SEISMIC BEHAVIOR OF SEMI-PRECAST CONCRETE ...SEISMIC BEHAVIOR OF SEMI-PRECAST CONCRETE SHEAR WALLS Golnesa Karimi Zindashti1, Barış Binici2, and Erdem Canbay3 1 Ph.D. Student, Civil

4th International Conference on Earthquake Engineering and Seismology

11-13 October 2017 – ANADOLU UNIVERSITY – Eskisehir/TURKEY

SEISMIC BEHAVIOR OF SEMI-PRECAST CONCRETE SHEAR WALLS

Golnesa Karimi Zindashti1, Barış Binici2, and Erdem Canbay3

1 Ph.D. Student, Civil Eng. Department, Middle East Technical University, Ankara

Email: [email protected] 2 Prof. Dr., Civil Eng. Department, Middle East Technical University, Ankara

Email: [email protected] 3 Prof. Dr., Civil Eng. Department, Middle East Technical University, Ankara

Email: [email protected]

ABSTRACT:

Successful seismic performance of well-designed and constructed precast concrete members is crucial for

sustainable, economical and high speed construction. The main disadvantage of precast concrete is the difficulty

of detailing the connection region between the vertical and horizontal load bearing elements to ensure structural

integrity. Double wall system, being a semi-precast approach, consists of two precast reinforced concrete layers

encasing a cast-in-place concrete layer which is usually employed to create a monolithic connection between the

wall and the slab. Seismic performance tests on double walls is scarce due to its limited use in earthquake prone

regions. To gain a better understanding of the seismic behavior of double walls, four full-scaled specimens with

different section shapes were tested subjected to cyclic horizontal loads at Middle East Technical University1. In

this paper, test results are presented along with the comparisons with simple analytical models and performance

limits presented in ASCE/SEI-41, Eurocode 8, and Turkish Earthquake Code 2007. It has been found that while

TEC2007 provided estimations with the highest standard deviation, ASCE/SEI-41-06 demonstrated a better

agreement with the tests results.

KEYWORDS: Double Walls, Hybrid System, Section Analysis, Seismic Performance Assessment.

1. INTRODUCTION

Precast concrete construction offers advantages such as better quality, high speed and economy. The major

challenge however is providing sufficiently ductile connections between vertical and horizontal load bearing

elements. Past research, mostly focused on investigating the seismic performance of beam-column connections

and improving them for better earthquake resistance. It is well known that structural concrete shear wall systems

exhibit good seismic performance providing stiffness, strength and deformability. The benefit of shear wall

systems has not been realized in precast concrete construction due to the difficulty of devising practical and high-

performance connections. The concept of double walls are good candidates for precast shear wall systems

benefiting from the advantages of shear wall systems and use of cast in place concrete in their connections. A

double wall is composed of two reinforced concrete shells connected with section through connectors (lattice

girders or point connectors) encasing a void layer, to be filled with cast-in-place concrete after erection. Double

wall systems are usually constructed with filigree slabs to enable monolithic behavior at a higher speed of

construction. After curing of concrete placed in the middle layer of walls and top of slab shells, a monolithic wall-

floor system is obtained. Seismic performance of double walls was examined for the first time in China by testing

three full scale hybrid shear walls and conducting numerical studies later in Illinois Institute of Technology (Xu,

Shen, & Shen, 2014). Those tests employed lattice girder as the connecting system. One disadvantage of the lattice

girders to connect adjacent double walls is the absence of sufficient development length to place connecting cages.

This requires casting connecting reinforced concrete cast in place columns between adjacent double walls slowing

mailto:[email protected]



down the construction. In 2008, late Erich Kastner, in Germany, patented a point connector that could make it

possible to place adjacent double walls with connecting cages. In fact, this invention made it possible to devise a

practical seismic resistant double wall system. In 2014, a comprehensive research program was initiated at Middle

East Technical University, with support from Oberndorfer International Company to investigate the seismic

performance of double wall systems. This paper summarizes the test results from that study and presents the

comparisons with simple analytical models and performance limits presented in ASCE/SEI-41, Eurocode 8, and

Turkish Earthquake Code 2007. The outcomes are believed to guide engineers in designing seismic resistant

double wall systems.

2. REVIEW OF THE EXPERIMENTAL STUDY

Four specimens were tested during the course of the experimental study. First two tests were conducted to compare

the seismic response of two insulated exterior double walls produced with single and two adjacent double walls,

respectively, in order to simulate insulated double walls (Fig. 1). The other two specimens were considered as

interior walls of a building; hence no insulation material was used. Sections of the specimens 3 and 4 were designed

as U-shaped and T-shaped, respectively (Fig. 2). These two specimens were constructed following the regulations

of Turkish Earthquake Code (TEC2007). Mechanical properties of the materials employed in each of them are

summarized in Table. 1.

Figure 1. Details of Specimen 1 (a), and Specimen 2 (b) (Binici and Canbay, 2014)

C* C

Insulation

51410

6

30 Ø 8 / 10 (x3)

22 Ø

8 /

10 (

x3)

300

225

295 cm

220 c

m

150

225

51410

6

30 Ø 8 / 10 (x3)

22 Ø

8 /

10 (

x3)

C* C

Insulation145 cm

220 c

m

30 Ø 8 / 10 (x3)

145 cm

150

Starter Bars

Ø 12 / 15 (x2)45

15

Starter Bars

Ø 12 / 15 (x2)45

15

(a) (b)



Figure 2. Details of Specimen 2 (a), and Specimen 3 (b) (Binici and Canbay, 2014)

Table 1. Properties of Materials

Specimen f'

c(s)*

(MPa)

f'c(c)**

(MPa)

ϕ6

fy, fu (MPa)

Φ8

fy, fu (MPa)

Φ12

fy, fu (MPa)

Φ14

fy, fu (MPa)

1 45 28 340, 470 380, 540 490, 610 325, 455

2 43 27 340, 470 380, 540 490, 610 325, 455

3 45 25 340, 470 380, 540 490, 610 325, 455

4 45 25 340, 470 380, 540 490, 610 325, 455

*: Compressive Strength of Shell Concrete, **: Compressive Strength of Core Concrete

All of the specimens were tested under lateral cyclic displacement reversals. Axial force was not applied during

the tests due to insignificant axial loads for walls in such buildings. Test setup employed for this experimental

study and the displacement loading history of specimens is presented in Fig. 3. The lateral load deformation

responses of all the walls are presented in the following section. According to the visual observations, specimen 1

and specimen 2 behaved in a comparable manner. Despite their squat dimensions, they behaved in a ductile manner

in both directions of loading and had a displacement ductility of about 6.5. Specimen 3 had a displacement ductility

of about 4.8 and 5.2 in the positive and negative directions, respectively. It can be stated that despite the observed

shear cracks, the specimen was able to behave in a very ductile manner. The displacement ductility of the wall

was about 4 (Binici and Canbay, 2014).

225

140

5

101010

10

25

39 39

100853525155

100902910

10 10

20

510

1010

25

170

1010

75

71

65

2035

625

1004011020

150

22

5

10

25

101017

(a)

(b)



Figure 3. Schematic Details of a Sample Test Setup (a), Displacement History During All Tests (b)

3. SECTION ANALISYS

3.1. Theory In general, it has been accepted that the first loop of the hysteretic response follows the same pattern with the

moment-curvature diagram under monotonic loading (Ersoy and Özcebe, 1997). Following the procedure

indicated by Priestley (Priestley et al., 2007), a set of spread-sheets were developed to construct the moment-

curvature relations and predicting the force-displacement relation of a cantilever wall member. For this purpose,

the flexural response of the section where the plastic hinge is expected was calculated and afterwards, a force-

displacement relationship was obtained using a simplified solution based on the concept of the "plastic hinge". For

material modeling, the unified stress-strain relationship proposed by Mander et. al. (1988) was used for confined

concrete. The model proposed by King et al. (1986) was used for the stress-strain relation for the reinforcing steel. Plastic hinge length, Lp, was assumed following Priestley (Priestley et al., 2007). Shear deformation of the member

was calculated considering regions before shear cracking, after shear cracking and shear yielding. Finally, the sum

of the flexure and shear deformation were added as a parallel spring model to obtain the total displacement of the

member. The recommendations provided by Priestley (Priestley et al., 2007) was followed for bilinear idealization

of the moment curvature diagram. The moment capacity of concrete members is known to be affected by the

presence of shear in the member (Arlekar, 2004). Consequently, the flexure capacity of the specimens was

recomputed considering the moment-shear interaction effects employing Modified Compression Field Theory

(MCFT). In order to demonstrate the mode of failure of the specimens, the shear capacity of each wall was

computed according to ACI318-11 and marked on the response curves (Karimi Zindadshti, 2016).

3.2. Results Comparison of the results is provided in Fig. 4 and Fig. 5. The results from section analysis of models provide

slight overestimation of the capacity of the specimens 1 and 2. It can be observed that the estimated shear capacity

for these two specimens were higher than the lateral load capacity determined based on flexural yielding. This

situation indicates the estimated failure mode to be flexure dominated. The moment capacities computed by

considering shear flexure interaction shows that the presence of shear reduces the moment capacity of the test

specimens by about 25%. Upon considering the shear effects on moment capacities, the estimations turned out to

be on the safe side compared to the experimental results.

Figure 4. Comparison of the Moment-Curvature of Specimen 1 (a), 2 (b), 3 (c), and 4 (d) with Analytical Model

-3000

-1500

0

1500

3000

-30 -15 0 15 30

-3000

-1500

0

1500

3000

-30 -15 0 15 30

-3000

-1500

0

1500

3000

-50 -25 0 25 50

-3000

-1500

0

1500

3000

-50 -25 0 25

(a) (b)

(a) (b) (c) (d)

Curvature (1/km)

Mo

men

t (k

N.m

)



Figure 5. Comparison of the Response of Specimen 1 (a), 2 (b), 3 (c), and 4 (d) with Analytical Model

For specimens 3 and 4, the results demonstrate that the behavior is shear-dominated owing to the smaller shear

capacity compared to the lateral load based on flexural yielding. The maximum bending capacity is preceded by

the shear capacity of the specimens 3 and 4, indicating a brittle mode of failure. Despite the shear critical nature

of specimens 3 and 4, these walls behaved in a ductile manner during the tests and the failure of these walls

occurred in a flexure-shear mode. The moment capacities computed considering M-V interaction shows that the

presence of shear reduces the moment capacity of the test specimens by about 10% and 25% in positive and

negative directions, respectively. Upon considering the shear effects on moment capacities, the estimations turned

out to be closer to the experimental results still being on the safe side. Table. 2 presents the comparison of estimated

to experimental lateral strength.

Table 2. Comparison of the Experimental and Estimated Capacities

Specimen Vtest VShear VFlexure Vmin V(MCFT) Vmin/Vtest V(MCFT)/Vtest

1 1069

-903

1645.2

-1645.2

925

-925 925

678

-678

0.87

-1.02

0.63

-0.75

2 1072

-1017

1645.2

-1645.2

925

-925 925

678

-678

0.86

-0.91

0.63

-0.67

3 732

-575

686.8

-686.8

836

-715 686.8

614

-514

0.938

-1.19

0.84

-0.89

4 639

-393

459.2

-459.2

927.6

-440.2 459.2

420

-358

0.72

-1.17

0.66

-0.91

The values are in kN.

4. PERFORMANCE ASSESSMENT

In order to estimate the seismic performance of buildings, maximum permissible damage states (performance

levels), considering certain levels of seismic hazard of the site and based on the observed damage states of the

building, are specified. For this purpose, three discrete “Structural Performance Levels” are used in different

provisions; Immediate Occupancy (IO), Life Safety (LS), and Collapse Prevention (CP). In this section, the

evaluation of three seismic assessment guidelines are provided in light of experimental results. These guidelines

are ASCE/SEI 41-06 and ASCE/SEI 41-13, Eurocode 8 (2005), and Turkish Earthquake Code (2007). ASCE/SEI-

41 defines performance levels for both flexure and shear-controlled walls, however in Eurocode 8 and TEC 2007,

the assessment is only performed for a ductile flexural mechanism. Hence, comparisons are presented for selected

specimens for TEC (2007) and Eurocode 8. Performance level of each specimen was additionally estimated from

experimental results considering the idealized elastic perfectly plastic response. In this study, estimated CP state

was considered as the ultimate point where the maximum strength dropped by 15 percent. Immediate Occupancy

was determined by connecting the origin with a line passing through 70% of the ultimate load on the initial loading

curve (defined as first yield point) and extending this line to 85% of the ultimate load. Accordingly, Life Safety

-1.6 -0.8 0 0.8 1.6

-1800

-900

0

900

1800

-40 -20 0 20 40

-1.6 -0.8 0 0.8 1.6

-1800

-900

0

900

1800

-40 -20 0 20 40

-2.4 -1.2 0 1.2 2.4

-1000

-500

0

500

1000

-60 -30 0 30 60

-2 -1 0 1 2

-1000

-500

0

500

1000

-50 -25 0 25 50

AVE.: 1.05 0.85

La

teral

Loa

d (

kN

)

Displacement (mm)

Drift Ratio (%)

(a) (b) (c) (d)



state was estimated as the 75% of the ultimate point (Binici and Canbay, 2014). The results are presented in Fig.6

for ASCE/SEI 41-06, ASCE/SEI 41-13, Eurocode 8, and TEC2007, respectively. Table. 3 summarizes the related

comparisons. In the elements controlled by flexure, it seems that ASCE/SEI-41-06 provided a better agreement

with the experimental results, while the updated version of this document, overestimated the damage limits

indicating that after complete analysis, collapse state should be determined. For shear controlled specimens,

comparing the related shear capacities specified by the codes, ASCE/SEI-41 (ACI318) provides the closest

estimate to test result. It may be concluded that the shear strength expressions of ACI318 were safe to compute

the capacity of double walls. Among these codes, TEC2007 provides the estimation with the highest standard

deviation. Eurocode 8 provided the best prediction of the failure mechanisms of specimens.

Table 3. Comparison of the Criteria Proposed by Seismic Guidelines with Experimental Response of Specimens

ASCE/SEI41-06 ASCE/SEI41-13 Eurocode 8 TEC 2007 Experimental

(-) (+) (-) (+) (-) (+) (-) (+) (-) (+)

VFlexure (kN) -1017 1077 -1017 1077 -925 925 -925 925

VShear (kN) -1645 1645 -1645 1645 -1101 1101 -1772 1772

IO (mm) -8.16 8.16 -8.16 8.16 -9.4 9.4 -7.15 7.15 -2.95 3.2

LS (mm) -13.06 13.06 -22.86 22.86 -17.44 17.44 -19.6 19.6 -13.65 15.375

CP (mm) -22.86 22.86 -40.01 40.01 -19.18 19.18 -23.31 23.31 -18.2 20.5

Failure

Mode Flexure Flexure Flexure Flexure Flexure-Shear

VFlexure (kN) -1017 1077 -1017 1077 -925 925 -925 925

VShear (kN) -1645 1645 -1645 1645 -1101 1101 -1772 1772

IO (mm) -8.16 8.16 -8.16 8.16 -9.4 9.4 -7.15 7.15 -2.7 2.2

LS (mm) -13.06 13.06 -22.86 22.86 -17.44 17.44 -19.6 19.6 -9.08 13.54

CP (mm) -22.86 22.86 -40.01 40.01 -19.18 19.18 -23.31 23.31 -12.1 18.05

Failure

Mode Flexure Flexure Flexure Flexure Flexure-Shear

VFlexure (kN) -873 863 -873 863 -715 836 -715 836

VShear (kN) -687 687 -687 687 -827 827 -1043 1043

IO (mm) -9.8 9.8 -9.8 9.8 - - -10.3 11.2 -5.5 8.4

LS (mm) -18.4 18.4 -36.8 36.8 - - -30 34.5 -16.1 23.1

CP (mm) -24.5 24.5 -49 49 - - -44.5 47.9 -21.5 30.8

Failure

Mode Shear Shear Flexure-Shear Flexure Flexure-Shear

VFlexure (kN) -597 918 -597 918 -440 928 -440 928

VShear (kN) -459 459 -459 459 -590 590 -635 635

IO (mm) -9.8 9.8 -9.8 9.8 - - - - -6.4 6.5

LS (mm) -18.4 18.4 -36.8 36.8 - - - - -15.2 17.2

CP (mm) -24.5 24.5 -49 49 - - - - -20.3 23

Failure

Mode Shear Shear Shear Shear Flexure-Shear



Figure 6. Comparison of Experimental Results with Damage States According to Provisions

4. CONCLUSIONS AND DISCUSSIONS

Utilizing the experimental results of double walls with three different section types along with section analysis

based models, performance of each wall behavior of walls were evaluated. Following conclusions can be drawn

from the study:

Comparing the moment curvature results with the experimental results shows that the strength and section

response of the double walls can be predicted with standard section analysis procedures of cast-in-place

reinforced concrete. This fact enables the use of existing analysis tools for structural design of double wall

systems.

Although specimens 3 and 4 exhibited significant shear strains during the tests and a brittle mode of failure

were diagnosed through performance assessment of the walls. These walls were able to sustain lateral load

considerably and a rather ductile behavior was detected until the ultimate capacity.

The range of the displacement ductility levels of the specimens were between about 4 and 7.5. It is obvious

that tested walls were squatter with respect to the walls incorporated in buildings. Therefore, it can be

easily realized that the seismic behavior of building walls will even be more ductile.

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-2.4 -1.2 0 1.2 2.4

-1000

-500

0

500

1000

-50 -25 0 25 50

-2 -1 0 1 2

-800

-400

0

400

800

-50 -25 0 25 50

-2 -1 0 1 2

-1200

-600

0

600

1200

-50 -25 0 25 50

-2 -1 0 1 2

-1200

-600

0

600

1200

-50 -25 0 25 50

-2.4 -1.2 0 1.2 2.4

-1000

-500

0

500

1000

-50 -25 0 25 50

-2 -1 0 1 2

-800

-400

0

400

800

-50 -25 0 25 50

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-1.2 -0.6 0 0.6 1.2

-1200

-600

0

600

1200

-30 -15 0 15 30

-2 -1 0 1 2

-1200

-600

0

600

1200

-50 -25 0 25 50

La

teral

Loa

d (

kN

)

Drift Ratio (%) L

ate

ral

Loa

d (

kN

)

La

teral

Loa

d (

kN

)

La

teral

Loa

d (

kN

)

Displacement (mm)

(1)

(1)

(2) (3) (4)

(2) (3) (4)

(1) (2) (3)

(1) (2)

AS

CE

/SE

I-4

1-0

6

AS

CE

/SE

I-4

1-1

3

Eu

roco

de 8

TE

C2

00

7



The detailed evaluation procedures of ASCE/SEI-41, TEC2007 and EC8-3 were performed. In the

elements controlled by flexure, it seems that ASCE/SEI-41-06 provided a better agreement with the

experimental results, while the updated version of this document, overestimated the damage limits

indicating that after complete analysis, collapse state should be determined.

For shear dominated elements, comparing the related shear capacities specified by the codes, ASCE/SEI-

41 (ACI318) provides the closest value to test result. It may be concluded that the shear strength

expressions of ACI318 are found to be safe to compute the capacity of double walls. Among these codes,

TEC2007 provides the most improper values.

REFERENCES

Binici, B., and Canbay, E. (2014). Component Testing of the Double Wall System for Seismic Qualification.

Department of Civil Engineering, Middle East Technical University, Ankara.

Xu, L., Shen, X., and Shen, J. (2014). Seismic Study of Hybrid Shear Wall. 10th National Conference in

Earthquake Engineering, Earthquake Engineering Research Institute.

King, D., Priestley, M., and Park R. (1986). Computer Programs for Concrete Column Design, Research Report

86/12, Department of Civil Engineering, University of Canterbury, New Zealand.

Mander, J., Priestley, M., and Park, R. (1988). Theoretical Stress-Strain Model for Confined Concrete, ASCE

Journal of Structural Engineering, vol. 114, no. 8.

Ersoy, U., and Özcebe, G. (1997). Moment-Curvature Relationship of Confined Concrete, First Japan-Turkey

Workshop on Earthquake Engineering, vol. 1, pp. 10-21.

Arlekar, J.N., and Murty, C.V.R. (2004). Shear Moment Interaction for Design of Steel Beam-To-Column

Connections, 13th World Conference on Earthquake Engineering, Vancouver, B.C., Canada.

Priestley, M., Calvi, G., and Kowalsky, M. (2007). Displacement-Based Seismic Design of Structures, IUSS Press.

Seismic Rehabilitation of Existing Buildings, Report No: ASCE/SEI 41 - Supplement 1, Reston, Virginia, USA:

American Society of Civil Engineers (ASCE), 2007

Seismic Evaluation and Retrofit of Existing Buildings, Reston, Virginia, USA: American Society of Civil

Engineers, 2014.

ACI Committee 318, Building Code Requirements, Structural Concrete and Commentary, American Concrete

Institute, 2011.

CEN (2005). Eurocode 8: Design of structures for earthquake resistance-Part 3: Assessment and

retrofitting of buildings, European Standard EN 1998-3-2005, Comité Europèen de

Normalisation, Bruxelles, Belgium

Turkish Earthquake Code 2007, Specification for Buildings to be Built in Seismic Zones, Ministry of Public Works

and Settlement.

Karimi Zindashti, G. (2016). Seismic Behavior of Semi-Precast Concrete Shear Walls. Master’s Thesis,

Department of Civil Eng., Middle East Technical University, Ankara, Turkey.

Documents

SEISMIC BEHAVIOR OF SEMI-PRECAST CONCRETE ...SEISMIC BEHAVIOR OF SEMI-PRECAST CONCRETE SHEAR WALLS Golnesa Karimi Zindashti1, Barış Binici2, and Erdem Canbay3 1 Ph.D. Student, Civil