GROUND LATERAL SPREAD EFFECTS ON SINGLE PILE USING UNCOUPLED ANALYSIS METHOD

Transcription

1 Journa of GeoEngneerng, Vo., No. San-Shyan, pp. 5-, Ln, December et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method 5 GROUND LATERAL SPREAD EFFECTS ON SINGLE PILE USING UNCOUPLED ANALYSIS METHOD San-Shyan Ln, Yu-Ju Tseng, Jen-Cheng Lao 3, C.H. Wang 3, and We F. Lee ABSTRACT Permanent ground deformaton or ground atera spreadng s observed to be the man cause for the dstress of pes embedded n quefed ground. The purpose of ths paper s to use uncouped method for anayss of ground atera spread effect on pes. The computer code, CYCLIC-D (Egama, et a., ) deveoped at Unversty of Caforna at San Dego and accessbe from the web, s used for atera ground deformaton estmaton. Subsequenty, the pe performance, treated as beam on Wnker foundaton, s studed consderng the effect of ground deformaton obtaned from CYCLIC-D. Three centrfuge tested exampes and one rea fed case were studed by the aforementoned method. Reasonabe agreement was obtaned between the predcted and the measured resuts. Key words: Pe, atera spreadng, so quefacton.. INTRODUCTION It s known that the effects of quefacton on pes are often damagng. The atera spreadng usuay s trggered at the sghty ncned sope wth quefabe sos embedded among the so ayers. When the so quefacton was ntated, the quefed sos tend to sde downward aong the ncned surface. Whe the pe s embedded among these movng quefabe sos, the pe can sustan atera force caused by the quefed sos. Serous structura damages can be produced, such as the 9 Ngata earthquake (Hamada, 99), the 995 Kobe earthquake (Tokmatsu, 3) and the 999 Ch-Ch earthquake (Hwang, et a. 3) had eft extensve damage to many pe foundatons of brdges and budngs. Bhattacharya, et a. () proposed an aternatve mechansm of pe faure n quefabe deposts durng earthquakes. It was consdered that the pe becomes unstabe under axa oad from oss of support from the surroundng quefed so, provded the senderness rato of the pe n the unsupported zone exceeds a crtca vaue. The nstabty causes the pe to bucke and cause a pastc hnge n the pe. In terms of so pe nteracton, the method assumes that, durng nstabty, the pe pushes the so. Hence, the atera oad effects are consdered to be secondary to the basc requrement that pes n quefabe sos must be checked aganst Euer s buckng. However, ths method can ony consder one pastc hnge nstead of two pastc hnges, whch were observed at the nterfaces of the quefabe so ayer sandwched between two no-quefabe so ayers. Meyerhosn (99) proposed that pes subect to atera Manuscrpt receved September, ; revsed October 3, ; accepted November 3,. Professor (correspondng author), Department of Harbor and Rver Engneerng Natona Tawan Ocean Unversty, Keeung, Tawan, R.O.C. (e-ma: ssn@ma.ntou.edu.tw). Graduate student, Department of Harbor and Rver Engneerng Natona Tawan Ocean Unversty, Keeung, Tawan, R.O.C. 3 Assocate Researcher, Tawan Constructon Research Insttute, Tape County 5, Tawan, R.O.C. Researcher, Tawan Constructon Research Insttute, Tape County 5, Tawan, R.O.C. spreads resutng from so quefacton mght cause two dstnct faure modes. The frst one s atera pe defectons nduced by ground atera spreads that may resut n the pe reachng ts bendng capacty and hence deveop a pastc hnge. Another faure mode s the combned acton of ack of suffcent atera support due to the reduced stffness of the quefed so and the atera defecton mposed on the pe, may resut n pe buckng. Whether bendng or buckng mode of a pe may deveop depends prmary on the stffness of the quefed so, ength of pe exposed to quefed so, axa oad mposed to pe, and bendng stffness of the pe. However, ony bendng faure anayss was conduced for the evauated case hstores. Ln, et a. (5) back studed possbe faure modes of three case hstores. Whether these pes were faed by ether bendng or buckng mode was re-evauated. The desgn procedures suggested by Tokmatsu, et a. (99) and by JRA (99) were aso used for case hstores evauaton and compared to avaabe observaton resuts. Two bendng faures and one buckng faure among the three studed cases were concuded. In order to understand the performance of the pe durng so quefacton wth the numerca anayses, the paper uses the uncouped numerca anayss to resove ths probem. Frst, the Wnker type mode s used to smuate the so-structure nteracton. The Bouc-Wen mode s used to mode so behavor. At the same tme, the Bouc-Wen mode s aso used to cacuate the pe structure ntegrty whe the pe fracture s trggered. Second, whe consderng the so-structure nteracton durng the so quefacton event, another mportant factor s the change of the excess pore water pressure. To obtan ths souton, the CY- CLIC-D (deveoped at the Unversty of Caforna at San Dego) s used to generate the acceeraton and excess pore water pressure vaues wthn sos, however the possbe effect of the exstence of pe foundaton s gnored, under the correspondng nput earthquake moton. Then the vaues of excess pore water pressures at varous depths can be combned wth the Wnker mode. In order to verfy ths uncouped approach, ths paper aso smuates the resuts of a centrfuge test and studes one rea fed case by the aforementoned method. Reasonabe agreement was obtaned between the predcted and the measured resuts.

2 5 Journa of GeoEngneerng, Vo., No., December. UNCOUPLED MODELING Ths study uses the Wnker mode to smuate the sostructure nteracton caused by the earthquake moton. Consderng the force equbrum between the surroundng so and the pe tsef, the equaton s descrbed as Eq. (). EPIP w( x, t) + m w( x, t) x t = F ( xt, ) + F( xt, ) () S d where w(x, t) s the atera pe dspacement n varous tme step t ; x s depth to the ground surface; m s eement mass of pe; E P s Young s Moduus of pe; I P s moment of nerta of pe; E S (x, t) s non-near so reacton force; E d (x, t) s radaton dampng force of pe. Two maor components of Eq. () are the E S (x, t) and E d (x, t). In the foowng, these two parts are defned based on the so modeng. After the ntaton of so quefacton, the atera force produced by the so movement may ncrease the bendng moment. Furthermore, when the bendng moment ncreases to a certan eve, ths may cause the matera fracture that eads to the strength reducton. Ths study nvoves the pe fracture phenomenon that s descrbed n the foowng secton of moment-curvature reaton of pe.. So Modeng Based on the Bouc-Wen mode, the force resutng from the nonnear sprng aone can be gven as (Wen, 95; Ln, et a., ) F ( x) =α K w+ ( α) K w ς ( x) () S where α s a parameter contros the post yedng stffness; K s a reference stffness; w s the pe defecton at the ocaton of the sprng; w s the vaue of pe defecton that ntates yedng n the sprng; and ς s a hysteretc dmensoness quantty that s governed by the foowng Eq. (3) (Wen, 95; Ln, et a., ) w ς +γ w ς ς +β w ς A w= n n (3) where A, β, γ and n are parameters that contro the shape of the hysteretc oop and are chosen such that the shape of the oop s reasonabe for the type of matera consdered. The maxmum vaue of ς s gven as /n A ς=, when dς / dw= () β+γ The sprng reactons of the pe for coheson-ess sos were gven by Badon and Makrs (99) as + snφs Fs( x) =μ γs d x ζ( x) (5) snφs where d s the pe dameter; φ s s the ange of the so nterna frcton; μ s a nonnear hysteretc parameter; γ s s the specfc weght of the so. Another maor concern of the so quefacton event s the ntaton of the excess pore water pressure. Kagawa, et a. (99) used a reducton factor F to descrbe the reducng the so strength as foows. u u F = ( σ'/ σ ') = ( u) () α α where α u s the experment parameter for sand s.5; u s the pore pressure rato. By combnng Eq. () wth Eq. (5), Eq. (7) shows the non-near so reacton force F S wth the effect of the reducton form the pore water pressure generaton. α + snφ u s FS = ( u) μγsd x ζ (7) snφs When the so-structure nteracton s subected to the sesmc force, the radaton dampng shoud be consdered. Accordng to Badon and Makrs (99), t can be descrbed as Eqs. () and (9)..5 d S S F = Qa ρ V dω( Δ w) () π Q = + π( νs ) where ν s s Poson s rato of so; a s non-dmensona frequency dependent parameter a = ; ω s frequency; d s ωd Vs dameter of pe; V s s shear veocty of so; ρ s s so densty; (Δw) = w (Δw > w ) for the non-near case; (Δw) = Δw (Δw w ) for the near case. Under the effect of the excess pore water pressure, Eq. (9) can be re-wrtten as Eq. (). / Vu L.5 Fd = ( u) + Q a ρsvs dω( Δw) ( Vs + Vp) (9) () where V p s veocty of pressure; V L s vscous veocty of quefed so. Transformaton from frequency-dependent F d to tmedependent F d s detaed n Badon (997).. Moment-Curvature Reaton of Pe The Bouc-Wen mode s aso used to mode momentcurvature reatonshp of the pe and s expressed as (Ln, et a., ) M =α ( E I ) φ+ ( α ) M z () m P P m y where M y s the yed moment; φ s the curvature; α m s a parameter controng the rgdty of the pe after yedng; and z s the hysteretc parameter, whch can be expressed as (Ln, et a., ) z = AMIP B z {sgn( φ z) + } φ () φy n whch sgn( φ z) = f φ z > ; sgn( φ z ) = f φ z < ; φ y s the yedng curvature; and A M and B s the parameters controng the shape of the hysteretc oop.

3 San-Shyan Ln, et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method 53 For concrete pes, once the moment nduced on the pe exceeds a certan magntude, the moment of nerta of the pe may be reduced due to concrete crackng. A sem-emprca moment versus moment of nerta reatonshp s used n ths paper (Ln, et a., ). The sem-emprca form s expressed as I I = I, ( M < M ) (3) ef cr II I II M cr Ief = I + ( I I ), ( Mcr < M < Mu ) M () where I ef s the effectve moment of nerta; I I = I p, the moment of 3 nerta of the non-cracked secton; I II s the moment of nerta of the competey cracked secton where the renforcement has reached the yed strength; M cr s the bendng moment correspondng to the begnnng of crackng; and M u s the bendng moment correspondng to I II. In order to take nto account the effects of fnte sze of pastc regons, the mode chosen for ths study s based on the goba frame member mode proposed by Roufae and Meyer (97), n whch the mode was aso successfuy used for concrete pe anayses (Badon, 997). The eement used s schematcay ustrated n Fg.. The ntermedate degree of freedoms can be factored out by the process of statc condensaton. The eement stffness matrx can be wrtten as [ K ˆ ] = b ( EI p p) ( EI p p) ( EI p p) ( EI p p) 3 3 I I ( EI p p) ( EI p p) ( EI p p) ( EI) p p I I ( EI p p) ( EI p p) ( EI p p) ( EI p p) 3 I I ( EI p p) ( EI p p) ( EI p p) EI p p I I ( EI p p) ( EI p p) ( EI p p) EI p p ( EI p p) EI p p ( EI p p) ( EI) p p I I ( EI p p) ( EI p p) p ( EI p p) ( EI p p) ( EI p p) EI p p ( EI p p) EI p p + + I I ( EI p p) ( EI p p) ( EpIp) EpIp ( EpIp) EpIp ( EpIp) EpI p I I ( EI p p) ( EI p p) EpIp EpIp ( EpIp) EpIp ( EpIp) EpIp + I I (5) whch s defned by M { F} = [ Kˆ ]{ Y} () b 5 E p I p (E p I p ) M y L n whch {Y} and {F} are the vectors of noda dspacement and noda oad, respectvey. Set the noda oad vaues for the nterna nodes to zero, the equaton can be rewrtten as Kbb Kbr yb Fb = Krb Krr yr (7) 7 3 (E p I p ) Fg. Eement used to mode varabe eastcty showng uncondensed of numberng and pastc zones (Badon 997) M y M n whch the subscrpt r ndcates the degrees of freedom to be condensed out. Hence, the modfed eement stffness after condensaton becomes n whch b bb br rr rb [ K ] = [ K K K K ] ()

4 5 Journa of GeoEngneerng, Vo., No., December ( EpIp) ( EpIp) 3 ( EpIp) ( EpI ) p [ Kbb ] = ( EpIp) ( EpIp) 3 ( EpIp) ( EpIp) (9) and [ Krr ] ( E I ) E I ( E I ) E I ( EpIp) EpIp ( EpIp) EpIp + + = EpIp EpIp 3 EpIp EpIp p p p p p p p p EpIp EpIp 3 EpIp EpIp ( EI) EpIp ( EpIp) EpI p ( EpIp) EpIp ( EpIp) EpIp + () ( EpIp) ( EpIp) 3 ( EpIp) ( EpIp) T [ Kbr ] = [ Krb ] = ( EpIp) ( EpIp) 3 ( E I ) ( E I ) p p p p () (E p I p ) and (E p I p ) n Eq. () are obtaned from the momentcurvature reatonshps at the correspondng ends. The pastc regon engths and for ends and are obtaned as M sgn( M) Mcr = L () M + M and M sgn( M ) Mcr = L (3) M + M Snce the pastc zone ength depends on the end-moments at the current oadng ncrement, an teratve technque s requred n whch s progressvey refned unt convergence s reached accordng to the cacuatng steps sted beow: For each oadng ncrement (Badon 997): (a) Cacuate,, (E p I p ) and (E p I p ) for a yedng eements based on the end moments at the prevous oadng ncrement; (b) Cacuate the modfed eement stffness matrx [K b ], gven n Eqs. (5) and (), for these eements and assembe the goba effectve stffness matrx and oad vectors; (c) Input noda forces {F b } b b b { Y } = [ K ] { F} () and cacuate noda dspacements {Y} T r rr br b { Y} = [ K ] [ K ] { Y } (5) (d) Update eement end-moments based on new noda vector; (e) Recacuate,, (E p I p ) and (E p I p ) and check for convergence. Repeat unt convergence s acheved..3 Numerca Procedures for Uncouped Anayss (a) Appy the CYCLIC-D program to smuate the ground response and compute the vaues of acceeraton and excess pore water pressure tme hstores at varous depths; (b) Substtute the acceeraton and excess pore water pressures data obtaned from procedure (a) nto Wnker mode of Eq. (); (c) Cacuate the atera dspacement and bendng moment at the correspondng depth of pe from procedure (b). 3. CENTRIFUGE TESTINGS In order to verfy the accuracy of the materas modeng

5 San-Shyan Ln, et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method 55 proposed n ths paper, the anaytca resuts w be compared wth the measured data from the centrfuge testng resuts. Abdoun, et a. (3) presented the resuts of the centrfuge testng that smuates a snge pe sustanng the atera spreadng force from the so quefacton. In ths study, three centrfuge test cases are compared. Case A As shown n Fg., the dmenson of the box s 5.7 m 5. m.39 m. The embedded mode pe s cm ong wth dameter of.95 cm and the so matera propertes are shows n the Tabe. Ths entre assembe s tested under the gravty vaue of 5 g. Under such gravty, t can smuate a fu-szed pe of m n ength wth 7.5 cm n dameter. To test the so quefacton-nduced atera spreadng effect, the ayout of the mode ncudng three ayers of sos: Top ayer: m cemented sand wth 3.5 of frcton ange and 5. kpa of coheson. Mdde ayer: m quefabe sand (Nevada sand) wth reatve densty of %, dry unt weght of 7.33 kn/m 3 ~ 3.7 kn/m 3. Bottom ayer: m cemented sand wth the same propertes of the top ayer. (Note: the dmenson shown above s the prototype mode.) Durng the testng, the bottom of the box s apped wth an exctaton to smuate the earthquake moton. The exctaton has frequency of Hz, maxmum magntude of.3 g, and cyces. In order to smuate the effect of atera spreadng, the entre box s tted by.. Free head and free tp pe boundary condtons are consdered n the numerca anayss. Case B As shown n Fg. 3, the ayout and the nput moton of the Case B are the same as those of Case A. The maor dfference s the prototype dmenson of pe s m, n whch the bottom of the pe does not extend nto the bottom non-quefed ayer. Free head and free tp pe boundary condtons are consdered n the numerca anayss. Tabe So matera propertes for Centrfuge Testng Mode (after Abdoun, et a. 3) Nevada sand Cemented sand Pe Shear box Prototype Reatve densty, D r (%) D (mm).9 D 5 (mm).5 Specfc gravty, G s.7 Max. vod rato, e max.7 Mn. vod rato, e mn.5 Max. dry unt wet., r dmax (kn/m 3) 7.33 Mn. dry unt wet., r dmn (kn/m 3 ) 3.7 Permeabty at g for D r = % (m/s). 5 φ ( ) 3.5 C (kpa) 5 Length, L(m) Dameter (cm).95 EI (kn-m) Gravty vaue (g) 5 Length, L(m) Dameter (cm) 7.5 EI (kn-m) Gravty vaue (g) Case C As shown n Fg., nput moton of the Case C are the same as those of Case A. However, the maor dfference s the so profes. In Case C, there s no surface ayer of the sghty cemented sand. Free head and free tp pe boundary condtons are consdered n the numerca anayss. To verfy the accuracy of the materas mode proposed n ths paper, the verfcaton procedure ncudes two steps, ncudng CYCLIC-D smuaton and the materas modeng cacuaton. The parameters used for the materas mode are shows n the Tabe 3. 5 g Mode unts Prototype unts cm cm cm A A cemented sand ayer Nevada sand ayer (Dr=%) PP PP PP3 SG SG SG 3 SG A5 A A7 LVDT LVDT LVDT 3 LVDT m cm A3 cemented sand ayer PP SG 5 SG A LVDT 5 LVDT m m α =. A Input Moton Fg. Centrfuge testng mode, Case A (after Abdoun, et a., 3)

6 5 Journa of GeoEngneerng, Vo., No., December Mode unts 5 g Prototype unts LVDT α cm cm cm cm A =. A A A3 cemented sand ayer Nevada sand ayer (Dr=%) cemented sand ayer PP PP SG SG SG 3 SG SG 5 Input Moton A5 A A7 A LVDT LVDT LVDT 3 LVDT LVDT 5 m m m Fg. 3 Centrfuge testng mode, Case B (after Abdoun, et a., 3) Mode unts 5 g Prototype unts cm cm cm A A Nevada sand ayer (Dr=%) cemented sand ayer PP PP LVDT SG SG SG 3 SG A A5 A LVDT LVDT LVDT 3 LVDT LVDT 5 α = A3 Input Moton Fg. Centrfuge testng mode, Case C (after Abdoun, et a., 3) Tabe Matera propertes used for the budng n Mkagehoma, Japan (Tokmatsu 3) Back studed case Pe E p I p (kn-m ) 5 M cr (kn-m) M y (kn-m) 75 M u (kn-m) φ y (/m).5 So Upper non-quefed ayer depth (m). ~.5 γ s (kn/m 3 ).5 φ ( ) 3 Mdde (or upper) quefed ayer depth (m).5 ~. γ s (kn/m 3 ).5 φ ( ) 5 Bottom non-quefed ayer depth (m). ~ 3. γ s (kn/m 3 ).5 φ ( ) 3. CYCLIC-D Smuaton CYCLIC-D was deveoped n the Unversty of Caforna at San Dego (Egama, et a., ). Ths program ams to sove so quefacton-nduced atera spreadng probems. By provdng the so propertes and so profes, t can estmate severa so responses (e.g., acceeraton, spectrum, excess pore water pressure, etc.) under the effect of the base moton. For further nformaton, pease access for detas. In ths step, ths study uses CYCLIC-D to smuate the acceeraton and pore water pressure tme hstores wthn varous depths of the prototype test, ncudng depth of m, m, m, and 9m. Fgures 5 and show the resuts of Case A; Fg. 7 shows the resuts of Case B and Fg. shows the resuts of Case C. Note that the acceeraton tme hstores excted at the bottom of the box of Cases A, B, and C are the same. The data generated n ths step w be mpemented wth the materas modeng to estmate the atera dspacement and bendng moment at the corresponded depth of pe.

7 San-Shyan Ln, et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method 57 A cceeraton (g) m. m. m 9. m Input moton Top cemented ayer (Dr=%) (Dr=%) Bottom cemented ayer 3 Tme (sec) Fg. 5 CYCLIC-D smuaton of acceeraton tme hstores at varous depths of Case A Excess Pore Pressure Rato m. m. m. m 3 Tme (sec) (Dr=%) (Dr=%) Fg. 7 CYCLIC-D smuaton of pore water pressure tme hstores at varous depths of Case B Excess Pore Pressure Rato m. m. m 9. m 3 Tme (sec) (Dr=%) (Dr=%) Excess Pore Pressure Rato m (Dr=%). m (Dr=%). m (Dr=%). m 3 Tme (sec) Fg. CYCLIC-D smuaton of pore water pressure tme hstores at varous depths of Case A Fg. CYCLIC-D smuaton of pore water pressure tme hstores at varous depths of Case C

8 5 Journa of GeoEngneerng, Vo., No., December 3. Materas Modeng Cacuaton Combned wth the resuts from the CYCLIC-D, the materas mode estmates the atera dspacement and the bendng moment of the pe. For the Case A study, Fgs. 9 and show the cacuaton resuts of pe atera dspacement and bendng moment, respectvey. Fgures and show the parametrc study resuts of bendng moment of pe wth or wthout pore water pressure consderaton at varous tme steps and at varous ground ncnaton anges, respectvey. In genera, due to the so quefacton effect at the mdde ayer, t produces the atera movement among the top and mdde ayer. However, the bottom ayer wth non-quefed sos that reman reatvey ntact. Ths acton of atera movement that produces force to appy on the pe, whch aso carry the upper part of pe to move wth so. The ower part of pe, on the other hand, reman reatvey unmovabe. Ths phenomenon produces arge bendng moment to concentrate at the boundary between the mdde quefed ayer and the bottom non-quefed ayer. Compared wth the Case A study, Fg. 3 shows the bendng moment of pe n Case B and Fg. shows the bendng moment of pe at varous ground ncnaton anges. Even though the pe does not extend nto the bottom non-quefed ayer, t st shows a concentraton of bendng moment between the top nonquefed ayer and the mdde quefed ayer. Tabe 3 Parameters used for the Uncouped Numerca Anayss Parameter Shape parameters of hysteretc oop Centrfuge testng A budng n Mkagehoma A. 3 β.9.5 γ.9.5 n.. Nonnear hysteretc parameter, μ. 3 Frequency, ω (/sec).. Shape parameters of moment- A M curvature hysteretc oop B 5 Yedng stffness parameter, α.95.3 So Depth (m) t=5 sec(ths study) t= sec(ths study) t=3 sec(ths study) t=5 sec(abdoun et a.,3) t= sec(abdoun et a.,3) t=3 sec(abdoun et a.,3) (Dr=%) Fg. Cacuaton of bendng moment of prototype pe at varous tme steps of Case A So depth (m) (Dr=%) (Dr=%) (Dr=%) wth wthout pore pressure pore pressure wthout pore pressure (a) 5 sec (b) sec (c) 3 sec Fg. Cacuaton of bendng moment of prototype pe wth or wthout pore water pressure consderaton at varous tme steps of Case A Depth (m) (Dr=%) t=5 sec(ths study) t= sec(ths study) t=35 sec(ths study) t=5 sec(abdoun et a.,3) t= sec(abdoun et a.,3) t=35 sec(abdoun et a.,3) So depth (m) (Dr=%) ncned ncned 3 ncned. ncned So Dspacement (cm) Fg. 9 Cacuaton of atera dspacement of prototype pe at varous tme steps of Case A Fg. Cacuaton of bendng moment of prototype pe at varous ncnaton anges of Case A

9 San-Shyan Ln, et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method 59 So Depth (m) Depth (m) (Dr=%) t =. sec(ths Study) t =. sec(abdoun et a.,3) sec(ths stdy) 35 sec(ths stdy) t= sec(abdoun et a.,3) t=35 sec(abdoun et a.,3) - Fg. 3 Cacuaton of bendng moment of prototype pe at varous tme steps of Case B.... So Dspacement (m) Fg. 5 Cacuaton of atera dspacement of prototype pe at varous tme steps of Case C t =.5 sec(ths Study) t =7.5 sec(ths Study) t =.5sec(Abdoun et a.,3) t =7.5sec(Abdoun et a.,3) So depth (m) ncned ncned 3 ncned. ncned (Dr=%) So depth (m) (Dr=%) Fg. Cacuaton of bendng moment of prototype pe at varous tme steps of Case C Fg. Cacuaton of bendng moment of prototype pe at varous ncnaton anges of Case B In the Case C, Fgs. 5 and show the cacuaton resuts of pe atera dspacement and bendng moment, respectvey. Fgure 7 shows the bendng moment of pe at varous ncnaton anges. Unke Case A, due to ack of the surface cemented sand, the dspacement and bendng moment dstrbutons are dfferent. However, the maxmum vaues st concentrated at the boundary between the quefed ayer and the bottom fxed ayer. By compared wth the actua measured data coected from the centrfuge data, a three cases show a good agreement between the cacuatons resuts and the actua measured data.. A BUILDING IN MIKAGEHOMA, JAPAN (Tokmatsu 3) A case hstory of 35 cm dameter and 3 m ong pre-stressed hgh strength concrete pes supportng a four-story budng n Mkagehama revewed by Tokmatsu (3) who used the pseudo-statc anayss method. After the 995 Kobe earthquake was aso back studed va the method by Ln, et a. (5) for ts possbe faure mode. The moment curvature propertes of the pes and other so propertes are gven n Tabe. In the cacuaton procedure, t aso consders the possbe concrete crackng effect when the bendng moment exceedng the maxmum vaue. Fgures and 9 shows the smuaton of acceeraton and pore water pressure tme hstores at varous depths, respectvey. Fgures and show the atera dspacement and bendng

10 Journa of GeoEngneerng, Vo., No., December So depth (m) (Dr=%) ncned ncned 3 ncned. ncned Fg. 7 Cacuaton of bendng moment of prototype pe at varous ncnaton anges of Case C Excess Pore Pressure Rato m m Lquefed Lquefed Lquefed 3.m Non-quefed.m Non-quefed 3.m 3 Tme(sec) Fg. 9 CYCLIC-D smuaton of pore water pressure tme hstores at varous depths of the budng n Mkagehoma, Japan (Tokmatsu, 3) 5 Depth(m) 5 Pe So Dspacement(m) Fg. CYCLIC-D smuaton of acceeraton tme hstores at varous depths of the budng n Mkagehoma, Japan (Tokmatsu, 3) Fg. Cacuaton of atera dspacement of pe and so at varous depth of the budng n Mkagehoma, Japan (Tokmatsu, 3)

11 San-Shyan Ln, et a.: Ground Latera Spread Effects on Snge Pe Usng Uncouped Anayss Method Fg. Cacuaton of bendng moment of pe at varous depth of the budng n Mkagehoma, Japan (Tokmatsu, 3) moment of pe at varous depth. Fed survey showed that the pes cracked near the pe head and near the bottom of the f causng ttng of the budng, as shown n Fg.. The predcted maxmum moment ocatons match we wth the observed concrete crack ocatons of the pe. 5. SUMMARY AND CONCLUSIONS In ths paper, a decouped numerca approach was used to study the ground atera spread effects on snge pes. The Wnker mode was used to smuate the so-structure nteracton between the pe and the surroundng quefed and non-quefed sos. The Bouc-Wen mode, whch had prevousy been successfuy used to represent statc and cycc so propertes for pe anayss (Ln, et a., ; Badon and Makrs, 99), has been extended here to mode the effects of concrete crackng on the pe performance caused by ground atera spreads. Three centrfuge tests produced by Abdoun, et a. (3) and one rea fed case (Tokmatsu, 3) were used to verfy the proposed mode. The CYCLIC-D (deveoped at the Unversty of Caforna at San Dego) s used to generate the acceeraton and excess pore water pressure tme hstores durng the so quefacton-nduced atera spreadng event. Fnay, the entre uncouped numerca anayses approach combnes the resuts of the CYCLIC-D and the Wnker mode to estmate the atera dspacement and bendng moment of the snge pe under the effect of atera spreadng force. The resuts show a good agreement between the measured and cacuated atera dspacement and bendng moment of pe. ACKNOWLEDGEMENTS Ths study was supported by the Natona Scence Counc and MOTC, Tawan, under grant number NSC 9--E-9-7 and MOTC-STAO-9-, respectvey. Gratefu apprecaton s expressed for ths support. REFERENCES Abdoun, T., Dobry, R., O Rourke, T. D., and Goh, S. H. (3). Pe response to atera spreads: centrfuge modeng. Journa of Geotechnca and Geoenvronmenta Engneerng, ASCE, 9(), 9-7. Badon, D. and Makrs, N. (99). Nonnear response of snge pes under atera nerta and sesmc oads. So Dynamcs and Earthquake Engneerng, 5, 9-3. Badon, D. (997). Nonnear response of pe foundatonsuperstructure systems. Ph.D. Dssertaton, Department of Cv Engneerng, Unversty of Notre Dame, Indana, USA. Bhattacharya, S., Madabhush, S.P.G., and Boton, M.D. (). An aternatve mechansm of pe faure n quefabe deposts durng earthquakes. Geotechnque, 5(3), 3-3. Egama, A., Yang, Z., Parra, E. and Ragheb, A. (). Cycc D-An nternet-based nonnear fnte eement program for executon of one-dmensona ste ampfcaton and quefacton smuatons. Unversty of Caforna at San Dego, for detas. Hamada M. (99). Large ground deformatons and ther effects on fenes: 9 Ngata Earthquake n case studes of quefacton and fene performance durng past earthquakes. Japanese Case Studes, Technca Report, NCEER-9-, NCEER, Buffao, NY, (), Hwang, J. H., Yang, C. W., Chen, C. H. (3). Investgatons on so quefacton durng the Ch-Ch earthquake. Sos and Foundatons, 3(), 7-3. Japanese Road Assocaton (99). Specfcatons for Hghway Brdges, Part V, Sesmc Desgn. Kagawa, T., Ta, Y., Sato, M., and, Mnowa, C. (99). So-pe-structure nteracton n quefyng sand from arge-scae shakng-tabe tests and centrfuge tests. Anayss and Desgn for So-Pe-Structure Inter Actons, GSP 7, 9-. Ln, S. S., Lao, J. C., Yang, T. S., and Juang, C. H. (). Nonnear response of snge concrete pes. Geotechnca Engneerng Journa, 3(3), 5-75.

12 Journa of GeoEngneerng, Vo., No., December Ln, S. S., Lao, J. C., Lang, T. T., and Juang, C. H. (). Use of Bouc-Wen mode for sesmc anayss of concrete pes. Deep Foundatons, GSP, ASCE, Ln, S. S., Tseng, Y. J., Chang, C. C., and Hung, C. L. (5) Damage of pes caused by atera spreadng-back study of three cases. Sesmc Performance and Smuaton of Pe Foundatons n Lquefed and Lateray Spreadng Ground, GSP 5, ASCE, -33. Meyersohn, W. D. (99). Pe response to quefacton-nduced atera spread. Ph.D. Dssertaton, Corne Unversty, USA. Roufae, M.S.L. and Meyer, C. (97). Anaytca modeng of hysteretc behavor of R/C frames. Journa of Structura Engneerng, ASCE, 3(3), 9-. Tokmatsu, K., Oh-oka, H., Satake, K., Shamoto, Y., and Asaka, Y. (99). Effects of atera ground movements on faure pathenns of pes n the 995 Hygoken-Nambu earthquake. Geotechnca Earthquake Engneerng and So Dynamcs, 3, 75-. Tokmatsu, K. (3). Behavor and desgn of pe foundatons subected to earthquakes. Keynote Speech, th Asa Regona Conference on So Mechancs and Geotechnca Engneerng, Sngapore,, 5-9. Wen, Y. K. (95). Response and damage of hysteretc systems under random exctaton. Proceedngs of th Internatona Conference on Structura Safety and Reabty,, 9-3.