout, slender columns with largely spaced stirrups, plain reinforcing bars, lack of shear reinforcement in beam-to-column joints, inadequate anchorage of stirrups, column lap splices in potential plastic hinge zones. The full-scale RC structure was subjected to bi-directional pseudo-dynamic tests at the JRC ELSA Laboratory. The tests were performed on the “as-built” structure under the Montenegro Herceg-Novi record scaled to two different peak ground acceleration (PGA) values (0.15g and 0.2g). After the damage, the structure was retrofitted using RC jacketing on



selected columns and then subjected to a new series of two tests with the same input accelerogram but scaled to PGA values of 0.2g and 0.3g.

Figure 1 Overview and plan of the SPEAR structure with indication of beam and column sections.

The aim of the rehabilitation strategy was to increase both the strength and stiffness of the “as-built” structure by the RC jacketing of selected vertical elements. The columns to be strengthened were selected in order to minimize the torsional effects due to the doubly non-symmetric plan configuration of the “as-built” structure and, thus, to reduce the displacement demand on the external columns. The original cross-section of columns C1 and C4 was increased from 250x250 mm to the jacketed 400x400 mm, Figure 2. The reinforcement of the jacketed columns was designed as three 16mm diameter bars for each side of the column and 8mm diameter stirrups; the stirrup spacing was 100 mm at the top and bottom of the columns at each story and 150 mm for the remaining column length. The choice of the columns to be retrofitted were performed in order to relocate the centre of strength (CP) in plan-wise irregular structure to reduce the torsional component of the response. The centre of strength (also called plastic centre) is assumed the location of the resultant of the yield moments of the columns at each floor. In the inelastic range of the response, torsional effects are mainly governed by strength eccentricity, rather than stiffness eccentricity. The coordinates of the centre of mass (CM), of stiffness (CR) and strength (CP) for a typical story in the case of “as-built” and RC jacketed structures are shown schematically in Figure 3. The eccentricity of the centre of strength and stiffness at each story of the structure was significantly reduced by the enlargement of columns C1 and C4. Such intervention was considered effective to obtain a reduction of the torsional response of the structure, also taking into account the feasibility and simplicity of the selected retrofitting, involving only two columns.

Figure 2 Plan of the retrofitted structure and cross-section enlargement.

Experimental tests carried out at JRC Ispra and numerical analyses performed in this study on the “as-built” configuration showed an excessive increase in the edge column drifts with respect to those of the CM and interaction effects between the eccentricities in the two directions, amplifying the torsional effects induced by the same value of eccentricity acting in one direction only. The damage was mainly concentrated on the columns, as the structure was characterized by a lack in the hierarchy of strength.



Figure 3 Location of the centre of strength (CP), stiffness (CR) and mass (CM) in the “as-built” (left) and in

the rehabilitated structure (right), (dimensions in meter).

3. NUMERICAL MODELLING OF THE TEST STRUCTURE

The finite element analysis computer code SeismoStruct was utilized to perform analyses for the assessment of the seismic response of the test model. The spread of inelasticity along the member length and within the member cross-section is modelled following a fibre modelling approach, allowing for an accurate estimation of damage distribution. The sectional stress-strain state of inelastic frame elements is obtained through the integration of the nonlinear uniaxial stress-strain response of the individual fibres into which the section has been subdivided. The discretisation of a typical reinforced concrete section is illustrated in Figure 4. Idealization of the structure is based on frame elements placed at mid-depth of the members and connected at the nodes. Each member has been further subdivided in more elements; the length of the element is critical to effectively capture the expected inelasticity behavior in dissipative zones of the structure. Four-six elements, with smaller elements at member ends, were used to model beams and columns to ensure inelasticity could be accurately modelled.

Figure 4 Discretisation of a reinforced concrete section into fibres.

Concrete was modeled by employing a uniaxial constant confinement concrete model based on the constitutive relationship proposed by Mander et al. (1988) and later modified by Martinez-Rueda and Elnashai (1997) for reasons of numerical stability under larger displacement analysis. The confinement effects, provided by the lateral transverse reinforcement, are incorporated through the rules proposed by Mander, whereby constant confining pressure is assumed throughout the entire stress-strain range. The model is defined by four parameters: the peak compressive strength of unconfined concrete (f’c), the tensile strength (ft), crushing strain (εco) and the confinement factor (K), which is defined as the ratio between the confined and unconfined compressive stress of the concrete. In the case of the SPEAR building, according to the reinforcement detail, the amount of transverse reinforcement of all members is very small to produce an effective concrete confinement. Due to the insufficiency of stirrups, the confinement factor K is calculated to be close to 1 for all members and thus approximated to be 1.01 in the analytical model. Figure 5 shows the stress-strain curves for unloading and reloading branches in the model of Mander et al. (1988) and in the model of Martinez-Rueda and Elnashai (1997). Longitudinal reinforcement steel is modelled employing a bilinear elasto-plastic model, Figure 6. In this model, loading in the elastic range and unloading phase follow a linear function defined by Young’s modulus of steel. In the post-elastic range, a kinematic hardening rule for the yield surface defined by a linear relationship is assumed (Elnashai and Elghazouli, 1993). Actual material properties, which were



obtained from the real test structure under or after construction, were adopted in the developed analytical models. In order to model slabs as rigid diaphragms, each corner of a slab was diagonally connected to the opposite corner by horizontally rigid members.

Figure 5 Stress-strain curves for unloading and reloading branches in the model of Mander et al (1988) and in

the model of Martinez-Rueda et al. (1997).

Figure 6 Bilinear stress-strain curve of steel.

The reinforced concrete jacketed rectangular section available in Seismostruct libraries was used for the modelling of rectangular columns retrofitted by means of reinforced concrete jacketing. Different confinement levels for the internal (pre-existing) and the external (new) concrete materials were defined. To evaluate the properties of the retrofitted elements, the following assumptions, according to Eurocode 8, were adopted: the jacketed column behaves monolithically with full composite action between old and new concrete, the concrete properties of the jacket apply over the full section of the element, the axial load is considered acting on the full composite section.

In order to understand the overall response of the “as-built” and retrofitted structures, periods and mode shapes were obtained from eigenvalue analysis. The modal periods and corresponding participating masses along with deformed shapes of the retrofitted structure changed with respect to the “as-built” structure. The elastic period of the rehabilitated structure decreased (from 0.66 sec to 0.53 sec) with respect to the “as built” structure, Table 1. The RC jacketing of columns C1 and C4 allowed decreasing the effects of the plan irregularity as the participating masses related to the first two modes of vibration were, in such case, represented mainly by the mass in the X and Y direction.

Table 1 First three periods of vibration of the “as-built” and reinforced structure.

1st mode 2nd mode 3rd mode

Period [s] of the “as-built” structure 0.662 0.587 0.505

Period [s] of the reinforced structure 0.534 0.496 0.441

4. NUMERICAL ANALYSES

Non-linear time-history simulations of the pseudo-dynamic tests of the three-story full-scale building carried out at the JRC Ispra were performed. Montenegro Herceg-Novi records for longitudinal and transverse components scaled to different PGA values (0.2g and 0.3g) were applied. The main results of the numerical analyses with 0.2 PGA intensity on the bare and RC jacketed structures were reported in this study.



Figure 7 and Figure 8 show the displacement time histories of the centre of mass of the third story of the model in X and Y directions under bidirectional earthquake with 0.2g PGA, for the “as-built” and the retrofitted structure, respectively. A good agreement between numerical and experimental results was observed, above all in Y direction; the correlation was less accurate in X direction, especially for the retrofitted structure. It can be noted that the test on the RC jacketed structure was carried out after the tests on the original structure and the rehabilitation with FRP and some damage was accumulated in the previous tests. A difference in terms of maximum top displacement was observed between the two analyzed numerical models. The retrofit intervention increased the stiffness of the RC jacketed structure, reducing the maximum top displacement and the interstory drift at all levels.

X Direction

-120-100-80-60-40-20

020406080

100120

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time [s]

Top

Dis

plac

emen

t [m

m]

Experimental Numerical

Y Direction

-120-100-80-60-40-20

020406080

100120

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time [s]

Top

Dis

plac

emen

t [m

m]

Experimental Numerical

Figure 7 Center of mass (CM) top displacement time histories in X and Y direction for the “as-built”

structure: comparison between numerical and experimental results.

X Direction

-100-80-60-40-20

020406080

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time [s]

Top

Dis

plac

emen

t [m

m]

ExperimentalNumerical

Y Direction

-100-80-60-40-20

020406080

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time [s]

Top

Dis

plac

emen

t [m

m]

ExperimentalNumerical

Figure 8 Center of mass (CM) top displacement time histories in X and Y direction for the retrofitted

structure: comparison between numerical and experimental results.

In Figure 9 and Figure 10 the base shear-top displacement curves related to 0.2g PGA analyses for the X and Y directions are presented for the “as-built” and the retrofitted structure, respectively, (top displacement is referred to the centre of mass of the third story). By comparing the average slopes of the curves, it was possible to assess the stiffness of the structure in the longitudinal and transverse direction. The comparison shows that the stiffness of the structure was greater in the Y direction than in the X one. This result is consistent with the arrangement of the wall type column C6 placed with its strong axis in such direction. Moreover, a very significant stiffness increase provided by the RC jacketing of the two column C1 and C4 was observed. The maximum base shear was about 300 kN and 400 KN in the X and Y direction, respectively, against the values of 200 kN and 270 kN recorded in the X direction and Y direction on the “as-built” structure.

X direction

-800-600-400-200

0200400600800

-100 -80 -60 -40 -20 0 20 40 60 80 100

Top Displacement [mm]

Bas

e Sh

ear [

kN] Y direction

-800-600-400-200

0200400600800

-100 -80 -60 -40 -20 0 20 40 60 80 100


Bas

e Sh

ear [

kN]

Figure 9 Base shear-top displacement relationships in X and Y direction for the “as-built” structure:

numerical results.



X direction

-800-600-400-200

0200400600800

-100 -80 -60 -40 -20 0 20 40 60 80 100


Bas

e Sh

ear [

kN]

Y direction

-800

-600

-400

-200

0

200

400

600

800

-100 -80 -60 -40 -20 0 20 40 60 80 100


Bas

e Sh

ear [

kN]

Figure 10 Base shear-top displacement relationships in X and Y direction for the retrofitted structure:

numerical results.

The comparison of the base torsion-rotation curve shows that the rotational component of the response of the RC jacketed structure was strongly reduced as expected to the rehabilitation design, Figure 11. The maximum rotation was about 11mrad in the case of retrofitted structure, while it was about 15mrad in the case of the original structure.

-1500

-1000

-500

0

500

1000

1500

-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30

Top Rotation [mrad]

Bas

e To

rsio

n [k

Nm

]

-1500

-1000

-500

0

500

1000

1500

-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30

Top Rotation [mrad]

Bas

e To

rsio

n [k

Nm

]

Figure 11 Base torsion-top rotation curve for the “as-built” (left) and retrofitted (right) structure: numerical

results.

Experimental tests and numerical analyses showed the maximum values of interstory rotation at the second level of the “as-built structure”. A significant reduction of the interstory rotation (IR) at the second level was observed for the RC jacketed structure, Figure 12. The maximum interstory rotation was about 4 mrad in the case of retrofitted structure, in the case of the “as-built” structure it was about 8 mrad.

Numerical results indicated that the intervention was effective in reducing the effects of torsion in the development of the failure mechanism at the second floor, improving the structural behaviour. The relocation of the centre of strength of the structure reduced the relative importance of the rotational degree of freedom with respect to the translational ones, thus effectively increasing the seismic resistance of the structure affecting the seismic demand.

-10-8-6-4-202468

10

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15Time [s]

IR [m

rad]

AS-BUILT STRUCTURERETROFITTED STRUCTURE

Figure 12 Interstory rotation (IR) at the second level of the “as-built” and retrofitted structure.

5. DAMAGE INDICES

The assessment of the damage of the structure resulting from the application of the bidirectional motion was performed using the ductility demand-to-capacity ratio (DCR) and the interstory drift ratio (IDR). In order to estimate accurately the damage inflicted to critical columns, the



bidirectional curvature ductility demand-to-capacity ratio (DCR) was used and calculated by the following expression:

22

, ,

yx

u x u y

DCRφφ

φ φ⎛ ⎞⎛ ⎞

= + ⎜ ⎟⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠ (1)

where Φx and Φy are the curvature in the X and Y direction, respectively, and Φu is the ultimate curvature. The ultimate curvature of each column was obtained considering an ultimate compressive strain of concrete equal to 0.0035 and an ultimate tensile strain of steel equal to 0.1. Figure 13 shows the DCR values calculated at the top of the columns of the second story under bidirectional 0.2g PGA earthquake for the “as built” and retrofitted structures. Experimental tests on the “as-built” and RC jacketed structure showed that the most significant damages were detected especially in correspondence of columns ends at the second story. Numerical analyses indicated that the highest DCR value was registered for the column C3 and this result agrees with experimental evidence. Poor local structural detailing involved lack of ductility at vertical elements, which was enhanced by the effects of high axial load in some columns. In particular, column C3, where the axial load was maximum and thus the rotational capacity was limited, was the member with major damages mainly concentrated at the top ends of both first and second story, where a heavy concrete spalling was found. The torsional influence was manifested in higher DCR values at the flexible edges and in smaller values at the stiff edges. Critical columns were columns C1, C2 and C4, which are located at the flexible edges and carry a relatively large axial load and column C3 with the largest axial load. The retrofit intervention on selected columns based on strength centre relocation strategy was effective in reducing the torsional effects with respect to the original configuration. The reduction of torsional influence was evident in a considerable decrease of DCR values at the columns of the flexible edges. The reduction of the strength eccentricity reduced the rotational components of the response, but the lack of ductility of the column C3 was not significantly improved.

0

0.5

1

1.5

2

2.5

DC

R

C1 C2 C3 C4 C5 C6 C7 C8 C9

As-built structureRetrofitted structure

Figure 13 DCR values at the top of the column of the second story under 0.2g PGA earthquake for the “as-

built” and retrofitted structure.

Numerical results suggested that the main parameters to be considered in the structural response were the local story drifts, above all the drift demand at the edge columns. The interstory drift ratio (IDR) was also used as damage indicators on the structure level. It was defined by the following expression:

1i i

i

IDRh

−Δ − Δ= (2)

where is the relative displacement between successive stories and is the story height. Figure 14 shows the interstory drift ratio (IDR) values at the second level of the structure. A reduction of the interstory drift ratio was observed for the RC jacketed with respect to the “as-built” configuration, above all for the column C1, C2, C4 and C7 of the flexible edges.

1i i−Δ − Δ ih



0

0.5

1

1.5

2

2.5

3

IDR

C1 C2 C3 C4 C5 C6 C7 C8 C9

As-built structureRetrofitted structure

Figure 14 Maximum interstory drift ratio (IDR) at the second level for the “as-built” and retrofitted structure.

6. CONCLUSIONS

Numerical models of a plan-wise irregular RC building tested at the JRC ELSA Laboratory were developed and validated with experimental results. The test structure was analyzed in the “as-built” and retrofitted configurations. The effectiveness of using RC jacketing as a global retrofitting technique for the seismic rehabilitation of RC building was evaluated. The design strategy was based on the decrease of the torsional components of the seismic response of the structure, which were highlighted in the response of the “as-built” structure, by means of the reduction of the eccentricity of the centre of strength (CP) with respect to the centre of mass (CM). The aim of the rehabilitation strategy was to increase both the strength and stiffness of the “as-built” structure by the RC jacketing of selected vertical elements. Different damage indices were used to evaluate the effectiveness of the retrofit intervention. Numerical results showed a significant reduction of the interstory rotation (IR) at the second level for the RC jacketed structure. In the “as-built” configuration, column demand to capacity ratios (DCR) were higher for columns at the flexible edges (higher demands) and for element with higher axial loads (lower rotational capacity). The reduction of the eccentricity of the centre of strength reduced the rotational components of the response and a considerable decrease of DCR and IDR values at the critical columns of the flexible edges was observed. However, the lack of ductility of the critical columns, where the axial load was high and thus the rotational capacity was limited, was only partially overcome and high DCR values were registered. Numerical results suggest that design retrofit criteria based on strength relocation requires an accurate knowledge of the response of the structure and the RC jacketing represents an effective retrofitting strategy; the enhancement of the element ductility remains one of the main criteria to be taken into account in conceiving retrofitting intervention, because the poor structural detailing affects adversely the seismic response of the structure.

7. ACKNOWLEDGEMENT

The author would like to acknowledge Dr. Paolo Negro of the JRC ELSA (European Laboratory for Structural Assessment) for his support and collaboration.

8. REFERENCES

Elnashai, A.S. and Elghazouli, A.Y. (1993). “Performance of composite steel/concrete members under earthquake loading, Part I: Analytical model”, Earthquake Engineering & Structural Dynamics, 22(4), 315-345.

Jeong, S.H. and Elnashai, A.S. (2005). “Analytical assessment of an irregular RC frame for full-scale 3D pseudo-dynamic testing - part I: analytical model verification”, Journal of Earthquake Engineering, 9(1), 95-128.

Mander, J.B., Priestley, M.J.N. and Park R. (1988). “Theoretical stress-strain model for confined concrete”, Journal of Structural Engineering, 114(8), 1804-1826.

Martinez-Rueda, J.E. and Elnashai, A.S. (1997). “Confined concrete model under cyclic load”, Materials and Structures, 30(197), 139-147.




Mola, E. and Negro, P. (2005). “Full scale PsD testing of the torsionally unbalanced Spear Structure in the “as-built” and retrofitted configurations”, International Workshop SPEAR, Ispra, Italy.

Negro, P., Mola, E., Molina, J. and Magonette, G. (2004). “Full scale bi-directional PSD testing of a torsionally unbalanced three storey non-seismic RC Frame”, 13th World Conference on Earthquake Engineering, Vancouver, Canada, 968.

SeismoStruct (2005). “Computer program for static and dynamic nonlinear analysis of frame structures”.



REHABILITATION OF EARTHQUAKE-DAMAGED AND SEISMIC-DEFICIENT STRUCTURES USING

FIBRE-REINFORCED POLYMER (FRP) TECHNOLOGY

W. K. Ong1

1FYFE ASIA PTE LTD, 8 Boon Lay Way #10-03 TradeHub 21 Singapore 609964 [email protected]

ABSTRACT: In earthquake-prone regions, buildings and other structures are designed with seismic considerations in accordance with guidelines from the international code of practice. However, earthquake tremors from such regions can be experienced in neighboring non-earthquake-prone regions. As such there is a growing concern towards the structural integrity of non-seismically designed (NSD) structures in these nonearthquake-prone regions. Due to little or no concern towards such threats, seismic considerations are not required under the building regulations in some of these regions. The low available ductility and lack of strength of such NSD structures are posting potential threats to public safety in the event of tremors from a neighboring earthquake. The advent of fibre-reinforced polymer (FRP) technology has provided a potential cost-effective solution to address the deficiencies in these structures. This paper shall look into the potential causes of failure in a NSD RC structure subjected to seismic impact and the techniques to rehabilitate these structures using the state-of-the-art FRP technology.

1. INTRODUCTION

Understanding the potential structural failures in NSD RC structures due to seismic impact is vital to address the deficiencies in these structures. In the event of an earthquake, there is tremendous amount of energy released which creates major ground movement. The cyclic loading resulted from the ground movement requires structures to be designed and incorporated with proper detailing to enable them to inhibit lap-splice failures in the plastic-hinge regions and also to have sufficient shear capacity to ensure ductile flexural response. Past records on structural failures in structures due to earthquake showed that brittle failures in structural members are common. There are various different types of failures commonly found in earthquake-damaged structures:

Shear and flexure cracks in RC beams Crushing of concrete in RC columns near upper/lower ends Failure of shear reinforcement ties in RC columns leading to buckling of steel longitudinal

reinforcements Shear failure in columns and walls Failure of beam-column joints Failure at construction joints Failure of staircases Sway mechanism in the structural frame Complete collapse or sinking of building due to soft storey

Based on an experimental study performed by Beres et al. (1996), there are at least seven critical structural details that could be potential causes of failure in a NSD RC structure subjected to an earthquake. The details are as follows:

1. Cross-sectional area of longitudinal steel reinforcements in RC column is less than 2% of the concrete cross-sectional area.



2. Lapped splices of the longitudinal steel reinforcements in RC column are above the construction joint.

3. Insufficient confinement provided by transverse steel reinforcements. 4. Construction joints above and below the beam-column connection. 5. Discontinuous positive steel reinforcement in the beam. 6. Lack of transverse steel reinforcements in the panel 7. Weak column-strong beam conditions

As observed, the most critical member of the structure in these common structural failures is the vertical structural member – the column.

With its many advantages over conventional strengthening techniques, the FRP technology has been identified as a more viable strengthening technique due to its ease of application, performance, overall time and aesthetical considerations. It offers easier, quicker and reliable application and does not cause distress or add weight to the member to be strengthened. It also forms a protective barrier against ingress of detrimental agents such as moisture, oxygen and carbon dioxide, which helps in arresting further carbonation and corrosion in the strengthened member. This research program series have been designated to understand the effectiveness of FRP technology in strengthening of members of NSD RC structures and rehabilitating earthquake damaged members.

2. OBJECTIVES

A series of research programs, initiated by Fyfe Co. LLC USA and partially funded by California Department of Transportation (CALTRANS), focusing on seismic rehabilitation of structural members of NSD RC structures was carried out. The program conducted seismic rehabilitation on RC rectangular and circular columns, RC beam-column junctions and un-reinforced masonry wall against future seismic impact. The aim of this paper is to look into the seismic upgrade of NSD RC structures using FRP composite system to enhance the ductility, flexural and shear capacities of RC columns. The paper will provide an overview of research programs focusing on seismic rehabilitation of RC columns against future seismic impact. The effectiveness of FRP strengthening to seismic-damaged columns after appropriate repair shall also form part of the discussion in this paper. This part of the research program will allow the understanding on the feasibility and technical effectiveness of the fiberglass/epoxy jacket system in a post-earthquake repair scenario.

3. RESEARCH METHODLOGY

3.1 Test Program

The program will look into the effect of seismic impact on the flexural and shear capacities of vertical structural members of NSD RC structures. A total of three (03) rectangular RC shear column test specimens, four (04) circular RC flexural column test specimens and a circular RC shear column will be subjected to cyclic lateral load and displacement input. These specimens will be tested under the flexural and shear test set-up shown in Figure 1 and Figure 2. The details of test specimens and their retrofitting configurations will be elaborated in the following sections of this paper.



Figure 1 Test set-up for flexural test.

Figure 2 Test set-up for shear test.

3.2 Specification of FRP Composite Materials

The FRP composite system selected for this series of research programs comprises of E-glass and polyaramid is known as the TYFO® Fibrwrap® Composite System. It is comprised of fibres and are stitched bonded by a process that is able to assemble a variety of materials into a single composite material. This composite meets the demands of the construction industry for a lightweight, easily applicable, structurally powerful, and reasonably priced retrofit material. This particular system has been thoroughly tested worldwide and conforms to ICBO AC125 (2003), ICBO material characterization, stringent system testing and durability requirements.

The properties of any composite are governed by the individual properties of the constituents. In particular, the properties of the unidirectional FRP are substantially higher in the longitudinal direction than in the transverse direction. It is the longitudinal properties of composites that are mentioned in this literature for comparative purposes. Properties of the E-glass/Polyaramid composite system used in this research are summarized in Table 1 below.



Table 1 Properties of the e-glass/polyaramid composite system.

Tensile Strength (MPa)

Elastic Modulus (GPa)

Coefficient of Thermal Expansion (10-6/°C)

Ultimate Strain (%)

575 (typ. test value)

26.1 7.74 2.2

3.3 Seismic Research on Rectangular RC Columns

Three (03) rectangular RC columns test specimens with clear height of 2.44m (8ft) and a cross sectional size of 610mm x 406mm (24in x 16in) were fabricated. The columns were reinforced with longitudinal steel reinforcements of diameter 19mm (fy = 414MPa, Grade 60 #6 bars) to give a longitudinal steel ratio of ρL = 0.025. Transverse reinforcements consisted of single peripheral hoops of diameter 6.35mm (#2) plain bars with 90° corner hooks lapped in the cover concrete. The nominal concrete strength at the time of testing was fc’ = 34.5MPa (5000psi). The dimensions and the reinforcement details of the test specimens are shown in Figure 3. These columns were identified as RC01, RC02 and RC03 with RC01 as the control specimen and RC02 and RC03 as the FRP composite strengthened specimens subjected to axial compression load of 507kN (114kips) and 1780kN (400kips) respectively. The higher axial compression load for RC03 is to create conditions more critical for confinement. RC02 and RC03 were identically strengthened with 3.42mm thick of the FRP composite over the entire column height and additional 3.42mm thick of the FRP composite over the top and bottom 600mm (2ft) of the columns. It should be noted that all layers were passive. The details of the test specimens are summarized in Table 2 below.

Figure 3 Dimensions and reinforcement details of test specimens.

Table 2 Details of test specimens – rectangular RC columns.

S/N Specimen Size (mm) Reinforcement (mm) Remarks

1. RC01 Control+507kN(V)

2. RC02 (P)+507kN(V)

3. RC03

610x406x2440 22-Ø19(L);Ø6.35-125(T)

(P)+1780kN(V)

(L)-Longitudinal; (T)-Transverse; (P)-Passive Retrofit; (V)-Vertical Load



3.4 Seismic Research on Circular RC Columns

A total of four circular RC column test specimens were fabricated with diameter of 610mm (24in) and clear height of 3.66m (12ft). The longitudinal steel reinforcements of diameter 19mm (#6 bars) were provided with transverse steel reinforcements of diameter 6.35mm (#2 bars) at spacing of 125mm (5in). The material strengths were fy = 315MPa and fc’ = 34.5MPa (5000psi). The dimensions and the reinforcement details of the test specimens are shown in Figure 3. The four circular column test specimens were identified as CC01, CC02, CC03 and CC04, having CC01 as the control test specimen and the rest retrofitted differently based on active/passive combinations of fibreglass/epoxy confinement layers. CC02 and CC03 were both actively retrofitted with 2.44mm thick and 1.22mm thick of the FRP composites and were also pressure-grouted to achieve an active confinement stress of 1.72MPa and 0.69MPa respectively over the bottom 1.22m height of the column. In addition, both CC02 and CC03 were passively retrofitted with the FRP composites having a nominal thickness of 3.25mm and over the lower height of 305mm (12in). CC04 was pressure-grouted with cement grout to an active pressure of 1.38MPa. The passive retrofit was the same with that for CC02 and CC03 but 1.38mm thick of FRP composite in the circumferential direction and 0.61mm thick of FRP composite with fibre oriented vertically were provided as the active retrofit over 0.91m height instead of 1.22m height of the column. The details of the test specimens are summarized in Table 3.

Table 3 Details of test specimens – circular RC columns

S/N Specimen Size (mm) Reinforcement (mm) Remarks

1. CC01 Control

2. CC02 (A-E)+(P)

3. CC03 (A-E)+(P)

4. CC04

Ø610 26-Ø19(L);Ø6.35-125(T)

(A-C)+(P)

(L)-Longitudinal; (T)-Transverse; (P)-Passive Retrofit; (A-E)-Active Retrofit/Epoxy; (A-C) Active/Cement

3.5 Seismic Research on Earthquake-Damaged Circular RC Column

In this part of the program, a circular RC column test specimen with dimensions and reinforcement details identical to that of the CC-series of test specimens in the above earlier research program was used. The specimen was subjected to fully-reversed increasing cyclic lateral load/displacement input until failure. The failed specimen was repaired with patching, epoxy injection and fibreglass/epoxy jacketing with total thickness of 3.88mm for the full height of the column. The specimen was also subjected to an axial compression load of 1780kN (400kips) in both tests carried out without and with retrofitting.

4. RESULTS AND FINDINGS

4.1 Test Results on Seismic Research on Rectangular RC Columns for Shear

Figure 4 shows the lateral force-deflection curve for the control specimen RC01. The rapid degradation after the shear failure should be noted. The shear failure occurred at a drift ratio (displacement/height) of 1.07%. By comparing with RC01, the performance of the two strengthened columns, RC02 and RC03, was remarkably good, as is apparent from the force deflection hysteresis loops of Figure 5 and Figure 6.

Strengthened specimen RC02 developed stable flexural ductile response with no signs of distress at ductility levels up to μ∆ = 4.5. At μ∆ = 6.0, first signs of distress in the plastic hinge regions at top and



bottom of the column were noted, with slight bulging of the FRP composite jacket on the compression face, indicating that the concrete cover had spalled inside the jacket, and the incipient reinforcement buckling was occurring. At μ∆ = 8.0, the bulging became pronounced, with tearing of the composite jacket at one corner in the bottom hinge region. Significant strength degradation occurred, during the three cycles to μ∆ = 8.0, but even after the three cycles, lateral forces resisted exceeded the theoretical flexural strength. At ductility μ∆ = 10.0, the composite jacket at the lower hinge tore vertically and horizontally resulting in a complete loss of confinement. Degradation was extremely rapid, with crushing of core concrete and buckling of longitudinal steel reinforcement.

In the final cycle, several reinforcements were fractured as a result of the low cycle fatigue associated with alternate bending and strengthening. The maximum shear force sustained by RC02 was 979kN (220ksi) at μ∆ = 8.0. This was 32.5% above the nominal flexural strength based on measured material properties. The yield displacement at 14.88mm (0.586in) was about 60% larger than predicted based on flexural deformations alone, indicating the strong influence of shear.

With reference to Figure 6, RC03 attained a peak load of 1498kN (262.5kips) at μ∆ = 8.0 which was 39% above the nominal flexural strength. Similar to RC02, the first sights of distress occurred at μ∆ = 6.0 with incipient bulging of the jacket on the compression faces of the top and bottom plastic hinge zones. At ductility μ∆ = 8.0, the composite jacket in the upper plastic hinge zone tore, resulting in comparatively rapid strength loss. However, the yield displacement was less than RC02 at 12.45mm (0.49in) despite higher yield force. This is due to the reduced significant of shear as a consequence of the beneficial action of the increased axial compression.

Figure 7 and Figure 8 show plots of jacket horizontal strain vs displacement at a height of 686mm (27in) above the base of RC02 and RC03 respectively. This location is just above the region of increase composite thickness and is typically a location of high strain. It will be seen that in both cases peak strains are about 3000x10-6 and that stable loops are obtained.

Typical strain profiles up the sides of the two columns are shown in Figure 9 and Figure 10. In both cases, strains are initially higher near the top and bottom of the columns. But as shear cracking extends into the central region, strains increase to similar levels as those in the plastic hinge regions.

Figure 4 Hysteresis loops of rectangular column RC01. Figure 5 Hysteresis loops of rectangular column RC02.



Figure 6 Hysteresis loops of rectangular column RC03. Figure 7 Typical strain-deflection plots for rectangular

column RC02.

Figure 8 typical strain-deflection plots for rectangular column RC03.

Figure 9 Typical strain profiles at different ductilities for rectangular column RC02.

Figure 10 Typical strain profiles at different ductilities for rectangular column RC03.

4.2 Test Results on Seismic Research on Circular RC Columns for Flexural

Experimental lateral force-lateral displacement hysteresis curves are shown in Figures 12, 13 and 14 for test specimens CC02, CC03 and CC04 respectively. Each plot includes the theoretical load-deflection envelope based on a nominal concrete compressive strength of f’c = 34.55MPa (shown as a dashed curve). The ideal strength based on f’c = 34.55MPa, fy = 315MPa and ultimate compressive strain of 0.006 and a model for confined concrete is also indicated as Vi.

The response of test specimen CC02, with the highest level of effective confinement, is excellent, with stable hysteresis loops up to the third cycle to displacement ductility levels of μ∆ = +8.0, -6.0. It will be seen that there is no sign of structural degradation associated with bond failure of the starter bars, apparent for control specimen CC01 (compare with Figure 11). Its behaviour is very close to that of a



steel jacket retrofit column reported by Chai et al. (1991). Strength and stiffness differences between CC02 and steel jacket retrofitted columns appear to be primarily due to differences in concrete compression strength. However, structural degradation with fibreglass/epoxy jacket retrofit did not occur until significantly higher displacement than with equivalent steel jacket columns. This apparent improvement in performance may have been a result of more effective confinement at the base of the column, combined with a spread of plasticity up into the column, resulting from the lower stiffness of the retrofit scheme.

The result of CC03, shown in Figure 13, is very similar to that of CC02 until displacement of approximately 150mm at μ∆ = ±6.0 when peak loads at each cycle degrade as a consequence of bond failure. It should be noted, however, that the degradation is very gradual and appears to be stabilizing at μ∆ = ±7.0. It is felt that this is a consequence of the clamping pressure provided across the failing lap-splice. Although this pressure was insufficient to eliminate eventual bond failure, it resulted in a dependable friction force across the failing lap-splice which resisted movement in both directions of loading. It will be noted that the width of the hysteresis loop, measured in the direction of the load axis, at zero displacement decreases after initiation of the bond failure and results in a reduction to the total energy absorbed per cycle.

Despite the higher effective confining stress of CC04, the force-deflection hysteresis behaviour, as shown in Figure 14, is very similar to that exhibited by CC03 except that degradation of CC04 after bond slip commenced seems to be more gradual than with CC03, and appears to be stabilizing at a higher force level for CC04. It should be noted that CC04 was taken to higher displacements than CC03.

Figure 11 Hysteresis loops of rectangular column

CC01. Figure 12 Hysteresis loops of rectangular column

CC02.

Figure 13 Hysteresis loops of rectangular column CC03.

Figure 14 Hysteresis loops of rectangular column CC04.



4.3 Test Results on Seismic Research on Earthquake-Damaged Circular RC Column

Figure 15 shows the force-deflection hysteresis behaviour of specimen without any retrofitting. The force-deflection hysteresis behaviour of the failed specimen retrofitted by steel jacket and FRP composite jacket are given in Figure 16 and 18. The test results indicated that the initial stiffness of the repaired test specimen was very similar to that of the original as-built column and the load-displacement response of the two columns was almost identical up to the displacement ductility μ∆ = 2.0. Thus, the repair measure was effective in restoring the original column stiffness despite the significant shear damage. The as-built column failed rapidly in shear at μ∆ = 3.0 but the repaired specimen sustained the cyclic lateral displacements up to μ∆ = 10.0 without any sign of lateral capacity degradation and with very stable hysteresis loops. The displacement at μ∆ = 10.0 corresponds to a column drift of 4.9%, which is significantly more than what can be expected under a maximum credible earthquake. At μ∆ = 10.0, the test was terminated due to limitations in the displacement capacity of the loading system.

A comparison with an identical damaged column repaired with steel jacket retrofit done in a separate research program is provided by Figure 19. Both the steel jacket retrofitted column and the FRP retrofitted column exhibited the same improved ductile response. This shows that FRP jacket retrofit is fully effective in improving the seismic behaviour equivalent to that of a well designed steel jacket retrofit.

Figure 15 Hysteresis loop of circular column before repair.

Figure 16 Hysteresis loop of circular column retrofitted with steel jacket.

Figure 17 Loading history. Figure 18 Hysteresis loop of circular column retrofitted with FRP jacket.



Figure 19 Comparison of load-deflection envelopes.

5. CONCLUSION

The shear tests on rectangular RC columns retrofitted with passive confinement from FRP jacket showed that shear failure of these shear-deficient columns can be inhibited and will also result in high level of ductility, converting brittle shear failure modes to ductile inelastic flexural deformation modes. This response was partly due to the highly elastic nature of the FRP material. The tests also showed that confinement continued to be provided by the FRP jacket even after the buckling of the longitudinal reinforcement with the increase in the elastic restraint of the FRP jacket as a result of the membrane action developing in the deformed jacket.

From the flexural tests, the failure of lap-splices under cyclic inelastic action can be inhibited with the provision of active confinement from the FRP jacket and epoxy/cement pressure-grout. Active confinement is expected to improve the seismic performance of a column than compared with passive confinement as dilation of the concrete core, which is necessary to activate confinement inna passive retrofit, is not essential in an actively confined retrofit.

The re-test of a failed circular RC shear column repaired with FRP jacket and epoxy injection showed that the employed repair technique was fully effective in restoring the original column stiffness characteristics, in transforming the brittle shear failure mode into a ductile flexural mode and in providing displacement ductility to the systems equal to that observed in a comparative full height jacket retrofitted test column.

Based on the results from the research program, NSD RC structures can be retrofitted to with stand potential seismic impact using FRP technology. The effectiveness of the FRP jacket has made it anviable technique for a post-earthquake structural restoration to seismic-damaged structures.

6. REFERENCES

Beres, A., Pessiki, S. P., White, R.N. and Gergely, P. (1996). “Implications of experiments on seismic behavior of gravity load designed RC beam-to-column connections”, Earthquake Spectra, Vol. 12, No. 2, May, pp. 185-198.

International Conference of Building Officials (2003). “ICBO acceptance criteria for concrete and reinforced and un-reinforced masonry strengthening using Fibre-Reinforced Polymer (FRP) composite systems”, ICBO AC125, June 2003.

Priestley, M.J.N. and Seible, F. (1992). “High Strength Fibre Rectangular Column Shear and Nolap Splice Flexural Tests”, SEQAD Consulting Engineers Report for Fyfe Co. LLC.

Priestley, M.J.N. and Seible, F. (1994). “Column Seismic Retrofit using Fibreglass/Epoxy Jackets”, SEQAD Consulting Engineers Report for Fyfe Co. LLC.



Priestley, M.J.N. and Seible, F. (1993). “Repair of Shear Column using Fibre Rectangular Column Shear and Epoxy Injection”, SEQAD Consulting Engineers Report for Fyfe Co. LLC.

Chai, Y.H., Priestley, M.J.N. and Seible, F. (1991) “Seismic retrofit of circular bridge columns for enhanced flexural performance”, ACI Structural Journal, 88 (5):572-584, Sept./Oct. 1991.



RELIABILITY ANALYSIS OF REINFORCED CONCRETE FRAME UNDER SEISMIC LOAD USING NON LINEAR PUSHOVER ANALYSIS

Muhammad Sigit Darmawan1 and Nur Ahmad Husin1

1Civil Engineering Department, Sepuluh November Institute of Technology, Surabaya, Indonesia Email: [email protected]

ABSTRACT: Earthquake is a major threat to building structures in Indonesia and many other part of the world. Hence, accurate determination of the effect of earthquake on reinforced concrete structures has become the subject of extensive research during the last decade. However, like other natural phenomenon, the determination of the effect of earthquake is subject to uncertainty and variability. Hence, the most realistic way to study the effect of earthquake on concrete structure is using probabilistic (reliability) analysis. This paper will present the reliability analysis of reinforced concrete frame subjected to seismic load. The structure considered is a four floor reinforced concrete frame with a total height of 17 m, and has two column’s span, each of 7 m length. The load considered in the study includes dead load, live load and earthquake load for Jakarta. The uncertainty and variability of material properties, dimension and loads will be taken into account in this study. This frame will then be subjected to earthquake load using non-linear push over analysis to determine the development of plastic hinges in the frame and the failure (collapse) mechanism of the frame. Based on these results, the structural reliability of the frame is formulated and determined. The analysis shows that the probability of failure of reinforced concrete frame increases as the number of plastic hinges formed increases.

1. INTRODUCTION

Earthquake is a major threat to building structures in Indonesia and many other part of the world. Hence, accurate determination of the effect of earthquake on reinforced concrete (RC) structures has become the subject of extensive research during the last decade. However, like other natural phenomenon, the determination of the effect of earthquake is subject to uncertainty and variability. Both loads (dead load, live load and earthquake load) and resistance of RC frame (dimensions of elements, yield strength of steel, reinforcement percentage, concrete strength etc) are subjected to uncertainty. Hence, the most realistic way to study the effect of earthquake on concrete structure is using probabilistic (reliability) analysis.

Previous study performed by Husin (2005) found that the probability of failure of four floor RC frame is 1,811x10-9, when it is not subjected to earthquake load. However, when the RC frame is subjected to an increasing earthquake load, the probability of failure of the frame also increases and reaches 4197,9381x10-4 at approaching failure. However, in this study Husin (2005) used first order second moment reliability index to determine the probability of failure of RC frame. This approach is not accurate since only mean and standard deviation of random variables is utilized and ignores the distribution of the random variables.

This paper considers the probability of failure of structural system of RC frame subjected to dead, live and earthquake load, for different level of earthquake load. In this study, the reliability analysis is determined using Monte Carlo simulation. Monte Carlo simulation is commonly utilized as it applies to any form of limit state, unrestricted to type of probabilistic description of random variables and combination of many elements (structural system). Therefore, this method is used to check the validity of other method in reliability analysis as mentioned by Melchers (1999). The earthquake load used in this study is the earthquake load for Jakarta based on SNI-1726.

2. RESEARCH METHODOLOGY

To obtain the probability of failure of RC frame, in this study the following steps are taken:




a. For a trial dimension of RC frame members, perform pushover analysis to get the development of plastic hinges in the frame, which satisfy the seismic provision, i.e. hinges formed first in the beams and not in the columns.

b. Perform structural system modeling based on the development of plastic hinges obtained in step 1.

c. Develop load function that covers dead, live and earthquake load and also develop resistance as a function of beam width (b), beam depth (h), effective beam depth (d), tension reinforcement area (As), compression reinforcement are (As

’), column width (bk), column height (hk), steel yield strength (fy), shear reinforcement area (Av) and stirrup spacing (S).

d. Determine the probability of failure of RC frame elements (beams and columns) against applied moment (M) and shear (V).

e. Determine the reliability of RC frame (as a system) for each stage of plastic hinge development.

3. STRUCTURAL MODELING

Structure studied in this paper is a 2 dimension of concrete frame, which is used as office. The structure has four floors with a total height of 17 m, and has two column’s span, each of 7 m length. Complete description of building configuration is shown in Figure 1, while dimension and reinforcement of beams and columns are summarized in Table 1. As shown in Table 1, the shear reinforcement required is very high to guarantee that shear failure does not precede flexural failure.

3 @ 4 m

5 m

2 x 7 m

Figure 1 Structural configuration of building. Table 1 Dimension and reinforcement detail of beam and column.

No Element Left-support Right-support Shear Reinforcement

1 Beam (30/60) Top 8D19 8D19

Bottom 4D19 4D19 3Ø10-100

2 Column (40/70) 16D29 16D29 3Ø10-100

4. COLLAPSE MECHANISM ANALYSIS

To determine the hinges development of RC frame at different level of earthquake load and the collapse mechanism, non linear static pushover analysis is utilized. The non linear static pushover analysis is performed using finite element program called SAP 2000. The development of plastic hinges is given in Figure 2 to 5. Figure 2 shows the initial condition of RC frame when it is not subjected to earthquake load (Initial Condition/step-0). As the level of earthquake is increased, the first hinge is formed on one of the beam at first floor. This stage is called Immediate Occupancy



Condition/step-1. With further increasing load, the number of hinges formed also increases. Finally, the RC frame fails when hinges have formed in all lower columns. This stage is called Critical Prevention Condition/step-3. Note that Life Safety Condition (Step-2) is defined as the condition in between Step-1 and Step-3, see Figure 4. Based on this hinges development, the reliability analysis is performed for each different level of earthquake load.

5. RELIABILITY ANALYSIS

A frame structure is a highly redundant system. Therefore, the failure of one of its element (component) does not lead in the failure of the frame system. Assuming that a frame behave as a perfect ductile structure, the frame will fail only when sufficient number of plastic hinges have formed to cause a collapse mechanism. In addition, there may be a number of possible collapse mechanism in a frame structure. All of these possible mechanism are to be identified and the system reliability is to be calculated. The reliability analysis of RC frame is performed into 2 (two) stages. First, the determination of reliability of each structural element of RC frame, i.e. beams and columns. This then followed by the determination reliability of RC frame as a system. To determine the reliability of each element, the performance function (limit state function) has to be developed first.

Figure 2 Initial condition (Step-0).

Figure 3 Immediate occupancy condition (Step-1).

Figure 4 Life safety condition (Step-2).



Figure 5 Critical prevention condition (Step-3).

6. PERFORMANCE FUNCTION

6.1 Flexural Failure of Beam

The performance function of beam against flexure is defined as follows: if ys ff <'

( )'.600'1'2.1.1.'..85.0)( dd

xdAsxdxbfcG MB −⎟⎠⎞

⎜⎝⎛ −+⎟

⎠⎞

⎜⎝⎛ −=

ββ ( )ELD MMM ++− (1)

if ys ff ≥'

( )'.'.2.1.1.'..85.0)( ddfyAsxdxbfcG MB −+⎟⎠⎞

⎜⎝⎛ −=

ββ ( )ELD MMM ++− (2)

where x is distance of neutral axis from top of compressive zone and determined as

)'85.0(2)''600)('85.0(4)'600()'600(

1

12

ββ

bfcdAsbfcAsfyAsAsfyAs

x+−±−−

= (3)

and )28'(*007.085.0 −−= fcβ ; 0.65<β<0.85 (4)

MD, ML, ME, are the applied moment due to dead, live and earthquake respectively, d’ is the distance from top of compressive zone to compressive reinforcement and fs’ is the stress in the compression reinforcement.

6.2 Shear Failure of Beam

The performance function of beam against shear flexure is defined as follows:

( ) ( ) ( ELDscVB VVVVVG ++−+= ) (5)

( ) ( ELDyv

cVB VVVS

dfAdbfG ++−⎟⎟

⎠

⎞⎜⎜⎝

⎛+=

...'

61 ) (6)

VD, VL, VE, are the applied shear due to dead, live and earthquake respectively. When plastic hinge is formed in the beam, concrete contribution against shear (Vc) is 0.

6.3 Columns Capacity against Axial Load and Flexural Moment

For a given axial load, the flexural moment capacity of column is: if ys ff <'



( ) =−MKGkk hxb

xdAsdAs

xhdAshAs ...8375.10'..5.13'..90.'..45'..300

2

++−−

( )ELDk MMMdfyAshfyAsxb− .212.9

f >'

++−−+ ...15.02

... 2

f

(7)

if ys

)( MKG − xhbdfyAshfyAskk ...8375.10.'..15.0

2.'.

+−= kk

k hAsx

hdAsxb ..300...300..212.9 2 −+−

dAsx

dAs..90− ..90

2+ )( ELD MMM ++− (8)

where x is determined as

)'85.0(2)690)('85.0(4)1200()1200(

1

12

ββ

k

ELDk

bfcPPPAsdbfcAsAs

x++++±−

= (9)

PD, PL, PE, are the applied shear due to dead, live and earthquake respectively. Note that for columns As’=As.

6.4 Columns Capacity against Shear

The performance function of column against shear flexure is defined as follows:

( ) ( ) ( ELDccGK VVVVVG ++−+=− )

( ) ( ELDyv

kcg

uGK VVV

SdfA

dbfA

NG ++−

⎟⎟⎠

⎞⎜⎜⎝

⎛+⎟

⎟⎠

⎞⎜⎜⎝

⎛+=−

...'

61

.1412 ) (10)

When plastic hinge is formed in the beam, concrete contribution against shear (Vc) is 0. The probability of failure of RC frame element (beam and column) is determined by Monte Carlos simulation. The step used in Monte Carlo simulation is as follows a. Generate randomly the variable defined in the performance function b. Calculate the performance function G(x) c. If G(x) <0, failure takes place d. Record the number of failure e. Repeat step-1 to step-4

The probability of failure is then determined by the following formula

NN

p ff = (11)

where pf is the probability of failure and Nf is the number failures in N simulation runs. Table 2 shows the probability of failure of beams and columns of the frame at different level of earthquake load (Step). See Figure 6 for frame and hinge numbering system used. Number in black is used for hinge number, while number in red used for frame and joint number.



1

2

3

4

5

7

8

9

11

12

13

14

15

16

25 26 27

29 30 31

33 34 35

37 38 39

10

6

17

18

19

20

21

22

23

24

28

32

36

40

Figure 6 Frame and Hinges Numbering System.

Table 2 and 3 shows that with increasing level of the earthquake load, the probability of failure of beams and columns of RC frame also increases. Note that for beams, at step-3 all the beams have failed, while for columns only lower columns have failed. This is expected since the RC frame has been designed to comply with the concept of strong-column weak beam.

Table 2 Probability of failure (Pf) of beam elements. Frame Step Pf Frame Step Pf Frame Step Pf Frame Step Pf

5-37 step-0 0.0 11-33 step-

0 0.0 13-29 step-

0 0.0 15-25 step-

0 0.0

5-37 step-1 0.0 11-33 step-

1 7.3000E-

05 13-29 step-

1 0.5889560 15-25 step-

1 0.9185150

5-37 step-2 0.9223865 11-33 step-

2 0.9329835 13-29 step-

2 0.9484140 15-25 step-

2 0.9489405

5-37 step-3 0.9633095 11-33 step-

3 0.9718760 13-29 step-

3 0.9803245 15-25 step-

3 0.9804145

5-38 step-0 0.0 11-34 step-

0 0.0 13-30 step-

0 0.0 15-26 step-

0 0.0

5-38 step-1 0.0 11-34 step-

1 1.4403E-

02 13-30 step-

1 0.4342685 15-26 step-

1 0.6186005

5-38 step-2 0.1588945 11-34 step-

2 0.8518040 13-30 step-

2 0.8858565 15-26 step-

2 0.8813780

5-38 step-3 0.9062840 11-34 step-

3 0.9307395 13-30 step-

3 0.9471700 15-26 step-

3 0.9474650

6-39 step-0 0.0 12-35 step-

0 0.0 14-31 step-

0 0.0 16-27 step-

0 0.0

6-39 step-1 0.0 12-35 step-

1 1.1500E-

05 14-31 step-

1 0.3222170 16-27 step-

1 0.6138460

6-39 step-2 0.5259055 12-35 step-

2 0.9148980 14-31 step-

2 0.9455335 16-27 step-

2 0.9398235

6-39 step-3 0.9537550 12-35 step-

3 0.9637160 14-31 step-

3 0.9791435 16-27 step-

3 0.9767510



6-40 step-0 0.0 12-36 step-

0 0.0 14-32 step-

0 0.0 16-28 step-

0 0.0

6-40 step-1 0.0 12-36 step-

1 2.3718E-

02 14-32 step-

1 0.4273085 16-28 step-

1 0.7043320

6-40 step-2 0.2513475 12-36 step-

2 0.8166825 14-32 step-

2 0.8725485 16-28 step-

2 0.8867065

6-40 step-3 0.9068040 12-36 step-

3 0.9078670 14-32 step-

3 0.9393395 16-28 step-

3 0.9475505

Note: 20-16: Number 20 is frame number and 16 is hinge number

Table 3 Probability of failure (Pf) of beam elements.

Frame Step Pf Frame Step Pf Frame Step Pf

1-1 step-0 0.0 7-23 step-0 0.0 17-9 step-0 0.0

1-1 step-1 0.0 7-23 step-1 0.0 17-9 step-1 0.0

1-1 step-2 0.1564435 7-23 step-2 0.0 17-9 step-2 0.3537920

1-1 step-3 0.6530700 7-23 step-3 0.0 17-9 step-3 0.6908710

1-2 step-0 0.0 7-24 step-0 0.0 17-10 step-0 0.0

1-2 step-1 0.0 7-24 step-1 0.0 17-10 step-1 0.0

1-2 step-2 0.0 7-24 step-2 0.0 17-10 step-2 0.0

1-2 step-3 0.0 7-24 step-3 0.0 17-10 step-3 0.0

2-3 step-0 0.0 8-21 step-0 0.0 18-11 step-0 0.0

2-3 step-1 0.0 8-21 step-1 0.0 18-11 step-1 0.0

2-3 step-2 0.0 8-21 step-2 0.0 18-11 step-2 0.0

2-3 step-3 0.0 8-21 step-3 0.0 18-11 step-3 6.6666E-07

2-4 step-0 0.0 8-22 step-0 0.0 18-12 step-0 0.0

2-4 step-1 0.0 8-22 step-1 0.0 18-12 step-1 0.0

2-4 step-2 0.0 8-22 step-2 0.0 18-12 step-2 0.0

2-4 step-3 0.0 8-22 step-3 1.5000E-06 18-12 step-3 3.3333E-07

3-5 step-0 0.0 9-19 step-0 0.0 19-13 step-0 0.0

3-5 step-1 0.0 9-19 step-1 0.0 19-13 step-1 0.0

3-5 step-2 0.0 9-19 step-2 0.0 19-13 step-2 0.0

3-5 step-3 0.0 9-19 step-3 0.0 19-13 step-3 0.0

3-6 step-0 0.0 9-20 step-0 0.0 19-14 step-0 0.0

3-6 step-1 0.0 9-20 step-1 0.0 19-14 step-1 0.0

3-6 step-2 0.0 9-20 step-2 0.0 19-14 step-2 1.667E-06

3-6 step-3 0.0 9-20 step-3 0.0 19-14 step-3 8.133E-05

4-7 step-0 0.0 10-17 step-0 0.0 20-15 step-0 0.0

4-7 step-1 0.0 10-17 step-1 0.0 20-15 step-1 0.0

4-7 step-2 0.0 10-17 step-2 0.2290730 20-15 step-2 0.0

4-7 step-3 0.0 10-17 step-3 0.6710645 20-15 step-3 0.0

4-8 step-0 0.0 10-18 step-0 0.0 20-16 step-0 0.0

4-8 step-1 0.0 10-18 step-1 0.0 20-16 step-1 0.0

4-8 step-2 0.0 10-18 step-2 0.0 20-16 step-2 0.0

4-8 step-3 0.0 10-18 step-3 0.0 20-16 step-3 2.0666E-05

Note: 20-16: Number 20 is frame number and 16 is hinge number



7. RELIABILITY OF RC FRAME SYSTEM

Having determined the probability of failure of RC frame element, the probability of failure of RC frame will be determined. Based on hinge development and collapse mechanism shown before, the probability of RC frame is determined. The sequence of hinges development is shown in Figure 7. Figure 7 shows that there is four step/stage of hinge development. Number in box is hinge number as defined in Figure 6.

Figure 7 Sequence of hinge development.

Based on Figure 7, the probability of failure of RC frame system is formulated as shown in Figure 8. The failure of RC frame system is defined as either a. Failure of all beams or b. Failure of all lower columns c. Failure of one of columns

Note that 25M in Figure 8 indicates hinge number and caused of failure. M is failure caused by flexure and V is failure by shear.

Figure 8 System modeling of RC frame.

25

25,26,27,28,29,30,31,32,33,34,35,36,37

25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,1

Step-0

Step-1

Step-2

Step-3




8. RESULTS

At initial condition (Step-0), RC frame is subjected by dead and live load only. No earthquake is applied at this stage. The probability of failure of RC frame is 0. This is expected since the RC frame is designed based on the combination of dead, live and earthquake load. When the earthquake load increases, the plastic hinge starts to form in the RC frame. The first plastic hinge form in the beam located at first floor at end side (see Figure 3).This stage is called Immediate Condition (Step-1). At this stage, the probability of failure of RC frame is around 2.021x10-10. The base shear force has reached 141% of design based shear. With further increasing load, the hinges form at almost all beams. This stage is called as Life Safety Condition (Step-2). The base shear force has reached 141% of design based shear and the probability of failure of RC frame is around 1,529x10-5. Finally, plastic hinges form at all the beams and lower columns base when the base shear force has reached 257%. The probability of failure at this stage is 0.519.

9. CONCLUSIONS

a. The probability of failure of RC frame increases with increasing number hinges formed (increasing level of earthquake applied).

b. RC frame can be considered in a safe condition up to Life Safety Condition (Stage-2). At this stage, the RC frame still has a an acceptable probability of failure of 1,2957426 x 10-4. However, at Critical Condition (Stage-3) the probability of failure of RC frame has reached 0.519.

10. REFERENCES

Husin, N.A. (2005). “Reliability of Two-Dimensions Reinforced Concrete Frame System”, Master Thesis, Sepuluh November Institute of Technology.

Melchers, R.R. (1999). ”Structural Reliability Analysis and Prediction, 2nd edition”, John Wiley & Sons, New York.

SNI-1726 (2002). “Tata-cara Perencanaan Ketahanan Gempa Untuk Bangunan Gedung (Guidelines for Designing Earthquake Resistance Building)“, Badan Standarisasi Nasional (BSN).

SAP (2000). “Linear and Nonlinear Static and Dynamic Analysis and Design of Three-Dimensional Structures”, Computers and Structures, Inc. Berkeley, California, USA.

Surahman, A. and Rojiani, K.B. (1983). ”Reliability based optimum design of concrete frames“, Journal of Structural Engineering, ASCE, Vol. 109, No. 3, 741-757.



BEHAVIOR OF COLLAPSED BUILDINGS CAUSED BY 2004 SUMATERA EARTHQUAKE IN BANDA ACEH

Agussalim1, Yulia Hayati1 and Samsunan2

1 Department of Civil Engineering, Syiah Kuala University,

Banda Aceh 23111, Indonesia 2 Aceh Society of Civil Engineers (AcSCE) Banda Aceh 23111, Indonesia


ABSTRACT: The objective of this study is to investigate the behavior of collapse of buildings caused by the December 2004 Sumatera earthquake in Banda Aceh region, and those were not affected by tsunami. Twenty structural buildings have been selected to know the characteristics of damages of frames (beams and columns, including beam-column joints) of collapsed buildings. The types of buildings vary from two to five-floor of reinforced concrete structures (RC) buildings. The result of study shows that the collapse of structural buildings were mostly caused by shear or shear-flexural column failures due to the weakness of reinforcement details, such as the content and size of column stirrups or longitudinal bars below those specifies by Indonesian Concrete Codes (SNI, 2002). Detail failures of columns of collapsed buildings are caused by three modes of failures, i.e.: lap splices, local column buckling, and cut of longitudinal reinforcements.

1. INTRODUCTION

December 2004 Sumatera earthquake and tsunami hit Aceh West coast and East coast. The earthquake significantly affected the districts along West coast of Aceh Province. This Sumatera natural disaster has caused the loss of more than 200,000 people in Aceh province. That was the worst natural disaster in Indonesia after Krakatau explosion around Sunda strait in 1883. Many houses, school buildings, health facilities, roads and bridges, airports, other infrastructures, and lifeline facilities were destroyed. The objective of this study is to investigate the behavior of collapsed structural buildings (buildings) caused by the December 2004 Sumatra earthquake in Banda Aceh City region, those were not affected by tsunami. Twenty collapsed buildings have been selected to know the characteristic of damaged frames (beams and columns, including beam-column joints). In the following section the method of investigation is explained. Then in the next section is explained about result of study and discussion, where at the end paper will be closed by the conclusion of the study.

2. METHOD OF INVESTIGATION

The method of investigation is as follows. This includes the definition of the category of damaged buildings, method of collecting data, and how the evaluation and analysis were done.

2.1 Category of Damaged Buildings

Damaged buildings are divided into four categories (AIJ Committee, 1995): collapsed, severely damaged, medium damaged, and slightly damaged.

1. A building is considered collapsed when one story or more collapsed, including one the first story of which has collapsed. Generally, in case of the top story collapse there are categorized as partially collapse, some time that can be repaired if the structural frame of lower story is still stable and save condition;

2. A building is considered severely damaged when the structural frames or columns are unstable due to permanent deformation. In general the permanent deformation caused by wide cracks at the structural elements. The severely damaged building can not be repaired at all;



3. A building is considered medium damaged when the structural frames are stable and safe. There are slightly cracks at beams and or columns. The medium damaged building can be repaired at certain beams or columns; and

4. A building is considered slightly damaged when the structural frames are stable and safe. There are not structural damages such as damage of wall, ceiling, and other non structural elements.

2.2 Method of Collecting Data

Data are primary and secondary. Primary data are collected directly at the location of damaged buildings, such as the measurement of size of structural elements, the existing concrete strength, etc. Secondary data are collected from the related authority offices, such as material specification and engineering drawing, including detail of steel bar arrangement at critical sections and joints. Data were collected by two teams, where each team led by a senior engineer and accompanied by two junior engineers or graduate students. The equipment used consisted of concrete hammer test, measuring tape, and toiletries.

The target study of damaged buildings consisted of twenty reinforced concrete buildings. The selected structural buildings are eleven 2F buildings, five 3F building, two 4F building, and two 5F buildings (See Table 1). Those are the governmental offices and public buildings, including private houses.

2.3 Method of Analysis

The aims of the analysis and evaluation in this study are to evaluate and analyze the damaged buildings and to show the cause of the damage of those, specifically. For example, why some buildings were totally collapsed and the others one were not. Probably it is due to the weakness of material quality or the minimum requirements are not satisfied. The other consideration is whether those buildings were built according to the Codes, the new Indonesian Concrete Code (SNI, 2002), in particular or not.


As the results of the study, in this section the collapsed buildings located in Banda Aceh were examined. Twenty collapsed buildings caused by December 26, 2004 earthquake and the earthquakes in March 2005 were investigated. The identities of collapsed buildings are shown in Table 1. As shown in the table, the condition of the collapsed buildings are as follows: two buildings were first story only collapsed, one building was first and second story collapsed, one building was second to five story collapsed (partially collapsed), and sixteen buildings were totally collapsed. In general, the typical failure types of structural frames of collapsed buildings are column shear and or flexural failures. In the following, the conditions of collapsed buildings and detail of column failures are explained.

Table 1 The identities of collapsed buildings in Banda Aceh.

Number Building Identity Owner Location (Roads)

1 Pante Pirak Supermarket, 4F Private Enterprise Pante Pirak

2 Traditional Shop1, 2F (2 units) Private Enterprise Sri Ratu Safiatuddin

3 Traditional Shop2, 2F (2 units) Private Enterprise Hasan Dek

4 Kuala Tripa Hotel, 5F Private Firm Masjid Raya

5 Traditional Shop3, 3F Private Enterprise Diponegoro

6 Traditional Shop 4, 3F Private Enterprise SA. Mahmudsyah

7 Traditional Shop 5, 2F (4 units) Private Enterprise T. Iskandar

8 Traditional House , 4F Private Enterprise T. Iskandar



9 Treasury Buildings, 3F Treasury Department Cik Di Tiro

10 Fakinah Nurse Academy, 5F Private Foundation Cik Di Tiro

11 Zikra Book Store, 3F Private Enterprise KH. Ahmad Dahlan

12 Naga Sakti Dormitory, 2F Private Foundation P. Nyak Makam

13 Gubernor Office, 2F Aceh Government T. Nyak Arif

14 Traditional Shop , 2F Private Enterprise Sukarno-Hatta

15 Traditional Shop, 3F Private Enterprise Sukarno-Hatta

16 Traditional Shop , 2F Private Enterprise T. Nyak Arif

3.1 Condition of Collapse of Buildings

Base on the number of collapsed story, the condition of collapsed buildings are two groups i.e.: the buildings were first story and or second story collapsed, and the buildings were totally collapsed or each story collapsed.

3.1.1 First and or second story collapsed

The collapse of first story only consisted of two buildings, those are Kuala Tripa Hotel and a traditional shop3, 3F (No.4 and No. 5 in Table 1). The collapse of Kuala Tripa Hotel is shown in Figure 1a. The collapse of first story is due to the weak story or soft story. Generally the first story of hotel or traditional shop is open frame type, such as for lobby, café, show room, etc. Second story and above story are survive with limited deformation. Banda Aceh is located at high risk earthquake zone of Region-6 of Indonesian earthquake region map with return period of 500 years (SNI, 2003). Therefore all buildings located Banda Aceh should be designed according to earthquake resistant buildings of Region-6 (SNI, 2003).

The second case is the collapse of first and second story is only one building, a traditional shop4 (No.6 in Table 1). Same as the above case, the collapse of first story is due to the soft story condition, while the second story collapsed caused by the gravitational impact suddenly after the collapse of first story.

a) Collapse of first story of five stories building b) Total collapse of two stories building

Figure 1 Collapse of structural RC buildings.

3.1.2 Total collapse of buildings

The others buildings (No.1-3 and No.7-16) are totally collapsed, except one building (No.10 in Table 1) was partially collapsed where the first story is surviving. Those mean that each story or floor of buildings fell down on the ground floor. The failure mechanism is initiated by first story than followed by second story and others above stories. Visually showed that this total collapse



caused by the following items i.e.: a). frames as the open frame types; b). the weakness of RC column strengths, where mostly the column shear failures occurred; and c). the RC beam-column joints of outer columns were weak due to the mistake of detailing at joints, particularly the anchorage weakness and no confining stirrup at joints. The typical total collapse of two story building is shown in Figure 1b.

3.2 Detail of Column Failures

In the following, the mode of column failures that caused the collapse of building is explained. There are three modes of column failures i.e.: lap splices, local column buckling, and cut of longitudinal reinforcements.

3.2.1 Lap splice of longitudinal bars

The first failure mode of collapsed buildings is lap splice failure as shown in Figure 2a. Here, the failure occurred due to the loss of bond at splice reinforcements. In this case, the weakness of splice is caused by the splice length which is shorter than that specified by Indonesian Concrete Code (SNI, 2002) and column stirrup content is smaller than that specified by the Code, also.

3.2.2 Local buckling of column ends

The second failure mode of collapsed buildings is local buckling failure of columns as shown in Figure 2b. As shown in the figure, the longitudinal bars buckled at the end of column and concrete crashed, as well. In this case, the buckling occurred due the weakness of confining strength at the column ends. The column stirrup content and size are too small. At the end of column, the stirrup diameter and its spacing are 8 mm and 200 mm, respectively. The Concrete Code (SNI, 2002) at least recommends that the stirrup diameter and its spacing for shear reinforcement are 10 mm and 100 mm, respectively. According to the Old Concrete Code (Peraturan Beton Indonesia, 1971) the matter of stirrup requirement is not specified clearly. Even though, mostly, the collapsed buildings were constructed according to the Old Concrete Code (Peraturan Beton Indonesia, 1971).

a) Lap splice of longitudinal bars failure mode b) Local buckling failure mode

Figure 2 The failure of ends of columns

3.2.3 Cut of longitudinal bars

The third failure mode of collapsed buildings is the cut of longitudinal bar as shown in Figure 3. As shown in the figure, the horizontal swing of building frame cut off the longitudinal bars at tie beam-column joint due to the weakness of longitudinal bar strength. The content of this main reinforcement is few and the figure shows that the stirrup diameter is small, also.




Figure 3 Cut of longitudinal bar failure mode (few of bar content).

Base on the above three modes of column failures, the failure of RC frames are caused by column shear failures and column shear-flexural failures. Local buckling column failure modes are typical of column shear failures, while lap splice of longitudinal bars and cut of longitudinal bars of RC columns are known as column shear-flexural failures.

3.3 Existing Concrete Strengths

According to the data recorded by concrete hammer test at the face of column and beam elements of several collapsed buildings, the existing concrete strengths vary from 20 - 40 MPa. Based on this result, concrete strengths are relatively sufficient according to the Old Concrete Code (Peraturan Beton Indonesia, 1971), while the New Concrete Code (SNI, 2002) specifies the concrete strength for concrete structural buildings at least 25 MPa.

4. CONCLUSIONS

According to the discussion in Section 3, the following conclusions can be drawn. a. Collapse of structural buildings were mostly caused by column shear or shear-flexural

failures due to the weakness of reinforcement details, such as the content and size of column stirrups or longitudinal bars below those specifies by Indonesian Concrete Codes;

b. Detail failure of columns of collapsed buildings are caused by three modes of failures, i.e.: Lap splices, local column buckling, and cut of longitudinal reinforcements;

c. Most of the collapsed buildings were built according to 1971 Indonesian Concrete Codes.

5. ACKNOWLEDGMENTS

The author would like to express his sincere gratitude to the survey teams of Department of Civil Engineering, Syiah Kuala University Banda Aceh and the members of Aceh Society of Civil Engineer (AcSCE) as data collectors. Also thanks to Head of Material and Structural Laboratory of Civil Engineering Department, Syiah Kuala University for their cooperation during this study.

6. REFERENCES

AIJ Committee (1995). “Preliminary Reconnaissance Report on the 1995 Hyogoken-Nanbu Earthquake”, AIJ-Japan, April 1995.

Peraturan Beton Indonesia (1991). Departemen Pekerjaan Umum, Badan Pembina Konstruksi dan Investigasi, Jakarta.

Standar Nasional Indonesia (2002). “SNI 03-2847-2002, Indonesian Concrete Code”, November 2002.

Standar Nasional Indonesia (2003). “SNI 03-1726-2003, Design for Earthquake Resistant Buildings”.

Ultra Cipta Consultant (2005). “Report on Limited Investigation of Damage of Governmental and Public Buildings in Banda Aceh”, November 2005.



DYNAMIC RESPONSE OF REINFORCED CONCRETE BUILDING FRAME WITH TRANSFER SLAB

Jamal A. Abdalla1

1 Department of Civil Engineering, American University of Sharjah, Sharjah, United Arab Emirates


ABSTRACT: This paper presents a study of the effect of transfer slab on the dynamic response of a multistory reinforced concrete building. Two frames extracted from two buildings – one with normal slab and the other with a transfer slab at its first floor are used as the case study. Modal analyses of the two frames were carried out and the dynamic characteristics of the two frames were compared to assess the effect of transfer slab on the dynamic characteristics of the frame. The two frames were then subjected to two scaled down dynamic loads – harmonic (sine function) load and El Centro 1940 earthquake load and linear dynamic analyses were carried out to compare the responses of the two frames to the two dynamic loads. The effect of the transfer slab on the dynamic behavior of the building frame is then assessed. It is observed from this study that the peak lateral drift of the roof of the frame with transfer slab is less sensitive to the type of the two dynamic loads as compared to the peak lateral drift of the frame with normal slab. Also, at the level of the transfer slab, the peak lateral drift of the frame with transfer slab is much lower than the corresponding peak lateral drift of the frame with normal slab for both dynamic loads.

1. INTRODUCTION

The use of transfer slab (transfer plate) and transfer beam (transfer girder) in reinforced concrete buildings has increased significantly in recent years in many low-rise to medium-rise buildings. The aim of using transfer slabs and transfer girders is to provide large open spaces, with few or without interior columns, in the floor beneath for different functions. Transfer slabs that carry columns of 10 to 15 stories are very common. The transfer slab is usually designed to carry very large concentrated loads that are transferred by the columns supporting the floors above the transfer slab level. Such large concentrated loads require very high punching shear capacity. Accordingly, the thickness of a typical transfer slab is usually several times that of the normal slab and generally ranges from 1.5 m to 3.0 m. In addition, the removal of interior columns results in large span coupled with heavy dead load of the thick transfer slab and the large concentrated loads of the terminated columns these factors may result in large deflection as well. Also, the use of transfer slab in reinforced concrete buildings introduces a very large mass at that floor and therefore large inertia forces will developed at the transfer slab level when subjected to earthquake excitation. Accordingly, the dynamic behavior of the building will be affected. The dynamic effect of transfer slab is usually not accounted for when buildings with transfer slabs were designed based on static analysis only (Abdalla, 2006).

In recent years there are several modern buildings, especially in Honk Kong, in which transfer slabs or transfer beams were used. Several investigators have tackled the issue of transfer slab and transfer beam. Su et al. (2002) assessed the seismic performance of transfer plate high-rise building with emphasis on higher mode effects and soft story mechanism in columns beneath the transfer plate. They concluded that: (1) the load-displacement characteristics of the building has indicated poor ductility of the columns beneath the transfer plate; (2) the seismic inter-storey drift demands at lower and upper zones are higher than those of wind; and (3) the columns beneath the transfer slab and the coupling beams are the most vulnerable components under seismic loading.

Li et al. (2003) assessed the seismic performance of a low-rise building with transfer beam. They reached conclusion similar to that of Su et al. (2002) in terms of low ductility of columns supporting the transfer beam and the seismic vulnerability of the short-span beams connecting the column under seismic load.



Kuang and Zhang (2003) also investigated the behavior of transfer plate-shear wall with in-plane loading systems in tall buildings, however, they included the interaction effect of the transfer plate and the supported shear walls on the structural behavior of the system. They observed that the interaction effect causes significant stress redistributions both in the transfer plate and in the shear walls. Kuang and Li (2005) developed formulas for interaction-based design for transfer beam using box foundation analogy. The formulas they developed predicted, very well, the bending moment and the axial force of the transfer beams that supported the in-plane shear wall when compared with the finite element results.

Li et al. (2006) conducted a shake table experiment on a high-rise building with a transfer plate system. They observed that most of the damage took place at floors above the transfer slab system and concluded that story drift correlates very well with the degree of structural damage.

In this paper a preliminary study was conducted to compare the dynamic characteristics and dynamic behavior of building frames with normal slab and one with a transfer slab. Modal analyses of the two frames were carried out and the effect of transfer slab on the dynamic characteristics of the frames was assessed. The two frames were then subjected to two scaled down time history loads – harmonic (sine) and El Centro 1940 earthquake and linear dynamic analyses were carried out to compare the responses of the two frames to the two different dynamic loadings.

2. DESCRIPTION OF THE BUILDING FRAMES

The frames investigated are hypothetical and they represent frames extracted from two low-rise buildings – one with normal slab and the other with a transfer slab. Only self weight and scaled down dynamic loads (harmonic and El Centro earthquake) were considered. Each of the 2D frames has five bays and eight floors of similar spans and height. A preliminary design with a design criterion that stress level should be kept below nonlinear levels for all members was carried out. The harmonic and the earthquake loads were scaled down to ensure linear behaviour of all members. Figure 1 shows a sketch of the two frames.

Figure 1 Building frames: (a) with normal slab; (b) with a transfer slab.

(a) (b)

3. DYNAMIC CHARACTERISTICS OF THE TWO FRAMES

From the modal analysis it is observed that the period of vibration of the frame with normal slab is higher than that of the frame with the transfer slab by about 40-65%. This is attributed to the fact that the frame with normal slab is less stiff than the frame with transfer slab which results in larger periods and smaller frequencies.

Table 1 shows the modal parameters of the frame with normal slab. Although most of the effective translational mass factor is associated with the first five modes (98.01%), however, for the rotational mass almost all rotational mass is associated with the first mode (99.21%).



Table 1 Modal parameters of a frame with normal slab.

Mode Number

Period

(sec)

Percentage of effective

translational mass factors

(%)

Accumulative effective


(%)


rotational mass factors

(%)



(%) 1 25.597 77.19 77.19 99.21 99.21 2 7.422 11.24 88.43 0.024 99.23 3 3.647 4.97 93.40 0.370 99.60 4 2.137 2.83 96.23 0.014 99.61 5 1.400 1.78 98.01 0.037 99.65 6 1.001 1.13 99.15 0.006 99.66 7 0.781 0.64 99.79 0.007 99.66 8 0.670 0.21 100.00 0.001 99.66

Table 2 shows the modal parameters of the frame with transfer slab. Again, although most of the effective translational mass factor is associated with the first five modes (96.74%), however, for the rotational mass almost all rotational mass is associated with the first mode (93.20%).

It is also observed that contributions of higher modes (mode 6, 7 and 8), for both translational and rotational mass, is very small and approach zero for the frame with transfer slab as compared to that of the frame with normal slab. Also, contribution of the fundamental mode of vibration for the frame with normal slab, for both translational and rotational mass, is higher than the contribution of the fundamental mode of the frame with transfer slab.

Table 2 Modal parameters of a frame with transfer slab. Mode

Number Period

(sec)



(%)



(%)



(%)



(%) 1 14.633 58.33 58.33 93.20 93.20 2 3.282 16.54 74.87 3.65 96.85 3 1.326 9.02 83.89 1.34 98.19 4 0.742 7.58 91.47 0.53 98.72 5 0.500 5.28 96.74 0.15 98.87 6 0.486 0 96.74 0 98.87 7 0.407 0 96.74 0 98.87 8 0.404 0 96.74 0 98.87

4. RESULTS AND DICSUSSIONS

Figure 2 shows the lateral drift time history of the transfer slab and that of the normal slab at the first floor level. The peak lateral drift of the first floor of the frame with normal slab are several times higher than that of the transfer slab for both harmonic and earthquake loads. These peak lateral drift of the frame with normal slab reaches approximately 10 times that of the frame with transfer slab. The peak lateral drift of the frame with normal slab at the first floor level is comparable for the two loads while that for the frame with transfer slab is higher for El Centro earthquake load as compared to that of the harmonic load.

Figure 3 shows the lateral drift time history of the transfer slab and of the normal slab at the roof level. The peak lateral drift of the roof of the frame with normal slab is approximately two times that of the peak lateral drift of the roof of the frame with transfer slab for the harmonic load while the peak lateral drift of the roof of the frame with normal slab is approximately two third that of the roof of the frame with transfer slab for El Centro earthquake load. The peak lateral drift at the roof level of the frame with transfer slab is comparable for the two loads while that for the frame with normal slab is higher for the harmonic load as compared to that of El Centro earthquake load.



0 20 40 60-10

-5

0

5

10

Time (sec)

Ux

(mm

)

Normal Slab

0 20 40 60-1

-0.5

0

0.5

1

1.5

Time (sec)

Ux

(mm

)

Transfer Slab

0 20 40 60-10

-5

0

5

10

Time (sec)

Ux

(mm

)

Normal Slab

0 20 40 60-4

-2

0

2

4

Time (sec)

Ux

(mm

)

Transfer Slab

Sine Load Sine Load

ElCentro Load ElCentro Load

Figure 2 Response history of the transfer slab level.

0 20 40 60-150

-100

-50

0

50

100

Time (sec)

Ux

(mm

)

Normal Slab

0 20 40 60-100

-50

0

50

Time (sec)

Ux

(mm

)

Transfer Slab

0 20 40 60-60

-40

-20

0

20

40

Time (sec)

Ux

(mm

)

Normal Slab

0 20 40 60-100

-50

0

50

100

Time (sec)

Ux

(mm

)

Transfer Slab

Sine Load Sine Load

ElCentro Load ElCentro Load

Figure 3 Response history of the roof slab level.




5. SUMMARY AND CONCLUSIONS

This is study is of limited scope and its conclusion is applicable to the case studies in this investigation. A wider study with more real life building examples that take into consideration the nonlinear static and dynamic responses of these buildings is currently underway to develop a more general conclusion with wider range of applicability. Nevertheless, it is concluded from this study that:

(a) The peak lateral drift of the roof of the frame with transfer slab is less sensitive to the type of the two dynamic loads (harmonic and El Centro earthquake) as compared to the peak lateral drift of the frame with normal slab.

(b) At the level of the transfer slab, the peak lateral drift of the frame with transfer slab is much lower than the corresponding peak lateral drift for the frame with normal slab for both the harmonic and El Centro earthquake loads.

(c) The frame with transfer slab dampened the lateral drifts for all floor levels and for both harmonic and El Centro earthquake loadings much faster than the frame with normal slab.

6. REFERENCES

Abdalla, J.A. (2006). “Effect of transfer slab on the dynamic response of high-rise reinforced concrete building”, Abstract ID 815, First European Conference on Earthquake Engineering and Seismology (a joint event of the 13th ECEE & 30th General Assembly of the ESC), Book of Abstracts, 3 - 8 September 2006, Geneva, Switzerland.

Kuang, J.S. and Zhang, Z. (2003). “Analysis and behaviour of transfer plate-shear wall systems in tall buildings”, The Structural Design of Tall and Special Buildings, Vol. 12, No. 5, 409 – 421.

Kuang, J.S. and Li, S. (2005). “Interaction-based design formulas for transfer beams: box foundation analogy”, Practice Periodical on Structural Design and Construction, Vol. 10, No.2, 127-132.

Li, C.S., Lam, S.S.E., Zhang, M.Z. and Wong, Y.L. (2006). “Shaking table test of a 1:20 scale high-rise building with a transfer plate system”, ASCE Journal of Structural Engineering, Vol. 132, No. 11, 1732–1744.

Li, J.H., Su, R.K.L. and Chandler, A.M. (2003). “Assessment of low-rise building with transfer beam under seismic Forces”, Engineering Structures , Vol. 25, 1537–1549.

Su, R.K.L., Chandler, A.M., Li, J.H. and Lam, N.T.K. (2002). “Seismic assessment of transfer plate high rise buildings”, Journal of Structural Engineering and Mechanics, Vol. 14, No. 3, 287–306.



EVALUATION OF COLUMS’ FLEXURAL STRENGTH OF SPECIAL MOMENT RESISTING FRAME IN ACCORDANCE TO THE

INDONESIAN CONCRETE AND EARTHQUAKE CODES

Pamuda Pudjisuryadi1 and Benjamin Lumantarna1

1 Department of Civil Engineering, PETRA Christian University, Surabaya, Indonesia Email: [email protected] and [email protected]

ABSTRACT: To ensure “strong column weak beam” in Special Moment Resisting Frame System (SMRFS), SNI 03-2847-2002 requires that the nominal flexural strength of columns shall be more than 6/5 times the nominal flexural strength of the beams. This over strength factor is much more less than what is stipulated in the previous code. In this study the performace of two building, six and ten stories with three bays located at zone 2 and 6 on Indonesian seismic map is evaluated. The avaluation is done using the nonlinear static pushover analysis and dynamic nonlinear time history. The load patern used in the nonlinear static pushover analysis is based on the first mode, while for the dynamic nonlinear time history a spectrum consistent artificial earthquakes with a 200, 500, and 1000 years return period are used. It is shown that for buildings in zone 6, plastic hinges in columns appear when the structures are subjected to 200 years return period earthquake, while in zone 2, they only appears due to 1000 years return period.

1. INTRODUCTION

In capacity design of a Special Moment Resisting Frame (SMRF), “strong column weak beam” concept is used to ensure no plastic hinges will be developed in columns during severe earthquake, resulting in a safe failure mechanism called the “beam side sway mechanism”. This concept is adopted in Indonesian new concrete code, the SNI 03-2847-2002 (Badan Standarisasi Nasional, 2002). A value of 6/5 is set as minimum requirement of the ratio of total flexural strengh of the columns with respect to that of the beams, as follows :

∑Mc ≥ (6/5)∑Mg (1)

∑Mc is the total nominal flexural strength of columns, while ∑Mg is the total nominal flexural strength of beams. In the previous code, the SNI 03-2847-1992 (Departemen Pekerjaan Umum, 1992), this ratio is derived from “strain hardening” of material and “dynamic magnification factor”, a factor represents the shifting of structure behavior when plastic hinges are developed. The ratio from SNI 03-2847-1992 is much larger than 6/5, this raises some concerns of the seismic performance of structures designed by using the SNI 03-2847-2002.

2. STRUCTURES CONSIDERED

In this study, two symmetrical six and ten story buildings are designed by SNI 03-2847-1992 and SNI 03-2847-2002. The structures are assumed to be built on soft ground in Zones 2 and 6 of Indonesian Seismic Map according to Indonesian Earthquake Code, SNI 03-1726-2002. For simplicity, each building will be given an identification label. Example of the typical identification label is SRPMK-6-2, where SRPMK stands for Special Moment Resisting Frame, the following numbers are meant for number of stories (6 stories), and Seismic Zone (Zone 2), respectively. The building plan, and building elevation view are shown in Figures 1 and 2 respectively. Other technical data such as properties of concrete and steel reinforcement used, element dimensions can be seen in Tables 1, 2, and 3.



Figure 1 Typical building plan for 6 and 10 story buildings.

(a) 6 story building (b) 10 story building

Figure 2 Typical elevation view of the buildings.

Table 1 General data used for design. Floor to floor height 3.5 m

Slab thickness 120 mm

Compressive strength of concrete 30 MPa

Yield strength of longitudinal bars 400 MPa

Yield strength of transversal bars 240 MPa



Table 2 Element dimensions (in mm) used in 6 story building. Seismic Zone Story Beams Interior

Columns Exterior Columns

Corner Columns

1 400 x 700 560 x 560 500 x 500 475 x 475

2 400 x 700 560 x 560 500 x 500 475 x 475

3 400 x 700 530 x 530 475 x 475 450 x 450

4 400 x 700 530 x 530 475 x 475 450 x 450

5 400 x 700 500 x 500 450 x 450 425 x 425

2

6 400 x 700 500 x 500 450 x450 425 x 425

1 400 x 650 600 x 600 600 x 600 600 x 600

2 400 x 650 600 x 600 600 x 600 600 x 600

3 400 x 650 550 x 550 550 x 550 550 x 550

4 400 x 650 550 x 550 550 x 550 550 x 550

5 400 x 650 500 x 500 500 x 500 500 x 500

6

6 300 x 550 520 x 420 520 x 420 520 x 420

Table 3 Element dimensions (in mm) used in 10 story building. Seismic Zone Story Beams Interior

Columns Exterior Columns

Corner Columns

1 400 x 700 650 x 650 550 x 550 525 x 525

2 400 x 700 650 x 650 550 x 550 525 x 525

3 400 x 700 625 x 625 525 x 525 500 x 500

4 400 x 700 625 x 625 525 x 525 500 x 500

5 400 x 700 575 x 575 500 x 500 475 x 475

6 400 x 700 575 x 575 500 x 500 475 x 475

7 400 x 700 525 x 525 475 x 475 450 x 450

8 400 x 700 525 x 525 475 x 475 450 x 450

9 400 x 700 500 x 500 450 x 450 425 x 425

2

10 400 x 700 500 x 500 450 x 450 425 x 425

1 400 x 700 700 x 700 700 x 700 700 x 700

2 400 x 700 700 x 700 700 x 700 700 x 700

3 400 x 700 650 x 650 650 x 650 650 x 650

4 400 x 700 600 x 600 600 x 600 600 x 600

5 400 x 700 600 x 600 600 x 600 600 x 600

6 400 x 700 600 x 600 600 x 600 600 x 600

7 400 x 700 550 x 550 550 x 550 550 x 550

8 350 x 600 550 x 550 550 x 550 550 x 550

9 350 x 600 550 x 550 550 x 550 550 x 550

6

10 300 x 550 360 x 360 360 x 360 360 x 360



3. RESULTS AND DISCUSSION

Nonlinear static pushover analysis using ETABS software v.9.07 (Computer and Structures Inc., 2003) and dynamic nonlinear time history analysis using RUAUMOKO 3D software (Carr, 2001) are performed to evaluate the seismic performances of the buildings. The Moment-Curvature of beam and column sections are analyzed by ESDAP (Pono et al., 2003), a software developed by Petra Christian University. The first mode is used for load pattern of nonlinear static pushover analysis, while modified spectrum consistent ground motion is used for nonlinear dynamic time history analysis. Modification of El Centro 15th May 1940, N-S component ground motion is done by RESMAT (Lumantarna et al., 1997), to develop an artificial ground motion that produce response spectrum in accordance to SNI 03-1726-2002. Three levels of earthquake are used for evaluation the seismic performances, which are earthquakes with 200, 500, and 1000 years return period.

Figures 3 to 8 show the seismic performance of the buildings from nonlinear static pushover analysis as well as nonlinear dynamic time history analysis in terms of plastic hinges’ location, displacement, and inter-story drift ratio.

Return Period Exterior Frame Interior Frame Exterior Frame Interior Frame

200

years

500

years.

1000

years

Nonlinear Time History Nonlinear Static Pushover

Figure 3 Plastic hinges’ location in SMRF-6-2.



Return Period

200

years

500

years

1000

years

Exterior Frame Interior Frame Exterior Frame Interior Frame

Nonlinear Time History Nonlinear Static Pushover


ReturnPeriod

Nonlinear Static Pushover Nonlinear Time History


200

years

500

years

1000

years





Return Period

500

years

1000

years

200

years


Nonlinear Static Pushover Nonlinear Time History

The symbols show failure of plastic hinges (damage index > 1.0).

In Figures 3 to 6, it can be seen that location of plastic hinges obtained from nonlinear static pushover and nonlinear time history are quite similar. But in some location, results from nonlinear time history analysis show more severe damages, especially in exterior frames. An unstabe “soft story mechanism” is observed in exterior frame SMRF-10-6, when the building is subjected to earthquakes with 500 and 1000 years return period (see Figure 6).



SMRF-6-2

0

1

2

3

4

5

6

0 0.1 0.2 0.3

Displacement (m)

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000

SMRF-6-2

0

1

2

3

4

5

6

0 0.005 0.01 0.015 0.02

Drift RatioSt

ory

PO200 TH200 PO500

TH500 PO1000 TH1000

Figure 7 Displacements and inter-story drift ratios in SMRF 6-2.

SMRF-6-6

0

1

2

3

4

5

6

0 0.2 0.4 0.6

Displacement (m)

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000

SMRF-6-6

0

1

2

3

4

5

6

0 0.01 0.02 0.03 0.04

Drift Ratio

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000




SMRF-10-2

0

1

2

3

4

5

6

7

8

9

10

0 0.1 0.2 0.3 0.4

Displacement (m)

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000

SMRF-10-2

0

1

2

3

4

5

6

7

8

9

10

0 0.005 0.01 0.015 0.02

Drift RatioSt

ory

PO200 TH200 PO500

TH500 PO1000 TH1000


SMRF-10-6

0

1

2

3

4

5

6

7

8

9

10

0 0.2 0.4 0.6 0.8

Displacement (m)

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000

SMRF-10-6

0

1

2

3

4

5

6

7

8

9

10

0 0.01 0.02 0.03 0.04

Drift Ratio

Stor

y

PO200 TH200 PO500

TH500 PO1000 TH1000


According to SNI 03-1726-2002, the maximum inter-story drift in ultimate limit state should not exceed 2 %. From Figures 7 and 10, it can be seen that for both buildings (6 and 10 story buildings), the requirement is met for Zone 2, but not for Zone 6.




4. CONCLUSION

From the results, it can be concluded that the ratio of total flexural strength of the columns with respect to that of the beams (=6/5) according to SNI 03-2847-2002 does not ensure “strong column weak beam” criteria, especially in high risk seismic Zone (Zone 6 in this study). Unstable “soft story mechanism” begin to show in building subjected to Earthquake with 500 years return period, which is not acceptable. According to maximum inter-story drift criteria set by the code, the same conclusion can be made.

5. REFERENCES

Badan Standarisasi Nasional (2002). “Standar Tata Cara Perhitungan Struktur Beton Untuk Bangunan Gedung”, SNI 03-2847-2002. Jakarta, Indonesia

Badan Standarisasi Nasional (2002). “Tata Cara Perencanaan Ketahanan Gempa Untuk Bangunan Gedung”, SNI 03-1726-2002, Jakarta, Indonesia.

Computer and Structures Inc. (2003). ETABS Non Linear v. 9.07. “Extended Three Dimensional Analysis Of Building System”, Berkeley, California, USA, 2003.

Carr, A.J. (2001). “RUAUMOKO, Inelastic Dynamic Analysis, 3-Dimensional Version”, University of Canterbury, New Zealand.

Departemen Pekerjaan Umum (1992). “Standar Tata Cara Perhitungan Struktur Beton Untuk Bangunan Gedung”, SNI 03-2847-2002, Jakarta, Indonesia.

Lumantarna, B. and Lukito, M. (1997). “RESMAT, sebuah program interaktif untuk menghasilkan riwayat waktu gempa dengan spektrum tertentu [C]”, Proc. HAKI Conference 1997, Jakarta, Indonesia, 13-14 August 1997: 128-135.

Pono, B.R. and Lidyawati (2003). “ESDAP, Educational Section Design and Analysis Program”, PETRA Christian University, Surabaya, Indonesia.



COMPOSITE STEEL-CONCRETE AND RC COREWALL FOR THE 33-STORY CITY TOWER BUILDING

Davy Sukamta1,2

1Principal, Davy Sukamta & Partners, Jakarta, Indonesia 2President, Indonesia Society of Civil & Structural Engineers, HAKI

ABSTRACT: The City Tower is a 33-story modern office building with a total area of 72,000 sq.m, It is located in Jalan Thamrin, Jakarta. Built in a triangular 5,964 sq.m lot, the tower will provide 44,700 sq.m premium office space. The City Tower is far from the conventional office building. Its shape speaks of asymmetry, dynamic composition and sleek shapes. This building uses composite steel-structure and RC corewall as its structural system.

The gravity system is a composite concrete-metal deck floor with partial composite steel truss, spanning up to 16 m, supported by “super columns”, which are basically infill steel tubes with 50 MPa concrete. The steel tube gives very high confinement effect and acts compositely with the concrete to take the load of the building. The lateral system is RC corewall, with perimeters composite steel-concrete frame. Special details for the beam column connections are developed, to enable the transfer of force from the steel beam into the composite infill tube columns.

The building is designed to take 500-year earthquake with 10% probability of exceedance. A special study for Site Specific Response Spectra was conducted for this site. This paper described the design and construction of this project.

1. PROJECT DESCRIPTION

Located on a triangular site, the 33-story City Tower is far from the conventional and traditional building. Its shape speaks of asymmetry, dynamic composition and sleek shapes, by adopting the shape of a nautilus – a shape that suggests asymmetry yet centrality, order yet fluidity. The “nautilus” plan can be seen in sweeping lines and layered surfaces of the façade. With varying degrees of transparency and differing textures on the façade, City Tower would shine and glow, like a crystal nautilus. Figure 1 and Figure 2 show the artist impression on the project.

This building has 5-level basement, used for parking, and 33 stories of premium office space and sky-lounge on the top floor. On typical floors, 16 m free column office space, spanning from the core to the perimeter columns, will provide great flexibility for working space. On the other hand, having a big span is a challenge for the structural engineer and the services engineer (ME/P). With a pre-determined floor to floor height of 3.900 m and required net ceiling height of 2.800 m, the space for the structure and services is a mere 1000 mm.

The construction of the project started in March 2007, and the topping-off was conducted in December 2007. The project is slated for completion in August 2008. International Conference on Earthquake Engineering and Disaster Mitigation 2008.



Figure 1 The artist impression of Jakarta City Tower.

Figure 2 The artist impression of Jakarta City Tower.

2. STRUCTURAL SYSTEM

Given the limited space for the structural elements and services in the typical floor two schemes were studied at the beginning of the project. The first scheme was a prestressed concrete beam with conventional RC slab as gravity system, and RC corewall with RC perimeter open frame as lateral system. Since the foundation has been completed, this scheme allowed the developer to build as high as 25 stories.



The second scheme was using composite steel structure and RC corewall. The gravity system is composite concrete-metal deck floor with composite steel truss, supported by the RC core and “super-columns”. Given the span of the building, some trusses span up to 16 m. They are fabricated from high-strength ASTM grade 50 steel beam, cut into two, and laced with steel pipes as diagonal chords. This type of truss does not need gusset plates, simple to fabricate, and provides space for the mechanical/electrical services. Shear studs give the partial composite action. See Figure 3. HVAC ducts, cables, pipes, all pass through the truss. The ceiling can be attached directly to the bottom chord. Since the deepest steel member is the 900 mm truss, the net ceiling height of 2.800 m can be achieved with 3.900 m FFH.

Figure 3 Composite steel truss.

The super-columns are basically composite infill steel tubes. They use high strength concrete, grade 50 MPa. The steel tube gives very high confinement effect, and acts together with the concrete to take the load of the building. To transfer the forces from the steel-beam to the composite column, a special connection detail has been developed. The steel beam flange is inserted into the steel pipe, and tension compression force on the flange is transferred by bearing action to the infill concrete. This type of construction has never been used before in Indonesia, but has gained wide acceptance in USA. Figure 4 shows the steel pipe standing alongside the RC corewall. All steel pipe, beams and trusses are fire-proofed by vermicullite spray system to achieve two-hour fire rating.

Figure 4 The steel pipe standing alongside the RC corewall.

3. SEISMIC DESIGN

Indonesian Seismic Code requires a building to be designed for 500-year earthquake. Jakarta is located in seismic zone 3, and the peak ground acceleration will depend on the classification of the site soil. For this project, a special study was conducted to establish the Site Specific Response Spectra to be used in the seismic design. Downhole seismic test has been conducted to obtain the shear wave velocity for the project site. The result of the test was then used to conduct the site response analysis. The study was conducted by Masyhur Irsyam et al. and the process will not be



discussed in detail in this paper. Based on this study, it is found that the peak ground acceleration for the specific site is 0.215 g, while for T = 1.0 sec, the PGA is 0.352 g, for T = 0.2 sec, the PGA is 0.540 g. Figure 5 shows the design response spectra used in the project.

Figure 5 Response spectra.

The lateral system for this building is basically the cantilever RC corewall, and a seismic response modification factor value R is taken as 6.0 as per Indonesian Seismic Code. Analysis showed that 85% of the design beam shear will be resisted by the RC corewall, while the frame takes the balance of it. After some consideration, we decided to design the RC corewall as the main lateral system in this project. Total design value of seismic base shear is resisted by this RC core. Since the perimeter composite steel-concrete structure is designed as open frame with moment connection, then the perimeter frame is designed to take the seismic load according to its stiffness, although the RC core has been designed to take the full seismic load without the support of the perimeter frame.

4. CONNECTION SYSTEM

The main role of connections is to transfer forces between members. In this project, the most important connection is between the steel beam/steel truss and the composite infill steel pipes. In this case, the connection does not rely on welding the steel beam to the pipe alone, but by transferring the forces to the infill concrete of grade 50 MPa. A slot in the form of an I shape was made in the pipe walls. After passing the steel plates through the pipe, the slot was welded to the beam to achieve confinement of the concrete. The forces on the steel plates will be transferred to concrete by means of bearing stress, and the welds act as second layer of defense. See Figure 6.

Figure 6 Connection between steel beams and composite columns.



5. CONSTRUCTION ASPECTS

The work on the field started in March 2007. By that time, all the diaphragm wall, foundation and basement structure up to level B-2 has been constructed. Basically, construction work started on Level B-1 and above.

Steel structure fabrication started in 28 June 2007, and the first erection was performed in 5 July 2007.

The RC corewall is designed to proceed as a free standing element for its full height. The bare steel structure is allowed to go up to three stories. Than the steel pipe must be filled with concrete, followed by the casting of the metal deck.

The speed of steel erection was 130 days for 33 floor. Topping-off was conducted in December 2007. The steel fabricator for this project is PT. Jagat Baja Prima. Figure 7 shows the detail of steel truss, where no gusset plate is used to connect the diagonal element to the top and bottom chord.

Figure 7 The detail of steel truss.

Figure 8 The structure of the building near completion.




Figure 8 and Figure 9 show the structure of the building near completion.

The completion date for this building is in July 2008. When completed, it will stand proud along the main business road of Jalan Thamrin, Jakarta, joining other buildings in that area.

Figure 9 The structure of the building near completion.

6. REFERENCES

D.S. & P. (2006). “Structural Design Report of the City Tower”, Vol. 1, No. LPS-SA/JCT-R0/X-2006, Jakarta, Indonesia.



SEISMIC RESISTANCE OF LOW RISE BUILDING

Abdul Haris1 and Amrinsyah Nasution2

1Doctor Candidate at Civil Engineering Study Program ITB

2Professor in Structural and Construction Engineering, Faculty of Civil Engineering and Environmental – ITB

email: [email protected]

ABSTRACT: There are hundred-thousands low rise building which are built every years through out Indonesia. It might be designed by unfamiliar-with-seismic-design engineers, even by non engineers. Therefore, simplified seismic resistant design manual is absolutely a need. In this paper, requirements for seismic resistance in Indonesian Concrete Code (SNI) was extracted for more simple guidance manual.

Study of seismic resistance design is carried out by taking the advantages of limited choices of typically low rise building parameters in Indonesia such as maximum column-to-column span, build ability of structural members dimension, reinforcement bar diameter, economical constraint.

More effective way to understand how one can design a seismic resistance low rise building is by simple approach through illustration drawing which is completed by charts, tables and simple explanation.

1. INTRODUCTION

Earthquake is nature phenomena that is part of the nature itself. Delicate earthquake in general is not sensed by human although its frequencies reach thousand times in a year around the world. Moderate earthquakes that occur hundreds times in a year is less felt by human. Strong earthquake in part of populated area that instigates severe damages, happens only in numbers a year.

Do we need to design buildings that can resist a 200 or 500 year strong earthquake, while services of the building are only about 30 to 50 years ? Seismic resistance building design that will not be damaged even for rarely strong earth-quake produces a very large structure elements. It is not viable as well as uneconomical. Term “ seismic resistance” refers to design criteria of building that can be damaged during earthquake, but it will not be total collapse as for safety of life.

Seismic resistance is translated as tahan gempa in Indonesian language which has a certain connotation. Therefore, public and even engineers often misunderstand that seismic resistant building is about building that never damage, even under strong earthquake. Intensive socialization of what is seismic resistance meant is required by explaining proper information which is more easily understood and adopted.

One should realize that in developing country such as Indonesia, there are so many low rise buildings built in earthquake prone areas without any structural seismic resistance in order to stand strong earthquake. A number of strong quakes for the last ten years has triggered people to start to realize importance of preparing their properties for seismic resistance design. It is difficult to expect that engineers fully understand about seismic resistant design concept especially in conjunction with ductility concept. It is more realistic to give simply information which is more easily understood and adopted.

This paper is intended to explore a number of geometric alternatives that becomes simple practice guidance using charts and tables. Scope of work is limited for low rise building supported by open frame structure. There are hundred-thousands this type of building scattered throughout Indonesia which mostly are not designed by seismic resistant building approach.



a. Damage Structure b. Total damage structure

Figure 1 Damages of low rise building (source: Sigit, Wayan and Gunawan).

2. OVERVIEW

An overview of seismic resistant building design should be started from design philosophy. It involves uncertainty oft earthquake risk, economical constraint, architectural consideration and the most important is human life. So far, modern technology has not been exactly determined when and where earthquakes will occurs. What technology can do is to predict the risk level for some areas based on the earthquakes history record. Based on this record, codes issue seismic risk zone. Nevertheless, from seismic resistant design perceptive, it is more useful to have seismic risk zone than to know the D-day.

2.1 Design Philosophy

Design philosophy for seismic resistant building is developed from earthquake risk during its service life. At a delicate/weak earthquake, all structure’s elements in building will not be damaged. At a moderate earthquake, some structure’s elements in building may have minor damaged that need only renovation or retrofitting.

At strong earthquake, structure of building will not be totally collapse, although the elements are damaged. This condition is possible if the collapse is designed for ductile collapses mechanism, of which hinges joints at element’s joints are created at a certain hierarchy.

2.2 Frame Structure

In general, low rise building supported by open frame structure. We rarely find it is equipped by shear wall, especially in Indonesia. It is hiperstatic structure which has redundant force in its statically equilibrium equation. The advantage of hiperstatic structure is having many failure mechanism alternatives, so one can choose one alternative which will give maximum energy absorption. Frames with high redundancy can be modelled to develop plastic hinges near joints in a certain hierarchy to be a ductile structure. Exception for single element structure such as pole, chimney etc., the ductility concept is not prevail. In this type of structure, plastic hinges precisely are not allowed to be developed.

2.3 Ductility

Earthquake induce energy into flexible structure. Generally, induced energy of small earthquake is absorbed by developing elastic stress in structure material. In the other hand, moderate-strong earthquake induce significant energy into the structure. We have known that mostly, structure is designed remain elastic under gravity load. As a consequence, its dimension is very large under strong-earthquake loads to assure its material remain elastic. However, it is unrealistic to have elastic structure in all possibilities in conjunction with earthquake risk during its service life. Engineers must make a decision what they do in design in order to prepare to those possibilities. Since human life is the main concern, structural ductility become a main issue in seismic resistant building design. It means that building will damage when strong quake occurs but remain in tack/stand. Therefore, building ductility represent the capacity of the building structure to absorbs



induced energy by developing inelastic deformation. Many definitions refer to ductility term such as material ductility, rotational ductility, structural ductility, etc.

Material ductility should be understood before one can fully understand term of structural ductility. Without ductile material we never able to have ductile structure. Carbon steel is the most ductile of steel materials, even though it is less-strength compared with other kind of steel. Higher strength steel such as high-strength low-alloy, heat treated carbon steel and heat treated alloys steel, have lower ductility than carbon steel respectively. Ductility of homogen material is measurement by single action test. For instance, carbon steel ductility is determinated by comparing collapse and yield deformation in a tension test. In resisting moment, longitudinal reinforcement bar is dominant in tension or compression. In resisting moment, overreinforced concrete member is brittle material because its failure is initially by concrete failure. In other hand, underreinforced concrete is more ductile because its failure initially by yielding of reinforcement bar.

Ductility

Figure 2 Typical energy dissipation in structures.

Structural Ductility is not always automatically achieved by setting all member using ductile material. Structural ductile also depends on failure mechanism. No need to equipped all section member to become ductile. The most critical stresses of beams under earthquake loading will generally be at and near intersection with supporting column. This region required special attention to maintenance ductile against inelastic deformation. It is termed as plastic hinges. Large portion of induced energy is absorbed by plastic hinges and other small portion by elastic deformation. Therefore, in open frame structure critical region is modelled as plastic hinges. Special effort is intended to maintenance its strength or stiffness and ductility during earthquake. Plastic hinges will be developed during earthquake, one after onether depending on configuration of section capacity of structural elements beam and column.

Ductility measurement: There are many kind of methods to measure material, element or structure ductility. Steel ductility is measured by tension test, comparing longitudinal deformation at collapse and yielding. Ductility of structural member such as beam is measured based on its moment vs curvature curve. Structure ductility of low rise building is measured by comparing roof lateral deformation when it collapses and its first hinges developed.

Fmax

Δ

Energy dissipation of elastic action Energy dissipation of

Fmax

Δ

idealized elastic-plastic action



Figure 3 Failure mechanism in structures.

Structural Ductility

2.4 Capacity Design

Capacity design is intended to have ductile structure by setting section capacity of structural members in order to absorb induced energy as much as possible. Plastic hinge is initially expected at one of critical regions at end beams. We avoid formation of plastic hinges at columns end before as many as possible hinges plastic at end beam have been developed. Therefore, there are overstrength factor for flexural strength relative to that of beams meeting join. Overstrength factor is taken from 1.2 until 1.5 depend on which codes one adopts.

In capacity design, one does not only detail member section proportional to load level but also setting overstrength factor which is called as strong column-weak beam concept. This concept is absolutely effective, especially for low rise building with relative longer span such as school, auditorium, show room and multi purpose building. The next step is to equipped critically regions or plastic hinges by special detailing.

2.5 Beam

Earthquake design requires a specific function of beams, especially at beams end which are as well as plastic hinge regions. It is required more stringent lateral reinforcement than those for member designed for gravity loads or the less critically stresses part of members in earthquake-resistant structures. A closely spaced transversal reinforcement also support longitudinal compressive reinforcement against inelastic buckling and confine concrete core. A number of test has shown that confinement will increases concrete strength and ductility.

1 2

3 4

Ductile Failure Mechanism

Soft storey Mechanisme

(Brittle)

Ductile structure Strong column-Weak Beam

Brittle Structure (Soft Storey)

1 1

2 2



Capacity Design

1 2

34

Structural members is de-signed proportionally under grafity and wind loads, including reduced moderate -strong earthquake load. Columns has overstrength factor to beams, in order to assure hierarchy of plastic hinges deformation yields ductile failure mechanism Strong earthquake is ab-sorbed by developing plastic hinges as many as possible at beam ends. The next plastic hinges is designed at column ends near fixed sup-port.

Structural members is de-signed proportionally under gravity and wind loads, in-cluding moderate-strong e-arthquake load Strong earthquake is ab-sorbed by developing very large elastic deformation of material. There are no plastic hinges deformation.

Structural members is designed proportionally under grafity and wind loads except modera-te-strong earthquake load. Strong earthquake is absorbed only through li-mited plastic hinges at column ends as column im-mediately become un-stable before next plastic hinges developed.

Elastic or Proportional Design

Capacity Design

1 1

2 2

Figure 4 Elastic and capacity design of structures.

However, performance plastic hinge during earthquake loading is not only influenced by variables mentioned earlier but also level of nominal shear stress. Bertero indicates that when the nominal stress exceeds about '3 fc some reduction in ductility is initiated as well as stiffness when subjected to loading associated with strong earthquake response.

As consequence of applying capacity design concept, shear force is calculated based on the maximum probable flexural strength the beams ends.

2.6 Column

Capacity of column represented by P-M diagram interaction. For resisting moment frame, the flexure capacity of column is reduced by the present of axial load. Critically stressed regions caused by lateral load from earthquake loading is assumed will be developed at column ends near beam-column joints. It is not totally true by considering inelastic action as well as dynamic response where higher modes present, especially for tall frame building.

As mentioned earlier, column overstrength is required in order to avoid brittle failure mechanism such as soft storey mechanism. It is worth noting that in three dimensional we need to consider biaxial moment capacity. In monolithic reinforced concrete, flexural capacity of beam increases caused by participation of slab reinforcement over influenced width. Therefore, one has to calculate overstrength factor based on the flexural strength capacity of T or L beam.



Column, Beams and Joints

+

Figure 5 Pattern of internal forces elements.

Special transverse reinforcement is also required at critical stress parts as a consequence strong column-weak beam approach. As a result, confinement effect from transverse reinforcement increases concrete core.

After as many as plastic hinges at beam ends are developed, one can expect the next plastic hinges at column will be developed near fixed support in order to continue ductile inelastic deformation.

2.7 Beam-Column Joint

Joint is important member in conjunction with ductile behaviour. Its failure mechanism is dominated by shear. Under earthquake loading, joints disruption must be avoided by special de-tailing in order to maintenance join integrity and its stiffness caused by join crack and losing of bonding between concrete and longitudinal anchorage.

In the last decades, shear action is explained by struts and tie approach which is developed from what we call truss approach. Concrete cracks under stress tension form struts following such direction perpendicular to principal direction. Adequate reinforcement act as ties in struts and ties mechanism. Since column and beam longitudinal reinforcement bars are placed through the joints, detailing is carried out by inserting reinforcement bars which will completely forming struts and tie mechanism as well as to confine joint core concrete.

A number of test and experiments show that additional horizontal bar at mid joint increases joint shear capacity by developing truss mechanism. Slip between bar and concrete at joint can be reduced by using bar with small diameter. Stub is often used when anchorage regions is not wide enough.

3. SIMPLIFIED SEISMIC REQUIREMENT

Seismic resistant building technology has not completely socialized, especially in developing countries such Indonesia. Few engineers are familiar with earthquake resistant design. Concept of seismic resistant building design is totally different from ordinary low rise building design which dominated by gravity load. It has not easy to fully understood by practitioners. Therefore, one needs to simplify the requirement for low rise building by considering:



- Dimension of column or beam section is relative small and there are very limited bar configuration alternative which we can choose as well as no many choices in selecting bar size.

- As long as there is no extraordinary architecture expression, there are limited structure layout alternatives.

- Code of Indonesian Concrete 2002 likely is intended as a standard for high-rise building purposes. Therefore, there are too many check-points must be carried out even for simple low rise building design.

- In ATC 40 reports, in general, building with fundamental period less than 1.0 second such as lower rise building has its fundamental modal mass participation factor is dominant. As a result, we can apply simple lateral distribution load in static equivalent approach.

- In general, risk increasing as a consequence of unsymmetrical plan is more easily under-stood

The intensity of strong earthquake in Indonesia increases for last ten years. Low rise building collapses are reported from areas throughout Indonesia. School buildings, show room and other low rise buildings with relatively long span will have soft storey mechanism if it is designed according to elastic concept and assumption about building ductility factor is never reached.

4. EXAMPLE OF SIMPLE CHARTS, TABLES

In capacity design approach, beam’s longitudinal reinforcement and dimension are selected based on the moment output from structure analysis using trial beam section as proportional design does, except that it is included earthquake forces and assume ductility level. Ideally, earthquake forces is calculated based on time history record. Nevertheless, the computation is required special effort and consuming computational time. Instead, static equivalent and response spectrum methods are more often used. Transverse reinforcement is detailed referred to selected beam section. By using the table, one just selects appropriate section from the table. Additional reinforcement at the joint follows hoops patern of the column ends. Crossties is avoided in order not having congested reinforcing bar placement. Tables and charts example are carried out by numerical methods using MATLAB version 7.1. Example of tables are shown bellow.

Requirements check points are simplified as shown in Table 1.

Table 1 Simplified requirement.

No Requirement Simplified Requirement Note

I. General I.1 Strength Reduction Factor Table Included in charts & tables I.2 Dynamic Analysis No need 1.3 Push over analysis No need

Replace by static equivalent approach

1.4 Overstrength factor 1.5 Minimum compressive

strength Table and Chart

1.6 Minimum yield strength Table and Chart

It is limited by tables and charts, explicitely

II. Beam I.1 Transverse reinforcement Table Included in charts & tables I.2 Longitudinal reinforcement Table Included in charts & tables 1.3 Axial load Level No need Included in charts & tables I.4. Minimum or maximum bar

diameter No need Included in charts & tables

1.5 Axial Load Level No need III Column



III.1 Transverse reinforcement Table and Chart Included in charts & tables III.2 Longitudinal reinforcement Table and Chart Included in charts & tables III.3 Axial load Level No need Included in charts & tables III.4. Minimum or maximum bar


III.5 Minimum column size No need Included in charts & tables IV. Beam-Column Joint IV.1 Additional reinforcement Table and Chart Included in charts & tables IV.2 Axial load Level No need Included in charts & tables IV.3 Stress Level No need Included in charts & tables IV.4. Minimum or maximum bar


Table 2 Bare beam sections. Rectangular Bare Beam Sections, fy= 40 MPa and fc’= 20 MPa DETAIL

b (mm) 170 170 200 200 200 h (mm) 200 250 250 300 300 Top rfc 3D13 5D13 6D13 8D13 5D16

Bottom rfc 2D13 3D13 4D13 4D13 3D16 Trvs rfc φ8-100 φ8-100 φ8-100 φ8-100 φ8-100 d (mm) 20 20 20 20 20 ρ (%) 1.4667 2.0020 2.3062 2.4299 2.2745

M (kgf-m) 2.0631 4.5387 3.8940 7.0883 6.8296 DETAIL

b (mm) 200 250 250 250 h (mm) 350 250 250 300 Top rfc 6D16 4D16 5D16 8D16

Bottom rfc 3D16 2D16 3D16 4D16 Trvs rfc φ8-100 φ8-100 φ8-100 φ8-100 d (mm) 20 20 20 20 ρ (%) 2.4273 1.7773 2.3516 2.1671

M (kgf-m) 9.3981 4.1592 4.8951 8.2582

There are only bare beam sections in this paper. In the complete manual, there are more sections, including T and L sections. The use of T and L section are intended to accommodate monolith structure where well anchorage slab reinforcement will increases flexural capacity.



The next step in common design procedure is selecting column section using combination of axial force output and maximum flexural capacity of beam with overstrength factor and then checks and details beam-column joint. By using chart one just selects the appropriate column section. Example of charts is shown bellow.

Table 3 25x25 cm2 column sections. fy= 40 MPa and fc’= 20 MPa

DETAIL

b (mm) 250 250 250 h (mm) 250 250 250 Top rfc 3D13 2D16+4D13 6D16 Bottom rfc 3D13 2D16+4D13 6D13

Trvs rfc φ8-100 φ8-100 φ8-100

d (mm) 40 40 40

ρ (%)

0.9651 1.7013 1.9302

0 0.05 0.1 0.15 0.2 0.25 0.3 0.350

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

(Pu.

) / 0

.65.

A.'0

,85.

fc

(Pu.e/h) / 0.8.A.0,85.fc' Table 4 30x30 cm2 column sections.

fy= 40 MPa and fc’= 20 MPa DETAIL

b (mm) 300 300 300 h (mm) 300 300 300 Top rfc 2D19+4D16 6D19 2D22+4D19 Bottom rfc 2D19+4D16 6D19 2D22+4D19

Trvs rfc φ8-100 φ8-100 φ8-100

d (mm) 40 40 40

ρ (%) 1.5237 1.8902 2.1049 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

(Pu.

) / 0

.65.

A.0

,85.

fc'

(Pu.e/h) / 0.8.A.0,85.fc' 5. FUTURE WORK

Development an integrated seismic resistant design manual contains simple requirement, tables and charts especially for low rise building including steel and timber buildings. Ideally, it is supported by statistic and probabilistic analysis in conjunction with low rise building parameters trend in Indonesia. Pushover analysis is also required in order to have structural ductility of some simple typically low rise buildings.




6. CONCLUSIONS

1. There are hundred-thousands low rise building which are built every years through out Indonesia. It might be designed by unfamiliar-with-seismic-design engineers, even by non engineers. Therefore, simplified seismic resistant design manual is absolutely a need. In this paper, requirements for seismic resistance in Indonesian Concrete Code (SNI) was extracted for more simple guidance manual.

2. Study of seismic resistance design is carried out by taking the advantages of limited choices of typically low rise building parameters in Indonesia such as maximum column-to-column span, build ability of structural members dimension, reinforcement bar diameter, economical constraint.

3. More effective way to understand how one can design a seismic resistance low rise building is by simple approach through illustration drawing which is completed by charts, tables and simple explanation.

7. LITERATURE

ACI Commite 318 (2003). ”Building Code Requirements for Structural Concrete (318M-99) and Commentary (318RM-99)”, American Concrete Institute, Farmington Hills Michigan.

Applied Technology Council (1996). “ATC-40 seismic evaluation and retrofit of concrete buildings, Volume 1”, California Seismic Safety Commission, Report No. SSc 9601, November 1996.

Budiono, B. (1999). “Lecture Note of Earthquake Resistant Design”, Institut Teknologi Bandung. Derecho, A.T. “Seismic design of reinforced concret structures“, Chapter 9, The Seismic Design

Hadbook, 2nd Edition, Farzad Naeim (Editor). Imran, I. (1999). “Lecture Note of Advanced Reinforced Concrete Structure”, Institut Teknologi

Bandung. Nasution, A. (2000). “Lecture Note of Reinforced Concrete Structure“, Institut Teknologi

Bandung. Nasution, A. and Hasballah (2000). “Numeric Methods”, Penerbit ITB, Bandung. SK SNI 03-1726-2002 (2002). ”Tata Cara Perencanaan Ketahan Gempa untuk Bangunan

Gedung”, Badan Standardisasi Nasional.



THE DEVELOPMENT OF 490MPA CLASS HSLA STEEL FOR EARTHQUAKE RESISTANT STEEL STRUCTURES

Djoko Muljono1 and Basso D. Makahanap2

1General Manager of Research & Technology, PT Krakatau Steel 2Manager of Research and Product Development, PT Krakatau Steel

ABSTRACT: Steel is commonly used as one of structural materials for high-rise building and tower structure application. In some regions with high frequently occurrences of earthquake, requirements of steel material for this application become more complex. The requirements are not only high strength, high toughness and good weld-ability but also higher plastic deformation capacity to ensure structural safety at a big earthquake. These requirements can be fulfilled by chemical composition adjustment and appropriate controlled rolling during production.

This work concerned to the development of nickel bearing medium carbon steel for structural square pipe application. The addition of maximum 0.3% nickel into 0.18%C, 1.5% Mn steel combined with re-crystallization in a controlled rolling process resulted in an excellent impact properties (170 – 230 J at 0oC), high yield strength (350 – 420 MPa), high tensile strength (500 – 560 MPa) excellent weld-ability (0.44 Carbon Equivalent), fine ferrite grain size 7.26 – 10.96 µm (ASTM No. 10 – 11) and most importantly lower yield ratio properties (< 0.80).

1. INTRODUCTION

Durability is an important factor in material design for structural and construction application. Steel for this application – especially for structural in region with high frequency occurrence of earthquake – requires not only high strength, good toughness and good weld-ability but also excellent structural safety at the big earthquake by plastic deformation of the steel frame.

In some developed country such as in Japan, in order to anticipate earthquake risk, the earthquake resistant design code was executed in 1981. This requires better stiffness and stronger building structure, and could fulfilled follow characteristic:

− Building structure shall have an adequate elastic deformation capacity to anticipate the occurrence of middle earthquake, and structure shall be return into original condition.

− Building structure shall not be damaged even at a big earthquake and shall keep safety space for living persons.

To ensure structural safety even at occurrence of a big earthquake could be achieved by an adequate plastic deformation capacity of the frame. Therefore the steel products which were used for this frame shall have an adequate plastic deformation.

Since early 1990’s, Krakatau Steel has been producing steel for earthquake resistant structural application as specified by JIS G3136 SN400B standard in order to fulfill requirements of a Japanese pipe manufacturer. In 2007 Krakatau Steel has developed higher strength steel for this application as specified by JIS G3136 SN490B.

This paper briefly describes the experience of the company in developing a class 490MPa steel product for square pipe (Super Hot Column/SHC) for earthquake.



2. PRODUCT REQUIREMENTS AND DESIGN

The requirements which comprise chemical composition and mechanical properties are referred to JIS G3136 SN490B grade as summarised in Table 1 and Table 2.

Table 1.

C Mn Si P S CE Pcm

Min - - - - - - -

Max 0.18 1.60 0.55 0.030 0.015 0.44 0.29

Table 2.

YS (MPa)

TS (MPa)

El (%)

YS/TS (%)

Charpy at 0ºC (J)

Min 325 490 17 min (6≦t<16) 21 min (16≦t<50)

- 27

Max 445 610 - 80 -

The steel material considered both the uniformity of the targeted mechanical properties and the property of low temperature impact resistance. To achieve the specified strength, the chemical composition of the steel has been designed as shown in Table 3. A certain amount of nickel as solid solution strengthening were added to the steel containing 0.17% C and 1.4% Mn to achieve the tensile strength of 490 MPa. The target of mechanical properties of the designed product is shown in Table 4.

Table 3.

C Mn Si P S Al Ni

Max 0.17 1.4 0.3 0.015 0.01 0.05 0..3

Table 4.

YS (MPa)

TS (MPa)

El (%)

YS/TS (%)

Charpy at 0ºC (J)

Min 350 500 20 - 70

Max 440 600 - 80 -

3. PRODUCTION PROCESS

The production of the steel for JIS G3136 SN490B applications is carried out in the existing slab steel plant and hot strip mill with processing route as generally shown in Figure 1. In the Electric Arc Furnace (EAF), the raw material which consists of 20% scrap and 80% of sponge iron is melted. This liquid steel is then further processed in the Ladle Furnace to reduce the sulphur content of the steel and to adjust the chemical composition of the steel by addition of alloying elements. Subsequently, the steel is continuously cast in the continuous casting machine to produce steel slab.

The steel making and casting practices adopted for JIS G3136 SN490B grade can be summarized as follows: - Hot metal desulphurization - Calcium treatment by CaSi wire injection for inclusion shape control - Casting into tundish with argon gas purging and extended shroud to minimize re-oxidation

during set up - Controlling the cooling intensity, superheat and casting speed - Metal retention in ladle at ladle change



This practice is developed not only to achieve the required chemical composition but also to obtain steel with high cleanliness, free from internal and surface defect.

In the hot strip mill, the slabs which come from the slab steel plant are subsequently hot rolled to the specified thickness. Controlled Rolling Process is applied during rolling to get uniformity of microstructures of HRC and to achieve the required mechanical properties. Slab is reheated up to about 1200°C and soaked in reheating furnace then control-rolled with high reduction at a recrystallization temperature in a roughing mill in order to achieve a fine austenite grain size. This steel is then rolled into final thickness in the finishing mill. The steel strip is finally coiled at about 600°C.

In pipe manufacturer, the hot rolled coil will be cold formed into circular pipe first. The circular pipe is reheated up to 900ºC at reheating furnace and then will be hot formed into square pipe. This manufacturing process has a beneficial effect that the final square pipe characteristics will be the same as the hot rolled coil characteristics.

Steel Making Process

Hot Rolling Process

Figure 1 The production process line in steel making and hot rolling.



Figure 2 The production process line in pipe manufacturer.


4.1 Hot Rolled Coil Properties

4.1.1 Mechanical test results

The results of tensile tests which were taken from total production data of about 1,500 tonnes hot rolled coil in longitudinal and transversal direction are presented at Table 5. The yield strength of the steel with 12.0 mm thickness is from 350 up to 420 MPa, the tensile property is in the range of 500 to 560 Mpa and the elongation of the steel is 23 to 28 % in longitudinal direction. The value of yield strength and tensile strength in transversal direction are not significantly different with the value in longitudinal direction respectively. All these values for both 12.0 mm and 16.0 mm thickness are within the range of specified values for the JIS G3136 SN490 specification and achieve the targeted values.

Table 5 Tensile test result of hot rolled coil. Thickness

(mm) Yield

Strength (Mpa)

Tensile Strength

(Mpa)

Elongation (%)

Yield Ratio (%)

Testing Direction

12.0 350 – 420 500 – 560 23 – 28 70 – 76 Longtiudinal

360 – 420 510 – 560 21 – 26 71 – 77 Transversal

16.0 350 – 400 500 – 560 22 – 27 70 – 75 Longtiudinal

350 – 410 500 – 550 22 – 27 70 – 76 Transversal

Standard 325 – 445 490 – 610 17 min (t<16) 21 min (16≦t)

80 max Longitudinal

Figure 3 shows the Charpy Impact Test results at -40, -20, -10, 0 and room temperature. As it is shown in this figure, the value of charpy test results are 126 up to 170 Joule at room temperature. These charpy values slightly decrease to 132 up to 170 Joule and 112 up to 170 Joule at temperature of 0 and -10 °C, respectively. The charpy values decrease signicantly to the range 78 up to 150 Joule at temperature -40ºC. These charpy values indicate that the material has a very good toughness even at very low temperature.



0

50

100

150

200

-60 -40 -20 0 20 40

Temperature (ºC)

Cha

rpy

Abs

orbe

d E

nerg

y (J

)

Figure 3 Charpy test result of hot rolled coil.

4.1.2 Metallographic Observation

The typical microstructure at the quarter thickness of the steel is shown in Figure 4. Fine ferrite with grain size of 7.26 – 10.90 µm and fine lamelar pearlite were obeserved by means of optical microscopy.

Figure 4 Microstructure at quarter thickness of the steel.

4.2 Square Pipe Properties

4.2.1 Mechanical test results

The results of tensile test which were taken from production data of square pipe are presented at Table 6. All these values are within the range of specified values for the JIS G3136 SN490B specification and achieve the targeted values. Table 7 shows the charpy impact test results at 0 ºC which were taken from production data of square pipe. The charpy values were much higher than the minimum requirement of 27 J at 0 ºC.

Table 6 Tensile test result of square pipe.

Thickness (mm)

Yield Strength

(MPa)

Tensile Strength

(MPa)

Elongation (%)

Yield Ratio (%)

Testing Direction

12.0 390 – 440 500 – 580 25 – 30 74 – 78 Longitudinal

16.0 380 – 430 500 – 570 27 – 32 75 – 79 Longitudinal

Standard 325 – 445 490 – 610 17 min (t<16) 21 min (16≦t)

80 max Longitudinal



Table 7 Charpy impact test result of square pipe.

Thickness (mm) Charpy Absorbed Energy at 0°C (J)

12.0 120 – 210

16.0 150 – 230

Standard 27 minimum

It has been widely known that the addition of nickel as solid solution strengthening increases mechanical properties, and also improves the toughness material. Addition of certain amount of nickel into 0.17% C – 1.4%Mn Steel combined with recrystalization rolling and hot process during pipe manufacturing, is able to produce an excellent steel material with high strength, high plastic deformation capacity and high toughness properties. These material properties could be fulfill the requirement of earthquake resistant material which were used for building structural application, as can be observed in Figure 5.

Figure 5 Application of SN490B square pipe as building structure.

5. SUMMARY

The product design using nickel as solid solution strengthening elements coupled with recrystalized rolling process in producing steel for JIS G3136 SN490B has ensured materials with the targeted mechanical properties, especially in higher strength, higher plastic deformation capacity and high toughness.




6. REFERENCES Haneke, M. et al. (1993). “Low Temperature Steels (Steels with Good Toughness at Low

Temperatures)”, Edts. Verein Deutscher Eisenhüttenleute, STEEL, A Handbook for Materials Research and Engineering Vol. 2, p.274 – 301.

Heisterkamp, F. et al. (1980). “Niobium as a toughening element in pipe steel: influence on weldment properties”, 2nd Conference on Pipe Welding, The Welding Institute, London, Vol. 1, p. 307.

http://nakajima-sp.com



INELASTIC BEHAVIOR OF DUCTILE BUCKLING-RESTRAINED BRACED TRUSS-GIRDERS MOMENT FRAMES

H. Sugihardjo1, W. Merati2, A. Surahman2 and M. Moestopo2

1Department of Civil Engineering, Sepuluh Nopember Institute of Technology (ITS), Surabaya, Indonesia

2Department of Civil Engineering, Bandung Institute of Technology (ITB), Bandung, Indonesia Email: [email protected]

ABSTRACT: This paper evaluates the inelastic behavior of three Buckling-Restrained Braces (BRB) and a ductile Buckling-Restrained Braced Truss-Girder Moment Frames (BRBTMF) by analytical and experimental means. The BRB section is a strip plate core A283 grade C steel which has been treated with soft annealing, cased with rectangular hollow section A50 steel as a buckling-restraint material. The BRBTMF proposed in this paper is the modification of The Special Truss Moment Frames (STMF) X-diagonal-type by replacing the X-diagonal with the BRB inverted-V-type to improve its inelastic behavior. From the Nonlinear History Analysis using earthquake records, it can be concluded that in general, the BRBTMF has better inelastic behavior than the other ductile frames. The cumulative ductility factor of the BRBTMF subassemblages has met the requirements of the hysteretic system structure. Based on these preliminary studies, the use of the BRBTMF may be implemented to practice.

1. INTRODUCTION

Indonesia is located in the junction of three tectonic plates, namely Indo-Australia, Pacific, and Eurasian plates. This situation has caused most of the areas in the country are categorized in the high-risk tectonic earthquake zones. It is, therefore, required to design an earthquake-resistant building in Indonesia. Strength, stiffness, ductility, and the ability to dissipate the earthquake energy should be possessed by each earthquake-resistant building. These are the main aspects that should be considered in the building structures. Besides, the structural system, material used in the building as well as the connection system of each structural element are also very important and require serious attention.

A truss-moment frame is one of Moment-Resisting Frame System. If compared with the solid beams, the truss-moment frames have three economical advantages. Goel and Itani (1994b) stated that “It requires simpler connections, it can be applied for longer structural spans and the space within the truss elements can be used as a place for the utility”. The additional advantage is that if the truss system is implemented in Indonesia, where the labor cost is relatively low, the truss beams that require more labor work than the solid beams, is still preferable as a more economical choice. The only disadvantage of this system is that the occurrence of inelastic behavior in the columns and that the beams remain elastic due to the high seismic loading. In the full-scale testing, this type of frames had been proven to perform a sharp degradation on its hysteretic loops (Goel and Itani, 1994a).

To improve the performance of the truss system as the moment-resisting frames under high seismic loading, some Vierendeel’s panels are added in the midspans as the ductile segments. These segments work as fuses (warning system), that has a function to absorb the earthquake energy by means of inelastic flexural deformation. During the earthquake, the induced lateral forces produce the shear forces in the ductile segments resisted by the cords alone until the plastic hinges occurred due to the flexure and thus, forming the yielding mechanism in the frames as shown in Figure 1 (Basha and Goel, 1996).



X-bracing system Vierendeel-system

ductile segment

plastic hinge plastic hinge

ductile segment

Figure 1 Yield mechanisms of ductile truss-moment frames.

Another system is by using the additional diagonal X-shaped bars (will be called as X-diagonal subsequently) located in the panels around the midspans of the truss-frames (Goel and Itani, 1994b), as shown in Figure 1. The numerical analyses were carried out on the 4-story 7-bay frames. The response indicated that an asymptotical lateral displacement at the top roof and smaller inter-story drifts than those in the regular truss system. The material used in this study can be saved up to 30 to 40 percent. The quasi-static test with full-scale specimens has shown that a consistent behavior in the hysteretic loops. This study is adopted as the Special Truss Moment Frames (STMF) (AISC, 1997) or the Special Steel Truss Moment Frames (SSTMF) (NEHRP, 2000).

In the last decade, a new steel material was developed. It is known as Low Yield Stress Steel. The steel types that can be classified into this group are LY100, LY160 and LY235. It is the form of panel or bracing. The bracing is the shape of flat bars which is stiffened by the box structural steel. It is called as Buckling-Restraint Braces (BRB). This material is mostly used as a hysteretic damper. It has a low yield strength ranges from 90 to 245 MPa (up to one-third of A36). However, the ultimate strain can reach up to 1.5 to 2.5 times the A36 steel’s ultimate strain (Inoue, 2004b; Kamura et al., 2000). The implementation of low yield stress steel as a replacement to the normal steel assures that the plasticity process occurred at the small deformation without significant changes in the bracing and structural stiffness (Kamura et al., 2000). The BRB sectional configuration is in the form of strip plates (or other sections) from low yield or mild steel as core bracing inserted into the box case made from normal steel with or without unbonded material as the lateral stiffening elements. From the BRB configuration, the resulted compressive yielding capacity is relatively similar to the tensile yielding capacity. The application of the ASTM-A36 mild steel as the core bracing also produces consistent hysteretic loops, even though the cumulative ductility factor is lower than that in the low yield steel (Chen and Lu, 1990; Watanabe, et al., 1989).

In the latest code (AISC, 2005), a new structural system called Buckling-Restraint Braces Frames (BRBF), is introduced. The system comprises the solid beams stiffened by the BRB. From the analytical study carried out by Shimokawa and Kamura (1999) on the 11-story 8-bay steel frame structures stiffened by the BRB, it showed that the reduction of lateral displacement could be as high as 50% and the base shear was also reduced although it was relatively small. The study conducted by Kasai et al. (1998) on 14-story building retrofitted with the BRB made from unbonded material (Buckling-Restraint Unbonded Braces), it showed that the reduction of lateral displacement ratio could be up to 60% and up to 30% for the base shear. The conclusion of the study conducted by Clark et al. (2000) on a structure designed based on the target of the lateral displacement ratio and the stiffness, the base shear could still be reduced up to 50% with additional BRB of unbonded material as stiffeners.

From the above overview, it can be seen that the studies conducted was to stiffen the solid beam frame structure with the BRB. The idea was arisen, when the X-diagonal, which has small



compressive capacity in the post-buckling region, was replaced by the BRB with lower yield stress steel and has the compressive yield stress similar to the tensile stress, it is believed that it could improve the seismic response of the frame system. Hence, it is hoped that plasticity could happen earlier in the frame system with bulkier hysteretic loop but stable. The proposed structural system is called the Buckling-Restrained Braced Truss-girder Moment Frames (BRBTMF).

2. ANALYTICAL STUDY

The analytical model in this study was based on the study conducted by Goel and Itani (1994b), and Sugihardjo, et al. (2004), i.e., 4-story 7-bay frame with layout and analytical model of exterior frame in longer direction axes A and D, as seen in Figure 2. This model was modified from the STMF X-diagonal type by replacing the X-diagonals with BRB inverted-V, and has the height of 3962 mm (156 in.) for each story level. The cross-section of the frame elements and yield strength of the steel in this analytical study can assumed equal to those in the previous study, that is it made from A36 steel with fy = 248.2 MPa, as listed in Table 1. The BRB made from LY235 steel with fy = 225 MPa and the cross-sectional area designed using Eq. (1), where Py = tensile yield strength of the X-diagonals; Pcr = buckling strength of the X-diagonals; φ = ratio post-buckling to initial buckling strength of the X-diagonals (assumed to be 0.3).

)()( )( cryBRBycry PPPPP +≤≤Φ+

(1)

The most left-hand side columns of Figure 2(b) represents the stiffness of interior columns. First, the structural elements were designed with the static equivalent method according to the UBC 1988 and the load factors were adopted from AISC-LRFD 1986. These old standards were adopted to trace and compare with the previous study such that the static earthquake loading was equivalent (Goel and Itani, 1994b). The base shear: V = ZICW/Rw with Z = 0.4; I = 1.0; C = 1.94 based on the natural period of the building, T = 0.68 sec and the soil type S3 (Sfactor = 1.2). The total weight of the building W was calculated using the uniformly-distributed load of 3.83 kPa for each story. The structure was designed as SMRF with Rw = 12 (or R = 8.5 (ICBO, 1997)) since it is expected the ductile and stable hysteretic behavior.

The analytical method used was the nonlinear time history analysis. To analyze the yielding energy dissipated by the BRB, an inelastic motion system equation was adopted. In the form of energy equilibrium, this equation is in terms of the kinetic, damping, static (the sum of strain and yielding energies) and input energy equilibrium (Akiyama, 1985; Chopra, 2001) as given in Eq. (2):

∫ ∫ ∫ ∫−=++t t t t

gs dtutumdtuuufdttucdtutum0 0 0 0

2 )(),()()( &&&&&&&&& (2)

Table 1 Comparison of frame systems.

Element STMF X-diagonal BRBTMF

Column (stories 1 and 2) W 14x120 W 14x120

Column (stories 3 and 4) W 14x99 W 14x99

Chords 2L 88.9x88.9x12.7 2L 88.9x88.9x12.7

Outside diagonal 2L 63.5x63.5x7.9 2L 63.5x63.5x7.9

Outside vertical 2L 50.8x50.8x6.4 2L 50.8x50.8x6.4

Inside vertical 2L 31.8x31.8x6.4 2L 31.8x31.8x6.4

X-diagonal L 25.4x25.4x6.4 L 25.4x25.4x6.4

Core of BRB (stripe plate) 60x8

Casing of BRB (rectangular) 2L 50x50x3



A classic way of Rayleigh was used to define the viscous damping of 5%. By using the step-by-step procedure of Newmark, the inelastic motion equation can be solved with the aid of DRAIN2-DX software (Prakash and Powel, 1992). The structure was loaded under several scaled-earthquake records with the following methods: First, using the 1978 NS Miyagi-ken-oki earthquake with PGA 0.4g. This earthquake record was adopted since it could give the maximum dynamic response as in the study on the STMF X-diagonal type. The result obtained will be compared with the result of the previous study by Goel and Itani (1994b). Secondly, for validation purpose, the proposed system will tested under the 1940 NS Elcentro earthquake record with PGA 0,69g, Northridge (Newhall 0 degree) with PGA 0.42g and NS Kobe with PGA 0.58g. The scale factor was chosen such a way that the earthquake response spectra intensity is equivalent to the UBC-S3 velocity response spectra, as shown in Figure 3.

B

7x8484 mm

DC

1 2 3 4 5 6

3X914

4 m

m

7 8

A

BRB (YIELD IN TENSION OR COMPRESSION)

PLASTIC HINGE

(a) (b)

Figure 2 Analytical models (a) Plan (b) Frame A, D and inelastic activities (Miyagi record).

0

0.5

1

1.5

2

2.5

0 0.5 1 1.5 2 2

spectra acceleration

periode(sec).5

UBC-S3MIYAGI-0,40G

NRIDGE-0,42GKOBE-0,58GELCENTRO-0,69G

Figure 3 Scaled spectra of records to match UBC soil S3 spectrum.

3. EXPERIMENTAL STUDY

In the experimental work, three BRB and one full-scale BRBTMF frame specimens were tested. The longitudinal layout and cross section of BRB are given in Figure 4. The stiffener plate with thickness of 12 mm at the end of BRB represents a connection plate when BRB has been implemented in the truss. The base plate of 25 mm thick is used as a support to tie with the loading frame. All BRB use the A283 Grade C steel material which has been soft-annealing treated with the yield strength of about 265 MPa and the stiffener was made from a material similar to A50 steel, such that the dimension was smaller than the analytical study. The design of BRB dimension according to the stable hysteretic requirement from Shimokawa and Kamura (1999) was in the safe region of Inoue’s interaction diagram (2004a). The complete result can be seen in Table 2, whereas BRBTMF using BRB 50 × 8 mm with the dimension of the elements equal to the A50 steel, as shown in Figure 4. The cyclic loading history is implemented by strain to the BRB and the drift ratio of BRBTMF from 0.5 to 3% as in Figure 5. The applied strain was up to 3% to know the



BRB’s capacity, although from the earlier analytical study the strain in the BRB was not greater than 2%.

weld 6mm

casing A50

1/2s = 1.2

1/2s = 1.2

bi = 49.0

ti = 7.9

tp = 3.

3

bp = 48

.5

core A28312.0

6.0

gusset A50

base plate 25mm

795

las 630 60 30

40 60 3040 40

A

A

L=709

gusset 12mm

baut 3/4"B

B

A-A B-B

25

60

30 60 3040

40

120

3040 40

30

60

30

(a) Longitudinal layout (b) Section A-A

Figure 4 Details of BRB specimen.

C

plate 12mm

loadcell

2L 50x50x5

762

bolt 4x1in

2L 5

0x 5

0x5

762

roller

A

100

bolt

plate

4x1 in

20 mm

280

250

35

3535 200

35

STRONG W

ALL

2L 90x90x12

2L 60x60x6

STRONG FLOOR

D 2L 90x90x12

762 1778

B

167 in (4242 mm)

brb

brb

2L 5

0x5

0x5

2L 60x

60x6

940

BOLT A325

plate 2x20mm

10x7/8

940a

200

t=8

a

200

180

1373

BOLT 4x1 in

1/2 WF 350x350x12x19

2L 60x

60x6

las 6

6

1778

200

WF 3

50x3

50x

12x1

9

330

t=8

816

ACTUATOR

t=12

1573

250

156 in(3962 mm)210

roller

-4

-3

-2

-1

0

1

2

3

4

0 2 4 6 8 10 12 14 16 18 20 22drift ratio (%)

number of cycle

Figure 5 Test setup for BRBTMF.

Table 2 Dimensions of buckling-restraint braces.

CASING CORE BRB

Width, bi Thickness, ti Width, bp Thickness, tp

LENGTH (mm)

(1)-50x8 49,0 7,9 48,5 3,3 709

(2)-40x8 37,7 7,9 38,8 2,7 707

(3)-50x8 47,8 7,9 46,5 3,3 709


From the analytical study, the largest inelastic activity in the structure was due to the Miyagi earthquake. This is because the natural period of the structure was 1.0041, which was located near predominant period of Miyagi earthquake of approximately one (see Figure 3). All the BRB elements experienced the plasticity and the plastic hinges occurred at the end of the ductile segments in the first to third stories, as shown in Figure 2(b). The inelastic response, in the form of



comparison ratio between the story drift due to the Miyagi earthquake and the roof drift due to other earthquake, as shown in Figure 6. In terms of the drift ratio, the BRBTMF could reach up to 67% smaller than that in the X-diagonal-type STMF. It can be seen in general the drift ratio of the proposed frame was smaller than those in the X-diagonal type frame system, Vierendeel, solid frame, and the conventional truss system. In terms of roof drift, the maximum value in the BRBTMF was 275 mm, less 21.6% from the X-diagonal-type STMF (see also Goel and Itani, 1994b). The roof drift history in the BRBTMF was also stable due to the all considered earthquakes with various frequencies.

1

2

3

4

0 1 2 3 4 5 6 7

STOREY

DRIFT RATIO (%)

BRBTMF STMF X-diagonal SOLID FRAMES

CONVENTIONAL VIERENDEEL

(a)drif ratio (b)roof displacement

Figure 6 Inelastic response of BRBTMF.

From the experimental study, the relationship between the load and the displacement in the BRB and BRBTMF was plotted in Figure 7. The magnitude of the cumulative ductility factor, η, of the BRB and BRBTMF specimens can be computed from the comparison between the total energy to the elastic energy and the result are listed in Table 3. It can be seen that the cumulative ductility factor of the BRBTMF satisfied the recommendation for hysteretic structural system, where the practical value η was taken to be greater than 20 (Akiyama, 1985). For BRB, if the cumulative dissipating energy was computed until the failure, all BRBs have the η value greater than 100. This value is sufficient for the BRB elements as elasto-plastic structural components (Shimokawa and Kamura, 1999).

Table 3 Cumulative ductility factor (η) of BRB and BRBTMF.

BRB

Hysteretic energy, (ΣWi) (106 N-mm)

First yield load,Py

(103 N) First yield displacement, dy

(mm) η

BRB-1 20,61 89,67 0,61 377

BRB-2 15,58 66,67 1,11 211

BRB-3 22,83 90,90 1,52 166

BRBTMF 89,73 90,94 20,05 48

The ductility of the structural system can be calculated using the envelope curve obtained from the experiment. From Figure 7(d), by taking the peak points in each cycle and by mean of first yielding point position defined from the energy equilibrium in the strain hardening region, it was obtained the ductility, µ, of 5.98 (see Sugihardjo, et al., 2006). This value is 32.8% greater than that assumed earlier in the preliminary design of STMF X-diagonal type having the ductility of 4.5 when calculated based on 3Rw/8 (Uang, 1991).



(a)BRB-1-50X8 (b) BRB-2-40X8

(c)BRB-3-50X8 (d)BRBTMF

Figure 7 Hysteretic loops of BRB and BRBTMF.

5. CONCLUSIONS

1. From the nonlinear time history analysis, the BRBTMF has the story-drift ratio up to 67% and the roof drift up to 21.6% smaller than that in STMF X-diagonal type due to the Miyagi earthquake. The result satisfied the foregoing studies, that is, the use of BRB will reduce the roof drift and the inter-story drift ratio.

2. From the experimental study, it was proven that the BRBTMF has better ductility. Hence, the reduction factor of the first yielding load, R, could reach up to 11.38 (or Rw =15.94), which is 32.8% greater than the STMF X-diagonal type. The hypothesis that the use of BRB elements will increase the structural ductility is proven. This result also supports the NEHRP (2000) recommendation that the value of R for frame with BRB stiffeners can be increased.

3. From the experimental study on the BRB specimens, it is proven that the proposed model still has a stable hysteretic curve at strain of 2% (or 13,5δy) in cyclic loading. This value is sufficient since the resulted cumulative ductility factor, η = (166-377) > 100, has satisfied the requirement as an elasto-plastic structural element.

4. From the experimental study on the full-scale BRBTMF, it is proven that the structural hysteretic still occurs up to the drift ratio of 3% (or 6δy) in cyclic loading, even though there was stiffness degradation. This value is also sufficient because analytically, the BRBTMF had the drift ratio less than 2%.

5. In terms of the cumulative ductility factor of the BRBTMF specimen, which was 48, the proposed BRBTMF satisfies the requirements as a hysteretic structural system which requires a value greater than 20.

6. For the BRB and BRBTMF specimens, in terms of the comparison of the first yielding force and the cumulative dissipating energy on the relatively similar analytical model, it can be concluded that the model developed using DRAIN-2DX software was accurate.




6. REFERENCES

AISC (1997, 2005). “Seismic Provisions for Structural Steel Building”, American Institute of Steel Construction, Chicago.

Akiyama, H. (1985). “Earthquake-Resistant Limit-State Design for Building”, University of Tokyo Press.

Basha, H.S. and Goel, S.C. (1996). “Seismic-resistant truss-moment frames with Vierendeel segment”, The 11th World Conference on Earthquake Engineering, paper no. 487.

Bruneau, M., Uang, C.M. and Whittaker, A. (1998). “Ductile Design of Steel Structures”, McGraw-Hill, New York, 411-414.

Chen, C.C. and Lu, L.W., (1990). “Development and experimental investigation of a ductile CBF system”, Proceeding of the 4th National Conference on Earthquake Engineering, Palm Springs, Calif., Vol.2, 578-584.

Chopra, A.K. (2001). “Dynamics of Structures: Theory and Applications to Earthquake Engineering, 2nd ed.”, Prentice Hall, Upper Saddle River, NJ, 265-283.

Clark, P.W., Kasai, K., Aiken, I.D. and Kimura, I. (2000). “Evaluation of design methodologies for structures incorporating steel unbonded braces for energy dissipation”, The 12th World Conference on Earthquake Engineering, paper no. 2240.

Goel, S.C. and Itani, A.M. (1994a). “Seismic behavior of open web truss moment frames”, Journal of Structural Engineering, ASCE, 120(6), 1763-1780.

Goel, S.C. and Itani, A.M. (1994b). “Seismic-resistant special truss-moment frames”, Journal of Structural Engineering, ASCE, 120(6), 1781-1797.

ICBO (1988, 1997). “Uniform Building Code”, International Conference of Building Officials, Whtitier, Calif.

Inoue, K. (2004a). “Hysteresis-type vibrations dampers. Design of hysteresis type dampers”, Steel Construction Today and Tomorrow, The Japan Iron and Steel Federation, June, 4-6.

Inoue, K. (2004b). “Low yield-point steel for steel dampers”, Steel Construction Today and Tomorrow, The Japan Iron and Steel Federation, No.7, June, 7-8.

Kamura, H., Katayama, T., Shimokawa, H. and Okamoto, H. (2000). “Mechanical property of low yield strength steel and energy dissipation characteristics of hysteretic dampers with low yield steel”, US-Joint Meeting for Advanced Steel Structures, 1-4.

NEHRP (2000). “Recommended Provisions for Seismic Regulations for New Buildings and Other Structures”, BSSC, Washington, D.C., 43-75.

Prakash, V. and Powell, G.H. (1992). “DRAIN-2DX”, University of California, Berkeley, California.

Shimokawa, H. and Kamura, H. (1999). “Hysteretic behavior of flat-bar brace stiffened by square steel tube”, The 6th International Conference on Steel & Space Structures, Singapore, 1-4.

Sugihardjo, H., Merati, W., Surahman, A. and Moestopo, M. (2004). “Analytical study of behavior of ductile truss-girder frames with low yield X-diagonal-type as hysteretic damper”, Proceeding of Conference on Earthquake Engineering II, Indonesian Earthquake Association (IEEA), ISBN 979-95620-1-5, Yogyakarta, 67-79 (in Indonesia).

Sugihardjo, H. (2006). “Inelastic Behavior of Ductile Buckling-Restrained Braced Truss-Girders Frames as Component of Storey Buildings”, Dissertation, School of Postgraduate, ITB, Bandung (in Indonesia).

Uang, C.M. (1991). “Establishing R (or Rw) and Cd factors for building seismic provisions”, Journal of Structural Engineering, ASCE, 117(1), 19-28.

Watanabe, A., Hitomi, Y., Saeki, E., Wada, A. and Fujimoto, M. (1989). “Properties of brace encased buckling-restraining concrete and steel tube”, Proceeding of 9th World Conference on Earthquake Engineering, Tokyo-Kyoto, Japan, Vol. IV, 719-724.



BUCKLING-RESTRAINED BRACES FOR SEISMIC RETROFITTING OF STEEL FRAMES

Marco Valente1

1Dipartimento di Ingegneria Strutturale, Politecnico di Milano, Milano, Italia


ABSTRACT: The effectiveness of using buckling-restrained braces (BRBs) as global seismic retrofitting strategy for low-rise steel moment resisting frames was investigated. Buckling-restrained braces show the same load deformation behavior in both compression and tension and high energy absorption capacity. Detailed finite element models of buckling-restrained braces were developed and validated comparing numerical results with experimental data derived from laboratory tests carried out at the University of California at Berkeley and available in literature. Numerical analyses showed a good agreement with experimental tests and confirmed the ability of the developed numerical models to capture the behaviour of buckling-restrained braces. Then, a design procedure, based on ductility demand, for the seismic retrofitting of steel frames using buckling-restrained braces was presented. The design procedure was applied to retrofit a low-rise steel frame designed without earthquake provisions. The assessment procedure and retrofitting strategy of the steel frame were conceived making use of non-linear static analyses. Results from pushover analyses demonstrated the efficacy of the design procedure and the improvement achieved by the retrofitting intervention.

1. INTRODUCTION

Bracing is a very effective global upgrading strategy to enhance the global stiffness and strength of steel unbraced frames. However, the energy dissipation capacity of conventional braced frames degrades drastically under earthquake loading, due to buckling of bracings in the compression range. Buckling-restrained braces (BRBs) are an innovative retrofitting system, in which the buckling of the bracing is prevented and the energy is dissipated by plastic deformation under tension–compression cycles, with stable hysteresis. This study investigates the benefits of using buckling-restrained braces as global seismic retrofitting strategy for low-rise steel moment resisting frames. Detailed finite element models of the buckling-restrained braces were created using the computer code Abaqus. Experimental data from laboratory tests available in literature were used to calibrate the numerical models in order to represent the cyclic behaviour of buckling-restrained braces. Then, a ductility-based seismic design procedure was applied to retrofit a low-rise steel moment resisting frame designed without the “strong column-weak beam” provision. Numerical analyses were performed to study the seismic response of the bare steel frame without buckling-restrained braces. The sample moment resisting frame was then retrofitted by employing buckling-restrained braces. Numerical results show that buckling-restrained brace is a reliable and practical alternative to conventional framing systems to enhance the earthquake resistance of existing structures, capable of providing energy dissipation and the rigidity needed to satisfy structural drift limits.

2. BUCKLING-RESTRAINED BRACES

A typical buckling-restrained brace consists of a steel core encased in a steel tube filled with concrete, Figure 1. The steel core carries the axial load while the outer tube, via the concrete, provides lateral support to the core and prevents global buckling. The assembly is detailed so that the central yielding core can deform longitudinally independent from the mechanism that restrains lateral and local buckling. Since lateral and local buckling behavior modes are restrained, large inelastic capacities can be reached. The inelastic cyclic behavior of several types of buckling-



restrained braces has been reported in literature. Figure 2 shows a comparison of the hysteretic behavior of typical conventional bracing and buckling-restrained bracing. It can be seen that when the conventional brace is subjected to large compressive forces, it exhibits buckling deformation and unsymmetrical hysteretic behavior in tension and compression, showing substantial strength deterioration when loaded in compression. On the contrary, because of buckling prevention in BRB, no strength degradation or stiffness deterioration occur when BRB is subjected to strong ground motion and the energy absorption of the brace is markedly increased.

Figure 1 Details of a typical buckling-restrained brace.

Figure 2 Difference in energy dissipation between conventional bracing and buckling-restrained bracing

under cyclic loading.

3. FINITE ELEMENT MODELS OF BUCKLING-RESTRAINED BRACES

Detailed finite element models of buckling-restrained braces were created in order to reproduce and to study the local behaviour. The developed numerical models were calibrated and validated comparing the numerical results with experimental tests carried out at the University of California at Berkeley and available in literature, Black et al. (2004), in order to predict, with fidelity, the brace force–displacement behavior. The tested unbonded braces had a steel core with cruciform cross-section and were subjected to a loading program consisting of increasing displacement amplitude. Successively these tests were followed by additional tests which included large-deformation low-cycle fatigue tests. The performed experimental tests were simulated in the ABAQUS commercial code using non-linear finite element models. The steel core and the outer tube were modeled with shell elements, the concrete was modeled with eight-node continuum elements. Geometrical and material non-linearities were incorporated for the analyses. Structural steel was assumed to behave like an elasto-plastic material with hardening both in compression and in tension, using von Mises plasticity. Modulus of elasticity, yield stress and ultimate stress of steel were taken from the laboratory tests performed on the materials of the tested specimens. The “damaged plasticity” model was used for concrete. Details of the refined mesh of the buckling-restrained brace are reported in Figure 3.



Figure 3 Finite element model of the buckling-restrained brace: steel core, concrete, steel tube.

Numerical results indicate that the developed finite element models are able to capture quite well the overall behavior of the tested buckling-restrained braces. Figure 4 shows the recorded force in the brace versus the total displacement measured across the yielding portion of the brace and the comparison with the results of the numerical model subjected to monotonic displacement. As it can be seen, a good agreement is observed between experimental cyclic and numerical curves. The initial stiffness in the numerical model curve is identical to that of the test; after the steel core yielding, the numerical curve continues with almost the same slope as the experimental and the peak values of the recorded and the computed force in the brace are similar.

The force–displacement loops resulting from the low-cycle fatigue test performed on the tested specimen are shown in Figure 5. In order to validate the finite element model, the overall response was checked in terms of deformed shape and force-displacement diagrams and compared with the experimental results. The brace exhibited stable hysteretic behaviour for the entire test and in the compression phase the behaviour was not influenced by buckling. Numerical analyses showed a concentration of plastic strain in the steel core. The yielding of the BRB occurred at the same loading of the experimental test. At the end of the analyses, the whole yielding of the steel core occurred, while the connection remained in the elastic range, Figure 6. Numerical results suggested the effectiveness of the model, showing good correlation with experimental data.

-4000

-3000

-2000

-1000

0

1000

2000

3000

4000

-80 -60 -40 -20 0 20 40 60 80

s [mm]

N [kN]ExperimentalNumerical

Figure 4 Force-displacement curve of yielding portion of the steel core: experimental and numerical results.

-4000

-3000

-2000

-1000

0

1000

2000

3000

4000

-40 -30 -20 -10 0 10 20 30 40

s [mm]

N [kN]NumericalExperimental

Figure 5 Brace hysteretic behaviour under low-cycle fatigue test: experimental and numerical results.




Figure 6 Yielding of the steel core of the finite element model of the BRB.

4. SEISMIC RETROFITTING OF A LOW-RISE STEEL FRAME WITH BRB

A design procedure based on the ductility demand is presented for the seismic retrofitting of a low-rise steel frame using buckling-restrained braces. A steel moment resisting frame designed without earthquake provisions is considered. The geometry and profile sections of the analyzed frame are schematically shown in Figure 7. It is a three-story three-bay frame with an inter-storey height of 3.5 m for all floors with the exception of the ground floor, which is 4 m high, and a bay width of 5 m: the grade for steel is Fe430. The frame has to be retrofitted for a 0.25g PGA level. Figure 8 shows the BRB configuration adopted for the retrofitting intervention. The effectiveness of buckling-restrained braces as global seismic retrofitting strategy for steel moment resisting frames is investigated using static nonlinear (pushover) analyses. The nonlinear static analysis is considered to be a valid alternative to nonlinear dynamic approach, particularly in preliminary analysis and retrofit of regular, low-rise structures, for their simplicity in concept and convenience in application.

Figure 7 Geometry and profile sections of the frame at study (dimensions in meter).

Figure 8 Retrofitting of the steel frame using inverted V-type BRB.

4.1 Pushover Analyses on the Existing Frame

Non-linear static analyses were carried out to evaluate the seismic performance of the existing frame. According to Eurocode 8, Part 3, the state of damage in the structure was evaluated with reference to the following Limit States (LS): LS of damage limitation (DL), LS of significant damage (SD), LS of near collapse (NC). The relationship between the base shear and the top story displacement, generally called the pushover or capacity curve, is obtained by gradually increasing the lateral seismic story forces appropriately distributed over the stories. Two vertical distributions of the lateral loads were applied: a “uniform” pattern, based on lateral forces that are proportional to mass regardless of elevation and a “modal” pattern, proportional to lateral forces consistent with the lateral force distribution determined in elastic analysis. Figure 9 shows the base shear force – top story displacement curve for the frame. A soft-storey mechanism was observed at about 12 cm of top displacement of the frame. Several plastic hinges occurred at the Limit State of damage



limitation (LSDL). The displacement demands for the LSSD and LSNC were not satisfied by the structure.

0

50

100

150

200

250

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Top Displacement [m]

Bas

e Sh

ear [

kN]

Pushover Curve

Demand LSDL

Plastic Hinge

Figure 9 Pushover curve for the bare frame.

The sequence and distribution of the plastic hinges and the collapse mechanism for the frame are shown in Figure 10. The collapse mechanism was different for the two different loading conditions. A soft-storey mechanism was observed in both cases, but the ductility demand was concentrated at the first storey for the “uniform” distribution loading and at the second storey for the “modal” distribution loading. The most unfavourable result of the pushover analyses using the two standard lateral force patterns was adopted. The non-linear static analyses showed the seismic vulnerability of the frame.

1 2

6439

8

7

11 2 4 6

10 1 3

7 8

9

5

5

Figure 10 Collapse mechanism, plastic hinge sequence and distribution for two different loading conditions.

4.2 Design Procedure for Buckling-Restrained Braces

The sample moment resisting frame was then retrofitted by employing buckling-restrained braces. A design procedure, based on the ductility demand, for the seismic retrofitting of a steel frame using buckling-restrained braces was applied. The procedure has advantage of using nonlinear static analysis instead of nonlinear dynamic analysis. By performing pushover analyses, it was possible to determine the target displacement of the structure and then the maximum ductility demand on the frame.

The steel core strain εsc of the BRB can be expressed as:

cossc

sc

dL

φε ⋅= (1)

where cosd φ represents the brace axial elongation and Lsc is the steel core length, Figure 11. Given the steel core strain εsc, the steel core length Lsc can be evaluated in order to satisfy the relationship (1). On the contrary, when the ductility demand is known, one can determine Lsc which verifies the relationship (1) by varying the steel core strain. Consequently, by acting on the steel core length Lsc, the ductility can be controlled, independently from the choice of the steel type and from the core section.



Figure 11 Schematic of brace deformation versus inter-story drift relationship.

Figure 12 shows the graphical procedure for computing the target displacement. The effective damping ratio βeff can be obtained as follows:

( )2 1eff eq

μβ β β β

πμ−

= + = + (2)

where u

y

dd

μ = is the ductility demand and β is the inherent damping of the frame. The yield

displacement can be computed as follows:

,, cos

y sc iy y i

i

f Ld d

E φ⋅

= =⋅∑ ∑ 3)

where fy and E are the yield force and the elastic modulus of the BRB, Lsc,i is the steel core length and Φi is the slope of the brace of the ith storey. It can be seen that the yield displacement does not depend on the cross-sectional area of the brace, but on the yield strength and geometry of the brace. This simplifies the design procedure, because the yield displacement is known before the size of the braces is determined. The displacement du is the target displacement for the considered limit state. The unknowns for the evaluation of the performance point are the length Lsc and the axial strain εsc of the steel core. The proposed procedure aims at calculating by iterative procedure the value of the ratio Lsc/LBRB in order to find the performance point corresponding to the maximum ductility. The performance point is obtained from the cross-point of the demand and the capacity curves plotted in the spectral acceleration versus spectral displacement domain, known as the acceleration–displacement response spectra (ADRS).

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.05 0.1 0.15 0.2 0.25

Sd [m]

Sa [g

]

Demand Spectrum

Capacity Curve

Performance Point

Figure 12 Graphical procedure for computing the target displacement.

The main results obtained using the iterative procedure applied to the considered frame are summarized in Table 1.



Table 1 Parameters of the retrofitted frame.

storey Lsc/LBRB Lsc [m] dyi [m] dy [m] du [m] μ Tbrb [s]

1 0.8 3.77 0.012

2 0.8 3.44 0.010

3 0.8 3.44 0.010

0.032 0.185 5.78 0.81

Once the target displacement is computed, the cross-sectional area Asc of the energy dissipation core segment Lsc is determined as follows:

0.9sc

scy

PAf

= (4)

where the value of Psc is evaluated as:

cosi

sci

FP =Φ

(5)

and Fi is the story shear force. According to Tsai et al. (2004), the effective stiffness Ke of the BRB considering the variation of the cross-sectional area along the length of the brace can be accurately predicted by:

2 2c sc t

ec t sc sc t c sc c t

EA A AKA A L A A L A A L

=+ +

(6)

where Ac, Lc, At, Lt, Asc, Lsc are the cross-sectional area and length of the connection, of the transition region and of the steel core, respectively.

The required strength of bracing connections in compression and tension (including beam-to-column connections if part of the bracing system) can be determined as:

max y scP f Aβω= (7)

where β and ω are the compression and tension strength adjustment factors, respectively, AISC 2002.

The design parameters of the buckling-restrained braces are summarized in Table 2.

Table 2 Design parameters of the buckling-restrained braces.

Storey F [kN] Psc [kN] Asc [mm2] Pmax,t [kN] Pmax,c [kN] Ac [mm2]

1 140 265 830 403 444 1251

2 106 183 574 279 307 865

3 50 87 270 131 144 407

4.3 Pushover Analyses on the Retrofitted Frame

Non-linear static analyses were performed on the frame retrofitted employing BRBs. The aim of the application of the BRBs to the bare structure was the reduction and uniform distribution of ductility demand; a decrease of interstorey drift and a reduction of the plastic hinge rotation were observed. The analysis results showed that the collapse mechanism of the structure was preceded by the plastic deformation of all the BRBs which plasticized before the beams and columns. A significant increase of the stiffness and strength of the retrofitted frame was observed with respect to the bare frame, reducing the displacement demand on the structure. At the LSLD demand, no plastic hinges were observed. As the braces were designed in accordance with the proposed procedure, the yield displacement of each story resulted almost the same. In this case, the




structural damage was evenly distributed throughout the stories. Results from pushover analyses demonstrated the improvement achieved by retrofit and Figure 13 shows the efficacy of the retrofitting intervention. It can be concluded that the applied design procedure can be a convenient tool for the seismic retrofitting of a low-rise steel frame with BRB.

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.05 0.1 0.15 0.2 0.25 0.3

Sd [m]

Sa [g

]

Bare Structure

Demand Spectrum

Retrofitted Structure

Figure 13 Capacity curve for the bare and retrofitted structures.

5. CONCLUSIONS

The effectiveness of buckling-restrained braces as global seismic retrofitting strategy for steel moment resisting frames was investigated. The application of finite element models as tools to study the local behavior of buckling-restrained braces was presented. Detailed finite element models of buckling-restrained braces were developed and compared with experimental data available in literature. Numerical analyses showed that the experimental response of the buckling-restrained braces was accurately predicted by the developed numerical models. A design procedure based on the ductility demand for the seismic retrofitting of a steel frame using BRB was applied. An existing low-rise steel moment resisting frame designed without earthquake provisions was analyzed and the seismic response was evaluated performing non-linear static analyses. The sample moment resisting frame was then retrofitted by employing buckling-restrained braces. Results from pushover analyses demonstrated the efficacy of the design procedure and the improvement achieved by the retrofitting intervention.

6. REFERENCES

American Institute of Steel Construction (2002). “Seismic Provisions for Structural Steel Buildings”, Chicago, IL, USA.

Black, C.J., Makris, N. and Aiken, I.D. (2004). “Component testing, seismic evaluation and characterization of buckling-restrained braces”, Journal of Structural Engineering, ASCE, 130(6): 880-894.

Choi, H. and Kim, J. (2006). “Energy-based seismic design of buckling-restrained braced frames using hysteretic energy spectrum”, Engineering Structures 28: 304-311.

European Standard, prEN 1998-3, 2003, Eurocode 8 (2003). “Design of Structures for Earthquake Resistance”, Part 3: Strengthening and Repair of Buildings.

Huang, Y.H., Wada, A., Sugihara, H., Narikawa, M., Takeuchi, T. and Iwata, M. (2000). “Seismic performance of moment resistant steel frame with hysteretic damper”, Proceedings of 3rd International Conference STESSA, Montreal, Canada.

Mahin, S., Uriz, P., Aiken, I., Field, C. and Ko, E. (2004). “Seismic performance of buckling-restrained braced frame systems”, Proceedings of 13th World Conference on Earthquake Engineering, Vancouver, Canada.

Tsai, K.C., Lai, J.W., Hwang, Y.C., Lin, S.L. and Weng, C.H. (2004). “Research and application of double-core buckling restrained braces in Taiwan”, Proceedings of 13th World Conference on Earthquake Engineering, Vancouver, Canada.



PERFORMANCE OF ORDINARY STEEL MOMENT FRAMES DESIGNED BASED ON INDONESIAN STEEL BUILDING CODE

SNI 03-1729-2002

Ima Muljati1 and Hasan Santoso1

1 Civil Engineering Department, Petra Christian University, Surabaya, Indonesia

Email: [email protected] ; [email protected]

ABSTRACT: After Northridge- and Kobe-earthquake (1994 and 1995), Federal Emergency Management Agency (FEMA) investigation shows that building designed as Ordinary Moment Frames (OMFs) has performed well under moderate level of ground motion. OMFs are designed to have a large amount of stiffness and less ductility when carrying lateral loads. Therefore, “strong column weak beam” requirement can be neglected in its design. Unfortunately, this structural system has less attracted in the structural design development in Indonesia, especially for steel structures. In accordance with the new published steel structures design code for building, SNI 03-1729-2002, the objective of this study is to evaluate the performance of OMFs at zone 2 of Indonesian seismic map. Three buildings with symmetrically plan view, including 4-, 8-, and 12-story are evaluated. The structural performance of these buildings is analyzed using static nonlinear pushover and dynamic nonlinear time history analysis. The results show that the 4-story building fulfills the drift requirement but the 8- and 12-story buildings fail to fulfill the drift requirement. However, based on the performance matrix, the damage indices of all buildings are in acceptable value.

1. INTRODUCTION

In FEMA 450 Seismic Provision (2003), there are three types of steel moment-resisting frame structure. One of them is Ordinary Moment Frame (OMF). The proportioning and configuration of OMF are less restricted than Special Moment Frame (SMF) because it is designed to have limited ductility capacity. Consequently, OMF is expected to deteriorate at lower drift levels than that of an SMF. This is accounted for in design by prescribing a smaller response modification factor, R and “strong column weak beam” requirement does not need to be applied. Consequently, OMF is suitable for buildings in low seismic region.

FEMA 355F (2000) investigated three structures including 3-, 9- and 20-story buildings (12-m, 40-m, and 88-m) which were designed as OMF. The results showed that the 3- and 9-story buildings performed well under moderate earthquake. On the other hand the 20-story building showed a considerable drift value. Therefore, OMF is only recommended for structures with the maximum height of 30-m. Effendi and Sie (2001) reported that the study of OMF is relatively less in Indonesia, resulting minor application in the design practices. This study is aimed to evaluate the performance of OMF in the low seismic area which is designed based on the latest Indonesian Steel Design Code for Building, SNI 03-1729-2002 and Indonesian Seismic Code, SNI 03-1726-2002.

2. BUILDING DESCRIPTION

Three office buildings which represent low- to medium-rise building (4-, 8-, and 12-story), will be observed in this study. They are in symmetrical plan (Figure 1) and built on soft soil in the zone 2 of Indonesian seismic map (low seismic risk). The columns are placed so that the buildings have approximately equal rigidity both in x- and y-direction. All structural elements are made of steel grade BJ 37 with the yield strength, fy, equals to 240 MPa. The height of the first floor is 4-m, while the others are 3.5-m. The elevation views of the structures are shown in Figure 2.





Figure 1 Building plan.

Figure 2 Elevation view.

3. DESIGN

Some assumptions are made in the design procedures, they are: (1) connection are considered to be rigid connection; (2) secondary beams are modeled as a grid system; (3) all beams are laterally restrained due to the rigid diaphragm contributed by the floors; (4) buildings are modeled as 3D- structure.

Structures are designed as Ordinary Steel Frames (Struktur Rangka Penahan Momen Biasa, SRPMB) based on Indonesian Steel Structures Design Code for Building, SNI 03-1729-2002. The detailed calculation procedures can be found in Suwono and Juslim (2007). The required strength, U, are calculated based on the loading combinations:

EDUELDU

LDUDU

R

±=±+=

+==

9.012.1

6.12.14.1

(1)



where D, L, E are dead-, live-, and seismic-load, while LR is reduced live load. The seismic load is determined in accordance to the equivalent lateral force procedure as stated in Indonesian Seismic Design Code, SNI 03-1726-2002, using:

RIWCV t1= (2)

where V, C1, I, Wt, and R are seismic base shear, seismic response coefficient, occupancy importance factor, weight of the building, and response modification factor respectively. Then the beams capacity is checked using moment and shear interaction:

375.1625.0 ≤+n

u

n

u

VV

MM

φφ (3)

Finally, the columns capacity is checked using moment and axial force interaction:

0.12

then ;2.0 If

0.198 then ;2.0 If

≤⎥⎥⎦

⎤

⎢⎢⎣

⎡++<

≤⎥⎥⎦

⎤

⎢⎢⎣

⎡++≥

ncyb

ucy

ncxb

ucx

cc

uc

cc

uc

ncyb

ucy

ncxb

ucx

cc

uc

cc

uc

NM

NM

NN

NN

NM

NM

NN

NN

φφφφ

φφφφ (4)

M, V, and N are moment, shear, and axial force at the member; indices u, n, x and y represent ultimate, nominal, and global axes of the building; while φc, φb are reduction factors for column and bending.

4. ANALYSIS

The performance of the structures are evaluated to static non-linear pushover analysis (ATC 40, 1996) using ETABS-nonlinear (Habibullah, 1998) and non-linear time history analysis using RUAUMOKO 3D (Carr, 2002). The properties of hinges at the beams and columns are determined using XTRACT v3.0.5.

For the time history analysis, the study uses spectrum consistent ground acceleration modified from N-S component of El-Centro 1940. The modification is achieved using RESMAT (Lumantarna et al., 1997), a software program developed at Petra Christian University, Surabaya. As stated in the basic requirement for design, the ground acceleration is based on the 500-year earthquake return period. In order to obtain the ultimate performance of the building, the buildings are also checked to the 1000-year earthquake return period.

The building performance evaluations are determined based on the Asian Concrete Model Code, ACMC (2001). The drift ratio and damage index resulted from pushover and time history analysis is plotted in the structural performance matrix shown in Figure 3.

Figure 3 Structural performance matrix based on ACMC 2001.




The dimension of wide flange (WF) section used as main beam and columns are shown in Table 1 and 2. The resulting displacement and drift ratio are shown in Figure 3 – 5.

Both pushover and time history analyses give relatively equal value of displacement. However, the time history analysis detects a significant drift in the y-direction at the upper floor, especially for 8- and 12-story buildings. The large drift is caused by the use of different size of the column section at the 4th- and 5th-floor of the 8-story building and the 9th- and 10th-floor of the 12-story building. The maximum drift ratio resulted from both analysis are plotted in the performance matrix shown in Table 3 and 4. Subsequently, the damage indices are shown in Table 5 and 6.

Based on the evaluation of the drift ratio, it can be drawn as follows:

1. For the 500-years earthquake return period, the drift ratio of the 4-story building (in both direction) is in the safety limit state. However, for the 1000-years earthquake return period, it is in the unacceptable condition.

2. The drift ratio of the 8- and 12-story buildings is in the safety limit state, but only in the x-direction. While in the y-direction, they are in the unacceptable condition (detected by the time history analysis). Thus, the arrangement of column resulting unbalanced stiffness for the buildings. The buildings are less stiff in the y-direction than that in the x-direction.

3. In the observed cases, the time history analysis seems to be more conservative than the pushover analysis in predicting the drift of the structures.

Table 1 Beam dimension (WF section).

X-direction Y-direction Buildings Floor

Exterior Interior Exterior Interior 4-story All 300x150x5,5x8 350x175x7x11 300x150x5,5x8 350x175x7x11

1-4 300x150x5,5x8 350x175x7x11 5-7 250x125x6x9 8-story

8 300x150x5,5x8

350x175x7x11 300x150x5,5x8 350x175x6x9

12 Lantai All 300x150x5,5x8 400x200x7x11 350x175x7x11 400x200x7x11

Table 2 Column dimension (WF section).

Buildings Floor Exterior Corner Interior 1 350x350x14x22 350x350x13x13 400x400x13x21 2 350x350x19x19 300x300x11x17 400x400x15x15 3 350x350x13x13

4-story

4 350x350x13x13 300x300x12x12

300x300x15x15 1 400x400x13x21 350x350x16x16 400x400x18x28

2-4 400x400x15x15 300x300x10x15 400x400x13x21 5-7 300x300x10x15 350x350x12x19

8-story

8 300x300x9x14 250x250x11x11

250x250x11x11 1 400x400x18x28 350x350x12x19 400x400x20x35

2-5 400x400x21x21 350x350x10x16 400x400x18x28 6-9 350x350x16x16 300x300x12x12 400x400x18x18

12-story

10-12 250x250x9x14 250x250x8x13 350x350x13x13



Figure 3 Displacement and drift ratio of 4-story building (P = Pushover, TH = Time History).





Table 3 Building performance based on drif tion).

Table 4 Building performance b sed on drif tion).

ance Lev

t ratio (x-direc

Performance Level (%) Serviceabi maglity Da e

C lSafety ccepta Una ble

Ea e

period Building

PO T PO T PO T

rthquakreturn

H H PO TH H4-story 1,61 1,60 8-story 1,60 1,54 500-years

1 1 12-story ,4 1,71 4-story 1 ,95 2,03 8-story 2, 09 2,04 1000-years

1,87 12-story 2,04 Maximum drift (%) 0,5 1,0 2,0 >2,0

a t ratio (y-direc

Perform el (%) Serviceabi maglity Da e

lSafety ccepta Una ble

Ea e

period Building

PO T PO T PO T

rthquakreturn

H H PO TH H4-story 1,75 1,70 8-story 1,73 2,11 500-years

1,36 12-story 2,39 4-story 2,20 2,20 8-story 2,21 2,73 1000-years

1,74 12-story 2,40 Maximum drift (%) 0,5 1,0 2,0 >2,0

Table 5 Building performance ba ection).

ance Lev

sed on damage index (x-dir

Perform el (%) Serviceabi mag afety cceptality Da e S Una ble

Ea e

period Building

PO T PO T PO T

rthquakreturn

H H PO TH H4-story * 0.614 8-story * 0.573 500-years 12-story * 0.667 4-story * 0.739 8-story * 0.824 1000-years

ry * 7 12-sto 78Damage index 0,1-0,25 ,25-0,4 0,4-1,0 >1,0 0

Table 6 Building performance ba ection).

ance Lev

sed on damage index (y-dir

Perform el (%) Serviceability Damage

C lSafety ccepta Una ble

Ea e

period Building

PO T PO T PO T

rthquakreturn

H H PO TH H4-story * 0.642 8-story * 0.787 500-years 12-story * 0.726 4-story * 0.751 8-story * 0.978 1000-years

ry * 84 12-sto 0.7Damage index 0,1-0,25 0,25-0,4 0,4-1,0 >1,0




arthquake return period, the able damage. Similar to the drift consideration, the time history analysis is

edicting the damage indices of the buildings.

able for low rise

of pushover analysis is applied only for low-rise buildings. For nalysis is recommended due to its wide range applicability.

d to our students Kenneth Ivan Suwono and Hans Yulius Rukma Juslim with he work and collaboration, and for patiently enduring the six months of

ncy.

gs”, Federal Emergency Management Agency.

Effe . “Analisa Performance Based Design untuk Sistem Rangka

Suw inerja Sistem Rangka Pemikul Momen Biasa

etra Christian University.

Lu M. (1997). “Resmat, sebuah program interaktif untuk menghasilkan riwayat waktu gempa dengan spektrum tertentu”, Proc. HAKI Conference 1997, Jakarta, Indonesia, pp. 128-135.

Based on the damage index criteria, all buildings are in the safety limit state both for the 500- and 1000-years earthquake return periods. However, for the 1000-years ebuildings are in considermore conservative in pr

6. CONCLUSION

Based on the analysis of the observed buildings, it can be concluded:

1. The design of ordinary moment frame on soft soil in low seismic area based on the Indonesian Steel Design Code for Building, SNI 03-1729-2002, is conservative and suitbuildings (up to 20-m). For medium-rise building (20- to 40-m), some caution should be taken especially in the less-stiff-axes of the buildings due to the drift requirement.

2. In the case of analysis, the usebetter accuracy, time history a

7. ACKNOWLEDGEMENT

This paper is dedicategratitude for contributing tpreparation with us.

8. REFERENCES

FEMA-450 (2003). “NEHRP Recommended Provisions for Seismic Regualations for New Buildings and Other Structures”, Federal Emergency Management Age

FEMA-355F (2000). “State of the Art Report on Performance Prediction and Evaluation of Steel Moment-Frame Buildin

ACMC (2001). “Asian Concrete Model Code, Level 1 & 2 Documents”, International Comittee on Concrete Model Code. ndi, C. and Sie, W.A. (2001)Penahan Momen (SRPM) yang Direncanakan Sesuai SNI”, Undergraduate Thesis No. 1161 S, Petra Christian University.

SNI 03-1729-2002 (2002). “Tata Cara Perencanaan Struktur Baja untuk Bangunan Gedung”, Departemen Pemukiman dan Prasarana Wilayah.

SNI 03-1726-2002 (2002). “Tata Cara Perencanaan Ketahanan Gempa Untuk Bangunan Gedung”, Departemen Pemukiman dan Prasarana Wilayah. ono, K.I. and Juslim, H.Y.R. (2007). “Evaluasi K(SRPMB)yang Didesain Berdasarkan SNI 03-1729-2002”, Undergraduate Thesis No. 11011526/SIP/2007, P

ATC 40 (1996). “Seismic Evaluation and Retrofit of Concrete Buildings”, Volume I. Applied Technology Council.

Habibulah, A. (1998). “ETABS, Three Dimensional Analysis and Design of Building Systems”, Computer and Structures Inc.

Carr, A. (2002). ”Ruaumoko Computer Program Library”, University of Canterbury - New Zealand: Department of Civil Engineering.

mantarna, B. and Lukito,



EFFECT OF CONCRETE SLAB TO THE RESISTANCE OF NON-MOMENT RESISTING FRAME

Herman Parung1

1 Civil Engineering, Hasanuddin University, Makassar, Indonesia


ABSTRACT: In designing a steel frame to resist seismic force, usually only parts of the frame designed to carry seismic load (usually called moment resisting frame). Other parts will be designed to carry only gravity load (dead and live load). This part is usually called non-moment resisting frames. The connection between beam and column is designed as shear connection which carries only the shear force. In reality concrete slab is always present and can add rigidity to the structure including to the shear connection. The connection will then not longer carry shear force only but also some of the induced moments (act as a semi rigid connection).

The aim of the research is to study the behavior of shear connection in resisting seismic forces. Results of test on beam-column sub-assemblages using shear connection will be presented. The obtained result is then used to describe the behavior of a steel-concrete composite frame which has various connections (rigid bolted and welded connection, and also shear connection).

Experimental result presented in this report has shown that the shear connection attached to the concrete slab would act as semi-rigid connection, and could carry between 10 - 1 5 % of the total moments. Hence, it can be concluded that the resistance of the semi-rigid connections under increasing deformations constitute about 1/8 of the resistance of the rigid moment connection. This fact shows that resisting frames should only be designed to carry 85-90 % of the seismic-induced moments. By applying this finding, the cost for constructing composite frames to resist seismic action could be significantly reduced.

1. INTRODUCTION

All composite construction for rigid frame (with rigid connection) has been fully studied numerically and analytically by a research team Bouwkamp et al. (1997). In reality, a 3 - dimensional composite construction would not only consist of rigid frame, but also non rigid frame. This type of frame is designed with so-called shear connection, and assumed to carry loads only. However, due to the concrete slab and steel beam interaction, the shear connection would not carry only gravity load, but may also carry part of the horizontal load. This papers present experimental results carried out to study seismic resistance of shear connections.

2. SPECIFIC PURPOSE OF THE RESEARCH

Designing a shear connection with concrete slab attached to it will not only increase the rigidity of the frame, but also will decrease the total construction cost. This is due to the smaller cross section of the beam and column for the moment resisting frame because part of the induced - moments carried by the shear connection as part of the non-moment resisting frame. At the present, there is no specific rule/guideline exists in Indonesia about seismic design of shear connection, or semi-rigid connection. It is expected that by combining the result from previous experimental studies and analytical studies using a non linear computer program, a guideline for designing shear connection to resist seismic force can be proposed.

3. RELEVANT PAST STUDIES

A full scale testing for composite frame was performed at the Joint Research Center of the European Commission, Ispra, Italy. The connections of the frame were rigid and semi rigid (shear) connection. The result of the rigid frame testing has been reported to the European Commission in 1998.



Semi rigid composite connections are partial strength, partially restrained connections that can be used to resist love, wind, and earthquake loads in low to moderate height un-braced frames. In 1994, Roberto Leon was assigned by the American Society of Civil Engineers, Structural Division Task Committee to carry research on Design Guide for Composite Semi-Rigid Connection, Leon (1995). In USA- this type of connection usually consists of two seat-angle on the top - of the beam, and a plate welded to the beam web. Around the connection, several reinforcing bars are usually placed with the aim of increasing the moment transfer to the column.

In Europe, semi rigid connection does not use any seat angle. In 1996, a team was set up by the European Commission to study the semi rigid behavior of structural connections for braced frames, and the results of the study have been reported to the EC in 1996 (European Commission, 1996). It must be emphasized the connections were tested separately as individual connection, not as part of a real frame.

4. METHODOLGY

4.1 Design of Semi-Rigid Connection

All connections were semi-rigid connection, where beams were connected to columns by means of shear tabs. Although the shear tabs are to be designed to transmit the gravity loads only, a limited ductile behavior may be developed under rotation if certain conditions should be fulfilled, namely:

1. bearing resistance < resistance of bolt in shear

2. Fnet (= Anet x fultimate ) > Fgross (= Agross x fy ),

where fultimate and fy were taken as 35.5 kN/cm2 and 23.5 kN/cm2, respectively. The bolts were specified as HV M20-10.9. If the bolt resistance in shear is larger than the bearing resistance of the shear tab, ovalization of the holes will take place before shear failure of the bolts. Also, if Fnet is larger than Fgross, the strap will yield before failure in tension in the net section. For the typical shear tab, the following resistant values have been calculated: bearing resistance of IPE 200 is 153.7 kN while the bearing resistance of plate 10 mm is 272.8 kN and bolt resistance in shear is 245.0 kN. It shows that the bearing resistance in the web of the beam (IPE 200) is smaller than the bolt resistance in shear; hence, the beam's web is the weak (and ductile) part of the steel connection.

4.2 Experimental Works

Three beam-column sub-assemblages with concrete slabs were tested. These sub-assemblages had the same beam and column sections and the same connection and slab design as those of the 3-D test structure. Those specimens were designed with semi rigid: two with an interior column and one with an exterior column.

The structural characteristic of the sub-assemblages, which are shown are presented in the following. Specimen no. 1 is a planar frame section with semi rigid connection typical to those used in the 3-d test frame. Also, column and beam section also identical to those selected for the test frame. The basic characteristic of specimen no. 1, are: column HEB 240 and beam IPE 200, semi rigid connection, basic reinforced Q 513 with additional ∅ 10 mm bar on either side of column to compensate for cutting mesh around column, 80 mm spacing between studs along length of beam, corrugation of metal decking perpendicular to girder. Specimen no. 2 is identical to specimen no. 1, except that the corrugation of the metal decking is oriented parallel to the IPE 200 girders. Specimen no. 3 is a planar frame with an exterior column and solid cantilever concrete slab on the exterior. The corrugation of the metal decking runs perpendicular to the beam, as in Specimen no. 1. The reinforcing mesh, extending into the solid cantilever slab and additional cantilever reinforcement are laid out to provide a in-plane resistance of the slab under negative moment bending of the beam.

On the common assumption that in a typical moment resistant frame under lateral loads, the moment inflection points in the beam and columns are located at the mid-span and mid-



height of beams and columns, respectively, the sub assemblages were designed with hinges at those inflection points. By applying a double-acting actuator load at the top of the column and providing the necessary hinged support conditions at the other inflection, the beam-column moment transfer can be studied. This reaction frame was designed to allow a total cyclic horizontal load application of maximum 2000 kN.

As the specimens were to be tested under displacement-controlled cyclic loads, a potentiometer -with a range of + and - 200 mm was installed to control the top displacement of the column versus a reference column located on the far side of the reaction frame. For measuring the applied cyclic loads, a load cell was attached to the 100 ton actuator. At different sections along the length of the loaded beams, strain gauges were typically placed at both sides of the top flange and in the middle of the bottom flange. In order to measure in-plane transverse bending effects, strain gages were also placed at both sides of the upper flange of one of the transverse beams.

In order to measure the angular change between the beams and the column in the connection zone under loading, LVDTs were installed on each side of the column. In addition, the specimens 1 and 2 were instrumented with inclinometers installed to measure the rotation of the column and beams. The beam rotations at three positions along each beam were measured by inclinometers -laced against the beam web. The column rotation was recorded by an inclinometer placed on the side of the column flange.

Figure 1 Test setup.

5. EXPERIMENTAL RESULTS

5.1 General

The basic results of the tests are presented in this section. Specifically, for each test, the cyclic, “Force versus displacement” (actuator load F1 versus top-of-column displacement D 1) and the “Total moment versus global rotation” are being presented in one figure. Hereby is the “total moment” defined as the actuator load times the column height H", with H being the length of the pin-ended column, and the global rotation as the horizontal top-of-column displacement D 1 divided by H. Pus total moment is of course identical to the sum of the beam moments (left and right) calculated at the column center-line.

5.2 Cyclic Behavior

In this section the basic cyclic response of the beam-column assemblages with an interior column (Specimen 1 and 2) and an exterior column (Specimen 3) are being discussed. All three specimens have IPE 200 beam sections as is typical for all semi-rigid connections in the 3D test frame structure. The difference between Specimens 1 and 2 is the orientation of the metal deck



corrugation. For Specimen 1, the corrugation is oriented perpendicular to the beams and in Specimen 2, parallel to the beams. Considering the test records as presented in figures, the cyclic response shows the typical pinched hysteretic behavior associated with the slipping mechanism of the bolts connecting the beams to the shear tabs. Under bolt bearing, signified by a steep increase in the horizontal-force and moment values, the connection resistance increases markedly. The total-moment resistance values for all three specimens are virtually equal for both positive and negative moment application.

Specifically, the results of Specimen 1 and 2 show positive and negative moment resistance values under slippage of between 20 and 25 kNm. Under increasing rotations the bearing becomes significant and causes a gradual increase in the moment resistance. At rotations of + and - 0.04 rad the corresponding moments reach values of + and - 100 kNm. The fact, that these values are the same for both specimens, indicates that the effective slab thickness, due to different orientations of the corrugation, has no influence on the joint moment resistance. The overall behavior is clearly determined by the behavior of the shear-tab connection. which resistance is governed, as designed, by the bearing resistance of the beam web. It can be noted that the increase in the moment resistance from about 25 kNm at 0 rad to 100 kNm at 0.04 rad as noted above is better than linear. For Specimen 3. the exterior beam-column joint, the results are fully comparable with results observed for Specimens I and 2, namely. the total-moment values for the same corresponding rotations are about 12 and-50 kNm, respectively. Comparing the above total-moment results of Specimen 1 and 2 at a global rotation of 0.04 rad with those of fully rigid connection in Bouwkamp et al. (1997) and Parung (2003), it can be noted that the 100 kNm values for Specimen 1 and 2 compare to total-moment values of 800 kNm for rigid connection1 at 0.04 rad. Hence, it can be concluded that the resistance of the semi-rigid connections under increasing deformations constitute about 1/8 of the resistance of the rigid moment connection.

Considering the separate beam-moment rotation response data, it is obvious, that the results of the right beam of Specimen 1 and the (single) beam of Specimen 2 show excellent agreement with the earlier observed total moment behavior (an equal division of the total moment values corresponding to the associated global rotations). For the left beam of Specimen 1 a slight shift in the moment response values is obvious. For Specimen 2, the single beam-moment discrepancies for both the left and right beams are quite large. A fact is, that the cyclic reaction forces of the pin-ended struts showed a gradual shift from the zero-load line. On the other hand, despite of this phenomenon, the moment ranges, being the sum of the positive and negative moments, associated with the corresponding global rotations, or the total hysteretic loops are almost identical for all beams of the Specimen 1, 2 and 3.

Figure 2 Test result of Specimen 1.


http://rad.as/




6. CONCLUDING REMARKS

To resist seismic force, usually only several frames are design as moment resisting frames, the others would be designed as non-moment resisting frames. The later are designed to carry the gravity loads only (dead and live load), so their connections carry shear forces only. In reality, however, due to concrete slab actions, the shear connections would no longer resist shear forces only, but also moments induced by horizontal loads, such as seismic forces. Experimental result presented in this report has shown that the shear connection attached to the concrete slab would act as semi-rigid connection, and could carry between 10 - 15 % of the total moments. Hence, it can be concluded that the resistance of the semi-rigid connections under increasing deformations constitute about 1/8 of the resistance of the rigid moment connection.

This fact shows that resisting frames should only be designed to carry 85-90 % of the seismic--induced moments. Shear connection should be designed to resist shear forces only, which are induced by gravity loads. This connection, however, would carry part of the seismic-induced moments as the result of concrete slab action which turns the shear connection into semi-rigid connection. By applying this finding, the cost for constructing composite frames to resist seismic action could be significantly reduced.




7. REFERENCES

Bouwkamp, J.G., Parung, H., Plumier, C. and Doneux, C. (1997). “Research on Energy Dissipation Capacity of Composite Steel Concrete Structures”, Final Report to the European Commission, ISIS.

Doneux, C. and Parting, H. (1998). “A study on composite beam-column sub-assemblage”, Proceeding 11h European Conference on Earthquake Engineering, Paris.

European Commission (1996). “COST CI - Semi-Rigid Behaviour of Civil Engineering Structural Connections”, Final Report.

Leon, R. (1995). “Design Guide and Commentary for Partially-Restrained Composite Connections”, Report to American Society of Civil Engineers.

Parung. H. (1999). “Seismic resistance of composite structures”, Proceeding Konferensi Nasional Rekayasa Kegempaan, Bandung.

Parung, H. (2003). “Test on composite beam-column sub-assemblages”, Jurnal Hi-Tech, Edisi O1/Tahun V, Fakultas Teknik Unhas.



IMPROVED PERFORMANCE OF BOLT-CONNECTED LINK DUE TO CYCLIC LOAD

Muslinang Moestopo1, Dyah Kusumastuti1 and Andre Novan2

1Associate Professor, Structural Engineering Research Division, Faculty of Civil and Environmental Engineering, Institut Teknologi Bandung, Indonesia.

2Graduate Student, Faculty of Civil and Environmental Engineering, Institut Teknologi Bandung, Indonesia. Email: [email protected]

ABSTRACT: Recent development in seismic resistant steel frames was directed to the development of an effective and efficient energy dissipater such as a link in eccentrically braced frames. This research was carried out to study the performance of the shear-type link due to cyclic load, especially to improve the energy dissipation as well as the ductility and the strength of the link itself. Experimental work has been carried out to two specimens of bolt-connected link. Additional side-extended plates were welded at top and bottom flanges of each end of the link to prevent early failure of the link. A cyclic load was applied under displacement control. The results show that the side-extended plate has significantly improved the performance of the link, both in the strength, ductility and energy dissipation.

1. INTRODUCTION

The recent development of seismic resistance structure considers the performance of the structure as the design base. By this approach, the strength of the structure is no longer the only design criteria. The new Indonesian Structural Steel Design Code (Badan Standar Nasional, 2002) as well as the Seismic Provision by American Institute of Steel Construction (AISC, 2005) provide the design code base on the performance of the structure in dissipating energy due to seismic load. A number of seismic resisting structural systems as well as the structural requirement of each structural system are included in the code as result of very intensive and extensive researches following the Northridge earthquake (1994).

Eccentrically braced frames has been developed and widely used as a seismic resisting steel structure. Its advantages are supported by both the diagonal braces that provide lateral structure stiffness, and the link that mobilize the energy dissipation by its yielding mechanism. Previous studied by Kasai and Popov (1986) showed that the shear link or short link provided better energy dissipation than the flexural link or long link. The good performance of the shear link depends on the cyclic yielding mechanism on the web that should ensure a stable and ‘fat’ hysteretic curve. The compactness of the link web contributes to this performance.

The connections between the link–end and adjacent beam or column are commonly provided by welded-connection due to its advantage over the bolted-connection. So far, the welded-connection provides more rigidity or stiffness and more capacity, i.e. strength, plastic rotation and energy dissipation. Study by the first author (Moestopo and Mirza, 2006) showed that the lack of stiffness and strength of the bolts as well as the lack of stiffness of end-plate contributed to the deficiencies of the bolt-connected link. As shown by Gobarah (Ramadan and Ghobarah, 1995) and Stratan-Dubina (2002), the strong bolt-connected link could provide better energy dissipation due to cyclic load. Although the strong bolt-connected link could not exceed the rigidity and capacity of the weld-connected link, its advantage over the weld-connected link are obvious, i.e. easy to assembly and easy to replace after the earthquake occurrence.

This paper presents the experimental work by the authors on the bolt-connected link to study the improved performance of the shear-link in providing more energy dissipation by the yielding of the web of entire length of the shear-link.



2. PERFORMANCE OF SHEAR LINK

As a seismic resisting component that is expected to dissipate the energy by the yielding mechanism, the link of the eccentrically brace frame behaves differently according to its length. Link with the length of e, and its plastic moment and plastic shear are Mp and Vp respectively, will behave as follows:

a. e ≤ 1.6 Mp/Vp : pure shear link, yielding is dominated by shear b. 1.6 Mp/Vp ≤ e ≤ 2.6 Mp/Vp : yielding by more dominant shear and flexure c. 2.6 Mp/Vp < e ≤ 5 Mp/Vp : yielding by shear and more dominant flexure d. e ≥ 5 Mp/Vp : pure flexure link, yielding is dominated by flexure

Performance of the shear link is specified the codes (Badan Standar Nasional, 2002). It focuses on the plastic rotation angle of the link at the condition when the inelastic drift ratio of the structure is at the limit, i.e. 2% for structure with fundamental period of T > 0.7 sec and 2,5% for structure with T < 0.7 sec. Figure 1 shows the definition of the plastic rotation angle of the link and the drift ratio. To ensure the expected performance of the link due to seismic load, a number of requirement are specified in the code, e.q. the web slenderness ratio, the dimension and the spacing of the link stiffener, and the design force for the link connection.

(a) (b) Figure 1 Plastic rotation angle, (a) Split K-Brace (b) D-Brace.

Study on intermediate link by Richards and Uang (2006) showed that the severe inelastic deformation occurred at end part of the link due to higher bending moment. The fact that higher bending moment locates at link-ends also occurs at the shear-link. The damaged bolt-connected link in the study by Stratan and Dubina (2002) showed that the energy dissipation on the link-web mostly took place at the end part of the link. This was clearly shown by the plastic deformed shape (skewed) of the web at the link-end while the web at the middle part of the link showed much less significant plastic deformation. For some extent this showed that the link has not been optimally mobilized to provide maximum energy dissipation through the web yielding of the entire beam length.

In this study, an improvement was proposed to increase the energy dissipation of the shear link by preventing early fracture at the weld connection of the tensile flange. Side extended plates (SEP) are welded to each flange ends to increase the flange area and thus to reduce the tensile stress due to flexural moment on the link. This is illustrated in Figure 2 where the average tensile stress σ2 < σ1, since area of the side-extended flanges is larger than of the original flanges, L2 t > L1 t. It is expected that strengthening the flange would not significantly affect the behavior of the link prior to yielding. However, since the yielding of the shear link is concentrated at the web, the delayed fracture at the flange weld is expected to mobilize more energy dissipation through higher level of load and larger plastic deformation of the link-web. More over it is expected that the delay would mobilize more yielding of the web toward the middle part of the link.



450

3097.5

3097.597.5 97.5

450

3097.5

3097.597.5 97.5

30

97.5

30

97.597.5 97.5

450

30

97.5

30

97.597.5 97.5

450

(a) (b) Figure 2 Average tensile stress at the welded link end (a) Ordinary link (b) SEP link

3. EXPERIMENTAL WORK

Two specimens of shear link were determined as component of a three-storey eccentrically braced frame structure type split K-brace of an office-residential building designed in seismic region I according to Indonesian Seismic Code. Each link is a 450 mm link of Wide Flange 200x100x5.5x8, grade BJ-41 (or A36), as shown in Figure 3. The link–web is divided into 4 equal segments by 10-mm thick vertical intermediate stiffeners at both side of web. Flush-end type of connection is designed for both link-ends using the actual value of yield-stress and six high-strength bolts (A490) with 25-mm diameter.

(a) (b) Figure 3 Link specimen (a) Ordinary link (b) SEP link.

Specimen 1 and specimen 2 are identical, except that tapered-extended plates are welded at the sides of top and bottom flange on each end of the specimen as shown in Figure 3. The specimen 2 is then identified as Side-Extended Plate (SEP) specimen. The shear-link behavior under seismic loading is modeled by supporting one of the link-end to a fixed strong frame, while the other end is connected to the vertically moving actuator by moment resisting connection (Figure 4). The cyclic displacement load was applied with a rate of 0.02 mm/sec by the actuator mounted to a 1,000 kN loading frame. The displacement-controlled loading was applied according to the loading pattern shown in Figure 5.

P L1

tLP

11 ⋅=σ

P L2

tLP

22 ⋅=σ



450

Actuator

Link

FRAME IFRAME II

(a)

-15

-10

-5

0

5

10

15

0 1 2 3 4 5 6 7 8 9 10 11 12 13

Loading Cycle

Dis

plac

emen

t ( δ

y )

7

8

14

16

15

17

9

18

19

3 6

12

45

11 10

13 12

(b)

Figure 4 Testing set up (a) Loading frame, (b) Position of LVDT (red) and post yield strain gauge (blue).

Figure 5 Loading pattern/loading protocol.

A number of LVDT is used to monitor the vertical deflection of the link at the end link, relative vertical movement of end plate and support and bolt elongation. A number of rosette–strain gauges is mounted at the link web and other strain gauges are placed at the top and bottom flange (Figure 4). For monitoring the yielding part of the link, post-yield type of strain-gauges are used. Some limitations of instrumentation are considered.



4. DATA AND ANALYSIS

4.1 Load-Displacement

The test results showed that at the early stage of loading, the behavior of two specimens is similar as shown by the elastic part of load-displacement history in Figure 7(a). Specimen 1 and specimen 2 (SEP) indicate very similar yielding load, i.e. 207 kN and 203 kN respectively at the vertical displacement of 8.1 mm. This showed that the elastic stiffness of both specimens is similar.

As the loading increases, the yielding of the web of both specimens increased as indicated by the strain-gauge reading. The cyclic loading was continued until failure occurred on each specimen. The complete hysteretic loop for each specimen is shown in Figure 6, while the hysteretic envelope and the back-bone of loading history is presented in Figure 7.

-400

-300

-200

-100

0

100

200

300

400

-100 -80 -60 -40 -20 0 20 40 60 80 100

Displacement Δ (mm)

Loa

d (k

N)

-400

-300

-200

-100

0

100

200

300

400

-100 -80 -60 -40 -20 0 20 40 60 80 100


Load

(kN

)

(a) (b)

Figure 6 Complete hysteretic curve (a) Ordinary link (b) SEP link.

-400

-300

-200

-100

0

100

200

300

400

-100 -80 -60 -40 -20 0 20 40 60 80 100


Load

(kN

)

Spc2Spc1

400

300

200

Load (kN) 100

0 0 -100 -80 -60 -40

(a) (b)

Figure 7 (a) Back bone curve loading history (b) Hysteretic envelope.

The failure of specimen 1 occurs in the 6th loading cycle while the SEP specimen fails in the 10th loading cycle. Figures 7(a) and 7(b) show that the side extended plate increases the ultimate load as well as the ductility but only slightly improves the stiffness of the shear link. The maximum loads are 351 kN and 392 kN, the maximum displacements are 56.7-mm And 72.9-mm, while the maximum plastic rotation angles of the link are 0.126 rad and 0.162 rad respectively for specimen 1 and SEP specimen. This showed that the strength of the shear link is increased by 11% while the ductility increased by 28%.

-400 -300 -200 -100

-20 20 40 60 80 100

Displacement (mm) Spc2 Spc2Spc1 Spc1



4.2 Energy Dissipation

Table 1 and Table 2 show the energy dissipation of each cycle of each specimen. Although the energy dissipation of each loading cycle is quite similar until 6th loading cycle, the cumulative energy dissipation of SEP specimen is 184% higher that the specimen 1.

Table 1 Dissipated energy and input energy of specimen 1.

Dissipated Energy (kN.mm)

Input Energy (kN.mm)

Ratio Dissipated/Input Cycle Rotation

(rad) Each cycle Cummulative Each

cycle Cummulative Each cycle Cummulative

1 0.036 4,347 4,347 6,472 6,472 0.672 0.672 2 0.054 10,574 14,921 13,393 19,866 0.790 0.751 3 0.072 18,224 33,146 21,610 41,477 0.843 0.799 4 0.09 27,072 60,218 30,864 72,342 0.877 0.832 5 0.108 37,076 97,295 40,915 113,257 0.906 0.859 6 0.126 46,509 143,805 51,162 164,419 0.909 0.875

Table 2 Dissipated energy and input energy of specimen 2 (SEP link).

Dissipated Energy (kN.mm)

Input Energy (kN.mm)

Ratio Dissipated/Input Cycle Rotation

(rad) Each cycle Cummulative Each

cycle Cummulative Each cycle Cummulative

1 0.036 4,579 4,579 6,613 6,613 0.692 0.692 2 0.054 10,616 15,195 13,478 20,092 0.788 0.756 3 0.072 18,265 33,460 21,730 41,823 0.841 0.800 4 0.090 27,201 60,661 31,173 72,996 0.873 0.831 5 0.108 36,852 97,514 41,269 114,266 0.893 0.853 6 0.126 47,370 144,885 52,163 166,429 0.908 0.871 7 0.144 57,694 202,580 62,839 229,268 0.918 0.884 8 0.162 68,379 270,959 74,029 303,298 0.924 0.893

The significant increase in energy dissipation of SEP specimen is obviously due to the more loading cycles that can be provided by the SEP specimen. Although the back-bone curve of the load-displacement of both specimen is almost similar until the 6th loading cycle, there is no further loading cycle was mobilized by the specimen 1 after its failure in the 6th loading cycle due to fracture in the welding connection of its flanges.

4.3 SEP Link

The previous section showed that the side extended plate has significantly improved the strength, the ductility, and the energy dissipation of the shear link. By considering the fact that the hysteretic loading history of both specimen is quite similar until the specimen 1 fails, the improvement performance by the SEP is described by the delay of the fracture of flange welding connection, as expected.

Figure 8 showed that the specimen 1 fails when the fracture occurs at the welding between link-flange and the end-plates due to high tensile stress. At this condition, the large inelastic shear deformation is obviously shown in the web of the near-end segment of the link, while the middle part of the link is still in good condition without any visual damage. The failure is also shown by the inelastic buckling of the flange at the end-part of the link. This condition is also shown in previous studies (Kasai and Popov, 1986; Ramadan and Ghobarah, 1995).



(a) (b) (c)

Figure 8 Failure of specimen 1 (a) Fracture at link end upper-weld (b) Fracture at link-end bottom-weld (c) Large inelastic shear deformation only at near end segment.

The failure of the SEP specimen occurs similarly at the welding of the flange, but it occurs at almost all of the welding between flanges and the intermediate stiffeners as well as the end-plates. Figure 9 shows that the inelastic shear deformation has been mobilized in the web of all segment of the link, including the middle part of the link. It is worth to note that the buckling of the flange at the end of link is not visually indicated. This fact explains how the further loading cycles could be sustained by the SEP.

(a) (b) (c) Figure 9 Failure of specimen 2 (a) Fracture at link end and SEP weld (b) Fracture at intermediate stiffener and

web (c) Large inelastic shear deformation at web of all web segment.

The test also confirmed that the design of the bolt and the end-plate based on the actual (not the nominal) strength of the link provides a relatively good performance of bolt and end-plate that could maintain their elastic condition until very severe loading is applied to the link.

5. CONCLUSION

The experimental work in this study showed that the performance of the shear link as a seismic resisting component in the eccentrically braced frame has been improved by adding side extended plates at the ends of the link. The improvement includes strength, ductility, and energy dissipation.

The extended plate has successfully delayed the early failure of the shear link due to the fracture at the welding of the tensile flange that occurred in previous link model. It also enables the web shear yielding to spread to the entire length of the link. The failure is also spread to all welding of the flange. As the result, the capacity of the shear link has been fully mobilized to provide maximum energy dissipation as demanded.

6. ACKNOWLEDGEMENT

This research is funded by Research Grant from Institut Teknologi Bandung Year 2007 for which the authors expressed their appreciation.

7. REFERENCES

AISC (2005). “Seismic Provision for Structural Steel Building”, Chicago, American Institute of Steel Construction.

Arce, G., Okazaki, T., Ryu, H.C. and Engelhardt, M.D. (2004). ”Recent research on link performance in steel eccentrically braced frames”, 13th World Conference on Earthquake Engineering. Paper No 302. Vancouver, Canada.




Arce, G., Okazaki, T. and Engelhardt, M.D. (2005). “Experimental study of local buckling, overstrength and fracture of links in eccentricallt braced frame”, Journal of Structural Engineering, ASCE, October, 2005.

Badan Standar Nasional (2002). “Tata Cara Perencanaan Struktur Baja untuk Bangunan Gedung”, SNI-03-1729-2002.

Broderick, B. and Thomson, A. (2001). “The response of flush end plate joints under earthquake loading”, Journal of Constructional Steel Research, Elsvier, September, 2001.

Kasai, K. and Popov, E.P. (1986). ”General behaviour of WF steel shear link beams”, Journal of the Structural Division, Vol.112, No. 2: 362-382, February, ASCE.

Malley, O.M. and Popov, E.P. (1983). “Shear links in eccentrically braced frames”, Journal of the Structural Division, Vol. 110, No. 9, March, ASCE.

Moestopo, M. and Mirza, A. (2006). “Kinerja sambungan baut pada link struktur rangka baja eksentrik”, Seminar & Pameran HAKI, Agustus 2006.

Moestopo, M. and Khairullah (2003). “On improved performance of eccentrically braced frames”, 9th East Asia-Pacific Conference on Structural Engineering and Construction, Bali.

Popov, E.P. (1981). “Recent research on eccentrically braced frame”, End. Struct. Vol. 5, January, ASCE.

Richard, P.W. and Uang, C.M. (2005). “Effect of flange width-thickness ratio on eccentrically braced frame link cyclic rotation capacity”, Journal of Structural Engineering, ASCE, August 2005.

Richard, P.W. and Uang, C.M. (2006). “Testing protocol for short link in eccentrically braced frame”, Journal of Structural Engineering, ASCE, August 2006.

Ramadan, T. and Ghobarah, A. (1995). “Behaviour of bolted link-column joints in eccentrically braced frame”, Can. Journal of Civ.Eng. 745-754.

Shih, H., Khandelwal, K. and El-Tawil, S. (2006). “Ductile web fracture initiation in steel shear links”, Journal of Structural Engineering, ASCE, August 2006.

Stratan, A. and Dubina, D. (2002). “Bolted Link for Eccentrically Braced Steel Frames”, The Polytecnica of Timisoara, Romania.



PERFORMANCE OF HISTORICAL MONUMENTS IN THE 2003 BAM'S EARTHQUAKE

T. Mahdi1

1Department of Structural Engineering, BHRC, Tehran, Iran Email: [email protected]

ABSTRACT: On December 26, 2003, an earthquake with magnitude of Ms 6.6 hit the city of Bam, southeastern of Iran. Many buildings in the city have been hit hard in this earthquake. Among the many historical monuments in the area, the citadel of Arg-e-Bam is the most oldest and important one. It is situated at the northeastern skirt of the city at very short distance from the earthquake epicenter. The monument was constructed mainly from mud brick. Its construction goes back to more than 2000 years and it was repaired many times in the past. However, and due to the absence of proper maintenance and strengthening methods, the quality of most of its buildings was quite low. Generally, it can be said that most historical buildings in Bam had lacked the required stability before the earthquake. Based on this general remark, it was quite amazing to observe that some of the tall and thick walls in Arg-e-Bam have survived the earthquake. However, other thinner walls inside and some towers on the skirt of the fortress had failed badly. Moreover, most roofs of these historical buildings are very thick, and this led to increase the lateral forces on these buildings. In general, all historical buildings have suffered some type or another of cracks and many of them have completely collapsed. In this paper, a general review on the performance of these buildings has been made with some detailed discussion on reasons that led to their failures.

1. INTRODUCTION

On December 26, 2003 at 01:56:56 UTC equal to 5:26:26 local time, an earthquake hit the city of Bam in Kerman province, southeastern of Iran. The town lies to the east of the Gowk fault on which four large earthquakes have occurred on 1981, 1989 and 1998 (Zare, 2004). Because most buildings in the citadel of Arg-e-Bam go back to the Safavid period (1491–1722), many were under the impression that no earthquake occurred in Bam for around five hundred years. However, some collapsed walls in the citadel revealed repairs that may have been made after previous small earthquakes (EERI, 2004).

USGS reported that magnitude of this earthquake was Mw 6.6 and its hypocenter was located at 29.00N, 58.34E (USGS, 2003). This implies that the epicenter of the earthquake was only a short distance from the city of Bam. The focal depth of the earthquake was reported to be about 7 km. (BHRC, 2003) to 10 km. (USGS, 2003). At first, it was supposed that the main shock had occurred in the geological Bam fault, which was well known before. However, and according to the seismological investigations carried out by a group of researchers (Nakamura et al., 2005) based on the analysis of hypocenters of the aftershocks and referring SAR results (Talebian et al., 2004), it has been found that most of epicenters of aftershocks are not located on the old fault itself but are distributed along a line parallel and about 3.5 km. to the west of the old fault. It is also concluded that the newly founded fault has a nearly north-south direction with three branching fault sections at its northern part. Comparing the heavily damaged area with the projection of this fault, it has been found that these areas are located just or nearly on the three fault branches at the northern part of the fault (Nakamura et al., 2005).

2. HISTORICAL BUILDINGS IN BAM

Many buildings in Bam have been hit hard in this earthquake. In this area, the traditional buildings account for more than 30 percent of the total number. They are mostly made of sun-dried mud bricks. Most historical buildings were also made of such bricks. The most important ancient monument in this area is the citadel of Arg-e-Bam. The monument consists of four main parts,



high walls, 38 towers and a ditch. It is situated at the northeastern skirt of the city at very short distance from the hypocenter of the earthquake and few meters from one of the three northern branches of the newly founded fault. The central part of the monument is situated on solid gypsum sandy rocks. The total area of this adobe masonry fortress is around 200000 square meters. Accordingly, this complex was the largest of its kind in the world. Its construction goes back to more than 2000 years and it was repaired many times in the past. This fortress had been used for residential and commercial purposes with some strong military facilities. In modern times, its residential and commercial functions had diminished gradually and seized to exist around the beginning of the 19th century. However, a small barrack remained inside till 80 years ago.

Among other historical buildings in the earthquake-stricken area, a two houses known as Khaneh Arsham and Khaneh Ameri. These two houses had been constructed in the beginning of the 19th century in the suburb of the city. Although they are located around 3.5 Kilometer from the epicenter, a noticeable damage has occurred in both buildings.

Another important monument located in the city centre is Qaisaria Zardishtiha. This monument consists of a marketplace, a public bath, a caravansary and a mosque. It has been constructed around 140 years ago and it has been heavily damaged in the earthquake.

3. MODES OF FAILURE IN HISTORICAL WALLS

Many types of failure have been observed in historical walls especially in the free-standing high walls surrounding Arg-e-Bam. The most dominant modes of failure were the in plane (shear) and the out-of-plane (flexural) ones. The primary cause of the most often observed damage is the low tensile strength of adobe walls. As shown in Figure 1, forces parallel to the plane of the wall cause diagonal cracking in the short end walls. On the other hand, forces perpendicular to the plane of the wall, cause flexural stresses and cracking that had led sometimes to complete failures. In historical buildings other than the free-standing walls, the relatively large displacements and the combination of flexural and tensile stresses had led to the failure of corners. A fourth mode of failure had been observed at the historical Arge-Bam Citadel. Most of the walls found in this ancient monument, consist of many wythes that have been constructed at different historical periods. Accordingly, and because of the weak coherence of these layers, peeling of wall wythes has been observed as shown in Figure 2.

Adobe walls with higher quality have also been observed in the historical buildings of Bam. Many of these walls survived the turmoil and shown better performance than the traditional plain sun-dried brick walls. The most important techniques used in these buildings are the followings:

1. The use of straw to improve the tensile capacity of adobe.

2. The use of well-burnt, highly-dense adobe bricks in the walls.

3. The use of some types of woven mats as a bed reinforcements between brick layers.

4. The use of wood beams inside adobe walls to improve and increase their stabilities as shown in Figure 3. This wall has survived the original turmoil, but failed partially in another minor quake six month after the original one.

5. The use of cob blocks in the walls.



Figure 1 Diagonal cracking in the high wall and one of the towers in Arge-e-Bam.

Figure 2 Peeling of wall layers in Arge-e-Bam Figures or Illustrations.

4. ADOBE WALLS WITH CURVED ROOFS

Among the many types of roofs used in Iranian traditional buildings, curved roofs were among the most popular ones until the beginning of the 20th century. With the exception of small number, the majority of historical buildings inside and outside Arg-e-Bam have curved roofs.

4.1 The Arch

The arch can have many architectural and structural functions. In the traditional Iranian architectural system, arches are used to face the central courtyard of the building, as shown in Figure 4. The small arches shown in this figure is merely a facade that is hiding the real structure behind it and merely worked as a shear wall. For the arches perpendicular to this direction; and due to forces perpendicular to the plane of the wall, a complete separation between the facade wall and original structure had occurred. This has led to partial failures of these walls.

Figure 3 Using wood inside adobe walls.



Figure 4 Arches in the military barrack in Arge-e-Bam.

4.2 The Barrel Vault

The number of shapes and forms of barrel vaults used in the Bam’s area is quite large. Good performance of such roofs is expected when forces applied parallel to their generators. In Bam's earthquake, many barrel vaults were severely damaged or completely destroyed in the earthquake, e.g. see Figure 5. However, others had shown good behavior even when the direction of the movement was perpendicular to their generators. Figure 6 shows another example of such roofs. It can be seen from this figure that bearing walls had minor cracks compared to those appeared in the roof. The pattern of cracks in the roof indicates clearly that some vertical forces were applied to it. That can be explained by the high vertical acceleration component of this earthquake. Furthermore, some new masonry buildings in the city of Bam had been constructed using barrel vaults as their roofs. These roofs were connected well with the wall by ring beams. With the exception of minor failures and some cracks, these roofs performed satisfactorily.

Figure 5 Barrel vaults subjected to forces parallel to their generators at Khane Arsham, 3.5 Km. from the

epicenter.

Figure 6 A barrel vault in Arge-e-Bam.



4.3 The Dome

Domes are one of the main features of the city of Bam. They can be seen in different sizes and shapes all around the city and inside the historical monument of Arg-e-Bam. The methods used to support these domes and transfer their loads to bearing walls are quite different from one place to another. The simplest method used to support small domes is by placing them directly on bearing walls. As it is well known, the existence of openings like doors and windows in these walls destabilizes the dome structure. To overcome this problem, early Iranian architects used arches to transfer stresses from the dome to either side of the opening as shown in Figure 7. This was the start of development of the dome on four arches that later took more complicated forms.

In the majority of cases, complete destructions of adobe buildings including their domes were observed. However, example of good performance of domes can be also seen as shown in Figure 8. The building shown in the figure is located at Konari neighbourhood in Arg-e-Bam. With the exception of some cracks at the Intersection of perpendicular walls, no other failure has been observed. On the other hand, many examples of massive and wide destruction had been observed in the stable courtyard few hundred meters from Konari neighborhood. An example of such destruction is shown in Figure 9.

Figure 7 Using arch panels to support a dome in Arge-e-Bam.

Figure 8 Domes located in Konari neighborhood in Arge-e-Bam.



Figure 9 Heavy losses in the stable courtyard of Arge-e-Bam.

The stable courtyard inside Arg-e-Bam has a large number of domes. As shown in Figure 9, the majority of domes were destroyed completely. Supporting walls were also destroyed partially or completely. However, it is interesting to note that parts of these buildings that are located in the right corner of Figure 9 had survived the turmoil by taking advantage of their adjacency to two relatively rigid walls. In the author’s opinion, the reaction applied by these two walls during the earthquake had increased the capacity of these parts to resist vertical forces applied by the vertical acceleration component. Prominent among the surviving domes in the stable courtyard, are those shown earlier in Figure 7. In this case, the dome is supported on two arches and two walls. The forces applied were perpendicular to the arched panels and parallel to the wall. Arches in this case had fulfilled the role of bracings in inactive walls and distributed the loads between active walls to prevent partial failure. As it is clear from Figure 7, one of the arches had suffered minor failures. This failure reflects the disability of the arch panel to transfer forces between active walls.

In most cases observed in historical buildings, it was found that the transition regions from the square below (walls) to the circle above (dome) were not well defined. As a result, approximately all surviving domes had sustained some cracks. The typical pattern of the cracks observed in most cases is that of a meridian one. Such a crack had been initiated at points closer to the intersection line of dome and flat roof.

5. DISCUSSION AND CONCLUSION

The study of the response of historical buildings during Bam’s earthquake showed that the seismic performance of such buildings was a function of many factors. The main factors can be summarized in the following points: height to thickness ratio of the adobe wall, spans of internal subdivisions, size of openings, roof masses, nature of continuity with adjacent buildings, quality of construction, distance to fault, and topographical and site effects. In the survived historical buildings, it was found that most of the factors mentioned above had been implemented in such a way that tensile forces in adobe were eliminated.

In general, Bam's historical buildings had sustained severe damages worse than that normally expected in such moderate earthquake. However, this was true for all other types of buildings. In such circumstances, it was quite amazing to observe that some of these buildings and tall walls have survived the earthquake. It can be concluded that the low tensile strength of the adobe material and mortar, combined with near fault effect (high vertical acceleration and high pulse) were among the main causes of such wide spread damages. Moreover, and due to the absence of proper maintenance and strengthening methods, the quality of most of these buildings was quite low.

The roles of arches and domes in transferring horizontal forces have been discussed in this paper. As it is shown, the proper choice of structural elements capable of transferring these loads efficiently is quite important to the survival of these buildings.




6. REFERENCES

Building and Housing Research Centre (2003). "The Very Urgent Preliminary Report on Bam Earthquake of Dec. 26-2003".

Earthquake Engineering Research Institute (2004). "Preliminary Observations on the Bam, Iran, Earthquake of December 26, 2003".

Nakamura, T., Suzuki, S., Sadeghi, H., Fatemi Aghda, S.M., Matsushima, T., Ito, Y., Hosseini, S.K., Jafar Gandomi, A.J. and Maleki, M. (2005). "Source fault structure of the 2003 Bam earthquake, southeastern Iran, inferred from the aftershock distribution and its relation to the heavily damaged area: existence of the Arg-e-Bam fault proposed", Geophysical Research Letters, Vol. 32, L09308.

Talebian, M. et al. (2004). "The 2003 bam (Iran) earthquake: rupture of a blind strike-slip fault", Geophys. Res. Lett., vol. 31, L11611.

USGS Web Site. "http://neic.usgs.gov/neis/eqlists/significant.html". Zare, M. (2004). "Bam (SE Iran) earthquake of 26 December 2003, Mw6.5: a preliminary

seismological overview", Orfeus Newsletter, Vol.6, No. 1.


http://neic.usgs.gov/neis/eqlists/significant.html


USE OF CFT BRACES FOR SEISMIC RETROFIT OF BUILDING STRUCTURES

Susumu Kono1, Mamoru Oda1 and Fumio Watanabe1

1 Department of Architecture, Kyoto University, Kyoto, Japan


ABSTRACT: This study introduces a new brace system using concrete filled tube (CFT) to enhance its deformation capability. A half scale portal frame was constructed based on the old Japanese building standard and strengthened with two types of CFT brace without using rebar nor bolt anchorage. The specimens showed two to three times larger lateral load carrying capacity than that of the bear frame and large enhancement in ductility. The CFT brace was a single line element, easily handled to place in a frame, and allows the ductile design procedure due to its deformation capability.

1. INTRODUCTION

The seismic upgrading of existing buildings has been attracting more attention than ever. Upgrading of seismic performance of buildings may be made possible by increasing strength or ductility but many existing buildings in Japan have not had seismic upgrading. It is because that retrofit construction is normally costly due to intensive labor work and long suspension of service. The authors have been developing a simple seismic retrofit method so that a proposed retrofit method has short construction period and low construction cost by not using rebar or bolt anchorage. For this purpose, an X-shaped precast prestressed concrete brace system was proposed in 2001 at Kyoto University (Watanabe et al, 2004). In this paper, a new brace system made of concrete filled tube (CFT) is introduced for better seismic performance and easy construction.

The original X-shaped precast prestressed concrete brace consisted of four precast units as shown Figure 1(a). The brace was assembled at a construction site and prestressing force was introduced to two lower legs. Spacings between brace ends and frame corners were filled with high strength no-shrinkage mortar. The prestressing force was released after hardening of mortar for the X-shaped brace to extend by itself and fix to a boundary frame. When a frame with an X-shaped brace is subjected to lateral seismic load, only one of diagonal members works effectively. The remaining diagonal member becomes free and the tensile diagonal member may come off from the surrounding frame. To prevent the free member to come out, a flat spring and steel pipe (FSSP) in Figure 3(c) is installed at the bottom end of each diagonal member. This device maintains a certain amount of compressive force in the diagonal member even if the brace experiences elongation under reversal seismic events.

Figure 1(b) shows the lateral load – drift relations of the braced frame and the original unbraced frame. The braced frame had three times as large lateral load capacity as the original unbraced frame. Since the peak load was reached by the compressive crushing of the brace, the postpeak behavior was brittle and the load dropped suddenly. Because of this brittle failure mode, the required lateral load carrying capacity in design has to be set higher than that of the ductile frame.



300300 2100 300 300

P restressing bar- 32 to sim ulatethe long term axial loading.

Prestressing bar- 13 2 toprevent beam yielding in tension

×

4@ 100

4@ 70

4@ 50

2075

1425

325

325P 1

P 2Beam section 275x325

C olum n section300x300

B race section100x120 for N o. 1120x150 for N o. 2

-600

-400

-200

0

200

400

-0.4 -0.2 0 0.2 0.4 0.6

Braced frameOriginal frameAnalysis

Loading point drift (%) (a) Specimen tested in 2001 (b) Load-displacement relation

Figure 1 Proposed retrofit method and the experimental result (Watanabe et al, 2007).

In this paper, a new brace using concrete filled tube (CFT) is proposed as shown in Figure 3. It was expected that the ductility of the braced frame increases so that the required lateral load carrying capacity in design can be decreased. The configuration was changed to a single diagonal line element instead of X-shape so that the brace aesthetically looks better and an opening can be made. Since a single diagonal brace can resist against force in one direction but does not in opposite direction, it may be necessary to place brace for the opposite direction at some other spans. However, placing simple line elements instead of X-shaped elements saves construction time and labor. In addition, the steel tube of CFT may be divided into several pieces, brought to the construction site using existing elevators, assembled at site, and placed in the existing frame. Then the inside of the steel tube is filled with grout mortar. In this way it is possible to exclude heavy construction equipment.

The way prestressing was also modified. In 2001, prestressing force was applied with prestressing rods embedded in the X-shaped assemblage. Since the one end of the rod was anchored at the central hollow concrete circle, prestressing action was not very effective and it was impossible to take out prestressing rods after the construction. In a recent system, the prestressing force is applied to the whole length of the brace using external prestressing rods. Rods can be reused and the brace itself can be moved to other locations by reapplying the prestressing force if necessary.

2. TEST SETUP

2.1 Specimen Configurations and Test Variables

One reinforced concrete portal frame, as shown in Figure 2, was constructed as a sub-assemblage of a four-story reinforced concrete school building, which was designed following the pre-1980 Japanese Building Standard (Architectural Institute of Japan, 1950). Since the loading was basically one directional and the surrounding frame was reused in other direction. CFT-S60 was tested first as shown in Figure 3(a), the damaged brace was taken out and a new brace was installed in other direction, then the specimen was tested as CFT-M18, as shown in Figure 3(b) (Oda, 2008). In both cases, the brace was installed in the reinforced concrete portal frame with a FSSP (flat spring and steel pipe) device in Figure 3(c) so that the brace would not come out of the portal frame when experiencing elongation. If the FSSP device is compressed by more than 32kN, the flat spring contracted inside the steel pipe and the steel pipe practically carried the whole axial load. The sectional dimensions of the two CFT braces are shown Figure 3(d) and (e). Mechanical properties of materials are shown in Table 1.

Before applying the horizontal load, prestressing force of 450kN (0.28f’cbD) and 324kN (0.20f’cbD) was applied to the beam and each column, respectively, with internal unbonded prestressing steel bars. The column axial force corresponds to the long-term axial force of the first story column. Beam was prestressed to avoid the tension failure of the beam. The lateral load carrying capacity of the reinforced concrete portal frame without the brace was 190kN. By installing the brace, CFT-S60 and CFT-M18 were designed to have 357kN and 343kN of lateral



load capacity at which the brace was to buckle. At this ultimate stage, the shear capacities of the beam, the columns, and column-beam joints were computed and shown in Table 2. The table also shows their shear capacities based on Architectural Institute of Japan (1999), and prestressing forces introduced to the brace. The lateral loads with equal magnitude were applied at either end of the beam controlling the drift angle as shown in Figure 4. Enforced positive displacement was five times larger than negative displacement so that the surrounding reinforced concrete portal frame did not suffer too much damage in the negative loading.

300 300 2100 300 300

24003300

325

325

1425

2075

φ4 100@

φ4 70@

16 D10-

8 D13-

Two PC bars prevent excessiveaxial elongation

to

One PC barintroduce axial force

to

Foundation 325x1000: 16 D10- φ PC bar23

275

325

8 D13-

φ sheath32

300

300

50 sheath

16 D10-

φ4 70@

φ 40 PC bar

φ4 100@ (a) Frame (b) Column (c) Beam

Figure 2 Reinforcing bar arrangement of the portal frame.

Gap was filled withno shrinkage grout mortal

-

FSSP device was placedat the bottom

.

Two pieces of steel tubewere connected withplates and bolts

.

A single piece ofwas used

. steel tube

Surrounding frame was reusedafter testing CFT S60

- .

(a) CFT-S60 (b) CFT-M18

Flat spring

Steel pipe

A set of flat springs

Bearing plate Bearing plate

Min

imum

leng

th

Free

leng

th

7 5

75

t 1 6= .

Inside is filledwith paste

.　

1 00

100

t 2 3= .

Inside is filledwith paste

.　

(c) Flat spring and steel pipe (FSSP) (d) Section of CFT-S60 brace (e) Section of CFT-M18 brace

Figure 3 Specimen configurations.



Table 1 Mechanical properties of materials.

(a) Concrete (b) Steel

Specim en M em berC om pressive strength(M Pa)

Tensile strength(M Pa)

Young'sm odulus(G Pa)

Fram e 29.6 2.91 25.1

Brace 62.7 4.04 20.8Fram e 31.4 2.32 26.3

Brace 12.6 1.30 10.2C FT-M 18

C FT-S60

TypeYield

strength(MPa)

Tensilestrength(MPa)

Young'smodulus(GPa)

D13 358 512 181

D10 371 527 179D6 415 531 179φ4 522 579 202

Steel pipe 338 448 221

Table 2 Ultimate shear capacities of each member and prestressing force introduced to the brace.

Shearcapacity, Qu*1

(kN)

Design shear force atbrace buckling, Qr

(kN)

Qu/Qr(%)

Prestressingforce of brace

(kN)

Column 86.6 49.0 177

Beam 67.7 36.3 187

Joint*2 444 182 244

Column 86.6 43.5 199

Beam 67.7 32.0 212

Joint*2 444 155 286

CFT-M18

CFT-S60

Specimen

32.6

42.1

*1: Capacity was computed based on Architectural Institute of Japan (1999) *2: Joint stands for the column-beam joint.

1200kNhydraulicjack

500kNhydraulic jack

500kNhydraulic jack

Loadcell

Loadcell

Load cell Load cell

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

0 1 2 3 4 5 6 7 8 9 10 11 12Cycle

Drift

(%)

Figure 4 Loading system and loading protocol (CFT-S60).

3. TEST RESULTS

Load – drift angle relations are shown in Figure 5. A solid line represents the total load and a lightly shaded line represents the sum of shear forces of two columns. Shear force of columns was obtained by subtracting lateral load contribution of the brace from the total lateral load. Lateral load contribution of the brace was the horizontal component of the axial force which is computed in Figure 6. In CFT-S60, the initial stiffness changed when cracking was observed at the north column-beam joint and the beam ends at R=0.2%. The load carrying capacity still increased until the yielding of the beam at R=0.6%. The number of cracks increased from R=0.4% at the column-beam joint but the maximum crack width was less than 0.1 mm. The axial force of the brace reached the maximum value at R=0.4% and buckling started resulting in the second stiffness change. Total lateral load reached the peak at R=0.6% when buckling deformation of the brace was visually observed. Load carrying capacity decreased gradually after the peak but brittle failure was not observed. Load–drift relation of CFT-M18 up to R=0.4% does not differ from that of CFT-S60 and the damage of surrounding frame experienced from loading CFT-S60 cannot be seen. After reaching the peak load at R=1.0%, the load kept constant until R=3.5% which is not shown



in the figure. At R=1.6%, the brace started buckling locally at the upper end probably due to incomplete grouting. After the cyclic loading at R=2.0%, the load was monotonically increased up to R=6.5% at which the load was terminated due to the limitation of measuring system. The load during the pushover loading became maximum, 590 kN, at R=3.5% and 421 kN at R=6.5%. The load–displacement relation of CFT-M18 was very ductile. It is probably appropriate to consider the enhanced ductility of CFT brace in a design procedure.

Axial compressive force of the brace–drift relations are shown in Figure 6. Axial force was computed from strain gages on four faces of the brace and stress-strain relations obtained from the material test. In the figures, Nb represents the axial forces at brace buckling and Nu represents the axial compressive capacity for short columns. Both braces showed the relatively stable and ductile behavior. If the brace of CFT-S60 had not had connections at the midspan, the axial capacity of the brace would not have degraded as Figure 6(a) and have been similar to Figure 6(b). CFT-M18 was stocky and might have reached Nu if CFT was grouted properly at the upper end.

Figure 7 shows the observed damage of two specimens at drift angle R=0.8% and photos of braces at the failure. The reinforced concrete portal frame of CFT-S60 had some minor flexural cracks and the specimen failed due to flexural buckling of the brace because the brace of CFT-S60 consisted of two pieces. CFT-M18 showed the initial cracking at R=0.4%. The yielding of CFT steel started at R=0.6%. At the same time, the CFT started to penetrate the bearing region of column-beam joint and the cracking at this location widened. At R=1.6%, the brace of CFT-M18 start buckling locally at the upper end probably due to the incomplete grouting. The deformation of the brace concentrated at this location until the end of the test.

-200

-100

0

100

200

300

400

500

600

-0.2 0 0.2 0.4 0.6 0.8 1Drift （%）

Lateral load （kN

）

Fram e and brace

Surrounding fram e

Q b=357kN

C ollapse m echanismform ation of thesurrounding fram e：190kN

C FT-M 18

-200

-100

0

100

200

300

400

500

600

-0.5 0 0.5 1 1.5 2 2.5Drift （%）

Lateral load （kN

）

Fram e and brace

Surrounding fram e

Q b=343kN

C ollapse m echanismform ation of thesurrounding fram e：190kN

(a) CFT-S60 (b) CFT-M18

Figure 5 Load – drift relations (Qb represents the computed lateral load at brace buckling).

0

50

100

150

200

250

300

350

400

450

-0.2 0 0.2 0.4 0.6 0.8 1Drift （%）

Brace axial force （kN

）

N b=307kN

N u=499kN

0

50

100

150

200

250

300

350

400

450

-0.5 0 0.5 1 1.5 2 2.5Drift （%）

Bra

ce a

xial fo

rce

（kN）

N b=300kN

N u=435kN

(a) CFT-S60 (b) CFT-M18 Figure 6 Axial compressive force of the brace – drift relations.



Buckling causedat midspan

- - .

out of planedeformation

North(-) South +( ) North(+) South(-)

Local b occurred .uckling

(a) CFT-S60 (b) CFT-M18 (Frame of CFT-S60 was reused.)

Local buckling

Buckled at connection

(c) CFT-S60 (d) CFT-M18

Figure 7 Observed damage at R=0.8%.

4. CONCLUSIONS

The experiment on reinforced concrete portal frame strengthened with CFT braces was carried out. It was experimentally shown that the CFT brace system can easily retrofit existing reinforced concrete frames without using rebar or bolt anchorage, possibly leading to short construction period and low cost. It was also demonstrated that the brace can be easily replaced or moved to other locations. Both specimens failed due to buckling of the brace but the lateral load – drift relation was ductile for CFT-S60 and extremely ductile for CFT-M18 even after buckling took place. CFT was very easy to handle and construct. The CFT bracing system has a good potential to allow the ductile design procedure due to its enhanced ductility.

5. ACKNOWLEDGEMENTS

A part of this research was financially supported by Japanese Ministry of Land, Infrastructure and Transport (PI, F. Watanabe), Kajima Research Fund (PI, S. Kono). The authors acknowledge Mr. Y. Kimura, a former student at Kyoto University for his contribution to the experimental work. It is also noted that many technical suggestions were given by Takenaka Co., Daiwa Precast Concrete Co., and Nagai Design Office.

6. REFERENCES

Architectural Institute of Japan (1950). “Japanese Building Standard”. Architectural Institute of Japan (1999). “Design Guidelines for Earthquake Resistant Reinforced

Concrete Buildings Based on Inelastic Displacement Concept”. Oda, M. (2008). “Seismic Retrofit of RC Structures Using Revised Precast Prestressed Braces”.

Master Thesis Submitted to the Dept. of Architecture and Architectural Engineering, Kyoto University, March 2008. (In Japanese).

Watanabe, F., Miyazaki, S., Tani, M. and Kono, S. (2004). “Seismic strengthening using precast




prestressed concrete braces”, The 13th World Conference on Earthquake Engineering, Vancouver, August, Paper 3406.



CONVEX MODEL ANALYSIS FOR SEISMIC DESIGN OF STRUCTURES

Irawan Tani1 and Louis Jezequel2

1 PT. GHD Indonesia (Engineering Consultant), Civil Engineering Department, Trisakti University

Email: [email protected] and [email protected] 2 Ecole Centrale de Lyon, France, 36 av. Guy de Collongue, Ecully

ABSTRACT: A convex model is used to estimate the maximum response of structural systems subjected to uncertain excitation. The used of convex model as a non-probabilistic representation of uncertainty is convenient when only limited amount of information is available. In the case of seismic design, the convex analysis is performed on the basis of bounded energy excitation. In this paper, a convex model is presented with assumption of known bounded seismic excitation energy. Some numerical examples are employed to give some clearances ideas of this method.

1. INTRODUCTION

Perhaps no other discipline within engineering has to deal with as much uncertainty as the field of earthquake engineering (DerKieureghian, 1996). The randomness in the occurrence of earthquake in time and space, the vast uncertainty in predicting intensities of resulting ground motion, and our inability to accurately assess capacities of structures under cyclic loading all compel us to make use of statistical method in order to consistently account for the underlying uncertainties and make quantitative assessment of structural safety. While the need for probabilistic method in earthquake engineering has long been recognized, their use required some information about their density function and other parameters. The question arises is how if only limited amount information available. This paper aims to present one method for predicting maximum response of structure due to seismic excitation while there is only limited information available.

Various researchers have developed method which estimates the maximum response of structural systems to earthquake excitations. Some literatures (Drenick, 1973) proposed the idea that highly variable but limited deterministic information characterizing the ground motion could be used to find the least favorable response of a structural system to an earthquake. The total energy which the earthquake is likely to develop at the structure’s location was used. A critical excitation was sought within the set of allowable dynamic force which maximizes the structural response. However, since only total energy had to be bounded, the result was rather conservative.

Shinozuka (1970) has suggested characterizing the earthquake excitation by specifying an envelope of the Fourier amplitude spectrum. This measure yields much less conservative results than Drenick’s since the information embodied in the measure constrains the set of the possible excitations considerably; as a result the maximum structural response magnitude is reduced. In addition to the above method for estimating the maximum structural response in an earthquake, statistical method as well as probabilistic method has been proposed.

Ground motion involve large variability due to source effects associated with the rupture process, path effects related to wave propagation between a source and a site, and site effects due to soil conditions and topography. The magnitude, intensity, duration, envelope of the amplitude spectrum, and epicenter distance of an earthquake are uncertain parameters that are usually unavailable for design and make the task of estimating the maximum structural response very difficult. A recent study of a large database containing 1500 records of earthquake ground motion (Naeim et al., 1993) has shown that instead of using elastic and inelastic spectra for design, which do not contain the effect if strong motion duration, it may be realistic to use input and hysteretic energy spectra.





Recently; Ben-Haim and Elishakoff (1990) used convex models to represent uncertain dynamic excitation. Convex models provide a non-probabilistic representation of uncertainty. They are especially applicable when only a limited amount of information is available, as in the case of earthquake excitations. The uncertainty in the excitation is represented as a bound on the Fourier coefficient of the expansion excitation, as a bound in the energy of excitation or an envelope function with upper and lower limits. The concept of convex modeling on the basis of an upper bound was used along with linear programming to obtain the least favorable structural response (Ben-Haim et al., 1996).

2. CONVEX ANALYSIS MODEL

The question arises in this decade: how to deal with uncertainty? Since one of the meanings of the uncertainty is randomness, a natural answer to this question was and is to apply the theory of probability and random function. Indeed, probabilistic structural mechanics has achieved a high degree of sophistication. The power of probabilistic methods has been demonstrated beyond doubt in numerous publications (Yuen et al., 2002). Probabilistic modeling requires extensive knowledge of the random variables or function involved. It leads to evaluation of the probability of successful performance by the structure, called reliability (Ellingwood, 2001).

Indeed, the probabilistic approach is not the only way one could deal with uncertainty (Ben-Haim, 1994). The indeterminacy about the uncertain variables involved could be stated in terms of these variables belonging to certain sets, such as:

1. The uncertain parameters x is bounded, x a≤

2. The uncertain function has envelope bounds,

( ) ( ) ( )lower upperx t x t x t≤ ≤ (1)

where ( )lowerx t and ( )upperx t are deterministic functions which delimit the range of variation of

( )x t .

3. The uncertain function has an integral square bound:

( )2x t dt a∞

−∞

≤∫ (2)

Ben Haim and Elishakoff (1990) stated that instead of precise information on the probability content of random events, it is possible to posse imperfect, scanty knowledge on the uncertain quantities. The description above of uncertainty is called a set-theoretic, non-probabilistic one.

This set-theoretic approach is independently and almost simultaneously proposed by Drenick (1973), namely modeling of the response of structures to earthquake excitation.

3. CONVEX METHOD IN SEISMIC DESIGN CALCULATION

The problem of the earthquake excitation has attracted many investigations. Much of this effort has utilized a probabilistic approach (Shinozuka, 1970), modeling the earthquake excitation as a random process. Another approach, the so-called critical excitation method, is pioneered by Drenick (1973). The main idea of his work is to use highly reliable but limited deterministic information characterizing the ground motion. This function is written as a constraint:

( )( ) 2gF X t M≤&& (3)

where ( )gX t&& is some characteristic of the earthquake excitation, F( ) is a functional and 2M is a positive constant. This is supplemented by a governing differential equation:



( )( ) ( )( )gP x t Q X t= && (4)

where P( ) and Q( ) are differential operators. One then looks for an excitation function ( )gX t&& which maximizes the response on the set of allowable excitation.

Information on strong ground motion is rather scant. Basically, earthquake can be characterized by the following quantities:

1. Duration of the strong ground motion.

2. Ground motion intensity.

3. Envelope of the amplitude spectrum.

4. Effects of the macro-zone and the micro-zone.

5. Effect of local distance.

Drenick (1973) suggest adding another item, namely bounds on the peak ground acceleration and velocity. Various constraints on the ground motion appear in the literature (Ahmadi, 1986).

4. GLOBAL ENERGY BOUND FOR CONVEX MODELING

A convex method is a method of quantifying uncertainty, in this case the uncertain nature of earthquakes, without resorting to probabilistic concepts; instead, the uncertainty is characterized by a set of functions with common global characteristics. The application of convex models to represent uncertainty is well suited in situations where only a limited amount of information is available, which is exactly the case for structural system subjected to uncertain excitation such as earthquake. The convex model constraints the uncertainty, inherent in earthquake excitation, within a bound which is defined in terms of a function of bounded energy of the earthquake.

The equation of motion for a multiple degree of freedom (MDOF) structure is used in its normalized form with respect to the mass matrix. Let the natural frequencies of the structure be

1, , Nω ωL , the corresponding mode shape be 1, , Nφ φK , and the corresponding modal matrix be

[ ]1, , Nφ φΦ = K , where is the number of modes. N

The equation of motion of one structure subjected to ground motion can be characterized:

( ) ( ) ( ) ( )gMx t Cx t Kx t M X t+ + = Θ &&&& & (5)

The normalized properties with respect to the mass matrix can be express as: T M IΦ Φ = (6)

21

2

0

0

T

N

Kω

ω

⎡ ⎤⎢ ⎥Φ Φ = ⎢⎢ ⎥⎣ ⎦

O ⎥ (7)

1 12 0

0 2

T

N N

Cω ξ

ω ξ

⎡ ⎤⎢Φ Φ = ⎢⎢ ⎥⎣ ⎦

O ⎥⎥ (8)

where I is a N N× identity matrix, and iξ is the th mode damping ratio of the structure; the equation of motion can be expressed for each mode as:

i



( )22 Ti i i i i i i gy y y Xξ ω ω φ+ + = Θ &&&& & t (9)

where is the th mode displacement response in normal coordinates, Θ is a vector of ones

and

( )iy t

( )g

i

X t&& is the horizontal ground acceleration.

Using Duhamel’s integral and assuming zero initial condition, the response of the structure in the th normal mode is: i

( ) ( ) ( ) ( )( ,, 0

1 sini i

ttT

i i g D iD i

)y t X e tξ ω τ dφ τ ωω

− −= Θ −∫ && τ τ (10)

2, 1D i i iω ω ξ= − (11)

The i th modal velocity and acceleration in normal coordinated for zero are then obtained from the following superposition equations:

( ) ( )x t y= Φ t ( ) ( ); x t y= Φ& & t ; ( ) ( )x t y= Φ&& && t (12)

In what follows, the ground acceleration ( )gX t&& is assumed to belong to a convex set bounded by a global energy bound model as proposed by Pantelides and Tzan (1996). The values of displacement, velocity and acceleration are evaluated based in the convex set of admissible ground acceleration. For the global energy bound (GEB) convex model, this set can be expressed as:

( ) ( ) ( )2

0

:t

GEB g g GEBS X t X d E tτ τ⎧ ⎫

⎡ ⎤= ⎨ ⎣ ⎦⎩ ⎭

∫&& && ≤ ⎬ (13)

The global energy bound, , varies with time; however, as time goes to infinity the bound

reaches a finite value E which is larger than any

( )2GEBE t

( )2GEB ∞ ( )2

GEBE t . Note that time is equal to

infinity at the end of the earthquake duration, i.e. ft∞ t= , where ft is the earthquake duration. The peak value of modal displacement, velocity and acceleration can be found using the theory of convex models.

For the global energy bound (GEB) convex model, the set of admissible ground acceleration is given in Eq. (13): the maximum displacement can be obtained using Duhamel’s integration for the

th normal mode as: i

( )( )

( ) ( ) ( )( ), ,, 0

sinmax i i

g GEB

tTti

GEB i g D iX t S D i

y t X e tξ ω τφ dτ ω τω

− −

∈

⎧ ⎫Θ⎪ ⎪= −⎨ ⎬⎪ ⎪⎩ ⎭

∫&&

&& τ (14)

Since is linear function of the ground acceleration, which is assumed to belong to the convex set of Eq. (13), the maximum displacement Eq. (14) occurs on the set of extreme points of the set . According to the Cauchy-Schwarz for arbitrary function

( ),GEB iy t

GEBS ( )1f t and ( )2f t

( ) ( ) ( ) ( )2

2 21 2 1 2

0 0 0

t t t

f f d f d f dτ τ τ τ τ τ τ⎛ ⎞ ⎛ ⎞⎛ ⎞

≤⎜ ⎟ ⎜ ⎟⎜⎝ ⎠ ⎝ ⎠⎝ ⎠∫ ∫ ∫ ⎟ (15)



with equality occurring only if ( )1f τ and ( )2f τ are proportional. The maximizing profile on

the interval [ ]0, t is from Eq. (13),

( ) ( ) ( )(1 sini i tD if e tξ ω τ ),τ ω τ− −= − (16)

( ) ( )2 gf Xτ τ= && (17)

and for equality in Eq. 15,

( ) ( ) (,sini i tgX e tξ ω τ )D iτ κ ω− −=&& τ− (18)

The constant is determined by substituting Eq. (18) in Eq. (14) with a strict equality sign; when is substituted back in Eq. (14) the peak modal displacement is obtained for the GEB convex

mode is:

κκ

( ) ( ),

,2

Ti GEB d

GEB iD i i

E ty t

φ

i

λω ξ ω

Θ= (19)

where

22 2 2,(1 ) 1 1 sin cos 2i it

d i i i D i i De tξ ωλ ξ ξ ξ ω ξ ω− ⎡ ⎤= − − + − −⎣ ⎦,it (20)

Meanwhile, the expression for velocity and acceleration response of the i th normal mode are given on the basis of the differentiation of Duhamel’s integration. In the same manner, the peak velocity can be found as:

( ) ( ),

,2

Ti GEB i v

GEB iD i i

E ty t

φ

i

ω λω ξω

Θ=& (21)

where

( ) ( )22 2 2, ,(1 ) 1 1 sin 2 cos 2i it

v i i i i D i i i D ie tξ ωλ ξ ξ ξ γ ω ξ γ ω− ⎡ ⎤= − − + − − − −⎣ ⎦t (22)

21 1

tan ii

i

ξγ

ξ−⎛ ⎞−⎜=⎜ ⎟⎝ ⎠

⎟ (23)

Similarly, the peak modal acceleration for the global energy bound convex model can be found as:

( ) ( ) ( ), 22 1

Ti GEBT i a

GEB i i gii

E ty t X t

φ ω λφξξ

Θ= Θ −

−&&&& (24)

( ) ( )

2 2

2 2 2, ,

1 1 4

1 1 sin 2 cos 2i i

a i i

ti i i D i i i D ie tξ ω

λ ξ ξ

ξ ξ δ ω ξ δ ω−

⎡ ⎤ ⎡ ⎤= − +⎣ ⎦ ⎣ ⎦⎡ ⎤− + − + − +⎣ ⎦t

(25)

21

2 2

2 1tan

1i i

i

i i

ξ ξδ

ξ ξ−⎛ ⎞−⎜=⎜ − −⎝ ⎠

⎟⎟

(26)



The peak values of the modal response as time t , which is considered as critical value of displacement, velocity and acceleration are given below:

→∞

( ), 2

Ti GEB

GEB ii i i

Ey

φω ξ ωΘ ∞

= (27)

( ), 2

Ti GEB

GEB ii i

Ey

φξ ω

Θ ∞=& (28)

( ) ( )2

,

1 4

2

Ti ii GEB

GEB ii

Ey

ω ξφξ

⎡ ⎤+Θ ∞ ⎣=&& ⎦ (29)

The quantity is defined as shown in Eq. (13) with ( )2GEBE ∞ ft t∞ = . Eq. (27) – (29) shows that

the only quantity required obtaining the convex model estimate of the response, in addition to the damping, frequency, and mode shape, is the global energy bound of the earthquake evaluated at the end of the earthquake record. All detail calculation can be found in work of Pantelides and Tzan (1996) in which the energy bound is considered known.

5. NUMERICAL EXEMPLE

The simple numerical example consists of a portal frame structure (see Figure 1) subjected to El-Centro earthquake which is shown in Figure 2. The displacement responses obtained by: (1) a time-history analysis employing the actual record, and (2) the global energy-bound convex model.

Figure 1 Simple portal frame structure under seismic excitation.

Figure 2 Record of El-Centro earthquakes.



Figure 3 Displacement time history for El-Centro earthquake

(AR: Actual Record, GEB: Global Energy Bound)

The fundamental frequency of this structure is 2.76 Hz and the damping level is 5 per cent. The maximum response obtained by the actual record analysis is 0.02 m and that by global energy bound is 0.09 m. The time history displacement for the structure is shown in Figure 3.

The response obtained by the global energy bound convex model gives a higher value than the actual record by a ratio of 3.61. This result is very conservatism, since the convex is applied based on the limited information available. Hence for the future work, it is suggested the used of reduction factor for adjusting the response due to actual record.

6. CONCLUSION

The estimation of structural response using the energy-bound convex model is rather conservative when compared to the time history of the response using the actual earthquake record. However, this result is utile since there is only limited of amount information is available. This conservatism is to avoid some uncertainty in predicting a maximum structural response.

Some future works are required especially for adjusting the maximum response of structure since this method gives the result very conservative. However, in engineering field, since there is only limited information available especially due to seismic, this method could be as one alternative solution for designing.

7. REFERENCES

Ahmadi, G. (1986). “Bounds on earthquake response of structures”, Journal of Engineering Mechanics, ASCE, Vol. 112, 351 – 369.

Ben-Haim, Y. (1994). “A non-probabilistic concept of reliability”, Structural Safety, Vol. 14, 227 – 245.

Ben-Haim, Y., Chen, G. and Soong, T.T. (1996). “Maximum structural response using convex model”, Journal of Engineering Mechanics, ASCE, Vol. 122, 325 – 333.

Ben-Haim, Y. and Elishakoff, I. (1990). “Convex Models of Uncertainty in Applied Mechanics”, Elsevier.

DerKiureghian, A. (1996). “Structural reliability methods for seismic safety assessment: a review”, Engineering Structure, Vol. 18, No. 6, 412 – 424.

Drenick, R.F. (1973). “A seismic design by way of critical excitation”, Journal of Engineering Mechanics, ASCE, Vol. 99, 649 – 667.

Ellingwood, B.R. (2001). “Earthquake risk assessment of building structures”, Reliability Engineering and System Safety, Vol. 74, 251 – 262.

Naeim, F. and Anderson, J.C. (1993). “Classification and evaluation of earthquake records for design”, The 1993 NEHRP Professional Fellowship Report, Earthquake Engineering Research Institute, Oakland, CA.




Pantelides, C.P. and Tzan, S.R. (1996). “Convex model for seismic design of structures – I: analysis”, Earthquake Engineering and Structural Dynamics, Vol. 25, 927 – 944.

Shinozuka, M. (1970). “Maximum structural response to seismic excitation”, Journal of Engineering Mechanics, ASCE, Vol. 96, 729 – 738.

Yuen, K.V., Beck, J.L. and Katafygiotis, L.S. (2002). “Probabilistic approach for modal identification using non-stationary noisy response measurement only”, Earthquake Engineering and Structural Dynamics, Vol. 31, 1007 – 1023.



PERFORMANCE OF CONFINED PARTIAL HEIGHT URM PARTITIONS UNDER LATERAL LOADS

R. Rezvaniasl1, T. Mahdi 2 and A. Kaveh 3

1 Faculty of Building and Housing, Tehran, Iran

2 Building and Housing Research Centre, Tehran, Iran 3 Iran University of Science and Technology, Tehran, Iran

ABSTRACT: In recent years, confined masonry buildings have been considered as good replacement for the traditional non-reinforced masonry (URM) buildings in areas of high seismicity. To assess seismic response of confined URM partitions, a research program involves construction and testing of four full-scale specimens has been carried out. Three partial height and one full height partitions with varying geometries and connection schemes have been tested under cyclic loads. The experimental tests have been performed on single bay, single storey specimens. Furthermore, linear finite element models were developed based on the equivalent strut and two-dimensional models. The linear-elastic models used in this paper are based on simplified crack assumptions that actually occurred at different stages of the actual cyclic test. The results obtained by the two-dimensional model in the elastic range were in good agreement with the experimental ones, while the one-dimensional equivalent strut model give good but less satisfying results. Furthermore, and to understand the behavior of partial height URM partitions, a parametric study has been carried out. Important geometrical parameters include length of the bay, thickness of the wall, and height of the wall has been studied. The influence of some mechanical properties includes the compressive strength of concrete and the compression strength of masonry prisms has also been investigated.

1. INTRODUCTION

Iran is one of the most seismically active regions of the world. Frequent earthquakes with different magnitudes yearly occurred resulting in significant economic and human losses. The large number of casualties in recent earthquakes in Iran is due to both high seismic activity and high vulnerability of the masonry buildings. The majority of these damages were due to the destruction of older buildings which had been poorly designed, constructed with very poor and unreliable materials and without using any of the available seismic standards. Lack of Integrity, lack of ductility, and heavy masses are some of the important weaknesses of these buildings. Figure 1 provides samples of masonry buildings that have been damaged in recent earthquakes. In order to prevent these damages, the Iranian Code of Practice (IS 2800) has suggested the use of confined masonry walls as an alternative system (BHRC, 2003). In such a system, it is assumed that walls resist both vertical and lateral loads while bond beams and tie-columns are merely used to make the system work together. In recent Iranian earthquakes, failures of such walls were also observed. In most cases, and due to bad connections between masonry walls and roofs, these partitions worked as partial height ones. In this paper, four experimental tests have been carried out to investigate the performance of such walls under seismic-like in-plane loadings and the effects of different types of connections on the overall behavior of these walls. Numerical models are also provided to test different geometrical and mechanical parameters that affect the behavior of such walls.



Figure 1 Performance of masonry buildings in previous Iranian earthquakes.

2. EXPERIMENTAL RESULTS

Many experiments of confined masonry walls under seismic-like in-plane loadings were carried in the past. Full-scale or reduced-size masonry wall specimens made of different materials and different geometries have been tested at different boundary conditions by many authors. Examples of these works can be found in the works of Tomazevic and Klemenc, 1997; Yoshimura et al., 2004; Zabala et al., 2004; and Barragan and Alcocer, 2006. In the present paper, a series of experiments is designed to investigate the performance of three partial height and one full height walls confined with concrete bond beams and tie-columns specified according to the Iranian seismic code of practice (BHRC, 2003). As shown in Figure 2, perforated bricks are used in the three partial height masonry walls while porous clay masonry units are used in the fourth full height wall. All the models have been constructed and fixed on the strong floor and the walls have been connected to the lower beam using vertical bars. Furthermore, toothed joints and horizontal reinforced bars have been provided for the second and third specimens.

The dimension of each of the four specimens is 300 cm × 330 cm. The thickness of the wall in the first and fourth specimens is taken as 100 mm, while walls having 200 mm. thicknesses are used for the second and third specimens. In each of the first three wall specimens, the height to thickness ratios is taken around (12). Furthermore, with the exception of the lower beam that need to be rigid, the cross sectional areas of all the other elements are taken as 25 cm x 25 cm. The lower beam is taken as 30 cm x 30 cm. The reinforcements for all elements are chosen to meet the minimum requirements of the Iranian standard of practice; i.e. 4#10 deformed bars. The minimum requirements for lateral reinforcements are also met. In all the four specimens, such reinforcements are taken as plain 6 millimeters stirrups. The spacing of the stirrups has been taken as 25 cm. in the beams, and 20 cm. in the columns. In the third specimen and at the joints, the spacing of lateral reinforcements is made closer than the other three. In all the four tests that have been carried out, the masonry specimens are subjected only to lateral cyclic loads and no vertical loads other than their weights are applied. The loading history is designed to be general enough to provide the full range of deformation that the structure will experience under the severe earthquake excitation. The resistance envelopes, i.e. the envelopes of hysteretic loops of all the four tested walls are compared in Figure 3. Comparing the second and third wall specimens with the first one, the increase in height and thickness for the last two specimens has resulted in increasing the strength and stiffness of the whole system. For specimen No.4, porous clay masonry units much weaker than the perforated bricks used in the first three specimens have been used. However, comparing the strength and stiffness of wall specimen No.4 with the corresponding ones of wall specimen No.1, it can be seen that by using full height, using special connections between wall and upper bond beam, and using cement mortar as a coating has increased considerably both stiffness and strength. Details of these experiments can be found in Mahdi and Aghabeki, 2007.



Figure 2 A partial height brick wall specimen No.3.

0

2000

4000

6000

8000

10000

12000

0 50 100 150 200 250Displacement (mm)

Load

(kg)

WALL NO.1

WALL NO.2

WALL NO.3

WALL NO.4

Figure 3 Envelopes of hysteresis loops of the four tested walls.

3. NUMERICAL MODELS

3.1 The Equivalent Strut Model

The equivalent strut model is one of the most common models used for solving infilled walls. However, and since the behavior of confined walls is similar to that of infilled walls, this model has been adopted in this paper. In this model, the wall is presented by a strut using an equivalent width and a thickness equal to the actual one. Among the many numerical models found in the literature, this model has been the most dominant one for the last five decades. In this respect, the most simplistic approaches presented by Paulay and Priestley (1992) and Angel et al. (1994) have assumed constant values for the strut width between 12.5 to 25 percent of the diagonal dimension of the infill, with no regard for any infill or frame properties. The adequacy of this simplified model has been approved by many researchers (Abrams, 1994). Furthermore, it has been used in many international codes like FEMA 273 (ATC, 1997) and FEMA 356 (ASCE, 2000). Later on, Al-Chaar, 2002 has developed an eccentric equivalent strut model that has been used in this paper. The method suggested can be used for full-height confined walls as well as partial-height ones and perforated masonry panels. Using eccentric struts in this global analysis will yield infill effects on the column directly, which will negate the need to evaluate these members locally.



3.2 Two-Dimensional Elements

The in plane stress and shell elements are available in most softwares. In this paper, 4-nodes two-dimensional elements are used to model the masonry walls.

4. PARAMETRIC INVESTIGATIONS OF THE CONFINED INFILL WALL

To understand the behavior of confined walls, the influence of different geometrical parameters on displacements and stresses need to be studied. This has been done for two types of masonry walls; i.e. perforated brick walls that have average compressive strengths of 53.47 kg/cm2 and porous clay tile walls that have average compressive strengths of 8.73 kg/cm2. Furthermore, the influence of the concrete's strength used in bond beams and tie-columns, on the total behavior of the walls, has been also investigated. Numerical calculations, for full-height perforated brick walls subjected to lateral forces are presented in Figures 4 to 7. Similar results have been obtained for the porous clay tile walls that are not presented in this paper.

The influence of concrete strength on displacements of walls with different geometries is presented in Figure 4. The compressive strength of concrete is taken to vary between 15 and 60 MPA. As shown in this figure, increasing the concrete compressive strengths has resulted in decreasing the displacements. It is clear that such increase is more effective at lower values of compressive strengths. This figure also shows that decreasing the thickness of the wall and increasing its length would increase the displacements noticeably. As example, increasing the wall length from 5m to 6m would result in an increase of around 35% in the system displacement and this show the importance of controlling the unsupported length in the confined system.

Figure 5 shows the effect of concrete's compressive strength on stresses developed in the struts (masonry walls). As shown in this figure, increasing the concrete compressive strengths has resulted in decreasing the stresses in the masonry walls. It is clear that such increase is more effective at lower values of compressive strengths. This figure also shows that decreasing either the thickness of the wall or its length, a corresponding increase in the wall's stresses would result.

Figure 6 shows the effect of concrete's compressive strength on the maximum shear stresses developed in the tie-columns. These stresses are observed to occur near the intersection points of the strut with tie-columns. Based on this figure, it can be concluded that an increase in the concrete compressive strength would slightly decrease the shear stresses in the tie-columns. Such an increase in compressive strength would also result in slightly higher shear capacity. Accordingly, it is always advisable to increase the shear capacities of the tie-columns by providing close lateral reinforcements or increasing the size of the column at this region. Furthermore, decreasing either the thickness of the wall or its length would decrease shear stresses in the tie-columns.

Fc - Displacement

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

10 20 30 40 50 60Fc (MPA)

Displ

acem

ent (

cm)

t=10 L=3 t=10 L=4

t=10 L=5 t=10 L=6

t=20 L=3 t=20 L=4

t=20 L=5 t=20 L=6

Figure 4 Influence of concrete strength on displacements.



AXIAL STRESS - FC

2

4

6

8

10

12

10 20 30 40 50 60FC (MPA)

STRES

S (k

g/cm

2)

t=10,L=3 t=10,L=4

t=10,L=5 t=10,L=6

t=20,L=3 t=20,L=4

t=20,L=5 t=20,L=6

Figure 5 Influence of concrete strength on stresses developed in the struts.

Shear Stress - FC

6.35

6.85

7.35

7.85

8.35

12.5 22.5 32.5 42.5 52.5 62.5FC (MPA)

Stre

ss (k

g/cm

2)

t=10,L=3 t=10,L=4 t=10,L=5

t=10,L=6 t=20,L=3 t=20,L=4

t=20,L=5 t=20,L=6

Figure 6 Influence of concrete strength on maximum shear stresses developed in the tie-columns.

5. COMPARISON OF NUMERICAL AND EXPERIMENTAL RESULTS

In this section, the experimental results of the four walls that have been reported earlier in this paper (Figure 3) are compared with the numerical ones. Using both, the eccentric equivalent strut model and the two-dimensional elements, results based on different simplified crack assumptions that actually occurred at different stages of the actual cyclic test, are considered. These assumptions include a) walls without cracks, b) the occurrence of full cracks in concrete members with no cracks in masonry, c) the occurrence of full cracks in masonry wall with no cracks in concrete members, and d) occurrence of cracks in both concrete and masonry. It is obvious that assumptions (b) and (c) are only introduced to investigate the first crack's occurrence and have no other practical applications. Accordingly, Figures 7 to 10 present the no-crack, the full-crack and one of the assumptions (b) or (c). In these figures, the symbol M refers to the two-dimensional elements while the "No wall" symbol refers to the bare frames without masonry walls. Results of the eccentric equivalent strut model without cracks are also given in these figures.



WALL No.1

0

200

400

600

800

1000

1200

1400

0 1 2 3 4 5 6 7 8 9

Displacement (mm)

Load

(kg)

Experiment M - Brick

M - Both M - No Crack

Al - chaar No wall

Figure 7 Displacements - loads relations for wall No.1.

WALL No.2

0

1000

2000

3000

4000

5000

6000

7000

0 2 4 6 8 10 12 14 16 18 20

Displacement (mm)

Load

(kg)

Experiment M - Con


Al - chaar No wall

Figure 8 Displacements - loads relations for wall No.2.

WALL No.3

0

1000

2000

3000

4000

5000

6000

7000

0 2 4 6 8 10 12 14

Displacement (mm)

Load

(kg)

Experiment M - Con


Al - chaar No wall

Figure 9 Displacement - loads relation for wall No.3.



WALL No.4

0

2000

4000

6000

8000

10000

12000

14000

0 5 10 15 20 25

Displacement (mm)

Load

(kg)

Experiment M - Con


Al - chaar No wall

Figure 10 Displacements – loads relation for wall No.4.

The above figures show that results obtained by the two-dimensional elements are more accurate than those given by the eccentric equivalent strut model (Al-Chaar) in the elastic range, especially in predicting the initial stiffness. Indeed, the results of the two dimensional model are quite close to the experimental ones. This is quite important step toward establishing numerically the analytical displacements – loads model for confined masonry walls and predict with certain degree of accuracy the three pairs of parameters suggested by Tomazevic and Klemenc (1997).

6. CONCLUSIONS

In the confined masonry wall system used in Iran, the bond beams and tie-columns are merely used to make the system work together. In such a system, walls must resist both vertical and lateral loads. Based on the present experimental and numerical results, this assumption has been found to be reasonably accurate. According to these results, it can be concluded that some changes must be made in the confined masonry system. These changes, however, are not limited to a specific country but have general nature. Based on this, it is recommended that: 1. Provisions must be made to increase the reinforcements in bond beams and tie-columns. 2. Numerically, the use of the two dimensional model in the elastic range is recommended.

However, the eccentric equivalent strut model can be useful for parametric investigations.

7. REFERENCES

Abrams, D.P. (Ed.) (1994). “Proceedings of the NCEER Workshop on Seismic Response of Masonry Infills”, National Center of Earthquake Engineering Research, Technical Report NCEER-94-004.

Al-Chaar, G. (2002). "Evaluating Strength and Stiffness of Unreinforced Masonry Infill tructures", US Army Corps of Engineers, ERDC/CERL TR-02-1.

Angel, R., Abram, D.P., Shapiro, D., Uzarski, J. and Webster, M. (1994). "Behavior of Reinforced Concrete Frames with Masonry Infills", Structural Research Series No. 589, UILU-ENG-94-2005, University of Illinois at Urbana, Illinois.

ASCE (2000). “Prestandard and Commentary for the Seismic Rehabilitation of Buildings, FFMA 356”, Federal Emergency Management Agency.

Barragan, R and Alcocer, S.M. (2006). "Shaking table tests on half-scale models of confined masonry made of handmade solid clay bricks", Proc. 1st European Conference on Earthquake Engineering and Seismology, Geneva, Switzerland, Paper No. 1147.

BHRC (2003). "Iranian Code of Practice for Seismic Resistant Design of Buildings", BHRC Publication No. S347.

BSSC (1997). “NEHRP Guidelines for the Seismic Rehabilitation of Buildings, FEMA 273”, ATC-33 project, Federal Emergency Management Agency.



Mahdi, T. and Aghabeki, H. (2007). “Performance of confined masonry walls under cyclic testing”, Proc. 1st International workshop on Performance, Protection & Strengthening of Structures under Extreme Loading, Whistler, Canada.

Paulay, T. and Priestly, M.J.N. (1992). “Seismic Design of Reinforced Concrete and Masonry Buildings”, John Wiley & Sons.

Tomazevic, M. and Klemenc, I. (1997). “Seismic behaviour of confined masonry walls”, Earthquake Engineering and Structural Dynamics, Vol. 26, 1059. Yoshimura, K., Kikuchi, K., Kuroki, M., Nonaka, H., Tae Kim, K., Wangdi, R. and Oshikata, A.

(2004). “Experimental study for developing higher seismic performance of brick masonry walls”, Proc.13th World Conference on Earthquake Engineering, Vancouver, Canada, Paper No. 1597.

Zabala, F., Bustos, J.L., Masanet, A. and Santalucia, J. (2004). “Experimental behavior of masonry structural walls”, Proc. 13th World Conference on Earthquake Engineering, Vancouver, Canada, Paper No. 1093.
