CONCRETE LIBRARY OF JSCE NO. 21, JUNE 1993

VISUAL SIMULATION FOR DYNAMIC RESPONSE OF REINFORCED CONCRETECOLUMNS SUBJECTED TO EARTHQUAKE MOTION

(Translated from Proceedings of JSCE, No.451/V—17, Aug. 1992)

Shoji IKEDA Takahiro YAMAGUCHI

SYNOPSIS

In order to generate the dynamic response of reinforced concrete structures ina visual manner without a shaking table test, a simulated dynamic responsevisualization system has been successfully developed. It utilizes a computer~controlled video tape recorder in a pseudo—dynamic testing. By playing thetape, the visual dynamic response can be observed in real time scale, althoughthe experiment is carried out using pseudo-dynamic testing. Visual dynamicsimulations of four types of reinforced concrete columns during severe earth—quake motion have been carried out using this system, demonstrating that it is auseful means of visualizing the characteristics of ongoing seismic failure.i

Keywords : reinforced concrete columns, pseudo-dynamic testing,visual simulation, dynamic response, video tape recorder

iii:iiiii-|—-a_|-|n-nu_nu-uuu-nu-unn-n|-uu-u¢-p|-|-4--_-_-_-—--n-|--n-|—n-_n—|n-n_|n-on-uq-uu-n-||—»¢|—-_-1-11--1---n-nun-:1-1unu¢n-magnum-u||—p;—uq—u11--1iiiiiiii

JSCE Member Shoji Ikeda is a professor of Civil Engineering at Yokohama NationalUniversity. He is a member of both the concrete and the structural engineer-ing committees of the JSCE. He was director of the Japan Concrete Institutefrom 1985 to 1987. He is now a director and vice—president of the Japan Pre-stressed Concrete Engineering Association. He served as president of the JapanGas—Pressure Welding Association from 1989 to 1991. He is a member of IABSE, afellow of AC1, and vice—president of FIP. He graduated from the University ofTokyo in 1960 and obtained a Doctor of Engineering degree in 1974. He special-izes in the design and mechanics of reinforced and prestressed concrete andcomposite structures.1111111iii:-n—-m-nan-u_nn_n|--n-no-anon-u-nun;-Q‘,-Iu-no-p1.-11--—--_---na-—-n—u@i-|_nn_nn_|a.-|o¢¢u-un~un-|—u|—;,—q_—-—------:1:-1ihnunn-mun-~n¢.|111----¢-iiiiixyii

T. Yamaguchi is a research associate in Civil Engineering at Yokohama NationalUniversity. His research interests include the seismic design of reinforcedconcrete structures and the response behavior of reinforced concrete membersunder static and dynamic loads. He is a member of the JSCE, JCI, and JPCEA.111111111-1-n¢—u11|--|-um-1»-nu-nmumun-up---—-|—---_|—-_-----I--nii¢—|-_-an-ntnn-unmann-up-pa-u_|q—||—u-—---—-a-u¢—n‘-1-|_|-1-aqupuuunu-up-|—n|@i|—--iiiiiixyi

1 . INTROI)UCTION

Pseudo-dynamic testing is an experimental method in which the dynamic responseof a structural system during earthquake motion is obtained by a combination ofsequential loading and on-line computer analysis. The pseudo-dynamic testingmethod was developed by Hakuno, et a1.[1] and since then has been used in muchresearch work[2]. The method is also referred to as computer-actuator on-linetesting, hybrid experimentation} and the on-line comptlter test COntrOl meth-od[3]. Since loading and on-line computer analysis are carried out alternatelyat each stepl PSeudo-dynamic testing takes much longer to yield the seismic dis-placement than an actual seismic movement would. Howeverl PSeudo-dynamic test-ing has following advantages over shaking table tests: (1) The method can treatspecimens of any size; (2) The natural period of the structural system can bearbitrarily assigned; and (3) Ongoing seismic failures can be observed in detailbecause they arise slowly and loading can be stopped at any time if necessary.On the other handl Since pseudo-dynamic testing.is a discrete quasi-static load-ing method, the dynamic response of the structural system cannot been observedon a real time scale.

In this study, the development of a simulated dynamic response visualizationsystem which generates the dynamic response of a reinforced concrete structuralsystem without shaking table tests was described. The system is based on acomputer-controlled video tape recorder(VTR) and pseudo-dynamic testing methods.Using this system, the visual dynamic simulation of four types of reinforcedconcrete column during a severe earthquake was carried out.

ai EaAn 9E BBB99Qu mG ARBu_N OF DYNAMICRESPONSE

As shown in Figure 1' pseudo-dynamic testing (PD testing) was carried out usingan acceleration record of an earthquake digiti2;ed in 0.02 second intervals(A t). Since loading is applied according to the analyzed response displacementat each A t in PD testing, the time required to obtain the designated responsedisplacement at each A tvaries as a result of the conversion process. Also, areinforced concrete specimen deforms greatly, so the specimen must be connectedto the actuator through a hinge to ensure accurate loading. Since it is diffi-cult to avoid longitudinal looseness at the hinge, however, a feedback controlsystem as shown in Figure 2 was introduced. This controls the displacement ofthe specimen using a displacement transducer attached directly to theUsing this systeml the time taken for the displa.cement to converge isat each A t due to the varying number of iterations required, and this

Earl.hquakeacceler-ation5i (t)

F(x(t))

Responseanalysis

mat(t)+ck(t)+F(x(t))=-m2(t.)

(Computer)

r - - - -- -- -- -

.[ i Kit)IIF(Ii(

I(IlIIIlI_ _ •`

x(t)

-•`----------•`-l

'

'W F

(Loading)----•`--d------

_ _ I .i

Fig.1 - Conceptualprocedurefor pseudo-dynamictesting

- 46-

IIJIa[l

ZIIII

J

specimen.differentis one of

( I ( t ) ) J I

ltt7j ,

Computer

Responseanalysis

Yes

T.D.

S.D.

r --------

l,No

error

ITargetdisp.

Targetvolt

; converter-Diff-e-r-encein disp.

Restoringforce

Disp.

T.D.:Target displacementS.D.:Specimendisplacement

Inner DLTof actuator

Applyingdisp.

disp.:displacement

Voltfor load

Specimen

Load

Voltfor specimencohverter disp.

A/D

Fig.2 - Feedbackcontrol

/-Outer

i

transl;

_ _ _ __ _ J

DLTattachedto the specimenDLT:displacement

transducer

the obstacles to visual simulation of the dyna,mic response using PD testing.

The authors have combined PD testing with a computer-controlled VTR to makemovies of the dyna.mic response of a specimen during an earthquake. The simulat-ed dynamic response of the specimen can be visuali2;ed in real time scale as theearthquake progresses by playing the tape.

3. RECORDINGUSING COMPUTERCONTROLLEDVIDEOTAPE RECORDER_•` - - - - -Repeated stopping and starting of the VTRis necessary during PD testing.Furthermor.e, the recording duration of one shot must be very short correspondingto the short interval between each digital value in the earthquake motion.There is also the resitriction that the total recording time must coincide withthe or.iginal earthquake duration. In the recording method the authors adopted,the video camera. was kept on continually and recor.ding was controlled by switch-ing the VTRon and off. The switch signal used to start the VTRwas 5 voltswhile a. pause was marked by 0 volt; these signals were transmitted by the com-puter on line. The VTRfor prLOfessional use was adopted due to requirement fordurability of the recording parts and for the fact that the VTR facilitated aconnecting terminal for remote control. The video timer used in this system hasa o.ool second resolution and, like the VTR, has a connector for remote con-trol. By using this timer, each time corresponding to the time of the earth-quake acceleration record could be shown on the monitor.

A videotape records images at the rate of 30 frames per second. When the re-cording time of the VTRis controlled by signals from the computer, the numberof frames fluctuates by one or two from this target frame rate. If there werefour frames or fewer in one shot, the picture often could not be recorded atall. Thus, it was concluded that the minimum possible recording time for oneshot wasrecordingwas aboutpublicity,reduce theof frames.

o .2 second, corresponding to six frames. The time taken to cause theof six frames, that is, from the pause state to the end of recording'

6.5 seconds. The authors adopted a VTRsystem in consideration ofbut a laser disc recorder might be more suitable because it would

minimum recording time and result in fewer fluctuations in the numberphoto i shows the video apparatus used in this system.

The visibility of cracks on the monitor is influenced by the quality of thevideo camera and the lighting conditions. If white poster color is applied tothe concrete surface, the minimum crack which can be identified on the monitor

- 47-

Photo.1 - Videoapparatus

Eorizontal disp.

AmplifiezZ X.Yrecordar ActuatorLoadcellE]C)•ỳ2hy)

O1•`arQ

IO

'C9Q)y)

FL<

A/D

DComptlter

p•`ct3•G̀jO•` OaB

D/ACOnV

E=C>

J..}

CQAU3

y4•`rTClhOOO

F3<

HE;; VTR

Data

7•` C>d5 O

'B •`.5

ter

DLT

+ I +

. I

Vb I

+.-.

- II

specimenI':

Actuator

ontro11 Horizontal disp.

/

Axialforce

Interfaceunit

Timer

VideoCamera

Fig.3 - Pseudo-dynamicloadingsystemwith recordingsystem

has a width of 0.1mm in a 50cmx 50cmobject area underconditions.

The combined PD loading and video recording system is shownloading system was designed to maintain the status quo whilesystem is in operation.

DLT:DisplaceJnenti ransdtlCer

laboratory lighting

in Figure 3. The PDthe video recording

AimgEmm

In order to verify the propriety of the combined PD and video recording systemla test using an H-shaped steel column wa,s carried out as shown in Figure 4[4].The stiffness of the column within the elastic range in a static loading test

- 48-

Fixedpoint

Disp. i0O

tL)T1

LOO)

Cq

transducer Actuator

E-shapedsteelcoltunn

200x 200

Reactionfloor

Reactionwall

Fig.4 - Loadingconditionofi-shapedsteel cohlmn

wa.s 4.8l kN/mm. An initial part of theeight seconds was used in the PD test.periodmade .?i lth:/cs9i?:n

was assumed to be i

iAqQ

J•`pU

a)Unay)a

8

0

-8

- - - - with recordingsystemwithout system

Tine (see.)

Fig.5 - ResponsedisplaceJnentof i-shapedsteel colum

wave of the E1-Centro 1940 (NS) forA twas 0.02 second and the natural

o second. The maximumacceleration wasorder to keep the loading within the elastic range of the

column. The video recording system was set up such that the VTR recorded foro.2 seconds for every 0.2 seconds of the E1-Centro record, so the total record-ing time would be eight seconds corresponding to the actual duration of therecord.

Figure 5 shows the response displacement-time curves for the H-shaped steelcolumn obtained from these experiments. The curve obtained from the experimentwith the video recording system coincides exactly with that taken without thevideo r.ecording system. This confirms that the newly developed system is notinfluenced by the frequent pauses in loading required by the r.ecording proce-dure. As for concrete nenbers, the experimental results night be influenced bycreep effects occurring during pauses in loading. However) since displacementcontrol is a fundamental process of PD testing, creep deformation is not expect-ed during these loading pauses. It was confirmed that the VTR was able torecord the shape of the specimen at the intended times and intervals and thatloading was a.ble to maintain the status quo during recording.

The simulated dynamic response of the column is visuali2ied by playing back thetape; the playing time was found to be nearly equal to the intended value. Thetime taken to complete the experiment for an original duration of eight secondsof earthquakeonds in theSince the VTReight seconds)

motion was 429case of no VTR

is switched 40the time taken

In this case) when the tape isper second, and the monitorspecimen's dynamic response.

seconds with the VTRsystem while it was 182 see-recording.times forto recordplayed atdisplays

The ratio of time is thus 2.4earthquake acceleration dataone shot was approximately llordinar.y speed, there are five

an almost smooth visuali2:ation

5 . VISUAL SIMULATION OF I)YNAMICRESPONSE OF REINFORCEDCONCRETECOLUMNS____ - - 4IIII- - •`

to i.lasting

seconds.pictures

of the

u &peQiBie& d BBBBEiB9ALai4eLhgd

The characteristics of each rein for.ced concrete (RC) column specimen used inthese visual simulations are summari2:ed in Table 1. Figure 6 gives details of

- 49-

one specimen and its

JSCE's Specification

loading conditions [5],[6]. The specimens are RC columnsstanding on footings. In order to prevent early shear failure, hoop reinforce-ment, the amount of which is equal to 104%-109% of the necessary values of

based on allowable stress design method of 1980 edition[7], was arranged having the spacing of less than o;e half of the effectivedepth (d/2) for each specimen. As indicated in Table 1, the specimens weredivided into four categories by their mechanical properties, such as ratio oflongitudinal reinforcement, the shear span-effective depth ratio (a/d), and theaxial compressive strength. Two specimens of each category were made, one forstatic testing and the other for PD testing; that is, for the visual simulation

Table i - Characteristics of specimens

*1 :

*2 :

*3 :

P S:Static test, PD:Pseudo-dynamictestSize is expressed in terms of JIS ; Dmeans Deformed"and the numberis the nominaldiameterValuesin ( ) showthe ratio of actual hoopreinforcementtothe value given in JSCE's Specifications , 1980 edition.

O0Cr)

IJ,Cl

e-Cql[C

h =300rl

E3A-A

D13D6

Unit (mm)D

OO•`

OOrJ1

0LLl

•`

Reactionfloor

Fig.6

150

- Details of specimenandloadingconditions (PS-1,PD1)

- 50-

Specimen Longitudinal reinforcement Hoop reinforcement Shear span

- effectivedepth rat io

Axia lcompressive

s i; ;2r hType* 1 Size*2 Amount Ratio Designation Space Ratio* 3

(cm 2) (A ) (size) (cm ) (% )

P S - iD 1 3 25.34 2 .82 D 6 7.0

0 .30

(lO6)2 .78 0 .98

P D - 1

P S - 2I) 1 3 25.34 2.82 D 6 13.0

0 .16

(109)4 .63 0.98

P I) - 2

P S - 3D 1 0 ll.41 1.27 D 3 2 .5

0 .19

(104)2.78 0 .98

P D - 3

P S - 4D 1 3 2 5.34 2 .82 D 6 4 .5

0 .47

(107)2 .78 4 .71

P I) - 4

O i sp lacem en tN

tram s du cer P

A L

,i

J A

J} rSl lefofoc:-h:;

:I! IllIiii

lJ[

i!::[;llJl1ll l rl

I Ill(～(I+～// I

of dynamic response. The static and PD test loads were applied using two actua-tors, as shown in Fig.6. One actuator applies a constant axial compressiveforce at the top of the column, while the other applies a hori2;Ontal force ordisplacement. Table 2 shows the mechanical properties of the materials used inthe specimens.

Table 3 gives calculated values and experimental results for the static test. Afiber model, in which the cross section is replaced by many fiber elements, wasadopted to carry out this calculation. Figure 7 shows the stress-strainof the concrete and reinforcement used in the calculation. A pattern ofreversed cyclic loading was as follows : After applying the axial force, azontal force wa.s introduced in one complete cycle of the calculated yield

Curvesstatichori-

load

pdf;,lTa::nmLntt?e6dyi;slha:er;:::tZ;n3;S;i:::hnoe:tw::d:Sregh:u;?e;S;::d;yC::e:;cl26oFTable 2 - Mechanicalproperties of materials

ReinforcementSize*1 Type Yield strength

(MPa)Tensile strength

(HPa)Application

D13 SD30 357 519 Longi. reinf. of No.1, 2, 4

D10 SD30 355 514 Longi. reinf. of No.3

D6 SD35 418 539 Hoop reinf. of No.1, 2, 4

D3 SD30 303 396 Hoop reinf. of No.3

Concrete (MPa)

Compressivestrength Tensile strength Young's modulus26. 5 2 . 26 0 . 24X105

*1 : Size is expressed in terms of JIS ; Dmeans " I)eformed"and the ntlmberis the nominaldiameter

Longi. : Longitudinal reinf : reinforcement

Table 3 - Calculated values and results of static tests

* i :Py is the load at which the tensile strain in the outside longitudinalreinforcement reaches its yield value.

- 51-

Spe c im e n

C a lcu la te d v a ltle S (kN ) E xp er im en ta l re su lt s

C ra ck in g Y ie ld * 1 M ax im um Y ie ld M ax im um D isp la cem en tloa d lo ad lo ad d isp lac em en t loa d at m ax imu m lo adP er P y P u (cm ) (kN ) ( 5 y )

P S - 12 6 . 5 1 3 6 . 0 1 7 3 . 0

0 . 4 6 1 7 0 . 0 3

P D - 1

P S - 21 5 . 7 7 8 . 5 9 9 . i

0 . 9 2 9 6 . i 2

P D - 2

P S - 32 4= . 5 7 3 . 6 9 4 . 1

0 . 2 5 9 1 . 2 3

P I) - 3

P S - 45 2 . 0 1 7 9 . 0 2 0 6 . 0

0 . 5 0 2 1 4; . 0 2

P D - 4

I

CT

(a) Concrete

Sn6ytha:d

Fig. 7 - Stress-strain curvesof materials

so forth until the specimen failed

196

5rCcE5O

r•`

r•`

B98aO•`

IY<

hO

FEj

0

ii.tt

Ii.:..i

•`

r T/I-rh 1

-6y

1 't•` =Sss:;--PS-3

'''''''' PS-4

0 5 10

HorizontaldisplacenKnt (Cm)

Fig. 8 - Envelopecurveson positive side

Figure 8 shows the envelope curvespositive side derived from the load-displacement curves. The- maximum

f::S::eaf::gth;:P;:ifhye:xa;steS::::;d;.rn:;fu6dfni;abdi:Si: t::c:ebSS::i;.e:: ;::maximumload-bearing capacity and failure mode of specimens differed greatly dueto differences in the mechanical properties of the specimens. In the case of

esvp::ium:;;ypS-:h::g;S;fLr:h:cgi:f:Za1!nCsrpa:5;m::t;;Tana:t:ernS1?ngfZ;;a:eg, w;;hdthe buckling of the longitudinal reinforcement in specimen PS-4. In the case ofspecimens PS-2 and PS-3, bending failure occurred due to the growth of cracks atthe bottom of the columns. As for specimen PS-2, evident deterioration was seenin the concrete over a wide area of the column bottom.

The initial data, used in the PD test are shown in Table 4. The adopted dampingratio was 0.05 to reflect ordinary RC members. The vibration mode was assumedto have one degree of freedom. In order to carry out pD testing atacceleration response of the E1-Centro wavel the natural period ofPD-1 to PD-4 was a,ssumed to be 0specimen was 187 tons, 54 tons,case, the value of magnification2.5. Here, the virtual lumpedmodulus at the yield loads ofstatic tests. The limit statestipulated in Chapter 9 of the[8]. v4 is a reduction factorability of the structllre afterability after an earthquake was

5 seconds and the virtual lumped mass186 tons, and 227 tons, respectively.in the acceleration response spectrum

I;A6Eu;;i;;8u6u;lt;;;;;; ;pyegif;1cl:;lil.:lslA;for earthquake forces accor.ding to thethe earthquake. In the Specification,classified into four levels of damage

the largespecimensfor eachIn this

was aboutmass was a value calculated using the secanteach load-displacement curve obtained in themodification factor (v^) shown in Table 4 is

Concreteservice-service-

according

c.t:::s,:nd!6t.ylf;Z;lsz::;:?!,6,figisf!a;:emenfp::;:ee;

to the maximumresponse displacement in such a waycorresponds to sound condition, a 26., displacementa 36 v-displacement c&r&66La; iT.-mYegiunL-a;;;;Ag':;p;nydgtuSv;i;;;;;AcA;n;Vi;;;vg;lruTL:

and amaximumacceleration

was adjusted so as to correspond to the significant damage (46y) lev-el and

- 52-

o 'c k

(23 .5W a )

(-0.000 1)0 8 e t E c t

i o .65 crJck

jJ :I

6 ,= (Z e ,c 8 - i ):a , o ; &

e /c^ 2i i ,

: o . 2E c.E c t] E Jc k e 'E" icT = E c.e (0.00 2)

0 3 y al I[ (I (l If lE . i

[ II I

_J1e )h

(0 .Ol5)

(b ) R einfo rcexKn t

TablJe4 - Data used in pseudo-dynamictests

* 5y : Yeild displacement

failure level, as shown in Table 4. Here, the failure level is defined suchthat it is an acceleration two times greater than the significant damage level;it results in severe failure. Loading up to the failure level was car.Tied outafter 46 y level loading was finished for each specimen.

The equation of motion for the spring-mass system was solved numerically by thecentral difference method with A tas the integration time interval. The feed-back control system as shown in Fig.2, with the allowable error in displacementset to 0.04mm, was adopted during loading using the actuator.

The recording system was operated such that a VTRwas set to record a stillpicture for 0.2 seconds for every 0.1 seconds of the earthquake accelerationrecord used in the PD test. Thus, the total recording time was 32 seconds,corresponding to twice the elapsed tine of the earthquake record. This 0.1-second recording interval was selected since it suited the natural period of thespecimen, which was assumed to be 0.5 seconds, and the picture was played backat double speed to generate a dynamic response lasting 16 seconds. This re-sulted in smooth motion since there were 10 pictures per second.

u B39BEiBBBiai EBBBiL&9f B9 L9BL d yiBBd siBgiaLi32B

Figur.es 9 and 10 show the relationshipobtained from PD tests at the failuretics of t,he response displacement arein which the diagonal cracks extended.columns at the times shown in Figures

between response displacement and time asloading level. The fr.equency characteris-almost identical except for specimen PD-1,

Photos 2-5 show the damaged features of9 and 10. These are still pictures taken

from the monitor using a video printer, and the numbers in the photographsrepresent the time (in seconds) into the E1-Centro (NS) earthquake wave. Hence,it wasgraph scrushingoccurredPD-1 had

thought that the minimumcrack width which could be seen in the photo-was larger than that on monitor screen. As these photographs show, the

of concrete at the bottom of the specimen PD-4 column had alreadyby 2.0 seconds. Within 4.5 seconds, the diagonal cracks in specimen

already greatly extended and crushing of the concrete at the bottom ofthe specimen PD-2 column had already occurred. Up to 16 secondsl Which was theend of the loading, failur.e mode of each specimen was not the same as in thestatic test. The damaged features indicate that not only the amount but alsothe spacing and diameter of hoop reinforcement at the bottom of a column are amatter of significance as regards ductility improvement. For example, theductility of a column under severe earthquake acceleration may be improved ifthe spacing of the hoop reinforcement at the bottom is reduced below d/4 accord-

- 53-

Spec iLnen P D - i P D - 2 P D - 3 P D - 4

Ear thquake wave EI Centro l940 (N S) 0.02- l6.0 (see .) A t=0 .02 see.

DaLnPing ratio h 0 . 0 5

N atural period T (see) 0 . 5

Secant stiffness at the yield load Ko(utV/nl) 2 9 . 6 8 . 5 2 9 . 4 3 5 . 9

V irtual luJnPed mass l1(ton) 1 8 7 5 4 1 8 6 2 2 7

Level of damage Limit statemodification factor v 4

Maximum accelera tion of earthquake wave (m/s2)

4 6 ,:i level 0 . 4 0 . 9 1 i . 8 2 0 . 4 9 0 . 9 9

Failure level 1 . 8 2 3 . 6 3 0 . 9 9 i . 9 8

i :oo30

i 20A 10

&e3 o

wE3;:2S&3-30

-40

-50

2.0 S•` 4/5.4 S5.5S

- PD-1-.-.--. PD-2

(see

:. :I'}i

4.4 4.5S

i....

Timefaiidre level

50

4030

)?S0

-10

-20

-30

-4.0

-50

2.0

5.iS

5.5S

PD-3

PD-4

(see. )

'.I

.iS

failure level

Fig 9 - Curvesof responsedisplacenx!ntagainst time(a)

Fig. 10- Curvesof responsedisplacementagainst time(b)

Table 5 - Experimentalresults of pseudo-dynamictests

* i : Maximumrestoring force occurs under 46Ly loading.The values were obtained from the P-6t curvesrecorded on an X-Y recorder.

OCCurr. : Occurrence

ing to the 1986 Specification. It should be noted that the maximum accelerationapplied to each specimen was adjusted to correspond to its mechanical proper-ties, as shown in Table 4. Hence, the maximumforce applied to specimens in thePD test depended on the specimen. Table 5 shows the experimental results of thePD tests.

The dynamictime scaleI;esponse atactual time

specimens PD-1, PD-3, and PD-4 in which a/d is 2:78, the displacement at the t.pof the column was merely due to the rotation at the bottom of the column and notby deformation of the column itself. In the case of specimen PD-1, which suf-fered shear failure, it was observed that the column moved horizontally withoutrotating at the bottom due to diagonal cracks in the concrete. when the tapewas played back at a speed coincident with the length of the earthquake acceler-ation.record) the dynamic response appeared realistic. Naturally, the simulateddynamic response Can also be watched slowly in order to,clarify the crack pat-tern and so on in detail.

Also, it should be noted that the column displacements seen in these picturesare displacements relative to the ground; that is> the observer is moving withthe.ground during the earthquake. Hence? the displacement of columns and theopening and closing of cracks can be seen clearly because the footings are fixedto the ground. If a system in which the video camera moves according to the

response Of a column during an earthquake can be visualized on anysimply by changing the playback speed of the tape. The dynamicthe failure level was visualized on the monitor in one half of thespeed, and this led to the following observations: In the case of

- 54-

Sp e cim en M ax iJnum re sp on se d isp la ce me n t(Fa ilu re lev e l)

M ax im um * 1r e sto r in g

f or ce4 6 L, le ve l

M ax imu m

sh ea rs tm 5 thT ime of D isp la ce me n t

o c cur r . (se e .) (mm ) P (k N ) (M Pa )

P D - 1 5 . 4 8 3 5 . 5 8 1 8 2 . 4 2 . 2 5

P D - 2 2 . 9 2 5 8 . 2 2 l o g . 8 1 . 3 5

P D - 3 2 . 9 8 1 5 . 4 7 9 8 . 1 1 . 1 7

P I) - 4 2 . 9 8 3 5 . 6 9 2 1 8 . 7 2 . 7 0

calctllated displacementh Of the ground is set up, the dynamic responsecolumn will includeof a shaking tablethe nonitoT may notIn other WOTds? thepoint of an absolute

the gTOund's displacement on the screen, just as intest. However, the displacement and crack patternbe clear, becallSe the coltlmn itself moves with theobserver would see the response of the column from

coordinate systeJA in this case.

Photo.2 - (PD-1)

of thethe caseSeen Onfooting.

the view-

Photo.3 - (PI)-2)

photo.2 , Photo.3- Featuresof deteTioTationat variotlStiws

- 55-

Photo.4 - (PD-3) photo.5 - (PD-4)

Photo84, Photo.5 - Features of deterioration at various times

It has been indicated in many investigations [9] that the response behaviorobtained fTOn a PD testing does not coincide exactly with that of a shakingtable test due to differences in the loading speed. However} it is recognizedthat any difference Plays a minimal role in the main issues of response behav-ior. On the other hand, if a shaking table test were carried out using the sameRC column specimens as in the PD testing to obtain pictures of the dynamic re-

- 56-

sponse, many difficulties would have tothe shaking tablel the need for a verygravitational force on the structuralcamera to obtain relative displacements.

be confronted, such as the capacity ofheavy mass which does not influence thesystem, and the placement of the video

Considering these difficulties, the PDloading test combined with this recording system appears t,omethod of visuali2;ing the dynamic response of RC structures.the loading speed of the PD testing can be greatly increased,become more coincident with the actual dynamic response during

6. CONCLUSION

The results of this study lead to the following conclusions:

(i) A successful PDdeveloped. Itthe earthquakethe specimen'sstep, with thethe earthquake

be a vcr.y usefulFurthermore, if

this system willan earthquake.

loading system connected to a computer-controlled VTR wasis able to record images of a specimen at any time duringacceleration r.ecord. To achieve this, a still picture of

response behavior. is recorded for a short time at each timetiming of the recordingacceler.ation record.

(2) The continuous dynamic response of thecan be seen on the monitor by playing

corresponding to the elapsed time of

specimen under earthquakethe tape containing the

still pictures.(3) Given the present capabilities of VTRs, the best recording time

conditionssequential

for thesestill pictures was set at 0.2 seconds. In the case of an RC specimen witha natural period of 0.5 seconds, a series of play-back pictures with a rateof 10 pictures per second offer.ed a smooth play-back of the dynamic re-sponse on the monitor.

(4) The peculiar behavior and differences in dynamic response among RC columnsof four different types were clearly demonstrated by playing the tape.

(5) An investigation of the influence of mechanical properties such as thediameter and spacing of hoop reinforcement at the bottom of the column byusing this simulation system was found to be extremely effective in helpingto improve the ductility of reinforced concrete structures during severeearthquakes.

(6) The hardware of the recording system, its reliability, and the quality ofthe image will be improved greatly in future. Eventually, it is hoped thissystem will find application in investigations of many different types.

ACKNOWLEDGEMENTS

Part of this study was supported by Grant-in-Aid for Scientific ResearchNo.01460174 given by the Japanese Ministry of Education. For their assistanceduring the experiment, thanks are due to the staff and students of the Depart-ment of Civil Engineering, Yokohama National University.

REFERENCES

[l] Hakuno, M., Shidawara, M., and Hara, T., 'TDynamic Destructive Test of aCantilever Beam Controlled by an Analog-Computer,.I proceedings of JSCE,No.171, Nov., 1969, pp.1-9 (in Japanese)

[2] Hakuno, Motohiko, et a1. , 'tApplication Manual of Hybrid Experiment," Reportof the Ministry of Education in the Form of Gr.ant-in-Aid for. Co-operativeResearch(A) No.63302042, 1990 (in Japanese)

[3] Takanashi, A. and Nakashima, M., 7.Japanese Activities on On-Line Testing,T7Journal of Engineering Mechanics, Vo1.113, No.7, July, 1987, pp.1014-1032

[4] Ikeda, S. and Yamaguchi, T., "Visual Simulation of Dynamic Response Behavior

- 57-

Using Pseudo-dynamic Te?tllT proceedings of the 45th Annual Conference ofJSCE, Vo1.5, pp.648-649, 1990 (in Japanese)

[5] Ikeda' Shoji, et a1., 'fEvaluation of the Hysteretic Behavior of Reinf.reedConcrete Columns with Consideration of Cross-Sectional Properties underSeismic Loading,.7 proceedings of JCI, Vo1.13-2, pp.601-606, 199l (in Japa-nese)

[6] Yamaguchi, Takahiro, et a1., ''Evaluation of seismic response.f RC c.1umnsand its visual simulationlTT proceedings of the 46th Annual Conference ofJSCE, Vo1.5, pp.14-15, 1991 (in Japanese)

[7] Japan Society of Civil Engineers, "standard Specification f.r Reinf.reedConcrete Design (1980 edition),I. 1980, pp.22-25 (in Japanese)

[8] Japan Society of Civil Engineers, "standard Specification for Design andConstruction of Concrete Structures Part i (Design)," 1986, pp.89-98 (inJapanese)

[9] Takanashi, A. and Oi, K., "A Correlation Study in Pseudo-dynamic On-LineTests vs. Shaking Table TestslTT summarize of Technical Papers of AnnualMeeting, Architectural Institute of Japan, 1984, pp.1413-1414 (in Japanese)

- 58-