415

19 Full-scale 3-D Shake Table Tests on Fail-Safe Performance of Reinforced Concrete Buildings against Extreme Motions
1

Toshimi Kabeyasawa and 2Toshikazu Kabeyasawa

1 2

Earthquake Research Institute, The University of Tokyo, Japan, kabe@eri.u-tokyo.ac.jp Faculty of Engineering, The University of Tokyo, Japan, tosikazu@eri.u-tokyo.ac.jp

Abstract

Full-scale three-story reinforced concrete buildings were tested at E-Defense, the world largest three-dimensional shake table, in 2006 to verify the input loss at the base foundation. As one of the highlight test results, the obvious reduction of damage to the building structures was observed owing to the slip behavior at the base foundation with concrete-to-concrete casting joint surface. The results indicated feasibility of development and practical application of building-foundation system, which would be effective to the fail-safe seismic design. In this paper, the full-scale test results are outlined and the new and simple building-foundation system is introduced, by which not only safety but also serviceability insensitive would be ensured under possible and inexperienced extreme motions in the future.

19.1 Full-scale Test Specimens on Concrete Base

Following the successful execution of the full-scale test on the six-story wallframe building at E-Defense in 2005, low-rise school buildings on flexible foundation were selected for the full-scale test 2006 among several possible test plans. Two three-story specimens are designed to be tested with the following specific objectives: (1) the simulation of progressive collapse of existing school buildings, (2) verification of strengthening effect by the attached steel frames, and (3) soilstructure interaction with flexible foundation[1]. The first specimen is a reinforced concrete school building designed following 1970' Building code of practice in Japan, which is called as the bare RC specimen, in correspondence with the second specimen strengthened with braces. The failure mode of the bare specimen under an extreme motion is expected to be shear and axial collapse of columns in the first story starting from flexure to shear failure of

416 the short columns, then inducing progressive structural failure with redistribution of column axial loads. The second specimen was originally designed and constructed as identical as the first bare reinforced concrete specimen, then strengthened after the construction with attached steel frames, simulating seismic retrofit of existing buildings. The failure mode of the specimen would change into an overall mechanism of rocking and sway at the foundation. Effectiveness of a conventional detail and a new detail for strengthening is to be verified and compared. Both specimens are constructed on a pool-shaped container with backfill soils without fixing at the base. The concrete is placed on the base concrete surface with the construction joint modeling the load-baring foundation so that the shear transfer at the joint would be friction and cohesion of concrete surface. The flexible boundary condition for the shaking table test would be the world first trial simulating the rocking and sway behavior with neighborhood soils, which could be realistic in the full-scale test. The structural floor plans, such as 1st floor plan, foundation level, and 2nd floor plan with the base container of the bare reinforced concrete specimen are shown in Figure 19.1. The structure has three spans in the longitudinal (Y) direction, two spans in the orthogonal (span, X) directions. The span length is 4m in Ydirection, 2m and 6m in X-direction. The specimen models an end part of typical Japanese schools in the longitudinal direction, where irregular location of columns has often been adopted to make a special classroom with wide span. Note that total number of columns is 11 and a column at X2 and Y2 is missing from regular location. Structural walls are located in 6m span of the outer two frames in X-direction, which is also a typical plan. The elevations of the frame structures are shown in Figure 19.2. The inter-story height is 2.5m each for the 1st to 3rd story, which is not corresponding to the fullscale but approximately by five-sixth. The scale of the specimen is selected considering the limitation of the crane capacity of 800ton and the area of the shake table 20mx15m. The steel weights of 370 kN were attached on the roof after setup on the table in order to adjust the scaling effect. As for the foundation, the height of the foundation beams and the footings is 0.8m. The backfill soils are filled into the surrounding area at the base of the specimens, 1.0m width from the footing or 1.4m from the foundation beam to the side face of the container beam. No reinforcement are placed across the construction joint at the bottom of the footings, while insert for bolts were imbedded at the base of the container for fixing the footings to the container in case of the test with the fixed foundation. The structural calculation was conducted based on the allowable stress method in accordance with the standard structural calculation method by the Architectural Institute of Japan in 1975 edition, and Building Standard Law and the Corresponding Enforcement Order in 1970. The earthquake lateral loading is rectangular distribution based on uniform seismic coefficient for inertia of 0.2. The calculated weights of the roof, 3rd, 2nd and base level including steel weight on the roof and

417 other instruments are 1103kN(roof), 789kN(3F), 789kN(2F) and 855kN(base) respectively assuming 2.4 for concrete density. The section sizes and the reinforcement details are shown in Figure 19.3. The concrete strength of the 1st story was estimated from the cylinder test as 31Mpa for the first bare specimen and 28 MPa for the second retrofit specimen. The strength of rebar of D19 was 384MPa.
1500 1000

CS1 F1 FG6 F1 FG5 F1

CS1 G6 C2

CS2 G5 C1

CS1

Y4

Y4
4000

C3

4000

FG2

FG1

FG1

G2

G1

CS1

S2

S1

S1

G1 C1 G1 C1a G1 C1

B1

CS1

Y3
17000 4000

FG4 F1 F1

FG3 F1

Y3
12000 4000

C3 G4

C2a

G3

FG1a

FG2

FG1

G1a

G2

B1

CS1

S2

S1

S1

CS1

Y2
4000

FG3a F1 F1

C3a

Y2
4000
G2 G1a

G3a

FG1a

FG2

FG1

CS1

S2

B1

S1

S1

CS1

1000 1500

F1

F1

F1

1000

Y1

FG6

FG5

Y1

C3 CS1

G6 C2 CS1 CS2

G5

CS1

1000 2000

4500

3000 6000 8000

5001000

X1
1000 1500 2000 X1 X2 4000 13000 1500 1000 X3

X2

X3

Fig.19.1 Structural floor plan of the specimens with the container at the base
Y4 frame
430
G5

3frame (class room)

2530 800 1200 530 2100 430
C1 C2 G6 C3 C1 G1 C1a G1 C1 G1 C1

W1 G5 G6

SW1 G1

SW1 G1

SW1 G1

2500

8700

W1 G5 G6

8700

2500

SW1 G1

SW1 G1

SW1 G1

400 800 2500

400 800 2500

W1 FG5 FG6

1800
SW1 FG1 SW1 FG1 SW1 FG1

6000 8000

2000

4000

4000 12000

4000

X3

X2

X1

Y1

Y2

Y3

Y4

1frame (corridor)
1200 800 530 1200 800 500
C3 G2 C3 G2 C3a G2 C3

Y1frame
2530 2100 430
C3 G6 C2a G5 C1

SW2 G2

SW2 G2

SW2 G2

W1 G6 G5

2500

8730

8700

2500

W1 G5

SW2 G2

SW2 G2

SW2 G2

G6

400 800 2500

400 800 2500

SW2 FG2

SW2 FG2

SW2 FG2

W1 FG6 FG5

4000

4000 12000

4000

2000

6000 8000

Y4

Y3

Y2

Y1

X1

X2

X3

Fig.19.2 Structural elevation of the bare reinforced concrete specimen

columns section
C1 C2 C2a 400 C1a 400 C3 400 400

walls section footing

C3a 120 150 F1 F2 (for load cell) BD=10001000 BD=14001400 H=800 H=800 X, Y, Z: D13@150 X, Y, Z: D13@150

400 400 8D19 10D22 main bar hoops 2D10@100 X:2D10@100 Y:2D10@100

300 8D19 2D10@100

400 8D19 2D10@100

slabs
SW1 SW2 W1 RF 3F 2F 1F

girders section
G1 G1a 500 G2 500 G3 G3a 600 G5 G6 400

150 mm D10@150 120 mm D10@150 120 mm D10@150 100 mm D10@100 FG main bar D10@200 D10@300w FG2 G4 600 G4a 800 600 800

main bar

250 300 250 250 300 300 250 300 top 2D19 top 3D19 top 2D19 top 2D19 top 3D19 top 3D22 top 3D22 top 3D19 bottom 2D19 bottom 2D19 bottom 3D19 bottom 2D19 bottom 2D19 bottom 3D19 bottom 3D22 bottom 3D22 middle 2D19 middle 2D19 middle 2D19 middle 2D19 middle 2D19 middle 2D19 middle 2D22 middle 2D22 stirup 2D10@200 2D10@200 2D10@200 2D10@200 2D10@200 2D10@200 2D13@200 2D13@200

Fig.19.3 Section size and reinforcement details of the specimens

2100 430

2530

800 1200 530

2530

2100

418 To validate the effect of seismic strengthening with attached steel braces, the second specimen, or the retrofit specimen, was tested. The specimen was constructed in the same way as the bare RC specimen and then the steel braces were attached to the center span of the two outer frames in the longitudinal direction as shown in Figures 19.3 and 19.4.
0 3 5 2
0 3 5 0 0 2 1 0 0 8 0 0 5 0 0 2 1 0 0 8 0 0 5 0 0 2 1 0 0 8
chemical anchor D19@150
300 100

200

beam column

200

0 0 5 2 0 0 5 2

200

stud 19@150
230 114 250

column

mortar

150

Fig.19.4 The steel braces attached to X1 frame

0 0 5 2 0 0 5 2

0 0 5 0 0 8 0 0 2 1 0 0 5 0 0 8 0 0 2 1

150 150

0 3 5 2

0 3 5 0 0 8 0 0 2 1

chemical anchor D19@150 beam

100 150

column mortar
230

230

column stud 19@150

Fig.19.5 The steel braces attached to X3 frame

A conventional detail of the connection to the existing frame was used in X1 frame on the corridor side, while a new detail was developed for the X3 frame on the room side. The conventional joint was designed following the Japanese design manual for attached frame using chemical anchors. The outer frame on the other room side was designed with a new detail using the inner steel frame and penetrating post tension bolts into the girders. Design lateral strength of the steel braces was 120tonf and 100tonf for X1 and X3 frame, respectively. The detailed design procedure will be reported elsewhere. The spandrel walls of the retrofit specimen were isolated from the columns using shear-type seismic slits. Lateral reinforcing bars in the walls were remained through the slits. The relations are based on the calculated strengths and deformability of the columns assuming the story mechanism. The cumulative seismic strength coeffi-

75

250

150

0 0 0 4
0 0 0 4

0 0 0 4

0 0 0 4
0 0 0 4

0 0 0 4

419 cient CT is the sum of the ultimate strengths of the first story columns, either major value of shear and flexural strengths, expressed in terms of story shear coefficient. The ductility index is taken as F=0.8 corresponding to the ultimate story drift of the short columns (1/500rad) in the shear failure, F=1.0 (1/250rad) for the shear dominant columns and F=1.27 (1/150rad) and more for the flexural columns, which are derived from the ratio of shear strength to flexural strength. In the sum of the strength index, the different yield deformations of the columns are considered with approximate assumption and associated strength contribution factor at the calculated ductility level. The seismic index of the designed specimen is Is=0.51(F=0.8, CTSD=0.63) due to the short columns in X1-frame on the corridor side, which is less than the standard objective performance levels: Is=0.6 for ordinary buildings and Is=0.7 or 0.75 for school buildings. The bare RC building should be retrofitted in Japanese practice. The calculated maximum base shear coefficient was 0.63 at the failure of short columns, which was a little less than or equal to the estimated friction coefficient at the base.

19.2 Shake Table Tests and Static Friction Tests

The dynamic tests for the two specimens were conducted at E-Defense from September to November 2006 as listed in Table 19.1. The test procedure and observed structural damage levels are outlined in this section. The first dynamic test on the bare specimen was carried out from Sep 22 to Oct 2. The target earthquake record of all the test runs was the same, JMA_Kobe 1995[2], while the amplitude ratio to the original level was varied from 0.1 to 1.0 as shown in Table 19.1. The base was not fixed but with construction joint to the container through Run-1 to Run-4 allowing sway and rocking mechanism. Although the bolts were used to fix the footings before RUN5, the sway mechanism has occurred because the stiffness and the pretension of the bolts were not enough. In the RUN6, the steel plates were placed in the backfill soils between the footing and the container to prop the sway movement. The response spectrum using the measured acceleration data on the container base in Y direction is shown in Figure 19.6. The dotted line in this figure shows the spectrum obtained from acceleration data on the 1st floor. The remarkable difference of spectrums between two records is found between 0.2 ~ 0.5 (s) at Run-4 in the bare specimen. The dashed line shows the target spectrum of the original earthquake record JMA_Kobe. The damage level was evaluated and identified based on the Japanese standard after each run as listed in Table 19.1. Representative measured response values are also listed in the table, the maximum inter-story displacement in terms of rotation angle and story shear in terms of shear coefficient. Minor damages were observed in the bare specimen after Run-4 under the 100% of JMA_Kobe. A story collapse occurred due to shear failure of short col-

420 umns in Run-6 as shown in Figure 19.7 under almost the same level as Run-4. The obvious different damage level between the two runs may be attribute to the sway dislocation at the base in Run-4.
4000 acceleration (gal) 3000 2000 1000 0 Bare frame specimen 0.5 1 period (s) 1.5 20 Braced frame specimen 0.5 1 period (s) 1.5 2 Run4 (Input) Run4 (1st floor) Run6 (Input) Elastic Period Run4 (Input) Run4 (1st floor) Run6 (Input) Run15 (Collapse) Elastic Period

(a) Bare RC specimen (Run-4)

(b) Retrofit specimen with braces (Run-4)

Fig.19.6 Response spectrum of the input accelerations Table 19.1 List of Shake Test Runs: Input Motions, Base Conditions, Damages and Responses
(a) Bare RC specimen Motion Base Damage Shear Kobe 10 N No 0.13 Kobe 25 N Slight 0.28 Kobe 50 N Minor 0.60 Kobe 100 N Minor 0.96 Kobe* 75 FB Moderate 1.08 Kobe**100 FBP Collapse 1.30 (b) Retrofit specimen with braces Motion Base Damage Shear Kobe 10 N No 0.103 Kobe 25 N Slight 0.307 Kobe 50 N Slight 0.494 Kobe 100 N Minor 0.866 Kobe 100 FCP Minor 1.14 Kobe 130 FCP Minor 1.33 Takatori120 FCP Moderate 1.56 BCJL2, 210 FCP Moderate 1.69 Sine wave FCP Collapse 2.00 Motion: Input records and amplitudes in terms of the percentage ratio to the original intensity, *: The target amplitude was 100% while the input in the test was 75% due to mistake, **: The effective amplitude was equivalent to 100% Base: N:Free, FB:Fixed with bolts, FBP:Fixed with bolts and prop against sway, FCP:Fixed with bolts at the corDrift ners and prop against sway 1/10000 Damage: Observed damage level 1/5000 identified after each run based 1/2000 on the post-earthquake damage evaluation standard by JBDPA 1/666 1/588 Shear: Maximum 1st story shear in Y-dir in terms of story shear 1/344 coefficient 1/250 Displacement: Maximum 1st story drift in Y-direction in terms of 1/125 inter-story rotation angle (rad.) 1/30 Drift 1/5000 1/2000 1/700 1/250 1/180 1/20

Run 1 2 3 4 5 6 Run 1 2 3 4 5 6 8 13 15

Date Sep25 Sep25 Sep27 Sep29 Oct2 Oct2 Date Oct23 Oct23 Oct23 Oct27 Oct30 Nov1 Nov1 Nov6 Nov6

The second dynamic test on the retrofit specimen was carried out 15 times from Oct 23 to Nov 6. The target earthquake records were the same with the first test until Run-4 while the amplitudes and the records were changed to verify the higher performance level. The footings of the central bays (Y2 to Y3) with attached braces were not fixed allowing rocking motion until the end of testing although the sway and rocking motions at the other outer frames (Y1 and Y4) were fixed after Run-5.

421 The retrofit specimen survived with minor damages the extreme severe motions from Run-7 to Run-14, such as 130% of JR Takatori record [3] and 210% of BCJL2 [4], which were the capacity limit of the shake table. Comparing with the collapse behavior of the bare specimen, the effect of the attached braces on upgrading seismic performance is obvious from the observed damages. To identify the ultimate capacity of the retrofit specimen, ultimate strength and deformability, sweep sine wave motion, which varies from 5Hz to 1Hz with 0.75G, was input in Run-15 to induce the resonant responses to collapse within the capacity of the shake table. The apparent damages due to uplifting deformations were not observed. On the X1 side with the conventional detail of the connections to the existing frame, the joint above the second floor were stable and undamaged. However, crushing at the joint mortar occurred associated with bucking at the foundation level and the steel beam at the base was distorted and detached from the foundation beam. The joint to the first story column was also heavily damaged. On the other hand, the new connection details on X3 side basically survived with minor damages although the bucking of the braces occurred as well. It may be concluded that the new connection detail is more efficient and stable even in the extreme ultimate stage.

(a) Minor damage after Run-4 (free base)

(b) Near collapse after Run-6 (fixed base)

Fig.19.7 Damages to the bare specimen after Run-4 and Run-6

The relations between the base shear coefficient and the first story drift angle are shown in Figure 19.8 for the bare RC specimen and the retrofit specimen. The maximum base shear coefficient in the hysteresis of the bare specimen on the left reached 1.0 at the maximum drift angle of 1/250 during Run-4 and 1.2 at 1/180 during Run-5 The hysteresis of Run-6 of the bare specimen is shown in Figure.39 at left compared with the retrofit specimen. The maximum base shear coefficient of the bare specimen was 1.35 at the rotation of 1/200. The maximum drift responses of the bare specimen attained over 1/25 rad due to brittle shear failure of the columns. Therefore, it may be concluded apparently that the reason of slight damages at Run-4 much less than those at Run-6 was owing to the dislocation of the foundation at the base. On the other hand, the unexpected slight damages during Run-5 were due to the input level (75%) less than the target level (100%). However, the effect of the base dislocation occurred in Run-5 should be investigated further,

422 which was much stable hysteretic behavior than those in Run-4. The stable hysteretic behavior is to be simulated in consideration with the properties of the bolts used for fixing the footing to the container. The responses of the retrofit specimen did not yield with small stiffness degradation, and the maximum drift angle was 1/600 at Run-5 and 1/400 at Run-6. The retrofit specimen could survive with stable behavior although progressing minor damages the much more severe motions of Run-7 through Run-13, such as JR Takatori 130% and BCJ 210%, the levels of which were upper bound of the shake table at E-Defense. The footings in the mid-span of Y-direction (Y2, Y3) were not fixed so that vertical movement could occur. However, the apparent rocking or uplifting behavior was not observed. The maximum uplifting deformation was measured as 9mm in Run-13. The high performance against these severe motions was verified. Therefore, to identify the ultimate strength and deformability of the retrofit specimen, the artificial resonance and sweep motions consisting of sine waves 5Hz to 1Hz with maximum acceleration of 0.75G was input at Run-15. The maximum base shear of the retrofit specimen attained 2.0 in terms of shear coefficient at the rotation angle of 1/50. The stiffness degrading was due to yielding and buckling of steel braces and not by rocking of the structure, so that the deformability of the structure was not much increased from the bare specimen.
1.5 Base shear coefficient 1.0 0.5 0.0 Run6 Run4 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 Run15 Run6 Run4

-0.5 -1.0 -1.5 Bare frame specimen (Run1-6) -0.04 -0.02 0 0.02 0.04 lateral drift ratio of 1st story (rad)

Base shear coefficient

Braced frame specimen (Run1-15) -0.04 -0.02 0 0.02 0.04 lateral drift ratio of 1st story (rad)

(a) Bare RC specimen (Run1-6)

(b) Retrofit specimen with braces (Run 1-15)

Fig.19.8 Overall hysteretic responses of the bare frame and the braced frame

Reduction of the input motion and the damages to the buildings in case of the free bases were induced by the nonlinear sway motion or slip behavior at the base. The observed slip behavior at the base is measured and expressed in the hysteretic relations between the friction force and the slip deformation as shown in Figure 19.9. The lateral base dislocation was measured using the displacement meters, which was the slip deformation at the construction joint between the foundation bottom and the slab surface of the container. The lateral shear acting on the surface, defined as the foundation base shear, was derived from measured horizontal accelerations multiplied by the masses at upper floors including the foundation level. The total mass of the specimen is 3538kN.

423 Static tests were also carried out to identify the static friction coefficient between the foundation base and the surface of the container slab including the passive resistance of the surrounding infill soils. The static test was conducted on the table after the test with free foundation and then the building bases were fixed to the container to conduct the dynamic test in case with the fixed base. Oil jacks were placed on one side of the container at the upper level of the footings and the foundation beams to push the upper structure in the other direction with the reaction of the container beam, and the static load versus dislocation relation under the constant vertical load was obtained from the test as shown in Figure 19.10. The dislocation has occurred when the loading force increased up to 2500 kN for both specimens, which was corresponding to 70 % of the total weight. The restoring force in case of the bare specimen falls down by 500kN probably with the steep progress of partial dislocation or slip at the base. However, the resistance gradually recovered and increased up to 85% with regular interval of these cut downs.
1.00 foundation base shear coefficient 0.75 0.50 0.25 0.00 foundation base shear coefficient 1.00 0.75 0.50 0.25 0.00

-0.25 -0.50 -0.75 -1.00 -20 Bare frame specimen 0 20 40 60 lateral drift of the foundation (mm)

-0.25 -0.50 -0.75 -1.00 -20 Braced frame specimen 0 20 40 60 lateral drift of the foundation (mm) 80

80

(a) Bare RC specimen (Run-4)

(b) Retrofit specimen with braces (Run-4)

Fig.19.9 Hysteretic responses between the friction coefficient and the slip at the base
1 0.8 0.6 foundation base shear coefficient 0.4 0.2 0 1 0.8 0.6 0.4 0.2 0 10 braced RC frame 1st loading test (positive) 2nd loading test (negative) 20 30 40 50 60 lateral drift of the footing (mm)
0.78 0.69 0.66 0.62 0.76 0.74 0.64 0.55 0.89

10
0.833

20

bare RC frame 1st loading test (positive) 2nd loading test (negative) 30 40 50 60

Fig.19.10 Static test on base friction coefficient

424

19.3 Fail-safe System against Extreme Motions

The input motion on the 1st floor was the same as that in the shake table in the levels of Run-1 through Run-3, while the dislocation occurred during Run-4 and Run-5 in the bare specimen and Run-4 in the retrofit specimen. The base was free until Run-4 while the fixing method was different in Run-5: only using the high strength bolts for the bare specimen and the bolts at the corner footings with the prop beams for the retrofit specimen. At Run-5 of the bare specimen, the foundation base shear keeps much more stable friction coefficient from 70% to 100% of total weight during the response without uplifting slip behavior of the base footings. The maximum drift of the base was 30mm. The stable behavior is because of the fixing bolts which might be much more effective to the prevention of the partial uplifting and detachment than to the increase of lateral resistance in friction. It might still be controversial whether the obvious slip behavior observed at the base of the specimen would occur or not at the foundation base, especially in existing building structures in site with actual foundation. However, the base detail with construction joint at concrete surface can easily be designed and constructed similarly to the specimen in case of spread foundation and pile foundation as well. If we take into account the elastic stiffness of underneath soil in addition, the effect of the input loss would be much more even under the lower level of acceleration. Therefore, the behavior could be positively taken into account as a fail-safe design against the extreme earthquake motion exceeding the design level, especially for low-rise or medium-rise typical reinforced concrete structures with relatively high strength and limited ductility. As verified as above by the full-scale shake table tests, the sway or the slip behavior at the construction joint of concrete base would obviously reduce the damages to superstructures under extreme motion compared to the case with fixed foundation. The reduction was not effective up to the design level (Level 2: 0.4G) or lower, but was obvious under the higher level such as twice the design level with maximum acceleration of 0.8G. Therefore, it would be feasible to construct a fail-safe system against extreme motions exceeding the design level, as shown in Figure 19.11. As shown in the figure, a flat concrete slab with standing wall, pool-shaped container, is to be constructed at the bottom of the base foundation supported by piles or foundation soil. The building may be designed normally but preferably with relatively higher lateral strength, or strength-dominant type using walls and columns with wing walls/ or with steel braces. The slab surface shall be leveled and smoothed to control the friction coefficient. The coefficient would be around 0.6 statically, which would be reduced in dynamic response down to 0.4 or lower as is observed in the shake table test. The slip deformation would be 50mm to 100mm even in case the motion like JMA Kobe, while the damage to the building would be minor owing to the slip behavior at the base. The slip deformation and the clearance may be 100mm or less, which is much less that required for base isolation system. The

425 back fill soil may be replaced with alternative and economical material such as styrofoam and so on. The damping effect could be expected by the surrounding soil, which should be investigated in detail in the future research. Also the flexible joint should be used for the joint of the piping system into the building so as to ensure the functional use even after an extreme motion. It would be much easier and economical to ensure the fail-safe performance against extreme motions than to design and construct with special devices such as with base isolation devices, dampers and so on. If the constant friction coefficient is assumed, the required strength can be calculated so as to make the response to be elastic under extreme motion. It has been derived also theoretically that the upper bound of the required could be formulated as constant value regardless the fundamental period of the structure, derived from the effective parameters, such as the friction coefficient and mass ratios of the building and the foundation[5]. It should be noted that the theoretical required strength of the building in term of shear coefficient is constant insensitive to the frequency components of the input motion which is generally higher than the friction coefficient at the base.
Superstructure: reinforced concrete or any other structures, normally designed, preferably as strength-dominant with walls etc. Foundation footings connected with stiff beams Backfill soil or clearance gap Concrete to concrete joint Concrete pool with flat slab Piles or supporting soil layer

Figure 19.11 The simple fail-safe system against extreme motion`

19.4 Conclusions
Full-scale three-dimensional earthquake simulation tests on three-story reinforced concrete school building structures with concrete-to-concrete bases were planned and conducted at E-Defense, from which the followings may be drawn: (1) The maximum response base shear of the specimens in dynamic tests attained over 1.3 for the bare specimen and 2.0 for the retrofit specimen in terms of base shear coefficient. These values were higher than the calculated lateral-load

426 carrying capacity based on the standard seismic evaluation and still higher than those estimated from the detailed preliminary frame analysis. One of the reasons for the underestimation is the effect of the spandrel wall on beam strengths. (2) The maximum response inter-story drift and associated damages of the bare specimen with flexible foundation was much less than those in case with fixed foundation. This is owing to the dislocation at the construction joint between the footing and the leveling concrete, which would have great effects on reduction of damages under extreme ground motion in general. (3) The foundation base shear coefficient in dynamic response of the bare specimen at the start of the dislocation was around 0.7. The value corresponded to or a little less than the friction coefficient verified from the static test. However, the dislocation occurred even at much smaller coefficient level during dynamic responses. This is also the case in the retrofit specimen. The observed behavior could be applied to the fail-safe seismic design of reinforced concrete buildings against possible extreme earthquake motions. The fail-safe system is proposed with higher strength and limited ductility, which incorporate the slip behavior at the base positively as a kind of safety fuse.
Acknowledgements The full-scale dynamic test was carried out by the research group with the authors as PIs at EDefense, Hyogo Earthquake Engineering Research Center, National Research Institute for Earth Science and Disaster Prevention(NIED) in Miki City, as a part of National research project on mitigation of major disaster in major city/Theme II Improvement of seismic performance of structures using E-defense/Reinforced concrete structures (DaiDaiToku/RC project) of MEXT, Ministry of Education, Sports, Science and Technology, Japanese Government. The joint efforts on the test by the research group and the support of the NIED and the MEXT are gratefully acknowledged. References [1] Toshimi Kabeyasawa, Toshikazu Kabeyasawa, Taizo Matsumori, Toshinori Kabeyasawa, Yousok Kim: Full-scale dynamic collapse tests of three-story reinforced concrete buildings on flexible foundation at E-Defense, Proceedings of 14th World Conference on Earthquake Engineering, Beijing, China, October 12-17, IAEE, S15-002, 2008. [2] Japan Meteorological observatory Agency, an earthquake record in the south part of Hyogoprefecture Jan.17.1995 [3] Nakamura, Y., Uehan, F. and Inoue, H.: Waveform and its Analysis of the 1995 Hyogo-KenNanbu Earthquake (II), JR Earthquake Information No. 23d, Railway Technical Research Institute, March 1996 (in Japanese) [4] Building Center of Japan, "An artificial earthquake record fitted with Japanese design response spectrum (maximum 60 kine, 356 gal) coded by Building Center of Japan (BCJL2)", The Building Center of Japan. [5] Toshikazu Kabeyasawa, Toshimi Kabeyasawa: Nonlinear soil-structure interaction theory for low-rise reinforced concrete buildings based on the full-scale shaking test at E-Defense, Proceedings of 14th World Conference on Earthquake Engineering, Beijing, China, October 1217, IAEE, S15-016, 2008.