Soils and Foundations 2006年
出版

soils and Foundations

表示件数

10 30 50 100
タイトル 表紙(Soils and Foundations)
著者
出版 soils and Foundations
ページ -〜- 発行 2006/04/15 文書ID 20891
内容
ログイン
タイトル Contents(Soils and Foundations)
著者
出版 soils and Foundations
ページ -〜- 発行 2006/04/15 文書ID 20892
内容
ログイン
タイトル An Improved Method for Estimating In-situ Undrained Shear Strength of Natural Deposits
著者 Takaharu Shogaki
出版 soils and Foundations
ページ 109〜121 発行 2006/04/15 文書ID 20893
内容 表示
ログイン
タイトル Reliable Land Subsidence Mapping Using a Spatial Interpolation Procedure Based on Geostatistics
著者 satoshi Murakami・Kazuya yasuhara・Kumiko Suzuki・Hideo Komine
出版 soils and Foundations
ページ 123〜134 発行 2006/04/15 文書ID 20894
内容 表示
ログイン
タイトル Reexamination of Mononobe-Okabe Theory of Gravity Retaining Walls Using Centrifuge Model Tests
著者 shinya Nakamura
出版 soils and Foundations
ページ 135〜146 発行 2006/04/15 文書ID 20895
内容 表示
ログイン
タイトル Modeling Granular Crushing in Ring Shear Tests: Experimental and Numerical Analyses
著者 s. Lobo-Guerrero・L. E. Vallejo
出版 soils and Foundations
ページ 147〜157 発行 2006/04/15 文書ID 20896
内容 表示
ログイン
タイトル Numerical Analyses on Consolidation of Clayey Ground Improved by Vertical Drain System Based on 3-D Elasto-viscous Model
著者 W. Beak・Takeo Moriwaki・Yasushi Sasaki
出版 soils and Foundations
ページ 159〜172 発行 2006/04/15 文書ID 20897
内容 表示
ログイン
タイトル A New Nonlinear Hysteretic Rule for Winkler Type Soil-Pile Interaction Springs that Considers Loading Pattern Dependency
著者 Masahiro Shirato・Junichi Koseki・Jiro Fukui
出版 soils and Foundations
ページ 173〜188 発行 2006/04/15 文書ID 20898
内容 表示
ログイン
タイトル Ground Behavior due to Tunnel Excavation with Existing Foundation
著者 E. Sung・H. M. Shahin・Teruo Nakai・Masaya Hinokio・Makoto Yamamoto
出版 soils and Foundations
ページ 189〜207 発行 2006/04/15 文書ID 20899
内容 表示
ログイン
タイトル Main Factors Governing Residual Effective Stress for Cohesive Soils Sampled by Tube Sampling
著者 Hiroyuki Tanaka・Masanori Tanaka
出版 soils and Foundations
ページ 209〜219 発行 2006/04/15 文書ID 20900
内容 表示
ログイン
タイトル Undrained Shear Strength of Cement-Treated Soils
著者 Kiyonobu Kasama・Kouki Zen・Kiyoharu Iwataki
出版 soils and Foundations
ページ 221〜232 発行 2006/04/15 文書ID 20901
内容 表示
ログイン
タイトル Cyclic Resistance of Clean Sand Improved by Silicate-Based Permeation Grouting
著者 Yoshimichi Tsukamoto・Kenji Ishihara・Keitaro Umeda・Tadao Enomoto
出版 soils and Foundations
ページ 233〜245 発行 2006/04/15 文書ID 20902
内容 表示
ログイン
タイトル Load Transfer Characteristics of Socketed Piles in Mumbai Region
著者 s. S. Basarkar・D. M. Dewaikar
出版 soils and Foundations
ページ 247〜257 発行 2006/04/15 文書ID 20903
内容 表示
ログイン
タイトル Generalized Coulomb's Criterion for 3-Dimensional Stress Conditions
著者 Mitsutoshi Yoshimine
出版 soils and Foundations
ページ 259〜266 発行 2006/04/15 文書ID 20904
内容 表示
ログイン
タイトル Two Case Studies of Consolidation Settlement Analysis Using Constant Rate of Strain Consolidation Test
著者 J.-C. Chai・Norihiko Miura・Shigenori Hayashi
出版 soils and Foundations
ページ 267〜268 発行 2006/04/15 文書ID 20905
内容 表示
ログイン
タイトル Two Case Studies of Consolidation Settlement Analysis Using Constant Rate of Strain Consolidation Test(closure)
著者 Koji Suzuki・Kazuya Yasuhara
出版 soils and Foundations
ページ 268〜269 発行 2006/04/15 文書ID 20906
内容 表示
ログイン
タイトル JGS NEWS
著者
出版 soils and Foundations
ページ I〜II 発行 2006/04/15 文書ID 20907
内容 表示
ログイン
  • タイトル
  • 表紙(Soils and Foundations)
  • 著者
  • 出版
  • soils and Foundations
  • ページ
  • -〜-
  • 発行
  • 2006/04/15
  • 文書ID
  • 20891
  • 内容
  • ログイン
  • タイトル
  • Contents(Soils and Foundations)
  • 著者
  • 出版
  • soils and Foundations
  • ページ
  • -〜-
  • 発行
  • 2006/04/15
  • 文書ID
  • 20892
  • 内容
  • ログイン
  • タイトル
  • An Improved Method for Estimating In-situ Undrained Shear Strength of Natural Deposits
  • 著者
  • Takaharu Shogaki
  • 出版
  • soils and Foundations
  • ページ
  • 109〜121
  • 発行
  • 2006/04/15
  • 文書ID
  • 20893
  • 内容
  • SOILS AND FOUNDATIONS Vol 46, i¥'o 2, 109 1'1 Apl '006Japanese (J*eo echnical Sociei)AN IMPROVED METHOD FOR ESTIMATING IN-SITU UNDRAINEDSHEAR STRENGTH OF NATURAL DEPOSITSTAKAHARUSHOGAKli)ABSTRACTAn equation for estimatin*' in-s'itu undrained shear strength (q ( )) of natural deposits is derived as c!・(1)/,-c.( )=1 .O-0.285 In p***/So through the unconfined compression test (UCT) and Ko consolidated-undrained triaxial compression test (CKbUC). The q*(i) of natural clay deposits can be estimated from the q value multiplied by the reciprocalnumber of q (I]/2c*( ) of the equation using the suction (So) and cl obtained f'rorn UCT for a specimen, ¥vhere c ( Isin-situ shear strength measured from CKoUC* and p** is t¥vo times the effecti¥'e o¥'erburden pressure di¥'ided by three.The q (1)/2c ( ) ¥'alues ¥vere unrelated to lp, q and p, /So and the mean value of' these ratios vas 0.98 in the range oflp= '_6 - 1 10 and q** = 1'_ - 178 kPa. The mean ¥'alues of the ratios for q , q' ( and q (1) to '-c*(1) ¥vere 0.629, 0.998 and0.977 and the standard deviation of those ratios lvere O.14, O. 10 and 0.16, respecti¥'ely. Therefore, it can be seen thatthe improved method is appropriate as ¥vell as Shogaki's basic method (q> (i))・ The mean values of q (1)1,-c.(i) Ivere O.94,O.99 and 0.91 for I¥vai organic and soft clay plus Kahoku*'ata clay, respectively. The coefficient of variations of the q*and q ( ) values were (13 - 14)o/o and unrelated to soils, q. or q (1) values. Therefore, the applicability of the impro¥'edmethod ne vly de¥'eloped in this study can be confirmed for' Kahokugata and I¥vai clays as ¥vell as I¥¥'ai organic soils.The proposed method is a simple and easy one for practical engineering usa*'e.Kel.' w'ords: clay, iil-situ undrained strength, organic soil, sample disturbance, suction, unconfined compression test(IGC: C6/D5)pioneers in this field of study. Lirnit state design and/orINTRODUCTIONperformanced based design methods have been applied tofoundation design standar'ds for railroads (Rail vayTechnical Research Institute, 1997) and are included inThe undr'ained shear strength is the rnost importantdesign parameter for short-term stability problems ofclay foundations. The unconfined compressive strengthEurocodes 7, 8 and Geocode 21 (Honjo and Kusakabe,(q.) is ¥videly used in Japan for stability analysis of clay2002), etc. The probability and statistical approach tofoundations under undrained conditions. This is mainlybecause the mean value of q /2 clearly describes thegeotechnical data is indispensable to these designmethods. Therefore, the UCT can be used in theseundrained shear strength on the failure surface in a soil(Nakase, 1967; Matsuo and Asaoka, 1976; Sho_ :aki et al.,approaches based on the advantages of the q described.This study also considered the treatment of geotechnicaldata for these approaches.For the approximately 60 factors influencing q* values,the effects of sample disturbance lvere analyzed from theresults of questionnaires given to senior engineers and1997) and in addition to this, the simplicity of the testin_"_.procedure always meets the requirements of engineerin_O..practices and is successful in investigating for design andin analyses of failure cases.It is ¥vell kno¥vn that the qintrinsic inhomogeneity of a soil and the degree of aresearchers vith a ¥vealth of practical exper'ience. Also,laboratory tests and field investigations ¥vere carried outtechnician's skill from in-situ Sampling to testing in aunder the simplified conditions of each factor (Matsuolaboratory (Matsuo and Shogaki, 1984; Shogaki andKaneko, 1994). From this and through internationaland Shogaki, 1988). Ho¥vever, there was not enoughcoordination on soil in¥'estigation, testing and founda-of disturbed samples in the study. Therefore, the degreeof sample disturbance for the measured q,, values couldvalue is changeable by thequantitative interpretation of the effective stress behaviortion design methods, there are crltics of using the q valuein practical design (Hanza¥va, 1996). Ho¥vever, the effortsnot be quantitatively e¥'aluated.The undrained shear stren2:th for the short-term stabil-involved in providing these irnperfect UCT proceduresare very important in order to take advantage of theity problerns of clay foundations does not take intovalue, which is an excellentconsideration the in-situ undrained shear strength since itdesign method systematized by a number of Japaneseis infiuenced by strength isotropy, strain rate, etc.c= O method that uses the qi' Associate Professor, Na ional Def nse Academ}. I- lO-20 Hashirimizu, Yokosuka 239-86S6, Japan (shogaki( *nda.ac.jp).The manuscript for ihis paper was received for revie¥v on Januar) 25, 2005; approved on ,January 13, 2C06¥ !rit en discussions on this paper should be submitted before November l, 2006 to he Japanese Geo echnical Society, 4-38-2, Sengoku,Bunk}'o-ku, Tok_vo 1 1'--oo l, Japan. Upou request the closing daie may be exiended one month.10(}_ SHOGAKIl lO5mmUl( ;rl "t'f""II' "¥"""""'¥f': : >:L::j (,: ¥ ' Yv!!1 'rJ' 't) fHachirogatar"vr/JI1 f ¥i':"""'r"'1¥-'i¥i"/I'¥:"'¥ b41' r'l:'¥: T'/!'f/ "I'tr-JL/;////SMakir; ltKorea #rie=1 :::: ¥ar ' r':' Kahokugatar'IY'!r 'FP'/ 7Ariake;/v':/ '!"r't'l': ' l-'1"1'.'I'I ¥'・tf;--.-- 'e'ee ------i________i e '________;-.--- ;Oe ------:e{5mme,315mre,;O:"-/Ift¥"y i//¥¥! t Aoumi'!rl('2lii; C¥ 1- ')hYait'--.- ,; "' Kobejt i;Kasaokali(a)(b)' "r'f'pitrlF":> ftrKumaruotoFig. 2.Location of specirnens for the piane of the sampling tube" hvakuniFle. l. Sempiing sites of useti soilsHo¥vever, it does not mean that the ill-situ undrainedshear su"ength is not estimated. Site investi_.*aation, Iabora-tory testing and design method regulatory efforts need toplace emphasis on performanced based design. Therefore, the studies on the estimatlon of the strength relevantto idealiz,ed shear conditions, e¥rah, arion of sampledisturbance and the relationship bet¥veen measured clueffective o¥'erburden pressure (cr,'.), mean ¥'alue (cl andSo) of q** and specimen suction (So) are sho¥vn in Table I .The l and q** ¥*alues are in the range of )_6 to 370 and 15kPa to 168 kPa respectively, ¥vhich are ¥'ery ¥vide ranges.S (Small) specimens 15 mm in diameter (c[) and 35 mmin height (h) are used for' the UCT and the triaxialcompression test. Ten S specimens, 75 mm in cl and45 mm in /7, vere obtained by a 75-mm sampler and fourS specimens, 45 mm in c! and 45 mm in /7 ¥vere obtainedby a 45-mm sampler, as sho¥vn in Fig. ?-. Shogaki et al.¥*alue and estimatec.1 in-sitz! undrained shear strength, etc.(1995b) sho¥ved that the strength and deformationare required.properties of ten S specimens obtained from a sample 75mm in c! and 45 mm in /1 ¥vere similar in an eng:ineering:sense. The specimen site, Iocated a fe¥¥' mm a¥vay fromIn this paper, an equation for estimating il7-s!tuundrained shear strength of natural deposits is deri¥'ed asc/,,ti)/,-c,,(1)= 1.0-0.?_85 In pm/So through the unconfinedcompression test (UCT) and Ko consolidated-undr'ainedtriaxial compression test (CKoUC). The il7-s!tu undrainedshear strength (cl*,(1)) of natural clay deposits can beestimatecl from the c/L, va]ue multiplied by the reciprocalnumber of q,, 1]/.-c,,(1) of the equation using the suctionthe tube ¥vall, as sho vn in Fi . 2, does not infiuence thesample disturbance caused by tube penetration andfriction bet.¥veen soil and tube during sample extrusion.This ¥vas also confirmed by using a Scanning ElectronMicroscope (Shogaki and l¥,Iatsuo, 1985) and a color' Iaser'three dimensional profile micr'oscope (Shogaki, 2006).(So) and q** obtained from UCT for a specimen. TheThe Portable Unconfined Compression Apparatusapplicability of the improved method ne¥vly developed inthis st.udy can be confirmed for Kahokugata and lbaraki(PUCA) for measuring the So and the q** is sholvn inFig. 3. In this apparatus, the load is applied by the linearclays as ¥vell as lbaraki organic soils. In this study, onlyhead and is transmitted through an AC/DC po¥veredone or two samples, 74 mm or 45 mm in diameter and 100motor. This equipment has a height of about 20 cm and amass of about 70 kN. Therefore, since the equipment isportable, it is practical for field use. The UCT ¥vasperformed on S specimens at a str'ain rate of lo/o/min,after the specimen suction ¥vas measured using a ceramicmm in height, obtained from different sites, ¥vere usecl forthe ¥vhole series of tests zo pro¥*ide small-size specimens.SOIL. SAMPLES AND TF,ST PROCEDURF,SThe undisturbed soils and their remolded samples usedin this study were obtained from the Holocene claydeposits located in the Holocene Plains of Hachirougata,Kahokugata, ivlito. I¥vai, Sakura, Aoumi, Nagoya,Kobe, Kasaoka, I¥vakuni, Ariake and Kumamoto in.Japan, as ¥vell as Bothkennar clay in the United Kingdomand Kimhae clay in Korea, as sho¥vn in Fig. i . Field sampling ¥vas performed ¥ 'ith fixed-pisron samplers ha¥'in_-'inner diameters of 75 mm (.JGS 1''_1-7_003) and 45-mm(Shogaki, 1997a). The 45-mm sampler gives a sampledisc plate. The air entry ¥*alue of a ceramic disc is about?OO kPa. The PUCA, using a perspex cylinder, measuresthe suction over one atmospheric pressure and the c!**value for hard soil vith latent hair cracks (Shogaki,i 997b) .The procedure for measuring the suction is the same asthat reported in other literature (Shogaki et al., 1995a;Shogaki and i¥,laruyama, 1998). It is important that thespecimen suction be measured quickl_v, not only for theshear test using measured specimen suction, but also for'shortenin_"._ of the testing time and cost. The suctionquality similar to or better than that of the 75-mm sam-measurements of specimens taken in advance ¥verepler (Shogaki and Sakamoto, ?-004). The mean ¥'aluesconfirmed as follo¥vs:(t '*,) of naturall) The So of a specimen, in which the suction becomesvater content ( t'**), plasticity index (1*), ESTl¥_ iAT NG l¥r_SITU SHEAR STREGTH111f'or remolding almost equaled the total amount ofisplacementtransducer¥fi¥'e S specirnens.2) 'The mouth of the pouch vas closed after the air ¥vasremo¥'ed by suction.3) The soil vas lvell kneaded by hand from outside thepouch. 'The q*, ¥*alue, e,g., the Aoumi ciay, did notLoad cellchange after a remolding time of more than 5 min.The value of cl** Ivas deterrnined to be the maximumPerspexcylinderst,ess corresponding to an axial strain of less than 150/0.SpecimenThere ¥vere no differences in suctlon and shear strength. d[15wn']characteristics bet¥veen the S and the ordinar.¥' size speci-h 35rnrnCeramicmens (35 mm in diameter and 80 mm in height), ¥¥'hichdisc platePressuretransducerpev'erSPeed la:npHeight200mmLength250rnmMass70kNSpecimend l5-35mmsizeh=3 5・*gommLoad cell500-5000klNLoadingO. 1 5-2.00mm/mincontrol moto: OLoad! Powerspeedunload sw{tehswitchCurrentDirector AitematingF g. 3. Layoui of the portable unconfined compression apparatusconstant, can be measured quickest ¥vhen thepiezometer measurement indicates the same ¥'alue asthe specimen suction.2) Ho¥ve¥'er', the possibility exists that suction greaterthan that of the piezometer measurement cannot belvere exarnined for soils having II' frorn 17 to 150 and cl**from '_O kPa to 1000 kPa and taken from t¥venty differentsites (Sakamoto and Shogaki, 2003).The CKOUC ¥vere performed ¥vith the precision triaxialtest apparatus (PTA) (Shogaki et al., 1999; Shogaki andNochika¥va, '-004) using S specimens. The CA'oUC testsfor the undisturbed and disturbed soils lvere performedaccording to the standards of the Japanese GeotechnicalSociety (JGS) for consolidated-undrained triax'ial compression tests (JGS 0525-1996) on soils vith pore lvaterpressur'e measurement. The specimen height decreases¥vith Ki)-consolidation, becoming about 35 mm at shearafter Kb-consolidation. The initial isotropic consolidationpressure before Ko-consolidatlon ¥vas 5 - 10kPa. Thepore pressure coefficlent values ¥vere greater than 0.98.The specimens ¥vere shear'ed under a shear strain rate ( *)of I .Oo/olmin after the Ko consolidation at a strain rate of0.0050/0/min. The undrained shear strength from theCKoUCequals half' of the principal stress difference.The oedometeF tests¥'ere performed using a loadaccurately measured if all air is not removed from thepressure transducer pipe.3) If the air in the pressure transducer pipe is rerno¥'edand the piezometer sensitivit), is high, the measuredSo values are unreiated ¥vhen the specimen is put onincrement ratlo of' unity and the duration of loading f'orthe ceramic disc plate to give piezometer measurements.4) Ho¥vever, the time in ¥vhich suction becomes constant is less ¥vhen the suction drops from a largervalue to specimen suction.one-dimensional consolidation properties of' soils (JIS Al'_17-1993).The maximum time for measuring suction ¥¥'as sixminutes. For the effect of measuring suction before andafter shear on strength and deformation properties ofeach load increment ¥vas one day. The ¥'alues of thecornpression index (C*) and the preconsolidation pressure((7f,) ¥vere determined from the e-10g cf( cur¥'e basedon the Japanese Industrial Standard for determiningREVIEW OF SI'UDIES FOR ESTIMATlr ITG 11¥r_SITUU_NDRAINED SHEAR STRENGTHRe!ated Studies a/7d Shogaki's Basic l ifetl70c!there ¥vas no eflbct for natural deposits. Namely, theThere ha¥'e been numerous methods proposed for theestimation of in-situ undrained shear stren th. Therecompression method, using the CKOUC test (Bjerrum,decreasing ¥vater content of specimens was less than lolo1973) and the SHANSEP method (Ladd and Foott,under suction measurerrrent and the streng:th anddeformation properties ¥vere independent of suction1974), are ¥vell kno¥vn. These methods are based on theidea that the undrained shear strength under the stresscondition of in-situ corresponds to the in-situ undrainedspecimens, it ¥vas confirmed by Shogaki et al. (1995a) thatrneasurement. Each specimen vas sheared at lo/o/minafter suction measuremeut using PUCT according to theJapanese Industrial Standard for unconfined compres-shear strength. Mitachi and Kudoh (1996) proposed anestimation method for in-situ undrained shear strengthsion tests of soils (JIS A1'_161993).from the results of Sa and standard oedometer tests. ThisThe procedures for preparing the remolded soilsamples, using Shogaki's basic method (Shogaki andmethod is based on the idea that streng:th reductioncaused by stress release and sample disturbance duringIMaruyama, 1998), ¥vere as follo¥vs:1) The undisturbed specimen ¥vas removed after the soilvas put into a plastic pouch. The amount of soil usedsampling is a behavior concerning the s¥velling process ofthe e-log (T(, relatlonship. Ho vever, the effect of sampledlsturbance on the consolidation pararneters, such as the SHOC.AKIll2(7and C*, is not considered in their' method.If the saturated clay is taken from in-situ pressure toatmospheric pressure, the negati¥'e pore pressure (u)¥vithin the sample is given by the follo¥ving Eq. (1)Table l.Indexes values and mechanica] properties of samplesSampleIt*)lpAoumi l905)( O.(7(.*cl**S(kPa)(kPa)(kPa)i 661636516s59(Noorany and Seed, 1965), in ¥vhich a,'.+, is the effectiveoverburden pressure, Ko the coefficlent of earth pressureat rest and A* Skempton's pore pressure coefficient.Aoumi 295s)_166Ariake 31386839)-))u = -A'ocr(. -A* ((T(. -Ko(T(*) (1)Ariake 5l 245446316The effective pressure ((T *) of the perfect sample (LaddAriake 7l 12*35)388and Lambe, 1963) subjected to the complete release ofHachiro ala 257lOl50337total stress is therefore equal to minus u and is determined by Eq. (2);Hachiro ata 4681 lO56337Hachiro ata 5158lOO6240lO(7 ,=(T(. (A' A (1 A'o)) (-))Hachiro ata 714084673810In Shogaki's basic method (Shogaki et al., 1995a;Shogaki and Maruyama, 1998) for estimating ill-situI¥ 'ai 1-563616718243undrained shear str'ength, t¥vo parameters ¥vere consid-Ilvai 1-7128741794ered. They are the ratio of effective pressure to thelwai ?_-5467, 2614234l¥vai 2-65893 / O143))hvai 2-9s341915,I¥vai 5-339830614235lwai 5-459328914293I vai 5-5I176118IS380593091maximum value of So as Eq. (3) and the ratio (Rq**) of cl*of other samples to that (q**(m**)) of q*, of the high qualitysamplevithin the sampling tube.cr ,/So (3)The ratio of the effective pressure of a perfect sample tothe So is equal to one. The Rq** of the sheared samples areplotted versus the ratio of effecti¥'e pressure expressed byI vakuni i6Eq. (3). The Rc/,* of the perfect sample for a ratio ofeffective pressure of one can be extrapolated using theKahoku a adata points in such a plot. In-situ q*, ¥'alue (q*,( )*) can beKasaoka 89obtained from the q**(,*,**) times Rq**. For practical use, theKobemean consolidation pressure (p***) is used instead of (T <,as sho¥vn in Eq. (4), in ¥vhich K is assumed as O.5 and A*is defined by uf/q,,. The u is suction under the axial straincorresponding to q**.CT(. + _,cr(*A'o ' (4)p,,, =3 .3 (T.There are problems in comparing the relationshipbet¥veen the p* and So values since the So value includesnot only the residual effecti¥'e stress but also the matrixsuction and seepage pressure suction. Holvever, it ¥vasdetermined through testing that the effects of matrixsuction and seepage pressure suction on the So values arenegligible, since the So values for remolded cla_v are ¥'er.¥'small anci about 0.3 kPa, unrelated to soil types.Sho*'aki's basic method as described above, in lvhichthe irl-situ Strength can be estimated by measuring q** and88747734402896Kumamoto 99146Kumamoto 159657¥,Iito 26339h,Iito 135,-l 04-1 02193)44i6944876631439735133l12831l 404433l 63Sakura 8l 1350lo8339Sakura 12-ll 085o8524Sakura 12-2l lO62745118Bothkennar650l 02l940Kimhae 4-74026959227l34l ll 046440l 54Kimhae 4-15Kimhae 7-174728So, is easier than lvlitachi's from a practical point of ¥'ie¥v.Table I , excluding I¥vai deposits, are plotted against the lpHo¥vever, these methods also require sever'al specimens¥vith different degrees of sample disturbance. It is impor'tant in engineering practice that the testing methodfor undrained shear strength of clayey soils be simpleand q** in Figs. 4 and 5 respectively, ¥vhere the c (1) is thein-situ undrained shear str'ength measured from CA'nUCunder ill-situ preconsolidation pr'essure ((7 ( )) estimatedbecause of the design parameters set in foundationfrom Shogaki's method (Shogaki, 1996). The c**(l} ¥'aluecan be considered as an appropriate value for the in-siruconstruction design procedures.shear strength from the examinations of the Ko valueunder Ko-consolidation, c,,/p and effective stress pathsl/7-situ Uilc// ail7ed Sllea/' Strengtlls Esrilllated ,fl"On7(Shogaki and Nochika¥va,S/70gak'i's Basic A4letl70clare almost constant in the range of lp = '_6 - 1 10 and cl** =The ratios of q*,,1)"' to 2c*,rl) for _?4 samples listed in-004). The q**( )= 1,-c**(1) values12 - 178 kPa and unrelated to lp and q,* ¥*alues. The mean ESTI*¥,IATING h¥LSITU SHEAR STRENGTH1101 .5( ) : Meanlalueofq (T) 1'cu(1)(1.01)eAE ACC, l)( 1f 1vo..xx:xHachir0 atas A Kumanloto9Hachiro ata7 ] Kumamot015 e BothkennarA Ariake5O Kimhae4-7hvakcuni6 y Mit02(D Kimhae4-15D Ariake7KahokugataMit013CD Kimhae7- 5Hachirogata ' o Kasaoka8 V Sakura8Aoutni2Ariake3<x ;<xx (xxXX ><xxx:x > < :xxx xxxO qu(D /2cu( )x qu/2cu( ) ( ) : Meanaluofq*( ) /2cu(l}olp* / SoPlasticity index, Ip ("/*)5 6 7 8910Fig. 6. Re]ationsllips between the ratios of q* and q,* ElRelationsivip between q ([}*/2c fl) and IFig. 4.o(r ).751)(.+:50 100ox xX) oqJ2cu(1)=1 -O.285 InpFn/So xx xx XXx xx x x. )::Sakural2Sakural2*2x>xC' 0,50.5o o o o (1.01)05 ooxX(xO Aoumi I V Hachiregata4 X Kobeoopto 2cand* /Soi.52( ) : Mean value of qts(1 *12cua;(l-3'5)+(3'5-5'5)(1.01)A* );'eAE<l1 .5CC(.¥v x>);'*¥O.5O Aoumil V Hachirogase4 X tobeHachirogatas A Kumamot09e Aoumi2C] Ariake7Hachirogata7Sakura1x2- ISakura 2*21 Kuma,noto 5 e Bothkennare Kimhae4*7lwakwi 1 6 y Mit02(D Kimhae4-15Kahoku ata <: Mito 3X: x<x0.5'x ( (I x> <X x I) (D] Kimhae7-15Hachirogata2 o KasaokaS V Sakura81 OO 200oXxxfxx l :x':' x'' : jX ' :Xx xx Xx > x ---'xxxjx;c ]A Ariake3A Ariake5---t (5.5-8'7( ):oJL*82Unconfined compressive strength, q* (kPa)10p* / SoFig. 5.Relatlonship between q ,ll*12cl) and qPig. 7.value of this ratio is 1.01 and q**(l ' gi¥'es a similarundrained shear strength as 2c (1) in the ¥vide lp and qvalues. Therefor'e, it can be seen that Shogaki's basicmethod is appropriate for the sample used in Fi**s. 4 andRe!ationship between q.( ,/2c. land p /SeIMPROVED METHOD FOR ESTIMA1'1NG IN-SITUUNDRAINED SHEAR STREN_ GTH BY SHOGAKI5. The relationship bet¥veen the ratios of q** and q*(r)* toThe re*'ression line for the plots ( x ) of UCT in Fig. 6 is2c ( ) and the lo*'arithm of p /So is sho¥vn in Fig. 6. Themean value of the ratio of q*(1)* to 2c (1) is 1.01, as shownin Figs. 4 and 5. Namely, the q ( )* are similar to 2c (I andalso unrelated to the p /So values or sample disturbance.gi¥'en as Eq. (5) and the correlation coefficient (r) is 0.751 .Ho¥ve¥'er, the q /2c*(1) values lineally decrease ¥vithincreasing p /So caused by sample disturbance. The qq l-,c*(1) (5)= i0.285 In p ,/ScThis equation is obtained from 231 specimens at nineteendifferent Japanese sites as listed in Table l, excludingKahokugata and lwai, Kimhae in Korea and Bothkenner(27_ 99)o/o of 2c (I and caused by stress release, sampledist.urbance, etc. from soil sampling to testing. To wit,in the United Kingdom. The relationship between q /,-c (Iand p*/So is unrelated to soil types and locations. Therefore, the in-situ undrained shear strength (q**( )1,-) ofthe change from c (j) is not unique for pm/So or samplenatural clay deposits can be estimated from the qdisturbance. Figure 6 sho¥vs the lo¥ver r'eliability of qformultiplied by the reciprocal number' of q /2c (F] of the Eq.design results and evaluatin*" sample quality by measur-(5) using the So and q values obtained from UC*T for aspecimen. This method for estimating in-situ undrainedshear strength of natural deposits avoids the problem(x) values obtained from UCT are in the range ofin*' specimen suction since the samples used in Fig. 6 ¥veretaken for development of practical construction designs.valuedescribed pre¥'iously for estimating cl*(1)""The relationship bet¥veen q (1)12c.(1) and p*/So is sho vnin Fig. 7. The q (r) values are estimated from Eq. (,5) usin*'the measured So and q . The q (1)/2c (I} values changed in SHOGAKIll4Table 2.Statistica] values of the ratios of q**, to 2c*1 .5shown in rig, 8( ) : Mesn value ofq (1/ c {1)Pm I So,iean value,tl.O-3.5O^95 1o 1597o^994o.163,l ,085o.24l o?_(O 98) <1sVQ)l3,55.55 58.7l.O-5.5o.977l 99.o-8 7iO.5O AoumilV H8chirogata4 A Kuman?ot09HEchiroga ta:'*Aoumi2Ariake3A Ariake5D Ariake7NI mber of specimens, s: Standard devia ion,1 :o>O.18,00723(+0. 1 6ec Xe. )''*+1 Kumamot015・_ Sakuia 2-lHachiro.lwakun 16ta7 Mit02Mit013o Kasaoka8Sakura8H chiro at 41 X Kobeo* Sskur ll?--'-e BothkenTlarO Kimhae4-7(D Kimhae4*15CD Kimhae7-!52001 OOUnconfined compressive stren_ th, q (kPa)-: 20:Fig. 10.:;c's:SRe!ationship betlveen q*, [ /2c*,tll and q*,lO,40o055 qu(1/2cu(1)05 1 q J( /2cu{l.5)l.5 O 5lqu( /2c (1):- 30;os 20Fig. 8.Frequenc)' and distribution curves of q,,1 12c* l]t oo1 .5o05 1q /2c*[1)( ) : Mean valueofq(1/ c ( )(D A_ 1c;0.5>A Ariake3?e>1),loe Bothkennare Kimhae4-7e) Kimhae4-15・Hachlrelzata!Na20va2Mit013lwak m 1 6A Ariake5 C Kas oka8 G Ns oya6Sakura8D Ariake7 X KobeO] Kirnhse7-15oH2chiro';ata2 A Kumamot09Saku a 2.lV HnchirQ ata4 1 Kurnamotol5 * Sakura 2-2o50 100Plasticity index , IpFig. 9.5lq ( }!2cufl)¥. ,_OO Aoumil Hachirogata5. ._ . it02O Aoumi2)3av< >lI ;S 05q l ) /2C {(O.98)X 2c *A)(¥a15 ORelationsl]ip between q,,jl,/2c,,t and I05i 15qJ2CL *' qt r/1 c i' qv li ilJc,- T}Fi('.11. F equency and distributiom c rves of the ratios of q,.q J, itand q*, l, to 2ct l):lass and soft iron (Yokobori, 1974). It ¥vas confirmed byMatsuo and Shogaki (1988) and Shogaki et al. (1995b)the range of 0.62 to I .'_8, Ivhere p.*/So ¥vas less than 3.5.Ho¥vever, the scatter of this ratio increases ¥vhere p.,/So isthat the coefficients of variation for undisturbed Holocene clay deposits and their remolded clay are (8 - 17)o/o:reater than 3.5. The frequency and distribution cur¥'esand similar to those of q {1)12c (1)' The small s value of q,,( )and the statistical values of the ratio of q (1) to 2c*,tl) arcJ7 ,/S0= (1 .O - 3.5), (3.5 - 5.5) and (5.5 - 8.7). The il and s12c**(j) means that the reliability of q (I} estimated fromEq. (5) is higher than the measured q**. Therefore, it canbe seen that the improved method for estimating q ll) byin Table 2 indicate the number of specimens and standardusing a regression equation, as sho¥vn in Eq. (5), isdeviation of q,,( )/2c,,(1)' respecti¥'ely. The s value for allappropriate, based on the samples used in this study.sho¥vn in Fig. 8 and Table 2 for each plot classified ¥vithspecimens is 0.18 and for the data of p /S0= 1.0 - 5.5,The ratios (q,,(1)/.-c,,(1)) of the mean value (q,,( )) of q,,(1) tois 0.16. It is cletermined that the distribution cur'ves inFig. 8 are regarded as normai through verification of thecongruence. Namely, the q,,ll)12c,,(1) ha¥'e a deviation ofO.980.16 as a mean value ¥vhere p ,lS0= I .5 - 5.5. This2c**(l are plotted against the lp and q** in Figs. 9 and 10,respectively. The q.(1)12c {1) values are unrelated to lp, q,,and p ,/So and the mean value of this ratio is 0.97, in therange of I* =26 - 1 10 and q** = 1)_ - 178 kPa.means that the estimated q**(1) value, using Eq. (5), is inThe distribution cur¥'es and statistical values of thethe range of (- 18- + 14)o/o of 2c tl)' measured fromCKOUC under the consolidation pressure of (J ll)' Theratios of q**, q (F)* and q**(i) to 2c tl) are sholvn in Fig. 1 1coefficient of ¥'ariation of q**(1)17-c,,(1) is similar to those ofand q (1) are O.629, O.998 and O.977 and the s ¥'alues ofand Table 3. The mean ¥*alues of the ratios for q**, q*,tl)"' ESTI,¥,IATING IN-SlrU SHEAR STRENGTH'Table 3. Statistical values of the ratios of q**, q [I * and q**jlto c*ff" *,Tabie 4.l 15Indexes and stFength properties for lwai deposits24 cla)s e¥. cluding 1lvai depositsBoreRatioq 12c *(1qut I(t' . p*' ;') (g /cm3 )q(kPa)so 629O. 1423o.998O 1045*mm 4 4 ) 1 374- 492 l 03 - l 0917 - 33199o.977O.1650-mm 4_4 - 4_7 406 -494 1 Ol I 0612-2,23* /2Cu I )qu(1)/2Cu(Iihole Sampler (m)h'iean value!ll 75-mm 4 4-4 8 387 487 1 O1 1 07SPTn: Number of specimens, s: Standard deviation18-214 ) ) O 350- 379 l.06 l_O/17 - 20l 75-mm !4 78 l'6 132 l.35-1.3813 - 2245-mm I .49 - I _6516 - 26Cone 7 4 / 8 104- 147 l.34- l 4215 - 237) 80 106 1 6 l.34-1_4113 - 197 5 - 8SPT15684method on estimating in-situ undrained shear strength isexarnined for highly organic and clayey Holocene soils.Soi! Samp!in*". Soi/ Samp!es a/7c! Test P/'oceduresThe undisturbed soil samples lvere obtained frorn theHolocene and Pleistocene clay deposits located in I¥vaiCity using the cone (Shogaki et al., 2004a; Shogaki et al.,2004b), 45-rnm (Sho*'aki, 1997a; Sho*'aki et al., 2002b;n Basrc method:ifil..Improved methodShogaki and Sakamoto, 2004), 50-mm and 75-mmsamplers (JGS-12,_1 , 1995) and the disturbed soil samples¥vere obtained from the SPT sleeve (Shogaki et al.,Fig 12. Estimating fiow forcl} from the basic and improved methods2004a). The sampling depths (z) ¥vere 4 m to 8 m belo¥vthe ground surface. These samplers ¥vere used at aboutt.he same depths in difi rent boreholes. The majorthose ratios are O. 14, O. 10 and 0.16, respecti¥'ely. There-machinery components lvere the same for these samplers.f'ore, it can be seen that the improved rnethod is betterThe sample reco¥'ery ratios ¥vere 1000/0 for all samplin_',_.from a time and economic standpoint than the basicmethod.The indexes and strength properties of these soils aresho¥vn in Table 4 together ¥vith the samples used in thisFigure 12 sho¥vs the flow estimation for c from thebasic and irnpro¥'ed methods. The improved method hasan advantage in testing time and cost over the basicmethod since the improved method can estimate similarsection. The v,',, and qc ( ) values as the basic method for the mean value, usin_-'soils are (42 - 73)o/o and (6. i - 6.,_), respectively.the q* and So values obtained from a specimen. Holvever,the basic method requires several specimens for estimat-The slee¥'es are brass¥vare ¥vith a 35 mm inner diameter,100 mm in height and ,_ mm in thickness and five of' themare contained in the split-barrel penetration tube. T¥vo S**( )are in the range of (56 - 494)o/o and(12-33) kPa respectively and these are classified ashighly or'ganic and high plasticity clays for the depthssho vn. The ignition loss and pH ¥'alues of lwai organicing c ({), btit 1las an advantage on t/1e estimated accuracys'ince the standard cleviation of the estilnated valtie isspecimens can be taken f'rom a sleeve, as sholvn ins!ightly slna!ler tllan that of tl7e improvec! metlloc!.Fig. 13.APPLICABILITY OF THE IMPROVED METHODFOR IWAI ORGANIC AND SOFT CLAYThe undrained shear strengths for a sample obtainedf'rom the penetration tube of the Standard PenetrationTest (SPT) are lower than those of tube samplers ingeneral due to sample disturbance. Therefore, it is mainlyused for determining basic content and index properties,etc. The SPT is ¥videly used all over the ¥vorld for siteinvestigation and if the st.rength and consolidationproperties of the sarnple obtained from the SPT can bemeasured, it is advantageous in a practical engineeringsense.In this section, the applicability of the improvedUnconfined Compressive Strength P/'opel'tiesThe results of UC*T on the sarnples obtained frorn thetube and SPT samplers for' the z of (4.5-5) m and(7.5 - 8) m are sho¥vn in Figs. 14 and 15. The boreholes I ,2 and 5 sho¥vn in Figs. 14 and 15 sho¥v the test results f'orthe Holocene soil samples obtained from the 75-mm,45-mm and cone samplers. It ¥vas confirmed that thesample quality obtained from all samplers, except theSPT, ¥ 'as similar (Shogaki et al., 2004a). The shadedplots, as sho¥vn in Figs. 14 and 15, have similar ,v*, and p*values, therefore the strength properties are compar'edfor the plots of these specimens.Figures 16, 17 (a) and (b) sho¥v the relationshipsbet veen the pore water pressure (u) measured at the base SHOGAKlll615mm40Specimenc:$ 30b20)(!)%c') 10a; OSamplec'D(,De)-Ocec v_2{Fi,,. 13.35nun>{ic $;:sohe)Location of spccimens for sample of SPT sleevee-4o4. 5510Axial strain, 6* (o/o)15Fig. 16. Relationships betlveen pore watcr pressllre, stress and axialstrain (I¥vai organic soi])4. 6st]. 4.7(:split-barrel penetrator sholvn in Figs. i7 (a) and (b) aresho¥vn as slee¥'es (:i) and (e;i) of the penetrator tube in) 4.8c49Fig. 15. The relationships bet¥veen (T and e* of the speci-)l 2 5SPTSOOBore bok400lvn (o/o)0.9 1 1 _p(gicm')6 8 Oeo (kP a)oSo (kPa)from sleeve (=e;). Ho¥ve¥*er the So, qand Ejo ¥'alues fromthe sleeves are smaller than those of the 45-mm andResuhs of UCT (1lvai organic soil)FieF. 14.mens obtained from the 45-mm sampler' and the penetrator tube sleeves are similar for the specimens obtained50-mm samplers for Figs. 16 and 17(a). These have largesample disturbance and ar'e caused by sampling.Triaxia! Propel'ties and Sa/7lp!e Qua!ityThe Ko values measured from Ko-consolidation andrate of strength increase (c,,/p) after Ko-consolidation fororganic soil and soft clay are plotted a*"ainst the (7 inFigs. 18 and 19 respectively, ¥vhere p is (T and cT (1) is,_/t.estimated il7-situ preconsolidation pressure frome,c:Shogaki's method (Shogaki, 1996). The Ko values conver) 5SPTBere hobSO 150Tt'n (o/o)Fig. 15.l.36p t (_ /em3)1234Oeo (kPa)lOSo (ki?a)Results of UCT (Iwai soft cla))ed to a certain value in the area ofcT tl) value. On the other hand, thesleeve sample are smaller than thoseunder the same (T . These phenomenaas stress change caused by a change:reater than theKo Values from theof the tube samplecan be interpretedof soil structur'eduring drainage as follo¥vs (Shogaki and Nochika¥va,2004);1)of the specimen, the stress ((7) and the axial strain (8*) forthe same specimens and the test results are also listed inTables 5 and 6. The suction under shear is represented asthe u in Figs. 16, 17 (a) and (b) since the suction undershear becomes plus, ¥vith small So Values. Therefore, the uvalues at the 8*=00/0 are So Values. The So, c/*= and E50values of each tube are given in Tables 5, 6(a) and (b).The test results of different sleeve samples in the same, )The specimen ( ), in which the soil particles arearranged isotropically by sample disturbance, isdeformed in the direction of draina9:e.If the soil rigidity incr'eases ¥vhen the void ratiodecreases ¥vith drainage, the A'o value increases ¥vithincreasing cr ¥'alue since the Ko condition can not bemaintained ¥vithout a corr'esponding increase in thelateral pressure, ¥vhich in effect increases the cr¥'alue. ESTlh,IATING IN-SITU SHEAR STREN* GTH30Specimen( ' ) (g /cm3 )2- I eb2-22-3 e2-4c') 10fCe=p.}t'*BH # -No. *,t :,_.O* 20, 'Resuits of UC'I (Iwai organic soil)'Table 5.cs117*E<tl(h'lPa)erSo(? ) (kPa)3 74l 0930.8o.46lO_63*94040933 2o.3315.03 .o44 71 .0718.lo 407. ll .5394l 0716 8o 3_8.9, ,53791 0719.9o 24ll_54*4SPT sleeveSymbols used in Fig.16,oq*,(kPa)efo(,D(1)e)C,4eTable 6.Specimen-4BH : : -No.e)OA*6o10Axial strain, 6a (olo)(a) =(7.5-7.6) m (sleeve i )(b)BH54.91251 .3813 3o 29663137l .3415.4o.813.,4.4l 20l .4222.7o 96293*9l136l 3413.3o 375.・2,0,131l .3717.4o.575.・, 55-・__1'*-3afo4SPT sleeve= (7.7 - 7.8) m5-1o393,8BH # -No. *o SPT sleeveO -1A SPT sleeve 3 -2fa34.2o 74Specimen(, 10o . 9616 9l .3SSvmbols used in Flg, 17(a),b21.6l.35i ,_fe So('.".) (KPa)i27;:s:' :_ 20E<(h"IPa)1 -2 c1 -3 c1-c sq u(kPa)1305-2Soft clay, z= (7.7-7.8)mp,t'(・,i) (g/cm )l-1 c5-130¥l:Results of UCT (hvai soft cla) )(a) =(7 ) 7 6) m-2"' ::._t-1,*・p,t'=,() (g/cm3)qE< (}(kPa)(i¥,1Pa)e f Sf *( /.) (KPa)l lOl .4218 9o 705.62.01 07l .4217 lo 69563*41 04l 4217 8o 675.73.0l081 4118 lo 634.42,0I 06l 39IS 70.605,0, ,5cl:,L)C** t'=_,, - _-2e)Symbols used in Fig 17(b),'c:$SPT sleeve-4a)Ol6o10Axial strain, 6a (o/o)are similar to those from tube sarnplers under the samecr . Namely, it can be seen from Figs. 18 and 19 that theundrained triaxial cornpression test after Ko(b) (= (7.7 - 7.8) m (sleeve tFig. 17. Relationships bet veen pore)vater pressure, stress a ld axialstrain (Iwai soft clay)consolidation, around the (:r (1) value, can eliminatesample disturbance caused by the SPT and penetrationtube sampling, as described above.The efi ctive stress paths of CKOUC and UCT under* = I .Oo/o/min for organic soil and soft clay are shown in3) For the undisturbed soil, the soil displays a plasticbehavior vhen ,becomes about the (T ( ) value andthe amount of void ratio chan*'e of' each specimenbecomes similar, unrelated to sample disturbance.The c**/p ¥'alues in the area of small crvalues are largerand have an overconsolidated beha¥'ior. However, theydecrease ¥vith increasing consolidation. The c /p valuesof the sleeve sample, in near or greater areas than (T (i),Figs. 20 and 21. The effective stress paths of the UCTsho¥v an overconsolidated behavior and changed to anormally consolidated behavior of CKbUC with increasmg (T values. The stress level of (T (i) is located at theintersection of their rupture envelope lines, unrelated tothe soils. Figures 20 and 21 sho v the validity of theestimated cT (1) value and the UCT method ¥vith suctionmeasurement, using an S specimen. SHOGAKi11S80c:ie*_v:$e'D(,,Organic, F (4.5-5.0)m c_ S Line ( c!1 o=56s'tc60oa"c'dp(n( I +2Ko)/3 = 1 4kPa)Io (UCTBHlfh:.. to 'le $dvo(1+2Ko)/38kPa ?40'oic')i)o.c').' O BH210 20 3020oD BH5c $C SPT sleeve''e)oc':cs)0.8e'::(Ko: O.3_') + SPT sleeveRoEffective stress, (da+2dr)/3, p' (kPa)O.6eJ)::)Fig. 20(,Effective stress paths (Iwai organic soil)0.4ee)::$eO 20 40 60 80O.2Effective axial stress, da (kPa)Relationships betTveen c,,/p. K ] and ,7Fig. 18.(1lvai orgnnic soil)fce 80Soft clay, z= (7.5-8.0)mj60dp{n(1+2K1 )/3 '3kPaO;d .(1+2Ko)/3=1lkPa401O.8c::s1::cej ・.oc"e)d'0=1 8kPa Soft clay, f (7.5-7.9)mf dp(D3 7kPa Ko )'43C S Lme ( Ncf39 )UCT ,rJL 'c')20' O8Hl[] BH5...・・・1 ''ce>)O.6e)VQoC:)oe'C SPT sieeve@,@(K0=0.43)- - O.4s; $(&)o* SPT sleeve r/330 401020Effective stress, (da+2dr)/3,p' (kPa)50:;o.* 0.2ooFig. 21. Effective stress paths (hvai soft cial.')(J:BHlO.6Ci BH5c/d=041u p(1)'c!i::O.4iJ;:*O; 1c'o 03e)cec'( 02qu(1)*/_'cu{D:o Sofi clay o organic(qu( ) s/2cts(1))(organic) =:1 02 (qu(1) s/2Cu(1))(Soft clalY) I OlC Ii),*(Tube sampler)lrnproved methodO.5*'H^1 .5e SPT sleeveR,@e JO 20 40 60 80Effective axial stress, da (kPa)x+x x"J"'JI -=' ; ::1rX? "-1"'1'j+and;ix't j ++ t/ o l(qu(1)/2cu O) (Olg:mic) : ;;0 94 (qu(D/2cu(O) (SOft clay): ;0'99;qu(D/_7 cu(O: AxPig. 19. Relationships between c*,/p. Ir,''t'T"""'1" "'1""'1'1'1'1")(,! '+¥O,.1:=';+Soft clay A OrganicImproved methodi (ITvai soft cla,')l 2 4 5 67893Pm / SoThe ratios of the mean value of the estimated q**(l} andq**( )* for the 2c**(1) under the (Jp{1) value from CKr}UC areplotted against the p ,/So and sho¥vn in Fig. '-2. Figure '-2also sho¥vs the estimated results (+ and x ) obtainedfrom the impro¥*ed method. The mean values of ratios ofFig. 22. Relationsl]ips between the ratios of q* l* and q*,(h 'to 'c**1*deposits)q (1) and q**(1)* values to 2 c**{1) are O.94 and 1.0)_ for thea conservative result for short-term stability analysis forI¥vai deposits and these ratios are unrelated to the p*/Sohighly organic soils and O.99 and 1.01 for the soft clay.¥'alues .To ¥ 'it, the mean values of q (1) and q**ti)* are similar to'_c*,(1) for the soft clay and organic soil, the q*,(l} values gi¥*eThe ratios of the q*, q**(1) and q** 1)* to the 7-c**(l} values areplotted against Z in Fig.-3. The mean ¥'alues of c/,,/2c,,(l} ESTI 'IAT=1119G I_+¥LSITU SHEAR STREN(3 TH203organic <0'69> <0'94> (1ubc sampler)4x )ec c0l pIc ,,I exc aqul2cu(1)e o qu(1)/2cu(1)x xe * qu(1)*/2cu(1)Ip}eioo oerF:xx)'x eo :1 )> xe( 1N(qJ2cu(D)(TS) i I :)e( Isexh t_lPlj6-7,,e>o lO,:I os,e, e , __ ¥x, I eel' o; ;j/ ;-(qu(E)/1cLs(D)("Tse @oo (SPTsleeve)oe)c:)8>;( xx oox x) xx Xle* qu(i)*/2c s(1)g eJ**O.5Oi I.5 ?O 5 1 1qJ2cu( }, qL (Ir!2cqu(lr/2cu{Il5 _,ie o e C qu(f)/2cu(1)@ jx xx<0'I57> Q <0'99>xxSofi Clay101,q J2c*tE]O qu/'cu(i)llI9O.5 1 1 S -1qu/2cu(1)' qu(D/2cu(D' qu(DFig. 24. Frequenc} and distribution curves of the qand c! ([1 to 2c*,([](lwai organic soil)1.5,20/2cu(D.¥'Fig. 23. Relationships between the ratios of q , q*,,1) and cl*,(I] *'to 2cl] l]and depth (Iwai deposits);ocrVTable 7. Statistical values of the ratios of q,,, q,, [t and q t to 2c,,{[,(hvai deposits)O( ¥vai organic)q J2c*(T),1cl, /2c,Clul i I=qui I51/2Cu( I/ 2 cui I liean valueso.69o.2051C!ufl)/2Ci tlo ^ 94K hokugata : XJean value45o.573l.O1sO^ 1 645o.995(!u( i l* 12cui I lq l il /2C'u{1),r :eQ,::;;:SO1o.23400qu(r 2cuti),:;_S300XXs23 lo_63O 1423l .OOO 10199o.98,e,XX200*,i ooxx cx )e(,O 120 100 O 50Tv('/.) q(kPa) So (kPa)Fig. 26. Results of UCT andql; ll*/2c*( ,1; llx +++! :x;+i*,j ( i: tS,{ean ¥'aluequ / 2cutl)qLs(I} / 2c*{1)<0j8> <0.9F:e ),lL'f IqJ2cu(I500o 27(Other 23 clav)qti 12C'O 5 i 1 5 2(Iwai soft clay)l:L:{1)15 -,qLXIV2cu( )Fig. 25. Frequenc) and distrrbution curves of the c/,* and q*,jli to 'c*( ,e* NC1",1(! u ' 12C'051 .02(I¥1'ai soh clay)c/L' /2 cu05 1 15 204))e(x + *+x>e }+ +0.5 1 1 .5qu/ 2cu{ ), qu(IV 2cu( )' qu(If/ 2cu(1)ratios of qu' qLfE and qul]* tO Icl tE](Kohokugata ctav)O 16l¥ umber or specimens, s: Standard deviationare 0.69 and 0.57 for organic and soft clay respecti¥'ely.On the other hand, the q (1)> < and q (I} values are alsounrelated to z.Table 7 sho¥vs the statistical ¥'alues of the ratios of q ,cJ,,( )* and cl (1) to 2c (1) values for organic, soft clay and 2,3different Japanese clays, including Bothkennar andKimhae clays, as sho vn in Table 1. Figures 24 and ',_5sho¥v the normal distribution curves of the ratios of cl**and q (1) to 2c**(1) values for organic and soft clay respectively, as sholvn in Table 7. Holve¥'er, Figs. (a) and (b) of¥vell as the other 23 clays. Therefore, the qt*{1} and q (I}'values improve design reliability since the reliability ofthe mean value of q** improves.APPLICABILITY OF IMPROVED METHOD FORKAHOKUGA'I'A CLAYShogaki and Nochikalva ('_004) examined the triaxialand consolidation properties of Kahokugata clay throu*"hthe PTA, asvell as this study, and the c (1)' a (1" C.{. ) andKb ¥'alues are reported as 95 kPa, 230 kPa, 1.8 and 0.46,respectively. The results of UCT and the ratios of c/ , q**(1)and cl {1}* to 2c (1) are plotted in Fig. 26. The mean valuesthose Figs. only sho¥v the normal distribution curves of'_3 different Japanese soils ¥vith I¥val deposits to avoidcomplication. It can be seen f'rom Table 7 and Fi**s. '_3the cj**(j)* and q (1) values are similar to 2c (l]' Ho¥vever, theand 24 that the q,, values are 310/0 and 430/0 of c (1) valuesq,, value is 580/0 of 2c,,(1)' The statistical values of the qfor I¥ 'ai organic and soft clay deposits respectively, asand q (1' are summarized in Table 8. The coefficlent ofof these ratios are 0.58, 0.91 and 1.04, respectively and SHOGAKI120Table 8. Statlst cal laiues of the ratlos of q and q [ (Kohokugataand q ( ) values. Therefore, the applicability of thecla .')improved method ne¥vly developed in this study canIZ,Iean Yaluequ15o)_12.5be confir'med for Kahoku_g:ata and 1lvai clays as ¥vellas lwai org:anic soils.It is essential that the effect of q tl) or q {1)* onqull)15l 7613^6undrained shear strength, in the for'mula for deter-( ,h)mining the safety factor, be revie¥ved by verifying then: Number of specimens, *: Coe c erit of I riat oneffect of sample disturbance on practical construction design.variations of the qand q (Iare I '_.50/0 and 13.60/0 respec-tively, and are similar. Shogaki et al. (,_002a) has reportedthe same result for Ariake clay. The coefficient ofvariation of the q ¥'alue becomes greater ¥vith increasingsample disturbance (Matsuo and Shogaki, 1988). It ¥¥'asdetermined that the similar values of the coefficient ofvariation for the q and q**(1) values of Ariake and lwaideposits are caused by the small number of specimens.Ther'efore, Sho*'aki's basic and improved methods forestimating il7-situ undrained shear strength can be usedfor 28 Japanese clay deposits plus Bothkenner andKimhae clays. The l and q ¥'alues are in the range of ,_6to 370 and i5 kPa to 168 kPa respectlvely, ¥vhich ar'e ¥'ery¥vide ranges. Ho¥vever', the q**( )* or q**(1) values cannot beused directly for practical stability analysis since theAC_KNOWLEDGF,MENTSThe applicability of an improved method for I¥vaiorganic and soft clay lvas performed as a part of theResearch Committee on Miniaturiz,ation, Accuracy andDesign Reliability for Geotechnical Investigations andL,ab Tests. The authors ¥vish to express their sinceregratitude to the members of this resear'ch committee, toMr. Yoshikaz,u Maruyama and Ryo Sakamoto, ¥vho ¥veregraduate students of the National Defense Academy, fortheir cooperation in experimental ¥vorks and also to theGeotechnical Survey Laboratory of the Port and AirportResearch Institute for the use of their Bothkennar claysample in his research.sample disturbance is one of the factors influencing:safety. Therefore, it is essential that effect of cl (1) or' cj**{1} 'on undrained shear streng:th in the formula for deter'mining the safety factor, be revie¥¥'ed by verifying the effect ofsample disturbance on practical construction design.NOTATION. in*siii! undrained shear strength measured from the C_A'oUC_c** I)'under (7;(I}c**/p: rate of' strength increaseE<0: secant moduiusC_ONCLUSIONSThe conclusions obtained in this study are summarizedas follo¥vs:1) An equation for estimating in-siru undrained shearstrength (q (1)) of natural deposits ¥vas derived as q (1)l'_c*(1)= 1.0-0.'_85 In p ,/So through the unconfinedcompression test (UCT) and Ko consolidatedundrained triaxial compression test (CA'oUC). Theq,,(1) of natural clay deposits can be estimated fromthe q value multiplied by the reciprocal number oflp: Plastici y indexKo: coefficien of eaFth pressure at restp**: mean consolidalion pressure defined as 2(T(. /3q**(1): in-sitti cl ¥'alue estimated by Shogaki's impro¥'ed methodin-situ q** ¥'alue estimated by Shogaki's basic methodq tl) 'q**(******,: maximum vaiue of unconfined compressi¥'e strengthr: correkuion coefiicieus: standard deviationS{): specimen suctiontt' , :c * (¥^stage8.: axial strainq,,Il)/2c (1) of the equation using the suction (So) and qobtained from UCT for a specimen, ¥vhere c { ) Isin-situ shear streng th measured from CA'oUC and p ,is t¥vo times the effective overburden pressure dividedby three._) The q tl)12c ( ) values ¥vere unr'elated to lp, q:-: strain at f'ailure*: axial shear strain ratevolumetric s rain:cf : effecti¥*e axial stress under CKoUC_ Leste*: preconsolida ion pressure(rand p ,/So and the mean value of these ratios ¥vas O.98 in therange of lp='_6-110 and q*,=12-178kPa. Themean values of the ratios for q , q (1)>' and cl (1) to 2c**(1)¥vere 0.629, 0.998 and 0.977 and the standard deviation of those ratios vas O.14. 0.10 and O.16, respectively. Therefore, it can be seen that the impro¥'edmethod is better from a time and economic standpoint than Shogaki's basic method.3) The mean values of q**(1)12c (j) ¥vere 0.94. 0.99 and0.91 for I¥vai organic, soft clay and Kahokugata clav_ ,respectively. The coefficient of ¥'ariations of the q**and q**( ) values ¥vere (13 - 14)o/o, unrelated to soils, q*,natural ¥vater content}: effective imernal fric ion angle under normary consolida ioni): in-sin! preconsolidation pressure estimated by Shogaki'smethoda(.: effective vertical pressure under incremental loading oedometer testcT( : effecrive overburden pressureRF,FF.RENCES1) Bjerrum, L^ (1973): Problems of soil mechanics and constructionon soft clays and struclurally unslable soils, Proc. 8th ICS1 (IFE, 3,l I ll59.2) Hanzawa, H (1996): Procedures o determine the shear strength ofclav for shorl-term stabili y analysis, 4js! Geotec/1. S_vnzp. JGS,89-94 (in Japanese)3) Honjo, Y. and Kusakabc, O, (2002): Proposal of a comprehensivefounda ion design code Geocode 21 , Ver. )-, Proc. IIVS Kamakura ESTIMAT【NG IN−5∫rUS置E.へR STRENGT阿  2002,Balkema Publishers,95−106.4) Japanese Geo[echnical Socie[》’(1996);N「1e【hod forκo consolida芝e−121  ∫CSAゾF五,慧amburg,201一一204,21)Shogaki,T、(1997b):Streng出proper[iesQ∫claybypor芝able㈱col1一  undrained £riaxial cor貰1presslon {esζ Qn soi王s wit猶 Po「e wa【e「  触ed compression apPara{us.P1『oc、∫ノ∼’、Co’ぴGθo!θ‘h.五ン1919.  r)ressure measurement(JGS O525一三996),349−358(ln Iapanese),  Coθ5’、0εvθ10ρ、,85−88、5)JapaneseGeαecねnicaISocieこy(1998a):Themαhodforobtaining22)Shogaki,T. (2006)= へ・licrostr穏cture,streng芝h and consolida【ion  U夏1d1SしUrbed SOi!Samr)1eS USingζh1n一、、・a照edエUbe Samp玉er With蕪Xed  罫)ropert三es ofl Ariake c玉ay depos1ts ob柔ained from samplers, /、  pis[on(∫GSl221−2003),3’αnゴαr450ゾ/妙翻θ5θ0θo∼θch1∼icα1  、4∫7M∫’π.,3(5)(£oappear),  Soc’8有y/101』So’13α1∼∼ρ1’ノ∼9−S’α’∼4α’一4∫α1∼ゴ石ぎρ10’∼甜io∼∼∫一‘五1∼91な1∼23) Shogaki,τ,and Kaneko,N{、(1994):Effect of salnr)1e disturbance  v8顔o’り,1岬.  on s汀engτh aBd co職solida芝ion paralτ1e芝er of soflt day,50〃∫α’1ゴ6) Japanese Sエandard Associadon (1993a)= 氏’let鼓od fQr “駐con負ned  compression韮ests,、ノ1S.4 1216−1993,1−11(in Jar)anese)、7) Japanese SIandard 、へssociation (1993b): Test metねod for one−  Foμ1∼ゴαrio’∼∫,34(3),1一一10、24)Shogaki, T、 and 氏’laruyama, Y、 (1998): Estima【ion of 11∼一∫”μ  uadrained shear streng[11 using dis【urbed samr)les wi芝猶重n 員1in−  dimensionai conso}idation properties of soils,.ノ13瀦 1217−1990,  wailed samplers,Geo∼θcノ∼、5’厘C1∼θ1・αc’θr,,Atla瓶a,419−424、  玉一13(in∫&panese)、8) Ladd,C、and Lambe,F、(1963):The s[reng{h of‘‘undis£urbed”25)Shogaki,T、andMatsuo,M、(1985):FaαoranalysisapProacluo  uncoΩsolida芝ed“ndra1nedsllearsエrengt1}ofclay、vl癒some  clay determined flrom undrainedζests,Laboratory she&r test玉ng of  considera毛bnonmicroscopicstructures,Pノ}oc.εvノηρ、Sα’ηP1”∼9,  so三1s,擁STA4ン57P,(361),342−371、  109一三16(i鷺Japanese).9) Ladd, C、 C、 and Foott, R、 (玉974): New design procedure ∫or26〉 Shogaki,T、and Noc註1kawa,Y.(2004):Tr呈axial strength pror)ert玉es  s【abi蓑ty ofl soft clays,/、刈5C五一,100(GT7),763−786.  of naturai deposi【s at 1く70 co脆solida[ion sしa£e using a precision10)Matsuo,M温nd、へsaoka,.へ、(1976):Asta韮is韮icals篭udyonaconven一  triaxial apParatus wiξh sma呈I size spec玉mens,50〃5 α1∼ゴFδμ刀ゴα一  巨onal safeこy flactor method,So’Z5α1∼4Fo∼’11ゴα’io115,16(1),75−90、  ∼io’∼∫,44(2),41−52、11)Maζsuo,M、and S負ogaki,T、(1984):Analyses of several flacじors27) Shogaki,τ、and Sakamo【o,R.(2004):1疑e apPlicabi}玉ty of a small  i湘uenclngσu−value,Oo1ηε5”ぐ烈、∫o’Z賃nゴFo姻ゴ副oノア5,24(3),  diame芝er sampler wlτh a two−chambered hydrauiic pistoll f−or  B9−150(inJapanese)。  Japanese clay der)os三ts,So’Z∫αノκノFヒ刀”∼ゴα∫101δ,44(王),重13−124、12) Nlatsuo,Ni.and Shogak1,T、(1988):E鐸ect of Plasticity and sample28)S員ogaki,T.,Kaneko,M,,Moro,H.and M搬ara,S.(1995a):  d量sturbance on sこaτ1sし玉ca量properζies ofF u【1dra王ned s}1ear s【rength,  Estlma[ing in−5’π’undrained shear s【rengτh by uncoa痕“ed compres−  εo’なα’∼4’α11∼‘1α1’01∼5,28(2),14−24、  s1ontesいvi:hsuαionmeasureme煎,.4η’∼.S∫〃∼ρ。50〃Sα’ηρ1”∼9,13) NI主tacねi,T.and Kudoh,Y.(玉996):NIet}10d for predicting〃1一∫”π  95−102、  undrained s【rength of clays based on[he suction value and unco鷺一29) Sllogaki, 丁, で〉loro, H. and Kogure, K. (1995b): S{a【1stical  自ned compressive strengt}1,滅 oθo’θcノ∼.五1∼9∼9、,JscE,(541/  P「oPe貰ies of soil data wi{hin t−in『、valled samPle「s,P1−o(♪.5〃∼∫∼π,  III−35),玉47−157(inJapanese)、  (2於ho”θαn4Po1α1一五1∼9、Co/1ブ,,駐ague,406−413、14)Nakase,A,(1967):τねeφu=Oanalyslsofs臓bili籍・and uncon負ned30)Sわogaki,T、,Moro,H、andMa【suo,M、(1997):Aslopesτability  co撒pressions【reng由,So1Z5α’∼6Fα”漁∫’o’15,7(2),35−50  &nalysis considering球ndra1“ed strengtねanisotropy of na蒙ural cla}「15) Noorany,夏、and Seed,}{、B,(1965):1nづ々μsτrength characteristics  deposits,75μch∫イo一ノ百30,/.ノGS,45(8),13−16(圭n Japanese).  ofsofζclay,ノ、‘4SC五,91(SM2),49−80、31)Sねogaki, τ.,Marロyama,Y、and Shirakawa,S.(1999)l Aprecision16)Railway Technical Research Insεltute(1997)=1∼α’ム、,の,5ぴ‘κ∫雄θ  tr1ax呈al apPara[us using sma韮1size specimens aad sεreng工h proper−  ∂ε豆9’∼,S’α’∼ゴαr4Fα’n4α’o’∼S’π’α£’1。θ501z41∼θ’α加i/1gS〃}μαμrε∫一,  ties of soft clay, Oeorθcノ∼、 五1∼grg、 Tヂα’∼5ρoπ、 ∫’∼ゾ】r‘∼c∼1ごrθ,  レ205(inJapanese)、蓋7)Sakamoto,R、andSねogaki,T、(2003):Effeαofspecimensizeon32)Shogaki,τ、,Nakano,Y、andShibata,A、(2002b):Sampierecovery  uncon負ned compressive streng由 prQper【ies flor natural clay  ratiosandsamplerpe識e【ra£io“reslsζancehuubesamplingfor  deposits,13〃∼1’π.ρ蔀ho1でPo1α”勲9∼9.Con∫、a励”.,  Honolulu,426−43重、  Niigata sand,50’Z∫α’∼ゴ」Foε’110「σ’∫o’∼∫,42(5),111一玉20、18) Sねogaki,τ.(1991):Sζrength properし五es of cl&y by portable uncon一  丘ned comρression apParatus,Proc. /1π. Co’∼∫ (フεorθch.五ngrg.  Coα5’.0θyθ10ρ、,Yokoねama,85−88、19)Shogaki,τ ,(1996):Ame由od forcorrectlngconsolidaτbn  parame【ers for samp呈e disこurbance using volumetric straまn,So’な  α11ゴノ=0μηゴθ∼’on5,36(3),123一三31.20)Shogaki,T.(1997a)IAsmalldiametersamplerwi由atwo−cham−  beredhydraullcpistonandthequall【yofitssamples,Pヂoぐ、14∼h  1151−1157、33)Sねogaki,τ.,Sakamoto,R.,Kanno,Y.,Kobayas短,}{.and  Momose,S.(2004a):Standard peRetration韮est sampie quali【y,  Pmc、石ノ∼9∼9、ρノ}αc’icεPθ’ブoノ}’n.Soゾ!∂θpo3”5,0saka,159−164.34) Sねogaki,T,Sakamoτo,R、,Kondo,E.and Tachibana,ト{.(2004b):  Smail diame{er cone sampler and its appllcabl翫y for Pleistocene  Osaka罫〉la12clay,So’Z∫α1∼ゴ1=ヒ)μ1∼4α”oη5,44(4),119−126.35) Yokobori,T、(1974)=S芝rengtね,fracture and faこ呈gue of maτer重als,  Gihodho Pubhshまng(⊃ompany,3−6(in Japanese),
  • ログイン
  • タイトル
  • Reliable Land Subsidence Mapping Using a Spatial Interpolation Procedure Based on Geostatistics
  • 著者
  • satoshi Murakami・Kazuya yasuhara・Kumiko Suzuki・Hideo Komine
  • 出版
  • soils and Foundations
  • ページ
  • 123〜134
  • 発行
  • 2006/04/15
  • 文書ID
  • 20894
  • 内容
  • ログイン
  • タイトル
  • Reexamination of Mononobe-Okabe Theory of Gravity Retaining Walls Using Centrifuge Model Tests
  • 著者
  • shinya Nakamura
  • 出版
  • soils and Foundations
  • ページ
  • 135〜146
  • 発行
  • 2006/04/15
  • 文書ID
  • 20895
  • 内容
  • SOILS A 'D FOUNDATIONS ¥'ol46. No. 2, 135-146, Apr. 2006Japanese Geolechnical SocieiREEXAMINATION OF MONONOBE-OKABE THEORY OF GRAVITYRETAINING WALLS USING CENTRIFUGE MODEL TESTSSH I 'YNAKA (URAi)ABSTRACTGravity retaining ¥valls are videly used in Japan because of their sirnpliclty of structure and ease of construction. Indesign procedure, the seismic coefncient method is 'idely employed, in ¥vhich the earth pressure and inertia force arecalculated by con¥+ertin*' the seismic force into a static load. Earth pressure is usually calculated by the MononobeOkabe formula, ¥vhich applies C*oulornb's earth pressure computed f'rom the eqLrilibrium of forces In the static state.Ho¥vever, the Hyogoken-Nambu Earthquake of 1995 prompted the need to reexamine seismic design methods forvarious ci¥'il engineering structures. Gravity retaining ¥vall is one of such structures vhose seisrnic design has to bereexamined and rationalized. At this moment there is no clear' empirical basis for converting the seismic f'orce into astatic load. The design method has to take into account the behavior of gra¥'ity retaining valls during earthquakes.At the Public ¥Vorks Research Institute, rnodel tests vere conducted on gra¥'Ity retaining ¥valls using a centrifuge.The acceleration and displacement of a retaining ¥vall and its backfill as ¥vell as the earth pressure acting on the ¥vallvere measured simultaneously together ¥vith the deformation behavior of the ¥vall and its backfill, using a highprecision high-speed camera. The data sho¥v that the hypothetical conditions of the lvlononobe-Okabe formula do notappropriately express the real beha¥*ior of backfill and gravity retaining ¥valls during earthquakes.Kewords centnfuge rnodel test, earth pressure,(gra¥'ity retaining ¥vall),(Mononobe-Okabetheory), seismicresponse, slip surface (IGC: E5)ments conducted since the 1980s ou vard has re¥'ealed thatIN1'RODUCl'lONit is necessary to consider dynamic earth pressure as asystematic dynamic phenornenon of the caisson and theIn Japan, gravity l'etaining ¥¥"alls ha¥'e been ¥videly usedground (Kazama and Inatomi, 1990). It ¥vas foundexperimentally by thern that the inertia force andfor r'oad construction and the likes because of their' sim-plicity of structure and ease of construction. In thepresent earthquake-resistant design method of gra¥'itydynamic earth pressure are out of phase. Therefore, thevalls, the seismic force is transposed to a staticactual behavior of caisson revetments does not satisfy theload, and the earth pressure and inertia force are calculated by the seismic coefficient method.hypothetical condirions of the Mononobe-Okabe theory.retainingMononobe-Okabe theory applies Coulornb's earthSince the 1995 Hyogoken-Nambu Earthquake, therehas been an urgent need to revie¥v the earthquake-pressure theor'y (Mononobe, 1929; Okabe, i924). Seismicresistant design for ¥'arious civil engineering structuresearth pressure has been of an important concern inand to examine and streamline the design of gravityretaining ¥valls. This has prompted lively discussionearthquake and foundation engineering for many years.In exper'imental research, retaining ¥valls ha¥'e movedconcerning earth pressure during earthquakes (Steedman,during shaking in accordance ¥vith the hypothetical con-1998;ditions of the Mononobe-Okabe theory (Ichihara andcoefficient method requires the seismic force to be trans-Vatanabe et al., 1999). The present seismicMatsuza¥va, 1973; Ishibashi and Fang, 1987; Sherif et al. ,1982) ¥vhich assumes that the backfill is in a state of plas-clear empirical basis, and it is not clear lvhether or not thetic equilibrium. Consequently, although the pressurepresent calculation technique gives the correct seismicdoes not necessarily ¥'ary lvith a triangular distribution inearth pressure. Any rational rnethod for designing gravityretaining lvalls must reflect their actual seismic behavior.formed into a static load. That transformation has nothe vertical direction, the total acti¥'e earth pressureTo this end, analysis vas made of data obtained ingenerally matches the theory, and the Mononobe-Okabetheory has become the established theory for assessingthe seismic earth pressure (Seed and Whitman, 1970;VVhitman, 1990). In Japan, research on caisson revet-centrifugal rnodel tests of gra¥'ity retaining ¥valls ¥vhich¥vere conducted by the Public Works Research Institute.The analysis sho¥ved that ¥vhen the inertia force becomesi' Director. Engincering Department, Japan Construction ivlethod and '{achiner) Research Insli ute, Japan (nakamurase.cmi.or jr)) ('ormerl)'SenioF Researcher of Public vorks Research Institute, Independem AdministraLi¥'e Institution).The manuscript for this paper ¥vas received for revie¥v on April 30, 2004; ar)proved on December 13, 2005.vritten discussions on this r)ar)er should be submi ed before November 1, 2006 io the Japanese Geotechnical Societ)', 4-38-2, Sengoku,Bunk),o-ku, 'Tok),o I 12-001 1, Japan. Upon request the closing date may bc extended one monih.l.J"O NAKAlv* iURA1361 5ao280 40 26040 260 40 28040 260M24c}A20 A21 A26Fig. l.Illustration of the centrifuge modei (dimensions in mm)Table 1.He'sllCaseN'o ¥. todel Proto{mm) (m)CsSe 17 iOO 9C Ise IS 300Case i9 ioo 9Conditions of centrifu"e model testFour]ialion Embcdmenl GreundInpUl wave¥¥idlh dep h lhickness,odei Prolo ¥,odel Proloodel{nlnl) (m) {nlm) ( ) (mm)Pro otm)llJoa 3j 11 3 7S 50 ;OO75i2o¥V Ye (** f*)Proto number : 2rouniodelSi n ; 20lSin 603 75 SO 1 5 100Case 21 300 9case 22 300 9125 3 75 50 1 5 100 3}(obc PredorT]il ) 3 75 50 iOO 3Kobe Predomi 40'3 75 1 5 100I'S 3 75 50 1 5 IaOLomorPredemi 60 211*Ci se '6 SOO ; 9l 2S 3 1 OO* 35 50 1 5 1005015;¥omori30(g l)600Dr}3060a20S7Dr30600SS6aoO )3 SS Dry 3O i S600O )3088: S8O 53Drl 30 1 S30600O S3 Dry 30 IS8S i1>re oml 40Sweep 60PTototvpeS760: 2Predo ni 120 4Case 15 300 9(G)20Kobe Predomi 20 S3 AomQri5aecel({+* (G}) todeii 3 OO. 5353r)r)l,30 1.Slc < 23 300 9c s 2s ioO 9dS7C se 20= 300 915IY(lll20 O S3 Dr} I SSill 40;_iel'c:betlveen den IY conlenlFrequenc (Hz)ShapelS 75 iaO50S}I kins ncceleralionFrictionalcoef71lcien Rel ;;live IV ler CenlriiO 53600600S)r}30600SSDr}30O - 6TOt e k by (O 5 G! Y l e) (incre se40 IYaves) 1 7 2lboui!'¥t ve)maximum, the earth pressure on the ¥vall is almost theearthquake motions such as those recorded during thesame as the static acti¥'e pressure and the thrust force actsat the point of 400/0 of the wall height (Nakamura, 2005).1995 Hyogoken-Nambu Earthquake (Matsuo et al.,In this paper, the behavior of backfill and gravitydeformation of a real gra¥'ity retaining ¥vall. A crossretaining ¥valls observed in the centrifuge model tests issection of the model used in the centrifuge tests is sho¥vncompared to the hypothetical conditions of them Fig. I and the test conditions are summarized inTable 1. The models were prepared in a rigid steelMononobe-Okabe theory, to make clear ¥vhy the seismicearth pressure is similar to the static active pressure.F,XPERIMF.NT MF.THODA series of dynamic centrifugal model tests ¥vasconducted in order to investigate the effects of inputshaking on the seismic behavior of gra¥'ity retaining¥valls. Tests ¥vere conducted using the dynamic geotechni-cal centrifuge in the Public Works Research Institute,.Japan, which has a radius of 6.6 m and the maximumpayload of 400 ton-G. The shakin*' table used, Ivith acapacity of 40 ton-G, can accurately reproduce stron_g1998). It can reproduce the earthquake stress andcontainer ¥vith inner dimensions of 1500 mm in length,300 mm in width and 500 mm in height, ¥vith both sidesmade of high-strength **lass for observation.The model ¥vall is divided into three parts horizontally,both sides being made of concrete, and load cells areinstalled in the central part. Frictional resistancemitigation bet¥veen those three parts and the container isachieved by using various Tefion products (Fig. 2(1)).Moreover', to measure the total earth pressure acting onthe model and its distribution, all parts that touch theground and the backfill are co¥'ered ¥vith plates to receivethe pressure at each section, and the earth pressures are REEXAlvIINATION OF ,IONONOBE-OKABE THEORYl 372 Io E+- HPTFte on coat s rBy)/ Cqde seBl (terlon sqft tspe).1RConcreteCenorete3el)ort seat S npereroundVesselR( I ) Friction mitigationby Teflon products6Photo 1.Completion of model7No.'iJi - ;;Load celisB44 11 5H+*-+1go@ C@S5 52:26 ,ia21 .5soISgC l_T13 ,sIS2s 25 2s(2) Cross-section of' model retain in_(T* ¥vall(dimensions in mm)rrg 2. Illustration of the gravity retaining waHat sever'al points in the retaining lvaH and backfill. Earthpressure acting on the retainin*" ¥vall was measured ¥vithdivided load cells. Displacernents of the r'etainin*' valland backfill soil ver'e measur'ed vith laser displacementtransducers (Fig. 1). Deformation of the backfill andretaining ¥vall ¥vas measured usin*' a high-precision highspeed camera by rneasuring the coordinates of the targetson the r'etaining lvall and sand (Photo 1). The behavior ofthe gra¥'ity r'etaining'all and backfill was exarnined usindata on the acceleration, earth pr'essure, displacementand coordinates of targets.In the tests, after applying a centrifuge acceleration of30G, horizontal shakin*' ¥vas conducted. A sinusoidal¥vave of' 20 cycles ¥vas applied to cases 17 to 19 varying thefrequencies, the ground motion recorded at the Kobemeasured with load cells behind those plates. As for theMaritirne Observatory during the 1995 Hyogoken-plates at the corners of the model, the ed**es are cut by 45Nambu earthquakedegrees to allow pressure measurement at those corners(Fig. 2(2)). The ¥vei**ht and center of gravity are madeone recorded at Hachinohe-harbor during the 1968to simulate those of a prot.otype *'ravity retaining wallmade of concrete. The coefficient of friction bet¥veent.he retaining wall and ground (also backfill) is set totan (2c/3) according to official guidelines (Japan RoadAssociation, 1999) by pasting sand paper on the plates.By running consolidated undrained triaxial compres-vas applied to cases 20 to 22, and theTokachi-oki earthquake was applied to cases 23 to 25varying the predominant frequency, to determine theinfluence of the input acceleration on ea 'th pressuregeneration. For each case, the model ¥¥'as shaken severaltimes by gradually increasing the accelerations. In case26, the arnplitude of input harmonic acceleration wasincreased gradually to the maxirnurn, in order to deter-the material used in the model experiments vveremine when the slip line appears and how it proceeds.Dry sand was used in all the experiment cases, so themeasured (Japanese Geotechnical Society standard JSF Tseismic behavior of a gravity retaining ¥vall may differsion tests on Toyoura sand, the strength character'istics of524-1990). Dry Toyoura sand was first poured into afrom the experiment r'esultsmodel container to a depth of 100 mm and vith asaturated condition.predetermined relative density (Drvhen the real backfill is in a880/0). After themodel retaining wall was placed on the sand, the backfill¥vas also prepared by pouring. Various sensors ¥vereinstalled at predetermined positions during the modelRESEARCH POINTSThe Mononobe-Okabe theory assurrres that the backfillground preparation, and a mesh for deforrnationis a perfectly rigid plastic body. Thus, it is thought that ameasurement was created ¥vith black colored sand at thefront of the container. Rivet-shaped targets that moved¥vit.h the sand were also embedded in contact ¥vith theretaining ¥vall and backfill vill behave during earthquakesin accordance ¥vith the follo¥ving three points.front. The completed experiment model is sho¥vn infill. It falls at a time of' active state, and rises at atime of passive state.Photo l.Acceleration was measured by accelerometers installed(1) Displacement: A rigid ¥vedge is formed in the back-(2) Acceleration:Seismic acceleration occurs simultane- NAKA ,1URAl ,38o 3520.SSO lcase26 ( r dual increase wave)o.30Sr? rTle,e70 'rQ*Q f-19 45cQo 25oo 201 g. 40,8.045 loii2 :* ]TS5c{ 2fIi2OL7e1 17 )25 jT**'***7015+'OIs6 Oe5o lo15.030l f1 4.S20Io 05I I ll 'f513oo-o 05B 252.*,*Oft?;?eet7_25'SLL t5.740 11 E 245o oooIi 22l3 S5 J--tSIS. SS _l S.2eJi B'7G5}7; ll12.S751 -*TfoN?**** j 9・755T j15ii25 _1:c:I jB.530 rTi 1 L15,24e1J -15-7a5j3^Bsrl14[ 42a5!{ 4.7so!600.**.- *=**400Ol 200If8 ol*<i l] l: - 200-l+-400-600Frg. 3. DiFection of data800Tl Tll:X-?Input Aec.2134on or in the retaining ¥vall and backfill ¥¥'ere analyzed indetail. For point (3), the earth pressures on the retainin_'._lvall measured by the divided load cells vere analyzed indetail. The validity of the hypothetical conditions of the15 16 17 181920 21Time (sec}ously and uniformly throughout the retaining lvalland backfill, and there is no phase difference.(3) Earth pressure: Earth pressure act.ing on the retaining*ali varies in a triangular distribution ¥ *hoseprofile does not change ¥vith time.For point (1), the displacements of the retaining ¥valland backfill measured by laser displacement transducersand high-speed photography ¥vere analyzed in detail. Forpoint (2), the values obtained by accelerometers installedTl(gai)Fig. 4.Horizon a! displacement and input acceleration at top of walidisplacement at the top of the ¥vall against time. The statebefore shaking is set as the starting point O in the figure.The lef't-side figure of the upper ro v (: ;) sho¥vs t,hedisplacement vector' during the time ¥vhen the top of theretaining: wall is movin from the O state to a localmaximum state (A). The middle figure of the upper ro¥v((._ ' '__;) illustrates the dlsplacement ¥'ector from (A) to theMononobe-Okabe theory lvas examined by analyzingcontinual local minimum state (B). The right-side figureof the upper ro¥v (('i) reveals the displacement ¥'ector'these three points.from (A) to the follolving local maximum (C_). Theleft-side figure of the lo¥ver ro¥v (, ,;) demonstrates theF.XPERIMF.NT RESULTS AND DISCUSSIONIn this section a]1 data is expressed in prototype scale,displacement vector from O to B. The middle figure ofthe lo¥ver ro¥v ( i ) indicates the displacement vector fromB to C. The right-side figure of the lo¥ver row ( ,) sho¥vsand the positive directions of the data are as sho¥vn inFig. 3. From Fig. 3, ¥vhen the model undergoes negati¥'ethe displacement vector from B to the follo¥ving localacceleration, the inertia force occurs in the activeContour lines sho¥ving vector sizes have been dra¥vn.The broken line in each fig:ure is "the ultimate residualslip Ime" (heremafter "the slrp line". Photo '_.).direction.Disp!acel 71 en tTime histories of horizontal displacement at the top ofthe retaining'all and input acceleration are sho¥¥'n inFig. 4 (case 26, gradual increase in shaking amplitude,minimum (D).The third figure (:'._.',,) ¥vithin each of Figs. 5 to 9 sho¥vsthe displacement vector from a local maximum displace-ment to the follo¥vinlocal maximum. Note that the' Hz, embedment depth 1.5m). Figures 5 to 9 sho¥vdisplacement vectors, ¥vhich indicate the size andcontour line along the slip line gradually becomes parallelto the slip line as time elapsed. As deformation progresses, the contour line interval along the slip line becomesdirection of displacement at each part of the ¥vall andnarro¥v (Fig. 9), implying that deformation is graduallybackfill at ¥+arious time intervals. The conceptual idea ofFigs. 5 to 9 is sho¥ 'n in Fig. lO, ¥¥"hich plots horiz,ontallocalized along the slip line. Holve¥'er, the second figure(( ;: from maximum to minimum) and the fifth figure ( i : REEXAN'IIN ATION_ T OF ¥,10NONOBE-OKABE THHORYlv "91086I4:T23"c+o21086:C(DQ1:4:y32o8 6 -4 -2 O 2 4 6 8 6 -4 2 O 2 4Distance (m) DistanceFig. 5.6 -8 -6 -4 -2 O22468D istan c e (m )(m )Displacement vector no. l (case 26)rTILL17 14.520 second1086:z:cD4gq23:ro21086:C(1)ol:432o8 -6 -4 -2 O 2 4 6 8 -6 -4 -2 O 2 4D istance (rrl ) D istance (m )Fig. 6.6 8 6 4 2 ODistance22468(m )Displacement vector no. 2 (case 26)from minimum to maximum) Ivithin each of Fi**s. 5 to 9sho v that the backfill soil oscillates even belohv the slipline throughout the shaking. It turns out, therefore, thata rigid ¥vedge is not formed and that displacement is notvithin each of Figs. 5 to 9 (cumulative displacementvector from the start of shaking to a local maximumdisplacement of the top of thevall). The displacementconcentrated along the slip line as the Mononobe-Okabetheory assumes.vector is horizontal in the early stage of deformation (inthe first 1 3.995 seconds of shaking). After slanting slightly at 14.5,_O seconds of shaking, the contour lines becomeNo¥v the attention is focused on the first figure (*_ 1))par'allel t.o the slip line at 15.525 seconds of shaking, after NAKA,lvIURAl 40l 5 .525second1086 n:4 d ;+2 So21086 :C4 iEi"rt'2o8 6 4 2 O 2 4 6 8 -6-4 -2 ODistance (m)DistanceFig. 7.2 4 6 -8 6 -4 2 O(m )24D ista n c e (m )26 8Displacement vector no. 3 (case 26)1086 n:4 crq:2 So21086 :E4 q :r2+o8 6 4 2 O 2 4 6 -8 6-4 -2 ODistance (m)Fig. 8.2 4 6 -8 -6 4 -2 OD ista n c e (m )Distance24(m )6 82Di_,,pla:cement vector no. 4 (casc 26)¥vhich the slope of the contour lines does not change. It isc.onditions of the Mononobe-Okabe theory, in whichnoteworthy that the contour lines above the slip line areparallel with nearly equal separations to¥vards the top ofthe retaining ¥vall. This means that the entire backfillabove the slip line deforms, follo ving the displacementonly the slip line portion experiences plastic deformationof the retaining ¥vall.A cce!el'atio nIn light of this, it can be stated that the hypotheticaland the other portion remains ri**id, do not properlyexpress the actual seismic behavior of backfill.This section discusses the acceleration response of the REEXA lli¥'ATION OF NiONONOBE-OKABE THEORY141129 20.070 second1086!:d i'4:yr+32o21086::(Dd: "4IfS2o2-8 -6 4 2 O 2 4 6 -8 -6 -4 2 O 2 4 6 -8 -6 -4 -2 ODistancem)24D istan c e {m )Dist,ance (m)Fig. 968Displacement vector no. 5 (case 26)(QLho'lo4coc a)dScoNoIContour interval O )Oe):O20mm,r3Jr6J :1 OmmTime (sec_)Fig. lO.Conceptual itlea of displacement vector (Figs. 5retaining wall and backfill.Three acceleration time histories for case 26 (input, topof wall and top of backfill) are shown in Fig. 1 1 . Case '_6used a model of 9m in height, 3.75 m in foundation¥vidth, and 1.5 m in embedment depth, ¥¥'ith excitation9)vall undergoes an inertia force in the active direction.Figur'e 11 demonstrates that from 19.'_15 seconds to1 9.,3,35 seconds,vhen the inertia force occurs in the activedirection, the acceleration at the top of the backfillchanges only about '_OO gal, from 700 gal to 500 gal,sinusoidal shaking with the frequency of 2 Hz). In the¥vhereas the acceleration at the top of the retaining wallchanges about 1800 gal, from 1300 **al to - 500 gal.Figure i2 sho¥vs the acceleration distribution at eachfi**ule 1lhen the acceleratron rs ne'ative, the retainin_"*.point shol 'n in the acceleration time history in Fig. 1 1 . Incarried out by gradually increasing the input acceleration(reachin*' 600 gal after 45 cycles and 30cycles of N. VAKA ,1URAl 422000case26 (Gradual increase wave,2Hz,6aogaD [ i(H::9m,Bz3 7 5m,C l .5rr Dr88P ) l15QQf '; ]{;Jf[ !¥jlOOoc soCoooocQ o ji;-jLQ)oor*!f<11e)t,]]]1lsOO<i'p =;i L ;r i.[npvt:11 ;J'-1Cxx}l iI, i":;;::!1Wail top* t'l fBack fiil to- 5co ,9.2159 2759 335 19.3 59 455 I9 515 19 575 19 835 19 695 19.755 1 9 875 19 875 1 9,935Time (sec.,)Fig. 11. Acceleratron tlme hlsto ¥Photo 2. Ult mate reslduol slfp I ne (case '6)Acceleration distribution (gal)10 l:r {9'215seo'Ei :T12i98 : : rlr'. ¥'t=C;eQ / /ee:10j('/: ' ¥Goet+Ef¥¥) .YI :¥ICfT*-2c:oel;: :_rr' {e: :: 1:,j e /r// ¥ 43-1Q{) f (I llr¥/-/- ei' '"-1-'2 5= _1 : :/I - ee//¥¥'O,' -1/' '-2-3rG: T19,365seoZilo r5: 1T;";:i19.330seo'('i:lii:!'//__8I/j,1:' ::-__'4 ',E75=h. 'l¥ ¥¥;: 1c' '*'e_20 ._e :: '" -300 / -:=-{1't ]¥_]t:2 j; ;(" j : ] --::- e o ---'i ' .11-/ jj".;r_._//1//'¥_-sIs5c ./ 'o( ' -=-1'-*2310 9:$-'/ ''iO_, r--_ --//"=1'/=-- '_ =Ol5 -709A85seo1 1:*"-== *+ * 'l'i r /5ee __I -e'_{/* '-40e/':3 - ll3 1 o{)rl1o ;-' /!"ee /t_ _ ' //T_ 20 )7il:tll'rli;iEF i14: g.605seo- !6 -i( Q-IofrlllO.r' r-::'/ tIII "'e " /-2 i:'=::< ";1r-__; "'+' :_: JC'-:; ;ie'___'_ Lfr'ic¥:; ,_____::;;;/TL/ 'J ;'LS/; /____:'.**I/1'/f '- = * ==* --Eoe/'* '==*':-'-'1 rv Qil::¥:r¥-/r"'*O-1-2-310e 19i 65secc )::;;'7! ¥ e fT/-1;_ /r *r---¥ t' ¥¥h / ¥ L/l ' f::Ec 7' L' '¥ f/8 . 1/_ cH)¥l , /h/llI 1rsc ¥} * Distanoe mAcceleration distribution54 E_3 ;O6 7 8 -43-2-1 O 1 2 3 4 5 G 7 8 -4-3-2-1 O 1 2 3 4 5 G 7 8 9Oiste"ceOisten*e(m)me2(' ¥ - //0 0rl: "Dist nce (m98=** ** i- L//;- 'i)l412 5<c:J__*__-- '¥¥i3ee'1 'o/// fl1 : ;]- :' j ieO ii i !¥l!2 3 4 5 G 7 8 -4-3-2-1 O 1 2 3 4987_lT//L;'1:/11'/1: 'VII::////,o )fi= i= ' 3sc } __='.Dist noe (m)'/t・4ee'-1-3:__'1./' ': '== ='--- /,/--tIIIIS:;/'t¥? . _ :::'/21;!5: i/y//: /(//'(//6l74;j _:.' i'/f/(+.* ==( ,:1seo_98 ¥ec ) - ¥!'_4eQ' '5_/¥-2c 117'1]_......'--", ' 1-' ;;; ';:- I eo2 r-200_-4-3-21 O32 5:1rfil1 2: TIT9.54esec.9 51 5sec_/' -/ /' "_'-70{]J l! /54 tl -__r/4-LJ] :c{ 1jl/eo J.l/e-2-3! i '3 : Ti-7c-eeQ9.455seo'10 rT7ilr1 3:9'07"¥=' ) //1C , ..l/rl-ec )j87-3_'Q Gc8-/ 'g87'-// r//'T i_r310rrT8L I li:;1ii.425s*-e"j ¥'_'_¥ -vl j;,: cL ir;1;2 ¥ ' _'*/r/= ' / '-'¥-13niiil7: 19.395seo.'- =/{i{'l/C'/ oe-:eioI i f/ l-/ -=:¥L-4ee;//4;:i'i*h' ¥' :' lll_ *+ c")_ly//'i!_i:¥2 t .-2/f ! ¥ 1l 2Oeg87e2eo: :: rltlli2C'e llYllgJ¥ v¥vv _ ;¥o ":il !rlll"""1/ /1'//;tf/' lco// /1coJ ; ;iO CJllr+4:c }/_ :/1 llr ,.co{'el! 7i J "SCe /.L //. _ I42rrll1 7r ]4: 19.30jsec9.275see8C(i ¥_ ¥7e'; :*"/. ^L/'eQ :)i;j'Ce5 PS/7cY/r;:F^o"/1:/I1:3 ',._"E:?;/(;/* l/ *7E3:9.245sec-1-23I i43沢EEX、へMIN.AτION OF MONONOBE−OK.叛BE T}{EORY800800case18(sinusoidal二2,0Hz.600gal) }600    case21(Kobe,predominant:2.OHz,600gal)600                           1400  400I﹃﹁亀 200㊤  0の①  0の﹃L島 2008制一200『 ■『 『 『 『7                「 ユ  7皿  雰皿 電︸r   l制一200a&一80021﹁一6001﹁     1⋮ヨ謡一400…     li4   6i…… 12  14  {6一8002  3     l             瞬45  6  7  8  9  10Time(sec,)Time(sec,)Fio 玉3. l      l   l   i一600… 8   霊O三一4001叩uねcce置e蘭on(c便se至8,s藍nusoidal2.OHz)F韮g,14、 Inpu霊accelera【ion(case21,Kobe moIion2。OHz)梱爬seG)6,貰嗣ぶbD6工椀・、粥Flu 裏5,EIユr吐h pressure distribu霊ion(case玉8,si日usoidε選12、0}lz)て、r∩e爬seξ)  43……飼ぶbO“δ4>工簾50Fiσ 16.Ear{hpressuredis{ribu{io臓(c艮se21,Kobe鵬o吐ion2.OHz)宅his 丘gure when the input accelerat圭on is at a Iocalbetweenretainingwallandback薮Ilissig涌cant(di仔er−mlnimum(atl9.335secondsofsbaking),theaccdera老ionence of about1000gal at19.335sec).Thus the reta重ninginside the whole retaining wall is nearly ulliform,beingwall starts芝o experience displacement slm麺ltaneouslyeq“a正to the input acceleration,so the lnl)ut accelerationwith the exertion of圭uertia force in the active d呈rection,lsinstantlytransmittedeveぎ1tothetopoftheretalningwall.Thisimpliesthatthedi貸erenceinacceleration  It is clear from these experimental nndings that theand3fterwards the backf崖1至starts displacemeRt. NAKA ,lURAl 449876E5+JJ:Ao321Oo I o 20 30 40 50 60 70 80 90 1001101 201 30Earth pressure (kN/m2)rig. 17E,arth pressuFe distribuiion at local maximum of input accelerationthat seismic response occurs simultaneously and uni-ly changes ¥vith types of base shakin .In the case of the seismic ¥vave, the earth pressure in thehypothetical idealization of the lvlononobe-Okabe theoryformly in the ¥vhole retaining ¥vall and backfill ¥vithoutlo¥ver part of the ret,aining ¥vall, ¥vhich greatly contributesphase difference do not appr'opriately express the realseismic behavior.to the total earth pressure, is not as significant as in thecase of the sinusoidal shaking. This is because the relari¥'eEa/'t/1 PressureFig:ure 13 sho¥vs the time histor¥' of a sinusoidaldisplacement bet¥veen the retaining ¥vall and the backfillis not so small as to generate large earth pressure In thetest, the direction of the seismic ¥vave acceleration alter-excitation for case 18. Figur'e 14 sho¥vs the time history ofnates in shorter time intervals than in the sinusoidala seismic input motion used in case 21 . The predominant¥vave, and there is not enough time for the backfill soil tocatch up ¥vith the displacement of the ¥vall. Since ther'e isfr'equency of the seismic motion is the same as that of thesinusoidal shaking ('- Hz). Figures 15 and 16 show earthpressure distributions corresponding to Figs. 13 and 14respectively.Fi**ure 17 illustrates the earth pressure distribution atsi.x' moments sho vn in Figs. 13 (J, and) and 14 ( 3,___ _ to:) ¥vhen the input acceleration achie¥*es local maximumsor minimums. At Ll, ii and,.. _, the iner'tia force acts inno sinusoidal shaking in reallty, greater emphasis shouldbe placed on seismic ¥vave experiments, e¥'en thoughexperiments using both seismic vaves and sinusoidal¥vaves need to be conducted.In the present seismic coefficient method that isemployed by the lvlononobe-Okabe theory, stabilityanalysis of gravit), retaining ¥valls is conducted as the; itearth pressure and inertia force simultaneously takestheir maximums. In this sense, a study ¥vas made of theAccording to the hypothetical conditions of therelationship bet¥¥'een the inertia force and earth pressureincrement, especially at the time that the inertia force isthe active dir'ection in t.he model,¥*hile at 2 , L , andoccurs in the passive direction.Mononobe-Okabe theory, the earth pressure distributionhas a triangular shape that does not change ¥vith time.Figures 13 to 16 and 17 in contrast sho v that the earthpressure distribution is not triangular, and that its sizeand shape change over time. In addition, in these figures,the dlstribution shape of the sinusoidal ¥vave substantial-maximum. The earth pressure increment stands for thepressure during shaking minus the one prior to shaking.Fiure 18 sho¥vs time histories for the inertia for'ce andtotal earth pressure incrernent. The inertia force iscomputed by multiplying the negative value of the mass 1婆5REEX、AM互NATION OF MONONOBE−OK、へBE T}{EORYz亀o300釜 300250護 250護 250200旨 200爵 200150α 150I oo印 100註 100 50ヨ  502  50  0ちも  oo −50き 一50〈  300q 重5022−50蓬9 −i oO黛一100−100300『  300}  300200  200  200z100  茎OO  1003一100一100一重OO−200−200−200−300−300−300−400−400E2∈234っ67892GOcase20(K。be,pred・ml;4Hz〉 1z 150100ro⋮1  01−50l l I︸㎜ 50o[  1 li   l[i   I i l−100   l i…   i l l−150−2003001200r   l  1100  0一100…   l I−20010翌 150100曽 100 500  50  0為  0−50£ 一50−100亀δ一100−150琶一150−200300鍵一2001 −100−200 −200−300 −300 −400−4002  3  415 20 25 30 35 40 45 505  6  78  9  10Umi1 …i t[11i口日i F 」1l l∬19 100 50α  50結  0  0ヨ ー50−5028 −50−100お一100−150琶一150−200300−  300200(  200自9o㌔δ一100o琶一150∈9 −200幽  300  200算  100zo結  02竃ε一200E\∼  100重OO0   00一100} 一100−200−200肇醜200−300−300一 _300−400−400 −400一100o2  3  4   5   6   70   24   6   810  12 Tlme($ec.) Time(seG〉Fiσ 18.4567891011 12131415      Tirne(sec)z  150ユ  50…l l I l i目  i i l  I旧3100之菖 10071511目11 1 口1目  i150\護 150一…  l I 門l ?  200……i l i ii l目200?  200…網目目i i l i1口目目i lTirne(sec〉     Tlme(sec〉1…1 1  i1雛閏i轟 1   目 i   0一玉00日I u i口口I I以 1 11i …幽  300  10018 20 22 24468{0121416     Time(sec)case22(K。be,prεd。mi:i.3Hz〉 i2z 100i   i…i−40Q(  200  200i l l−300200150\1「12200ヌ                  ㎜,−400340678910111213141516      Time(sec〉     Time($ec)010 12     Time(5ec)2  4  6  8 14 16 181脆erIiaεorcea賎dincremen亘ofearthpressureof the retaini119 wall by dle玉nput acceleration。Ia themagllitudeof出einel齢tlaforce.丘gure,the same ID number is assigne(i to the il1ertla force  In Fig. 18,when the retaining wal玉is excited by thealld eart}1 pressure increment for 亡he salne time.Thissinusoidal motlon,the earth pressure increment reachesεaclli重ates understanding the ear由pressure incl’ementwhen the i!1ertia force is at由e maxlmum.For the seismiclocal minlmums(most of the values are O)∼∼・hen theillertia force is at the m&ximum,while fα由e cases ofw&ves, t紅e nulnbers are arrangecl in the order of theactual seismic mo重lo陰,the ear乳h pressure incremen宣does NAKAlviURAl 46not necessarily achieve a local minimum at the localthe active direction. In reaiity, the earth pressure in-maximum of the inertia force. Ho¥vever, the earthcrement is around zero, so the earth pressure ispressure increment is around z,ero ¥vhen the inertia forcenearly equal to the initial value prior to shakin*¥vhen the inertia force is maximum.is maximum. Accor'ding to the lvlononobe-Okabe theory,in contrast, a retaining ¥vall is subjected to the seismicactive earth pressure ¥vhen the ¥vall and backfill areloaded by the inertia force in the active direction. There-ACKNOWl,EDGMF,NTSfore the hypothetical conditions of the Mononobe-OkabeThis paper is based on a doctoral dissertation submit-theory do not properly express the actual seismicted to the University of Tokyo. The author gratefullyresponse of rctaining ¥valls and backfill.ackno¥vledges the helpful discussions and su_9:gestions ofProfessor Fumio Tatsuoka.CONCLUSIONFor the earthquake resistant design of retaining ¥valls,it is not appropriate to use the seismic earth pressurecalculated by the Mononobe-Okabe theory, because theseismic behavior of a retaining ¥vall and backfill based onthe hypothetical conditions of the Mononobe-OkabeREFERENCF,SI) Ichihara, lvl. and ivfatsuzalva. H (1973): Earth pressure duringearthquake, Soils anc! Fou,1c!a!ions, 13 (4), 7586.2) Ishibashi, I. and Fang, Y. S. (1987): Dynamic earth pressures ¥vithdiff'erem ¥ 'all movement modes, Soi!s anc! Foundations, 27 (4),l 1-22theor'y do not match the ac.tual seismic behavior. Thefindin*'s to support this conclusion are summarized in3) Japan Road Association (1999): Road earth¥vork-retaining lvall¥vhat follo¥vs.4) Kazama, lvl. and Inatomi. T. (1990): i¥,10del vibration tes for theseismic earth pressure acting on the rigid caisson foundation. Pr'oc.(1) Under the hypothetical conditions of theMononobe-Okabe theory, a rigid ¥vedge is formed inthe backfill and slides down along the slip surface¥vhen excited in the active direction, ¥vhile it slides up¥vhen loaded in the passive direction. In reality, thepart of the backfill that follo¥vs the displacement ofthe retaining wall plastically deforms ¥vhile slidindo¥vn.( _) The Mononobe-Okabe theory assumes that nophase difference occurs bet¥veen the motion of theretaining ¥vall and backfill. In other ¥vords, theinertia force acts simultaneously on the retainin_",_wall and backfill. In reality,¥*hen they are excited inthe acti¥'e direction, the acceleration is transmittedinstantaneously throu*"h the retaining vall and thentransmitted into the backfill.(3) The Mononobe-Okabe theory hypothesizes that theearth pressure on the back face of the retaining ¥vallhas a triangular distribution. In reality, the distribu-tion is not triangular and changes ¥vith time.Furthermore, according to the hypothetical condi-tions of the Mononobe-Okabe theory, the earthpressure increases when the inertia force is loaded inguidelines, (3) (in Japanese),JSCE, (416/1-13), April, 419428 (in Japanese)5) lvlatsuo, O., Tsutsumi, K , Kondo. K. and Tamoto, S^ (199S): Thedynamic _ ;eotechrlical centrif'uge at P¥VRi. Proc. Int. C0,IfCentrtfl!* e 98, Tok"¥'o, Baikema, 2530.6) Mononobe, N. (1929): Eanhquake-proof cons ructiou of masonrydams, Proc1・flor!c! Eng/'*"^ Conf., 9, 275.7) Nakamura, S ('-005): C_larification of seismic behavior of gra¥'ityretaining vail by using cenlrifugal model tests and a proposal forrarionalizarion o f the seismic coefficiem method, Proc. JSCE, (785 /3-70), ,Iarch, 107-122 (in Japanese).8) Okabe, S_ (1924): General theory of earth pressure and seismicstability of retaining valls and dams, JJSCE, 10 (6), 1277-13239) Seed, H,, and ¥¥!hi man, R. ¥r. (1970): Design of earth retaining:struclures for dynamic loads. A SC E Specia!t.v Conf・ , Later'(T! Stressin !he Grotmc! ancl Design of Earth Retaiiling Stnlc!ures, 103 147.lO) Sherif, ,1., Ishibashi, . and Lee, C_. D. (1982): Earth pressuresagainst rigid retaining ¥valls, J. Geot. Eng.. ASCE, 108 (CJT5),6 79-695Il) Steedman, R^ S. (1998): Seismic design of retainingvalls, Proc.Inst. Cil'_ Fngrs. Geotecll,, Engrg., 131, Jan., 12-22_2) ¥¥ratanabe, K , Kobavashi, Y., Touhata. I_ and ,Iaeda, T^ ( 999):Shaking-table tests on seismic earth pressuFe exerted on retainin :¥vall model, Proc. 2nc! Int. C_onf. Eanllquake Engl ;"・, 297-302^3) ¥rhitman, R. V. (i990): Seisn ic desigrl and behavior of gravi yretaining ¥valls, Proc. Co,rf on Design anrJ Pe!:fornzance ofEar!hRe!ainin*" S,nlc!lu-es, ASC_E Geotechnical Special Publication,('_5), 817-842.
  • ログイン
  • タイトル
  • Modeling Granular Crushing in Ring Shear Tests: Experimental and Numerical Analyses
  • 著者
  • s. Lobo-Guerrero・L. E. Vallejo
  • 出版
  • soils and Foundations
  • ページ
  • 147〜157
  • 発行
  • 2006/04/15
  • 文書ID
  • 20896
  • 内容
  • FSOILS AND FOUNDATIONS¥f ol46,l¥Jo, ,147l57, Apr. 2006Japanese G eotechnical SocletyMODELING GRANULAR CRUSHING IN RING SHEAR TESI'S:EXPERl /IENTAL AND NUMERIC 'AL ANALYSESSHBASTIA ' LOBO-GUERRERoi) and LUIS E. VALLEJoii)ABSTRAC.1'A fault consists of a zone of heavily fragmented granular rock (gouge), ¥vhich is confined bet veen t¥vo rough ¥¥*allsmade of fractured rock. The granular gouge is the result of previous fracturing of the vall rock b), the combinedeffect of compressi¥'e and shear stresses. Through time, the granular fault gouge ¥vill experience ¥'arious episodes offurther fragrnentation (crushing) as a result of the mobilization by shear of the fault ¥valls. The e¥'olution of crushingin a sirnulated gouge material ¥vas studied using laboratory ring shear tests and DEM ring shear simulations. Thelaboratory ring shear tests ¥vere developed using su*・ar as a ¥veak granular material. It vas found that the residualfriction coefficient of this material maintained a constant value regardless of the se¥'ere degradation of the particles,This degradation lvas induced by increasing the angular deformation or increasing the applied vertical stress.Moreover, it lvas found that the grain size distribution of the original uniforrn material evolved to vard a fractaldistribution of sizes. The results from the DEM simulations confirmed those from the laboratory tests and pro¥'idedalso a visualization of the e¥'olution of crushing. Event though or'iginally DEM does not consider particle breakage,this was allo¥ved by replacing particles fulfilling a predefined tensile failure criterion lvith an equi¥'alent gr'oup ofsmaller particles.Ke¥.' words: discrete element method, micro mechanics, particle crushing, ring shear test, strength (IGC: D5/D6)The continuous mo¥'ement of the ¥vall rocks in someLobo-Guerrero and Vallejo ('_005). The previous studyfocused on the e¥'olution of granular crushing In thedirect shear apparatus. The granular samples during thenatural faults generates a finely crushed rock layer calleddirect shear testing ¥vere subjected to a maximumfault gouge. The behavior of these faults is stronglyhorizontal deformation of 9 mrn during the actual testinfluenced by the mechanical properties of this material.Previous research has sho¥vn that particle crushing resultsand 5 mm in the DEM simulation of the test (theseINTRODUCTIONdeforrnations correspond to shear' strains equal to 430/0and 170/0 respecti¥'ely). The present study focused on theevolution of crushing at large ¥'alues of deformation asas a consequence of increments on stresses and induceddeforrnations inside faults, and the grain size distributionmeasured in the ring shear test. The maximurn shearof the fault gouge material evolves to¥vard a fractaldistribution of particle sizes (Sammis, 1997). Previousstrain induced in the three-dimensional actual ring sheartests ¥vas equal to 53400/0. In the t¥vo-dimensional DEMresearch has also sho¥vn that changes in the gr'ain sizedistribution of a granular material are associated ¥vithchanges in its hydraulic conductivity and defor'mationsimulations of the ring shear test the samples ¥veresubjected to a maximum shear strain equal to 2500/0. Thecharacteristics (Vallejo, 2003). Therefore, is important topurpose of the 2-D DEM simulations ¥vas to understandunderstand how different properties of the fault gougechange as a consequence of ..*(;ranular crushing and howand visualize ho v crushing develops in granular materialsubjected to large values of shear strain similar but notequal to the ones induced by actual rin_ : shear tests. Byconducting the ring shear tests (actual and simulated),important aspects such as the mobilized shear strength atthese changes affect the satiability of faults. This studyreports the results obtained in laboratory and sirnulatedrin_ : shear tests de¥'eloped in order to analyze and visual-ize the evolution of crushing in a continuous shearingzone. Laboratory ring shear tests were conducted using a¥veak granular material (sugar), vhile the computelsimulations ¥vere developed using a modified version ofthe discrete element rnethod (DEM).This study is a continuatiori of thelarge ¥'alues of strain, the possibility of reaching a criticalstatevhere granular crushing stops, and the developmentof a fractal grain size distribution as a result of' crushingtaking place at large shear deformationsvere studied.vork presented by*' Ph.D_ Sludent, Depar ment of Ci¥'il and Environmental Engineering. Universit)' of Pittsburgl . Pi tsburgh, PA, USA (sel2(;,pili.edu).**,Professor, ditto (vallejo civ.pitt.edu)The manuscript for this paper ¥vas received fbr revie v on Februar) 15, '_)005; approved on January 13, 2006.¥vritten discussions on this paper should be submi ted before November l, 2006 o the Japanese Geotechnical Society, 4-38-2, Sengoku,Bunk}'o-ku, Tokvo 1 12-001 l, Japan. Upon requesl the closing date may be extended one month.147 LOBO -GUERRERO AND ¥,ALLEJO14SPREVIOUS WORKStudying crushing of gr'anular materials has al¥vayst.emperaturevere carefully controlled, so the sugar useddid not experience any visibie change in structure. Thesugar grains tested had an average diameter equal tobeen limited by the machines needed to de¥'elop the1 .015 mm (material passin*・ sie¥'e No. 16 and retained inconsiderable loads that can lead to g:rain frag:mentation.One ¥vay to sidestep this is to use standard geotechnicalsieve No. 20). The coefficient of uniformity of the sam-equipment lvith weak materials (Mandi et al., i977;McDo¥vell and Humphreys, 2002; McDo¥vell and Khan,equal to 1.5. The maximum and minimum void ratios2003). Pre¥'ious researchers have reported results fromring shear tests on ¥veak granular materials such ascarbonate sand (Luzzani and Coop, 2002; Coop et al.,2004). Two interesting features have been revealedconcerning the behavior of crushable granular materiais:a) the grain size distribution of the original materialevolves to¥vard an stable fractal distrlbution as a consequence of particle breakage; however, this stable fractaldistributlon of sizes depends on the magnitude of inducedstresses and the initial grading of the sample; b) Themobilized angle of shearing resistance does not si*・nificantly change regardless the severe degradation of thematerial. In this ¥vay, ring shear tests conducted on ¥veakgranular materials ha¥'e sho¥vn to be a valuable tool inorder to understand the evolution of c,rushing.ples tested ¥vas equal to I .2The specific gravity, G*, ¥vaswere equal to 0.68 and 0.48 respectively. The void ratio atthe beginnin_'._ of the tests ¥vas equal to 0.65. The naturalan*'1e of repose (friction angle) ¥vas equal to 40'. Thecoefficient of compressibility of this sugar as determinedfrom the slope of the normal compression line from anoedometer test is equai to 0.37 (1,0bo-Guerrero andVallejo, 2004). The tested samples had an outside radiusof 5 cm and an inside radius of 3.5 cm. The samplesheight ¥vas equal to 5 mm. They ¥vere prepared by placlngthe sugar inside the rin_ : shear by using the pluviationmethod. Thus, the samples had a loose state at thebeginning of the tests and crushing of the su*'ar grainswas avoided duri lg the preparation of the samples.Figure I sho¥vs an example of the initial state of thesamples. As shown on these pictures, the originalOne disadvanta9:e of usin9: ring: shear tests in theparticles had a prismatic rectan*'ular shape. The hei**ht ofa grain ¥vas around I .7 mm, the ¥vidth and thickness bothlaboratory is that they do not easily allo¥v a visualizationmeasured I mm. The cap sho¥vn in Fig. I ¥vas placed onof the crushing process. Numer'ical simulations in thetop of each sample after its preparation. Each sample ¥vasplaced on the r'ing shear apparatus and ¥vas subjected to aform of the Discrete Element Method (DEM) have beenused in order to overcome this constrain. Since theoriginal DEM developed by Cundall and Strack (1979)¥'ertical stress. Thr'ee different vertical str'esses ¥vere usedin the testing program: 198 kPa, '_96 kPa, and 394 kPa.does not consider particle breaka_._'e, different sohitionsThe an_*,"ular velocity of the upper part of the r'ing shearha¥'e been proposed and programmed using DEM codes.had a constant value of l.'_ degree/minute during theThe first solution to this problem is to de¥'elop severalring shear simulations considering different grain sizedistributions. In this ¥vay, each grain size distributionrepresents a different state in the evolution of crushin__'tests.(Morgan, 1999; Morgan and Boettcher, 1999). The sec-Mobi!ized Frictioll CoefficielltFi**ure '- sho¥vs the curves relating the mobilized friction coefficient (rlcf) and the level of horizontal deforma-ond solution is to treat each granular particle inside thetion for the normal stresses used on the tests (horizontalring shear as a porous ag_ lomerate built by bondingdisplacement= [average radius][angular displacement insmaller particles,rads]). As is explained later in this paper', it ¥vas foundthat the sugar achieved an almost constant grading beforevhich is defined as a cluster (Jensenet al., 2001). Follo¥vin*' certain criteria, this cluster canfully or partially disa*'gregate during the simulation.completing one revolution (horizontal displacement ofA third solution to this problem consists on replacin*' theparticles that are fulfilling a predefined failure criterion26.7 cm), and very small crushing ¥vas expected to occur¥vith an equivalent group of smaller particles (Lang,'_002). This study uses the third solution and a simplifiedtensile failure cr'iterion that can be easily implemented onDEM. This paper also presents and compares the r'esultsobtained usin_"*, rin_".. shear tests conducted on sugar andthe results obtained usin9: DEM.EXPF,RIMENTAl, ANALYSISRing shear testsvere carried out using sugar in a stand-ard Bromhead type apparatus (Fig. 1). The surface ofthe platens of the ring shear apparatus applying thenormal and shear stresses to the samples ¥vere rough.These surfaces had the form of a zigza*' patter'n ¥vith aheight equal to I mm and a distance bet¥veen peaks equalto I mm. During the testing program the humidity and airafter this deformation. Fi**ure 2 indicates little variationof the mobiliz,ed fr'iction coefficient ¥vith respect to thenormal stresses and the induced deformations It can beobserved that a constant friction coefficient for all thevertical stresses varied bet¥veen 0.55 and 0.6 (corresponding to angles of shearing resistance betlveen 28.8'and 31'). An average value for this constant frictioncoefficient is O.575 that corresponds ¥vith an angle ofshearin_"* r'esistance of 29.9'. This angle of shearresistance is considerably lo¥ver than the angle of frictionbefore crushing (40') measuredvith the angle of repose.Tl7e Evohrtioil of CrusllingFi*・ure 3 sho¥vs the average size compositions of thethr'ee samples subjected to vertical stress of 198 kPa,296 kPa, and 394 kPa after one revolution in the ringshear. Figur'e 3 also sho¥vs ho¥v the ori_*"inal uniformj ¥. 'lODELING OF G RANULAR CRUSHiNTGProvmg Rin s j ertic.al dialLoad Dial ¥";""' ";'; """ 10:1L'ever"' ' ;""""'.. " *' ' .,=,=.4/_""*'/= "'**'*' " *"- -*Torque Arm Specimen ContainerLoad HFig. 1.{・-o,*el'Description of the ring shear apparatus ,7ith sugar samples149 LOBO-GUERRERO AND VALLEJO150101 aOO/oo.90= 360**o.8::[! Uncrushed originalm teri8]e (degrees)809!o0g,o. 7-A A"A A A A A A A '- '-A- -A_ ,06J 05Pat cr* = 296 kPaga *yQle,_2( + (kPa)0_4-H - 19803- -o, 27001'o90600/0E5DO/oe,409!oEi 180270IS*- 296o,::e+'T'-A-- " 394o.s:e,30'*loooe,200/0:LO OO O a5 O. I O 0. 15 O 20 O,25 O 301 f:e::・+Oo/oHorizontal displacement (m)360:e:OVIS:eOo/oo:,FO 638 a 2881 015Fig. 2. Fricrion coefi cient vs, honzontal tieformatlon laborator¥all3o 038Aver ge diame-ter (mrn)resultsParticie size distributton of the laboratory samples at (T. = 296rig. 4.kPa1 oo /,at 360". cFv (kPa)=9a9[! Uncrushed80010e,>::E700,60a/o,VU,a,509/0at (1 = 296 kPa O1984a9/03aoj'oCD200/eI! ' /v- O 100・. : tDp::2,a303 f IA//" // c- - - - Power (80)' - - ' Power (270)CSQ(X/ - //A" //'* /y = O 71 i5xO *'' ' " c11R2 = O 9536 ) .-ela9!o -Dp=2 O/y = O 69i 5xl1 0150.638O.288 OAverage diameter {mm)13 O^038Fig. 3. Particle size distribution of the samples after one revolution onr2l/ R2 = O 9261DFSSI 6789R2 = O 9543 E:Dp=1 8474o oloa oiothe laboratory ring shearO 1001 oaor/rLFig. 5.material became a ¥vell・-2G,t/// y :s O 55 IXI4.. . ":Oo!'oe9T I "!l''" //R2 = O 9669 /'"IPower (90)h/':y :: O 783lxr_- - - ・ - Power (360)i 394C:,U-, p/;/A 270x 360:+"'5:CDl;E 90o 1 80ra 296,:,S1 oOoariginal rr}aieriaie)zraded mixture of sizes after the(1 *Fragmentarion fractal dimenstons of the laboratory sampies at= 296 kPatests. It can be noted that the amount of crushing increased ¥vith the magnitude of the applied vertical stress.A comparison of Figs. 2 and 3 sho¥vs that even though thesamples had different level of crushing they all exhibited aconstant friction coefficient. This is in a'-reement lvith thelaboratory results recently reported b_v Coop et al. (2004).Figure 3 does not provide any information regardingcrushing ¥vas expected to occur.Using the results from the sie¥'e analyses, the fragmentation fractal dimension of the samples ¥vere calculatedusing the follolving equation (Tyler and ¥ fheatcraft,1 992) :the e¥*olution of crushing inside the samples ¥vith respectto the horizontal deformation since it only sho¥vs the finalM(R<,/')_() /' 3 D= (1)configuration of the sugar particles. Thus, four moretests ¥vere conducted on the ring shear apparatus. Thefour tests had vertical stresses of 296 kPa, but ¥vhere¥Vhere J (1(R <, r) is the cumulati¥re mass of the particlesended at different values of angular deformation (O', 90',180', '_70'). Sie¥'e and photographic analyses ¥¥'ereconducted after these tests. Fi_・..ure 4 sholvs the averageMT rL¥vith size R smaller than a given comparative size /'. M:Trepresents the total mass of particles, /' is the sieve sizeopening, and /・L is the ma.x imum particle siz,e as defined bythe largest sie¥'e size opening used in the analysis. Df is thesiz,e compositions of the four samples at the end of eachfra mentation fractal dimension. Fi ure 5 sho¥vs thetest (the results obtained at 360' ¥vere also included in thisfigure). It can be observed that most of the crushing ¥vasobtained results after applying Eq. (1) to the results ofthe sieve analyses. Also, po¥ver regression lines ¥vereproduced before an angular deformation of 90'.added to the laboratory data and the fragmentationMoreo¥'er, there lvas only a small amount of crushin_"._(almost 30/0) bet¥veen angular' deformations of 270' and360'. It can be concluded that the sample at 360' tendedfractal dimensions of the samples ¥vere calculated. It canbe noted that the fra zmentation fractal dimensions of theto reach a stable g:rain size distribution ¥vhere almost nosamples corresponding to angular deformations of 270aand 3600 vere similar (2.014 and ?.03) ¥vhich confirmsij ¥_ (ODELING OFGRNULAR CRUSHINGl)151P: pb} J pc);J.,(7:= 2p!'TTLDP> p.¥'// ¥ : l> pj I f';Fig. 7. Idealizatton of the induced tensile stressand arranoement ofthe produced fragmentsinduce tensile stresses in the particles if the coordina-tion number is low (lo¥v confinement, coordinationnumber less or equal than 3 for a t vo-dimensionalFig. 6. Photograph of particles from the iaboratory samples at(T. = 296 kPaDEM analysis). These tensile stresses could break theparticles (Fig. 7). If the coordination number is high(high confinement, coordination number greater than3), the confinement str'esses induced on a particle bythat the sample reached a stable grading ¥ 'here no furthertheir neighbors generates a hydrostatic state of stressesmaking a tensile f'allure of the particle less likely (Ladecrushing ¥vas expected to occur. Ho vever, these valueset al., 1996; Tsoungui et al., 1999; Nakata et al.,are bello¥v the maximum fragmentation fractal dimen-200 1 a) .sion (around ,_.5,_.6) reported in the literature (Turcotte,-For those partlcles having a coordinatlon number1986; Sarnrnis, 1997; lvlcDo¥vell and Bolton, 1998;smaller than or equal to 3, the real loading configuration such as the one presented in Fig. 7(a) is assumed toMcDo¥vell and Daniell, ,_OOi). This is in agreement viththe fact that more crushing ¥vas produced ¥vhen increas-be equivalent to the one obtained in a diametricaiing the ¥'ertical applied stress (Fig. 3).compression test such as the Brazilian test, as sho vn inFigure 6 sho¥vs a photograph of selected particies fromthe samples after the tests. Each colurrm sho¥vs particlesFig. 7(b). By using this simplification the inducedparticles. It is imerestln O to see ho ¥' the original particlestensile stress, a*, can be approximated lvith the expression presented on Fig. 7(b), Ivhere Pl is the ¥*alue of thehighest contact for'ce acting on the particle, L is thethickness of' the disk (unit thickness for the simulatedfractured into t¥vo big particles and some srnall frag-case), and D is the diameter of the disk. The inducedments. Also, It can be noted ho v some particles lost theirangularities as a result of crushing.tensile stress in Fig. 7(b) could be more severe than thestress induced in Fig. 7(a), but the authors belie¥'e thatfrom samples subjected to different values of angulardeformation, ¥vhile each ro¥v sho¥vs the size of theseDEM ANALYSISThe PFC D program produced by Itasca lvas used tothe implemented simplification is justlfied due to thesirnplicity in the calculation of the induced tensilestress.-The tensile strength of a particle having a radius of¥'isualize the evolutlon of crushing in granular materialsl mm is predefined as (T*,**1*,**= 3 x 106 Pa. This valueunder ring shear test conditions. In the PFC2D program,particles are idealized as discs that interact ¥vith each¥vas arbitrarily chosen by the authors. It represents thestrength of' the ¥'irtual material. By changing this value,other at their contacts. This interaction is mainlygoverned by three models: the stiffness model, the slipthe necessary stresses to produce crushing ¥vill alsochange. It is assumed that the tensile strength of amodel, and the bonding model (Itasca, Theory andpar'ticle ¥vith a r'adius r, (7*,,**(r), is related to (7*,***1*,,=**Background, 200?_). Only the first t vo models vere usedin this simulation. The PFCLD program does not considerparticle breakage.expressed in mm):according to the follo ving r'elationship (¥vhere /' is(T*,**(r) (2)= cF****1*,*, [/・]-Partic!e B/'eakage Crite/'ionIn this way, particles ¥vith a radius greater' than I mmA subroutine using the FISH Ianguage (Itasca, FISH inPFC, '_002) ¥vas programmed in order to allo¥v particlebreakage. A simplified failure criterion adopted duringthe simulation considers:-Only particles ¥vith a coordination number equal to orhave a tensile strength smaller than cr****1*, *, andsmaller than 3 are able to be broken. Particles subjected to external loads resist them through interparticledifferent materials such as quartz fibers and sand grainscontact forces (force chains). These contact forcesparticles ¥vith a radius smaller than I mm have a tensilestrength greater than cr***! **' Other researchers ha¥'ereported experimental results in order to describe thechange in tensile stren*・th as a function of size for(Billam, 1971 ; Nakata et al., 200lb). They found thatthe experimental results could be described using an LOBO-GUERRERO AND ¥,ALLEJOl 5_equation similar in form to Eq. ('-). As it is expected,the values of the t¥vo constants in Eq. (2) depend on theforces generated by their nei*・hbors located at the-Every particle ¥vith a coordination number smalleropposite side. Moreover, particles are allo¥ved to exitfrom one side of the sample entering at the same heighton the other side. These particles pass from one side tothan or' equal to 3 is allo¥ved to break if (y*> (T*,***(/').the other conserving their original velocity. Thus, the useIt ¥vas assumed that if a particle is fulfilling theof t¥vo-dimensional periodic boundaries is a techniquepreviously established failure criterion, it is allo ved tobreak into a group of 8 particles having 3 different siz,es,as sho¥vn in Fig. 7(c). The breaking of a particle into 8used to represent lvhat happens to a g:_ranular material in afr'agments as shown in Fig. 7 resembles the obser¥'edcodes (Mor_ :an and Boettcher, 1999; Jensen et al., 1999;breaking phenomenon in sugar and real aggregates(Lobo-Guerrero and Vallejo, 2005). The number andJensen et al., 2001; Lan :, '-OO?_).analy. zed material.distribution of particle sizes after breakage could varydepending of the material tested (Takei et al., '_OO1;Lobo-Guerrero and Vallejo, 2005). Thus, the author'sthree-dimensional rin_9: shear apparatus. This techniquehas been used before by many researcher using DEMThe samples used on the t¥vo simulations came fromthe same original sample. This original sample ¥vascreated **enerating 230 circular discs with a radius of3 mm and density of 2500 kg/m3 inside the simulated ringbelieve for computer efficiency of the model, the use of 8particles is reasonable. Also, the model used in this studyshear. Their positions ¥vere randomly generated by the(Fig. 7) considers only one type of crushing, that isnormal and shear' stiffness ¥vere set to I x 10s N/m, andparticle fragmentation due to tensile stresses. The othertheir friction coefficient ¥vas set to O.7. The values of thesetype of particle crushing, that is particle abr'asion, is notproperties are similar' to those used by L,obo-Guerreroprogram ¥vith no overlaps bet¥veen discs. The usedconsidered by the model used. Since in the model circularand Vallejo (2005) ¥vhen simulatin_g: the crushableparticles lvere used, crushing due to abrasion is less likelybehavior of ¥veak granular mater'ials subjected to directshear test conditions. The discs were allo¥ved to setrleand reorganize under the action of a gravity field (g = 9.8m/s2). Thus, the original sample had a loose structure.to occur.In order to implement the failure criterion, a ne¥vsubroutine ¥vas programmed using the FISH Ianguage.This subroutine automatically checks if a gi¥'en particle isfulfilling the failure criterion. If it does, the simulationstops, and the par'ticle that broke is automatically deletedA vertical force of I x 10> N Ivas applied by movingand replaced by the set of particles sho¥vn in Fig. 7(c).The subroutine does not restrict smaller particles fromthe first simulation. For the second simulation, thecontinuing to break. Thus, particles representing¥vas kept constant for the remainin*' part of that simulation. In both simuiations the shear stress ¥vas induced bymoving the upper ¥vall of the simulated ring shear to theleft ¥vith a constant velocity of 5 x i0-7 m/step.different ・_enerations of crushincan coexist inside thesample.down¥vards the upper ¥vall of the simulated ring shear.This applied force ¥vas constant for the remaining part ofapplied force ¥ 'as increased to a ¥'alue of I5 x lOs N, andConfigu/'ation of tlle Saj7lp!esThe irst step was the construction of the virtual ringshear apparatus. First a simulated box having i5 cm inlvidth and 6 cm in height ¥vas created. The top andbottom ¥valls simulate rou_'_.h ¥valls ha¥'ing a ¥vave lengthCrusl7ing. Pol'osit.y, alrd S/7ear Strel7gt/1 Evo!utiol7 of theSan7p!e Subjected to a Ve/'tica! Force of J x JO; NFigure 8 sho¥vs the applied shear force, the a¥'erageporosity, and some snap shots at different le¥'els ofof 2.5 cm and a difference in height between peaks ofdeformation of the sample subjected to a ver'tical force of0.5 cm. Their coefficients of normal and shear stiffness¥vere set to I x 109 N/m, and their friction coefficients¥'ere set to 0.7. The assumed properties of these ¥valls1 x 105 N, It can be noted ho v the applied shear forceexhibited a constant value regardless the se¥'ere andprogressive degradation of the sample. This is in agreement ¥vith the experimental r'esults obtained in the sugarrepresent ¥vell their r'oughness and the interactionbet¥veen the lvalls and the particles simulating the su_ ar insamples and those obtained by Coop et al. (2004). Anthe actual ring shear apparatus. Or'iginally these t¥voaverage value for the applied shear force ¥vas equal to¥valls ¥vere separated by a ¥'ertical distance of 6 cm beforethe generation of the particles. L,ateral per'iodic bounda-0.3 x 105 N; thus, the average residual friction coefficientlvas equal to 0.3 (0.3 x 10s N/1 x 105 N).ries ¥vere programmed by the authors using the FISHThe porosity of the sample significantly decreased fromuage (Itasca, FISH in PFC, 2002). The later'al period-0.175 to 0.153 due to particle crushing and particleic boundaries and the t¥vo horizontal ¥valls represent thefrontal surface of an unfolded hollo¥v cylindrical samplerearrangement bet¥veen horizontal deformations of O andenclosed by the rin*' shear apparatus. Thus, ¥vhat0.153 to 0.157) ¥vas produced due to the particle rear-happens in the frontal surface of a closed cylindricalrangement between horizontal deformation of 5 cm and10cm. Finally, the sample slightly contract bet¥veenhorizontal deformations of 10 cm and 15 cm (porositieslan_section can be vie ved in its unfolded state as a rectang:u-lar ¥vindo¥v. When using periodic boundaries, particleslocated at one lateral side of the sample interact lvithparticles located at the opposite lateral side. Particleslocated at the lateral boundaries experience the contact5 cm. After this, a ¥'ery small dilation (porosities bet¥veenof O. 157 and O. 149 respecti¥'ely). The interaction bet¥veencontractancy due to particle crushing and dilatancy dueto particle rearrangement is far from simple, since bothl 153MODELING OF GRANUL、へR CRUS賊【NGn讐0.1631n漏0、17511瀟0.1530604ShearForceら F疑 聖 r ・裏 1 転〆(xlO5N) 『        配 r02000204   06           08      〆「  1 01214H・riz・ntalde蜘ati・n(xlO叫1m)     /   /n篇0,157  /               n篇0。149           I        Details →Originaゆa纏icles∠ 》罫irs硬ener&tioII oflcruShl葺鞭   }                                             〕SeCOnd gel篠eration of−cms拒【19      }                                             〕   Th1rd.ibu曲,and勲h黛enera一                    }tbll ofcrushil19           呼Fig.8.Cr睦shing,P・rosi重y,謎ndshearf・rceev・lu重i・n・fIl・esimulaIedgr狂nul窺rm飢eri麓1(verIicalforce玉x105N)phenomeaa takes place at the same time.1t is compllcatedto conclude if the sample reachecl a cr量tica正state by or1正yIooking重he porosities of裳he sample since an almostconstant value of porosity cou豆d be the ba豆anced effectbetween contraction due to c茎一ushing and dilation due toparticlereαrr&ngement. 1t can be observed how  di9奄rePt generations ofcrus勤ing coexist illslde the sample,and some of the LOBO- GIJERRERO AND VALLE JOl 541 oOoo4criginali -u):5::R2=0.9841 1 Power- (e5cm j1 5crn)J:u'' - 2.25cm・--A,・y=C.2314 l0571 --x 5cm- OcmI {c5o-・ Ocm (afteF vertical load)I aoo9 1.2u'sG,E 11'5'..O.8hY.et **'. ':,,¥s:.9 06'o._oc,DE 0,4a,:tEc ,¥¥ ,:'ZL O.21 ooO OOio ooolo olooParticie radius r (m)O.05 O. 1 5O.1Horizontal displacement (m)Fig. 9. Evolution of the grain size distriburion of the simulated sarnple(verticaforce I x 10*Fig. 10. F, olut on of the fra'lr mentatlon fractal dLmenslon (1ertlca)force I x lO' +)original particles ¥vere surrounded and protected byCl'us/7in*". Pol'osity, a/7c! S/7ear Stre/7*"t/1 Evo!utioJ1 of t/7esmaller particles. The grain siz,e distributions of theSan7p!e Subjectec! to a Vertica/ Force of I.5 x 105 Nsample at differ'ent values of horizontal deformation ¥vereanal),zed using the follo¥ving equation (Turcotte, 1986)Figure 1 1 sho¥vs the applied shear' force, the averageporositv. , and some snap shots at different le¥'els ofN(R > r) =D K(r)(3)Where N(R > r) is the number of particles ¥vith a radiusR (could be other linear' dimension) bigger' than a givenvalue /'. K represents a constant and Df is the pre¥'iouslydefined fra :mentation fractal dimension. The results arepresented in Fig. 9. Also, a po¥ver regression line ¥vasadded to the data corresponding ¥vith a horizontaldeformation of 15cm. The obtained frag:mentationfr'actal dimension ¥vas equai to 1.057. This ¥'alue offra mentation fractal dimension in this t¥vo dimensionalarrangement corresponds ¥vith an equivalent value of'_.057 in a three dimensional case (Df3D=Df2D+1)(McDo¥vell and Daniell, ,_OO1; Sammis, 1997). Thus, thesimulated sample achieved a fragmentation fractaldeformation of the sample subjected to a ver'tical force of1 .5 x 105 N. The applied shear force exhibitecl an almostconstant ¥,alue of 0.45 x 105 N reg:ardless of the severe andprogressi¥'e degradation of the sample. Thus, the a¥'erageresidual friction coefficient lvas again equal to 0.3 (0.45 xl05 N/1 .5 x 105 N). This is in a_._"reement lvith the resultsfrom the first simulation, the experimental resultsobtained in the sugar samples, and those obtained byCoop et al. ('_004).The por'osity of the sample significantly decreased from0.166 to 0.139 due to particle crushing and particlerearrangement bet¥veen horizontal deformations of O and5 cm. After this, the porosity continued to decrease but ata very small rate (from O. 1 39 to O. 130 in 5 cm of horizontal deformation). It seems that the higher applied ¥'erticaldimension equi¥'alent to the maximum ¥'alue obtainedforce precluded the occurrence of dilatancy. Figure 1)in the laboratory tests on sugar (Fig. 5). Figure 10 showssho¥vs the results of applying Eq. (3) to the grain siz,edistributions of the sample at different values of horizon-the evolution of the fragmentation fractal dimension.It sho¥vs that the sample presented an almost constantfractal dimension bet¥ 'een horizontal defor'mations oftal deformation. A maximum value of frag:mentation7.5 cm and 15 cm. Crushin9: seems to have ceased at these¥*alues of deformation^ Ho¥vever, a very small amount ofin 3D) ¥vas achieved by the sample at the end of thesimulation. More crushing occurred during this secondcrushingvas still occurring in the sample, as is sho¥ 'n inFig:. lO.fractal dimension equal to I .1484 (Equi¥*alent to-.1484simulation than durin : the first one. This is reflected inthe fact that the frag:mentation fractal dimension ¥vasThe analysis of the results from the porosities athig:her than the one obtained during: the first simulationdifferent ¥ralues of horizontal deformation and the results('_. 1484 vs. '_.057). This is in agreement ¥vith the laboratory results presented in Fig. 3.from Fig. 10 sho¥vs that the sample tended to achieve analmost steady state after a horizontal deformation of7.5cm since the changes in porosity and __."rain sizedistribution ¥vere ¥*ery small. Ne¥'ertheless, it should benoted that a critical state ¥vas not completely achie¥'edsince the sample continued to experience a ¥'ery smallamount of crushin_",_ even at very iarge deformations.Figure 13 sho¥vs the evolution of the fragmentationfractal dimension during the second simulation. It can benoted that the sample achieved an almost constant frag:mentation fractal dimension at a horizontal displacementof 5 cm. A comparison of Figs. 13 and 11 sho¥vs that thesample reached an almost steady state at a horizontaldeformation of 5 cm since the changes in porosity andfragmentation fractal dimension ¥vere considerably small 「155 MODEL正NG OF GRANUL、へR CRUS}{INGI1瓢0。152n罵0.16611瓢0。139G806Shear fOI℃e(xlO5N)  0402O G02  04     06〆/〆   08Ho「⑳1t糟「王癒iOI’(x1α…m)1011=0,130Il=0、圭36      DetailsOriginalpaロlcles翫st generatiOn oi Cl■lshil1黛   }                                             》Iil       、Second generatio驚10ヂc田shiI19     噛                                       }芝ion ofcrusilin9          贈Flσ 11.i 曳il    Tllird.重bu貫h.a毫1d tコ£’th gellera一                   }塵    ノ㊨・、  (Crushing,porosl粟y,我ndshearε・rceevolutlon・εtl・esimulatedgran賎畳翫rma芝eri撰1(ve震ic謎ぼorce1.5×105N) 156乱OBO−GUERRERO AND VALLEJOfrom the laborαtory tests and provi(ied also a visualiza−,0000   o薩ginaltion of the crushlng evolution,DEM is a valuable tool+Ocm(a費erve由calroad)八駕that can be used to anderstand complex phenomena that一ゆ一一25cmoりコ℃に ・△一5cmoccur量n the evolution of crushing such as毛he interaction一×弓Ocmbetween sample contraction generαted by particle break−    Power(10cm)素琴の£  1000.9age an(i d量lation due to part圭cle rearrangement,黍 甲s ロ ヘ ¥■響’ACKNOWLEDGMENTS   ◇\、          ・”484    版嚢 y謬轟,,鳴ユo This work was supported by Grant No.CMS−0301815①ρto the University ofPittsburg勤from the National Science﹃昏E3zFoundatlon,Washington,D,C.Thls sllpport is gratefully^襖・  嶺   壌acknowledged.10000001001    0001Particle radius r(m)REFERENCESFiσ 12. EvoIution of Ihe or&in size dis{ributio鶏 of Ihe sinlu韮段Ied1)Biliam,,∫.(1971)ISomeaspec【softhebeめavlourofgranular   sample(verticωf・rce1。5×105N)  mater玉alsa鵬lgllpressures,S11・ε∬一5’11α’1∼βθhαv’o’μoゾ30’Z∫rP・o‘,  Ro∫coθル∫θ〃∼or1α15♪,1nρo∫1μ〃∼,Cambridge University,69−80、2)Coop,NL R・,Sorenser1,K・K,Bodas Freitas,丁.and Georgoutsos}  G,(2004)=Particle breakage d柱ring shearing of carbo目aIe sand,  Gθo∼θ‘1∼η姻θ,54(3),豆57−163!.43)Cundall,P.A.andStrack,0.D.L(1979):Ad1scrαenumerical  mode!for granu玉ar assemb1玉es,(フθ01θc11η’(71’θ,29(1),47−65、ζ0 1.2葛4)1τascaConsu1顛gGroup,lnc.(2002):PFC20〆Pα’「 がぐノθElowCoゴθ”1冨  Tl”0111∼θ〃,∫io〃5/vθ’510113.0≠rhθ011γ‘7114βαごたglu’174,α’2ゴFISH.塁 1  ln PFC.三£5)Je鷺se鷺,R、P.,Bossc鼓er,P,」.,Ples}遷a,M、E a目d I…di1,丁。B、呂α8  (1999)l DEM simu!飢ion of granuiar media−structure i瓶erfacelに  e鐸ecIsofsurfaceroug晦nessandpartlclesllape,∫1π.,1.醗’1η澱1∼α1.,90、6  ル∫θ’h.Gθ01ηθごh,,23,531づ47.お報6).}ensen,R.P.,Plesha,M.E.,Edi1,丁,B.,Bosscber,Pのεα4 L an(i  Kahla,N.B.(2001)l DEM s1mulatio陰ol parIicle damage in瑳  granularmedia}struc田re醸erfaces,/’π./.Geolnθc1L,1(1),2ト39.と  0.27)Lade, P. V., Yamamuro. J. A an(玉 BopP, P、A. (1996):  Signi員ca鷺ceof partlcle crush1ngin gran副arma[erlals./.Gθofθご1∼.0  万ηgrg。,ASC鷹,122(4),309−316.0      0.02     0、04     006     0.08      0I         Horizontaldisplacement(m)8)Lang,R.A.(2002):Numerica蓋simulaほo跳of comm1nution in  granular materials w臨an apPlica面n to fault gouge evoludo巨,  Mα5’αT1∼θ5Z5,Texas A&M U頃versily.Fig,13、Evolu重lonof“簾efrag田enIat韮onfracIaldimension(verdc田   force1.5×105N)9)Lobo−Gロerrero,S、a職d Valiejo,L.E.(2004):Modeling of mater玉al  cms歴ng ln granuIar road bases,Pヂoc.UNβ、4R VI Co11ゾ・,1》θvθ・  1ηθ1π5Uηδα〃マ4,University of No面ngham,England,33−4L10)Lobo−GuerrerQ,S。and Vallejo,L.B、(2005):Crush圭ng a weakafter this defomation.Nevertheless,it should be noted  graRularmaterial:exper1me競aレnし【merlcalanalyses,G80!θごh吻∼紹,  55(3),245−249,that a critical state was not completely achleved since the11)Luzzanl,L、and Coop,M.R.(2002}=O睦由e relatio離shipbetweensample contlnue to experience a veτy small amount of  partiClebreakageandtheCritiCa1Sta【eOfSandS,SO’Z5αηげcrushing even at very large deformations.  Fα〃∼ゴα∫’o/15,42(2),71−82.12)Mand1,G.,Jong,L.N、  J.andMa1由a,A,(1977):Shearzonesin  granularmaτer1al,Ro欲ル1θ(加11’c∫,9,95−144.CO蔑CLUSlONS  The evolution of crushing ln a granular m飢erial sub−jected to a ring shear test was studie(i using laboratorytests and DEM slmulations.The ring shear tests werecarried out圭n the Iaboratory using sugar.The residualfriction coe伍c量ent of this material maintained a constantvalue regardless of the severe degra(iation of the particles、This degradation was induced by increas量ng the angulardeformat量on or increasing the αPPl重ed vertical stress.置〉loreover,it was follnd t致at the gra量n size d三stribution of出e original uniform material evolved toward a fractalB)McDowell,G.R.and Bolton.M、D、(1998)10鷺出e mlcromec簸an−  icsofcru勲bleaggregates,Gθ01θc11吻∼だ,48(5),667−679.14)McDowel1,G.R.a鷺dDaniell,CM,(200正):Fractalcompresslonof  soil,(3θo’θぐh吻∼’θ,51(2),173−176、15)McDowel1,G.R。and}{umphreys,A。(2002)=Yield1ngofgranular  materials,Gヂσπμ1αrハ1απθr,(4),豆一8,16)McDowell,G.R.and K飯aR,工 L(2003)l Creep of granular  ∫naterials,θノ’01∼i’1α1’ル伽∼θノ。,(5).115−120、1フ)Morgan,」.K.(1999)INumericalsim慧1at1onsofgraaularshear  zonesus1ngthed1s置incteleme飢me出od,2、田ectsofpart1clesize  dis頃buξlon a鷺d1nterpar【lcle fr1ction on mechanical bebavlor,/.  Gθoρ妙∫’‘α1Rε蜘κ11,重04(B2),272レ2732.18)Morgan,」.KandBoettc紅er,M.S、 (1999)INumer1calsimuiations  of granular shear zones uslng[he distlnαeleme蹴method,L Sheardistribut呈on of sizes,  zone kine臓atics and tlle鵬icromecbanics of locallzation./,  The results from the DEM simulations con負rmed由ose  Gθoρノ∼♪’∫’cσ1Rθ5θσ1’‘ノ∼,104(鶏2),2703−2719、釜 FMODELING OF GRANULAR CRUSHING15719)Nakata,Y、,Hyodo,M、,Hyde,A.F、L、,Ka芝o,Y、εmdMurata,H.  【eStS,So〃5α1∼ごFα’1∼面”oノ∼5,41(1),97−121  (200王a):Microscopicparticlec田shingoflsandsubjecledto猶igll23) Tsoungu1,0,,Valle【,D.and Charmeミ,」、C.(1999〉:Numerical  pressure one−dimeasiol》al con》pressionラ 50’1∫ αノ∼4 」Foε〃∼ゴα∼io/1∫,  model of crush至ng of grains inslde {wo−dimensiollal granular  4玉(1),69−82、  ma芝e嫉als,1⊃o“て1θ1’T2ch1∼0109』}’,105,190−198、20)Nak飢a,Y,,K&to,Y、,Hyodo,M、,Hyde,A.F、L.and Mura[a,更{、24)Turcot[e,D.L、(1986):Fractalsandfragmentaτion,ノ.Gθoρ妙∫’cα1  (2001b): One−dimensional compression be}ヨaviour of uniformly  R(∼5θα1’c1∼,91(B2),192i−1926、  gradedsandrela芝edτosingleparエiclecrushings{reng乞h,50i15α1∼425)Tyler,S・NV、and XVheatcraft,S,、V、(隻992):Fraαal scaling of soil  Fα〃∼面r’011∫,41(2),39−51.  pardcie−s玉ze d1sエribu葦io員 anai}・sis and Iirnilaエiolls, So’1 ∫(ゴeηcθ21)Sammis,C、G、(王997):Fraαal fragmen【aエlon and栢cτional s【abil−  50ciθり,oゾ湘111θヂ’cθ/、,56(2),47−67  iζy in gran岨ar ma{erials,1UTl4M⑤t・1刀ρ、Mθc1∼α’1’c∫oゾG’・αノ1正’1αヂ26) Va量1ejo, L. E. (2003): Crush玉ng o匠 granular bases: flracζal and  αηゴPo1’o£’5Aイ10’θ1層’σZ5,23−34、  laboratory analyses, Pヂoc、 1 /δθ1へoα〃1θパごαη 511『’刀ρ, Pαvθ1ηθ’∼’22)Takei,1〉蓋、,Kusakabe,O.and Hayashi,T、(2001):Time deやende諏ζ  五∼∼9”∼θθr”∼9/xlV Co’01ノめiα1∼εv〃1ρ.Pαvθ’nθn’五∼19’nθθ1−1n9,  behav玉or of crushab!e materials 王n one−dimensional compression  Poρ¢vαn Colo1ノめ1α,C∠)一ROA/、
  • ログイン
  • タイトル
  • Numerical Analyses on Consolidation of Clayey Ground Improved by Vertical Drain System Based on 3-D Elasto-viscous Model
  • 著者
  • W. Beak・Takeo Moriwaki・Yasushi Sasaki
  • 出版
  • soils and Foundations
  • ページ
  • 159〜172
  • 発行
  • 2006/04/15
  • 文書ID
  • 20897
  • 内容
  • ログイン
  • タイトル
  • A New Nonlinear Hysteretic Rule for Winkler Type Soil-Pile Interaction Springs that Considers Loading Pattern Dependency
  • 著者
  • Masahiro Shirato・Junichi Koseki・Jiro Fukui
  • 出版
  • soils and Foundations
  • ページ
  • 173〜188
  • 発行
  • 2006/04/15
  • 文書ID
  • 20898
  • 内容
  • SOILS Al¥j'D FOUNDATIONSV o l46, No., ,i73-188, Apr, 2006Japanese Geotechnical Societ)A NEW NONLINEAR HYSTERET'IC 1 RULE FOR WINKLER TYPESOIL-PILE INTERACTION SPRINGS THAT CONSIDERSLOADING PATTERN DEPENDENCYMASAHIRO SHIRAToi) JUNICHI KoSEKlii) andJIRO F=UKUliii)ABSTRACTWe propose a ne¥v hysteretic rule f'or ply curves to be used in the dynamic analysis of deep foundations. V,re firstexamine the results of past laterai cyclic pile load experiments to clarif'y the characteristics that are to be modeled in theload transfer hystereses in ply cur¥'es. We then undertake an analytical study of soil element beha¥'ior ¥vhen subject tocyclic passi¥'e (compressive) and acti¥'e (extensile) def'ormation. ¥ re f'ound that the soil resistance intensity to pilesvaries ¥vith different cyclic loading patterns, as the stress-dilatancy beha¥'ior in soil ¥'ariesvith cyclic loading patterns.We developed a ne¥v hysteretic rule that satisfies the obser¥'ed dominant characteristics. Although the proposedhysteretic rule has its background in the peak-oriented rule, it is further extended to be a function of the loadin_",_pattern. Numerical tests using the proposed model sho¥ved that the model is capable of reproducing obser¥'eddifi:'erences in the beha¥'ior of piles subjected to fully-reversed cyclic loading and one-sided cyclic loading, even thoughthe typical peak-oriented rules are unable to predict these outcomes.Key lvords: dynamic interaction, earthquake, horizontal load, pile, (pTy cur¥'e) (IGC: El'_/E14/H1)considering that the pile shaft may become plastic,INTRODUCTIONA future alternati¥'e to the current seismic design of pileSeismic design methods and structural design codesfoundations is likely to involve extending the currentdesign using pushover analysis in the Beam on Nonlinearhave been developed on the basis of' Iessons learned frompast damage associated ¥vith large earthquakes. A recenttrend in seismic design to protect against the effects ofWinkler Foundation concept to¥var'd the nonlineardynamic analyses of foundations in the rime domain. Inlarge earthquakes in earthquake-prone regions such assuch a case, a nonlinear' hysteretic p-y curve is necessary.Japan is to control structural damage so that the structure can quickly return to service, considering not onlythe strength of the structure but also the ductility ¥vhenp-y curves have previously been developed empirically(see re¥'iew by Reese, 1993). Many monotonic ply curveshave been proposed based on the results of lateral loadexperiments of piles (e.g., Reese et al., i974; Matlock,1970; Kubo, 1965; Fukui et al., 1997). Many studies havesubjected to large earthquakes. A practical method ofassessing the ductile r'esponse of foundations to largeearthquakes is therefore required.In the code of practice for the design of high¥vay bridgefoundations in Japan, nonlinear ductile behavior' of thefoundation is examined by the pushover analysis usin_-'also sought to incorporate the dynamic effects of stift:nessthe Beam on Nonlinear Winkler Foundation concept.Other studies have attempted to develop the nonlinearThe Winkler hypothesis herein is taken to mean that thehysteretic p-y curve (e.g., Nogami et al., 1992; Boulangerfunction of the soil resistance p at a particular depth isindependent of the soil properties and pile displacementet al., 1999; Curras et al., 2001; Kondou et al., 1998;and damping into p-y curves, based on elastic ¥vavetheory and finite element calculations (e.g. , Novak, 1974;Nogami and No¥'ak, 1980; Gazetas and Dobry, 1984).Maki, 2002). In terms of soil nonlinearity, most pr'e¥'iousstudies have simply adopted conventional hysteretic rulessuch as Masing's rule, and ha¥'e ¥'erified them by compar-at other locations (Poulos and Davis, 1980). Theso-called p-y curves express the lateral soil-pile interaction at each depth, ¥vith p representing the soil resistanceing the numerical pile responses vith experiments thatinvolve piles subjected to harmonic loading or fullyreversed cyclic loading. Ho¥vever, it is considered thatthere is insufficient data to confirm that such a con¥'en-stress to a pile, and y representing the correspondlngdisplacement of' the pile relative to the far-field. The pileshaft (or deep f'oundation body) is modeled as a beam,i) Senior Researcher, Structures Research Group. Pu:blic ¥¥'orks Research nstitute, Japan (shirato ,plvri.go,,jr,).*" Prof ssor, Institute of 11 dustrial Science, Universit ' of 'Tokyo, Japan.iii) Dlrector. Structures Research Grour). Pubiic ¥'orks Research Insiiiute, Japan,The manuscrir)t for this papervas received for reviell on Septeurber 13, 2004; approved on Januar)' 13, 2006¥Vritten discussions on this r)aper should be submii ed before November l, 2006 to the Japanese Ceotechnical societ)', 4-3s-2, sengoku,Bunk}'o-ku. Tok}'o I 12-001 i. Japan_ Upon request the closing dale may be extended one month.17 'J lSHIRATOl T4ET AL _tional hysteretic rule represented by lvlasing's rule mostlypro¥'ides rele¥*ant results for nonlinear soil-pile interactions ¥vhen piles are subjected to random cy. clic loading.West"LEastIn reality, not only are earthquake motions random, butstr'uctural aspects can also cause various loading patterns.For example, in highlvay bridges and over'passes, thetransition from re¥'ersed-c .'clic loading to one-sideci cyclicloading is anticipated in the situation ¥¥'here P-A effectsappear, resulting from dama (Te to the bottom of a pier' orthe top of piles; the P-A effect results from secondaryo¥'erturning moment loads to the pier and foundationassociated ¥vith the mo¥'ement of the point of gra¥'ityvertical loads applied to the superstructure. Althoughsome past studies have compared the results of experiments invol¥'ing piles subjected to earthquake motions¥ *ith numerical results calcukued via a hysteretic p-ycurve (Boulanger et al., 1999; Curras et al., 2001),additional systematic research is also required to determine hysteretic mechanisms for various loading patterns.The aim of the current paper tackles to investigate thehysteresis of ply curves, based on both an experimentalstudy of pile behavior and an analytical stud¥_' of soilelement behavior. ¥Ve propose a ne¥v hysteretic rule thatcan be used in numerical simulations ¥vithin a frame¥vorkof total stress analysis. First4000Fig, 1. Illtrstration of the tcs setup (Llnits: mn )ve re¥'ie¥v past experimentsthat involved piles subjected to different patterns oflateral loading. ¥Ve focus on the hysteretic p..;' cur¥'esobtained by Shirato et al. (,_006) from the pile load testsconducted by Kimura et al. (1998) and Fukui et al. (1998).Ve then conduct an analytical in¥'esti_gation using asimple model at the soil element le¥'el. An importantfinc,ling of this study is that the h.¥'steresis of p-y ¥*arieswith different loading patterns; ¥ve then model thisphenomenon. A ne¥v hysteretic rule for p-y curves isThe sand deposit ¥vas very loose because of the construction procedure, ¥vith an a¥'erage relative density of 17.30/0and a¥'erage dry unit mass of 1.51 t/m3. The inter'nalfriction angle obtained by drained triaxial compressiontests ¥vas 39', ¥vith a relati¥'e density of 300/0 and a confin-ing stress of 29.4 kN/m2. A small strain shear modulus Go¥vas obtained from cyclic triaxial compression laboratorytests:developed that accounts for the modeled characteristicsGa= 199.63 ((7 /98)06 x 10 (kN/m ) (1)described above. Finally, ¥ve compare experimentalresults ¥¥*ith the numerical accuracy of the behavior of¥vhere cr is confining pressure expressed in k ! /ml.single piles embedded in sand and subjected to eitherlateral reversed cyclic loading or lateral one-sided cyclicThe experiment invol¥'ed steel pipe piles ¥vith adiameter of 318.5 mm and an embedded depth of 8 m^10ading; this is considered in terms of overall pilePiles ¥vith t¥vo different ¥vall thicknesses ¥vere used: 10.3behavior ancl load transfer hyster'esis bet¥veen soil andmm and 5.6 mm. A pinned device ¥vas attached to thepile.bottom of each pile to clarify the boundary condition.1,0ADING PATTF.RN DEPF,NDr.NCY OF p-yStrain gau*"es ¥vere arranged on the sides of the piles toestimate p-y curves. Lateral cyclic loads ¥vere applied tothe piles at an a¥'erage rate of loading of 60 mm/min. TheLoc7dii7(point of application of the lateral cyclic loading ¥vas? pattel'n Depel7clent Soi! Resis!ance Obsei'ved inPreviouS E.¥'pel'!n7ents of Cyclic Latera! Pi!e I,oacliilgFrom pre¥'ious cyclic lateral loading experiments ofO.7m above the initial sand deposit surface (groundle¥'el).piles, ¥ve first revie¥v an experiment of single piles subject-The experimental cases examined herein are listed ined to either fully-reversed or one-sided lateral cv_ clicTable I . Case S1 used a pile ¥vith a larger ¥vall thickness10ading. This experiment was originally conducted bythan the other tested piles and ¥vas subjected to fully-Kimura et al. (1998) and Fukui et al. (1998); e¥'.'per'imentalreversed cyclic loadin*'. Cases S'_ and S3 used piles ¥vithdata analysis ¥vas refined by Shirato (2004) and Shiratoet al. ()-006). Figur'e I sho¥vs the experiment setup.the same details and ¥vere subjected to fully-re¥'ersedA saturated soil deposit vas made from Kashima sand(D50=0.67 mm and U*=2.66) in a deep test pit at theFoundation Engineering Laboratory in the Public Workscyclic loading and one-sided cyclic loading, respecti¥*ely.As these experiments dealt ¥vith t¥vo extreme cyclic loadResearch Institute in Tsukuba, .Japan. The deep test pitpatterns in random loading, ¥ve expected more detaileddata than that generated in past tests; these data couldthen be used to model the hysteretic rule of p-y. Thehas a length of 4 m, amaterial elemem test results on the pile specimens ar'e alsovidth of 3 m, and a depth of 1 1 m.s SOIL-PILE INTERACTIO¥. Ttabulated in Table 1. In all cases, the amplitude of thespecified displacement ¥vas gradually increased f'rom 6 ton >< 6 (n=' 3 4, . . .), in ¥vhich a= 15 mm Three cycles' ,¥vere repeated at every displacement amplitude. Durin_*-IT5loading case (Case S3),・the mobilized soil r'esistances are clearlv smaller thanthose for the fully-re¥'ersed cyclic loading cases (CasesS1 and S2) at the same displacement levels; theeach experiment, part of' the soil surf'ace around the pilecon¥'entional hysteretic rules are not able to reproducethis tendency.subsided markedi.¥' and ¥vater appeared inside the sub-The load-displacement curves at the loading point alsosided area.Figure 2 sho¥vs observed p-y cui"¥'es at a depth of GL-0.96 rn f"or all cases. The results sho¥vn in Fig. 2 aresourced from earlier stages in the experiments because ofsho¥v the same trends varying ¥vith cycllc loadingpatterns.Similar loading-patter'n dependent differences in lateralsoil resistance are e¥'ident in the results of the experimentsthe increased reliability in data processing for p-yconducted by Agaiby et al. (199'_) and lvlayne et al.(1992). The drained and undrained lateral beha¥'ior of'relationships. In the fully-reversed cyclic loading case,・the pl)' path al¥vays approaches the pre¥'ious peakdisplacement point on the envelope curve, as predicted by lvlasing's rules and other conventional peakoriented hysteretic rules; the envelope curves aredrilled shaf'ts ¥vere investi :ated in various densitles ofsand deposits and various consolidation conditions ofclay deposits; the model drilled shafts ¥vere subjected tomonotonic and ¥'a "ious patterns of one-sided cyclicdenoted by dashed lines in Fi :. _.'unloading paths are stiffer than the reloading paths,and the unloadin gradient appears to be maintainede¥*en ¥vhen the peak displacement le¥'el increases.10ading:. The lateral resistance of the drilled shafts in one-sided cyclic loading tended to be smaller than that inThe term 'en¥'elope cur¥'e' is defined as a curve that tracesmonotonic loading. Theref'ore, it is considered a generaltendency that the lateral soil resistance varies ¥vlth thethe peak points of' the first loops. The term '10ading'cyclic loading pattern, independent of' soil classification.refer's to the case ¥vhen the increment of lpj is positi¥'e,Source oJ' Loadiilg Patrel',1 Depenc!el7cy and Moc!e!ingof P/'edomil7ant Clla/'acreristics in Locrding Patre/'nDepenc!enc_vTo examine the soi} r'esistance rnobilized in the near-¥vhile the term 'unloading' refers to the case ¥vhen theincrement of ipj is negati¥'e. In the one-sided cyclic- -- envelope curveRI Reloading, U: Unlaadingfield as simply as possible,ve used a model that simplifiesthe soil-pile interaction systern in a ¥vay as sho¥¥'n in'OOFig. 3. A ¥'ery thin slice and small depth rigid body・aosand viched bet¥veen tlvo very thin slice and small depthsoil blocks in a plane str'ain condition ¥vas subjected tocyclic loading under a ¥*ertical compression stress, assum-z;zing a pile shaft, surrounding soil, and an o¥'erburdenstress in the ground. The major and rninor principle stress- cO40 10 O 'O 40-200ax'es are xj and x2, respecti¥'ely; the xl-direction cor-40 20 O 10 40responds to the horlzontal direction in the field. Each soil' (mm)y (.l lm)block is modeled usin_"._ one element. The boLmdaryFlg. 2. Comparison of p-J' h) steres is between the three experimentalconditions are gi¥'en such that strains in the xi and x2directions can be freely mobilized. Accor'dlngly, onl"vcases (GL -0.96 m)X2 1 Pilei X2X3ij xlLi.! t. x Overburdenyressure::::Plane strainl condition¥j atadepth ofx2Front v'eWFig. 3.Simple model llsed to examine, P an VieWC}clic loading patternslFully-reverseds2Fully-reverseds3One-sidedDescription of experimental casesPile ¥vail thicknessl0.3 mm5.6 mmL____i Model front viewle infiuence of soil non inearitl on soil-pile interaction (from SI]irato et al.. 2006)rable 1.CasePiieYield stress562 N/ rl589 N'/mm:YOullgsll oduluS227 kN/mm238 kN/mm 1sSHIRATO ET ALl 763ao:Monotonic loadin; ff//f2f [ / /1/';/-l//ji 1i ¥Tayour s nde O 654(200l 1/_!olff f I. jIl!1iti i statei2 / I !fMonotonic( 1 s;78 5 kNlm(;:22=785 kNlm , ! ! cam ression]tOO Mono onic i e=0 eS5I1_1li!1i''Ii'! /!/'l!=/' Ji JJ[i! j ! j2- _ I : '/!/ / J / if ----- iv' tonotonlc loatilng!/ IJ '/ / 'l!otensione 0 655O1 OIts 1 (o/o)lO O(_1 L)!/.!ItO IOO(1'fL)/r." 02Fuily-reversed eyc ic loadings¥""'JE'J7-OI;e-sided cyclic loadin_o; LFig. 5. Anal .'tical results of the effects of stress*diiatancy beha:vior of4sand on p-y relationships:: -o 2Jt'04,O/= gll-e22 (O/o)1(T 2, cmab is the mobiliz,ed internal friction angle, y is shearstrain, and 8ij is the (i,j) component of strains. The_backbone curve (or the monotonic loading curve) isl' N¥,symmetric about the origin:- 05ky /__ y,Osin cn"b= "l + yly,Y-0.5-1i;O1/= g l-g_1'2 (D/o)fig. 4. Plane strain cyclic comprcssion-ex"tension test results forToyoilra sand from Masuda et al. (1999)x sin cma , (3)in lvhich c ,.x is the maximum internal friction angle andy* is the reference strain. ¥¥re assume that cm**= 300 and;'*=0.000,_. The stress-dilatancy relationship is Ro ve'sstress-dilatancy la¥v (Ro¥ve, 1962), ¥vhich Is expressed¥vith a combination of:Rl =K DI R crll, Dj c!8 .l oi・・ c!e""" ' ' Pl{:, ( vhen A(Tll >0), (4)horizontal displacement is fixed at both ends. The loaddisplacement responses of the rigid body sho¥vn in Fig. 3are calculated for many displacement histories of therigid body.Soil in the near-field around the piles can bc assumed toandR A' D R= (7 2D =c!ed8 1,(¥vhen A(7{1 <0) ())l lbe primarily subjected to cyclic passive (compression)in ¥vhich the ¥'alues of the coefficients of dilatancy, K1 andand active (extension) deformation because of the difference in the lateral displacement bet¥veen soil and pile, notonly in cyclic load experiments but only in the field duringK2, are assumed to be 3.5. Note that Shirato et al. (2006)document the results of more detailed investigations ¥vithseveral combinations of sand parameter values and sever-earthquakes; hysteretic p-y curves are assumed to beal loading patter'ns.associated ¥vith the nonlinear beha¥'ior of soil subjectedto cyclic compression-extension defor'mation. ¥Ve there-analysis. The left-hand diagram sho¥vs results for fully-fore use a constituti¥'e model basecl on the cycliccompression-e.¥'tension plane strain element test resultsfor sand derived by Masuda et al. (1999). Their tests ¥vereconducted under saturated and drained conditions.Figure 4 sholvs typical test results obtained by Masuda etal. (1999), in ¥vhich (T11 and (T22 ar'e principal stresses, elland 82? are corresponding strains, ,, is shear strain, 8. i is¥'olumetric strain, and c**,b is the mobilized fr'ictionangle. As indicated by lvlasuda et al., the stress-strainrelationship of the sand blocks can be described by acombination of a sin c,**.b-y relationship and a str'essdilatancy relatlonship. In the present paper, the sin c *.b-yrelationship is gi¥'en ¥vith a hyperbolic function backbonecur¥'e and lvlasing's hysteretic rule, in ¥vhich;sin c**.b = (R - 1)/(R + 1), y = 81- e(2)¥vhere R is the principal stress ratio defined by R = crl lFigure 5 sholvs some of the results from the presentreversed cyclic loading, ¥vhile the right-hand diagramsho¥vs r'esults for one-sided cyclic loading. The dashedlines represent the corresponding monotonic loadingcurves. In particular, the p-y curve for the case of onesided cyclic loading trends is considerably lo ver than theprevious peak states.Figure 6 sholvs the calculated sin c ,. -y relationships inthe sand blocks, in ¥vhlch a plane strain sand block on thepositive side ¥vith respect to y=0 is referred to as apositive-side sand block, and that existing on the negativeside with respect to ;'=0 is referred to as a negative-sidesand block. The center points of loops of sin c*no y shiftto the negative side of y because the negative dilatancy(i.e. ¥'olume contraction, le,el>0) is moblliz,ed at e¥'erydisplacement re¥'ersal. The magnitude of the shif't ¥'aries¥vith loading pattern: the mobilized value of sin c in thepositi¥'e-side sand block in the one-sided cyclic loadingij+.. SOIL-PILE INTFRACTIOiNcase, i.e. the passi¥'e-side sand block, is smaller than themobilized value in the sand blocks subjected to fullyre¥'ersed cyclic loading. Accordingly, the mobilized soilresistance stress, p, in the one-sided cyclic loading case177reversal point lvhen the plus and mlnus intensities of p atthe preceding and most recent re¥'ersal points are alrnostequal (trend I in Fig. 7(a)); as the dift rence in lp j of theportant role in the observed dependence of loadingpreceding and most recent reversal points increases, thereversed p-y curve trends are lo¥ver (trends 2 and 3); andif the most recent reversal occurs in the vicinity of thepreceding reversal point, t.he p-;' curve returns to thevicinity of the preceding r'eversal point (trend 4). Thesepattern on soil resistance. The 8.*l-y relationship sho¥vn inphenomena can also be reinterpreted in the mannerFig. 4 indicates the followin*・ points. For fully-reversedillustrated in Fig. 7(b). The backbone cur¥'e of p-ycyclic deformation in both cornpression and extension, alarge part of the incremental negative dilatant strain(A8..1>0) that is generated during a compressive defor-virtually drifts from the original position; the degree ofdecreased to a value clear'ly smaller than that in the fullyre¥'er'sed cyclic loading case.Soil stress-dilatancy characteristics played an im-mation phase (passive phase) is countered during thefollo¥ving extensile deformation phase (active phase). Fordrift is related to the de*'ree of eccentricity in the loadingdirection of the previous cyclic loading history. While theinterpretation in Fig. 7(a) is based on a vielvpoint of soilresistance stress, that in Fig. 7(b) is based on a vie¥vpointone-sided cyclic deformation only in cornpression, theresidual negati¥'e dilatant str'ain accumulates and thenegative dilatancy effect becomes stronger as it is notcountered during the extensile phase. A Iarger axialof' displacement.deformation is therefore required to mobilize large shearstrain and shear strength in the one-sided cyclic loadingcase. Ultimately, the mobilized soil resistance to a pile isassumed to be dependent on the degree of eccentricity ofthe cyclic loading direction.Based on the calculated results ¥vith various displace-served even in cohesive soil. Althou*'h stress-dilatancyment histories of y in the model,increase in excess pore pressure on the o¥'erall pileve recognize thefollolving trends, as illustrated in Fig. 7(a). In a p-y cycle,As described in the above revie¥v of pre¥'iouslypublished pile load experirnental results, variation inlateral soil resistance ¥vith cyclic loadin*" pattern is ob-behavior is still considered to play a key r'ole in loadingpattern dependency in cohesi¥'e soil, ¥ve encourage futurein¥'estigations that seek to confirm this hypothesis viarelevant element test results. Note that the above analysisdoes not consider the states in vhich the influence of thebehavior cannot be neg:lected,the p-y curve returns to a point close to the precedingPROPOSED HYSTERETIC RULE FOR p-J' CURVESFOR SINGLE PILES:': o');Negativef Positivt!f'=*' '* Isid8 j ! Ijside;-05o:o 5r-i;il ':'r-ol- !!Here ¥ve propose a ne¥v hysteretic rule for p-y as afunction of' the loading pattern. This proposed model canbe ¥videly used 1¥'ithin a frame¥vork of total stress computation, in which the generation of excessive pore pressurein the soil can be ignored.:':'160 7/-- y*o -40-2 oy/ lo *40 --, o-60oFull.v-reversed cyclic loading05o )Neg tiveTo describe the various conditions of the nonlinearo )ff/ Positive_______/' side ';relationship of p-y, we describe the hysteretic rule usin_"._[ side/o;; iI II -: or_L:':' :f ' Cf,combinations of piece¥vise straight branches. We dis-lregard inertial eft cts on the soil:J+-40 20 20 O05O-'edge in the interactionzone, because considering temporal variations in theinertial effect rnakes the present problem complex andmay also mask the essential elements of the presentproblem.JOne-sided cyclic loadin_Frg. 6. Calculated sin c ,,1' y relatronstups for the sand bhocksconsidering stress-dilatancy bc laviorp (1) (2) (3) (4)/ ^! f! 'pfif;pfyy: Preceding reversal]: Last reversa! point(a) Observed featuresFig. 7."? ;<;1TointFully-reversed cyclic ioadingOne-sideti cyclic loading(b) An alternati¥'e assumption based on observationCharacteristics of tlegradntion in soil resistance stress p with loading pattem 1 !- *SHIRATO ET AL.17Spsurrounding soil in front of a horiz,ontal loaded pile.i pu YBoth methods ¥vill be compared in the follo¥vin_",_numerical examples.The initial gradient kH is set by multiplying the reference gradient ko by an empirical modifier, ak, so thatkH = c kko, (8)" o) ' x (k0=- and,''(9)where Eo is the small strain deformation coefficient ofsoil, n is a constant that represents the loading ¥vidthdependency of subgrade reaction coefficients, B is thefoundation ¥vidth (i.e. pile diameter in the case of piles),Y'Fig. 8. Backbone curve of the pi,' relationshipBo is the reference ¥vidth in terms of the loading ¥vidthdependency, and c is a modifier used to fit the referencesubgrade reaction ko to the assumed bi-linear curve. TheJapanese Specifications for Highway Bridges employB0=0.3 m and n= -3/4, based on the experimentalMocle!ing of Backbone Curvestudy of Yoshida and Yoshinaka (1979). The ¥'alue of akThis paper simplifies the backbone curve as an elastoperfectly plastic bilinear curve, as sho vn in Fig. 8, incan also be determined by a load test of a single pile.¥vhich p is the ultimate soil resistance stress and y+, is thereaction coefficient equation in the .Iapanese Specifica-displacement level that the soil resistance stress achievestions for Highlvay Br'idges (Japan Road Association,at the ultimate value on the backbone curve. We expect2002), however, Eqs. (8) and (9) are also reasonable interms of dynamic Interaction. Gaz,etas and Dobly (1984)and Kavvadas and CJazetas (1 993) conducted comparativefinite-element studies of harmonic pile-head loading andthat the hysteretic rule, ¥vhich ¥vill be introduced later inthe text, can be applied to any shape functions of thebackbone curve; the choice of the shape of the backboneEquations (8) and (9) are similar to the sub_ radereported that as a first approximation, the subgradep-y curve is optional.The ultimate soil resistance stress can be defined as afunction of the passive resistance to the pile. Therefore, asimple expression of the ultimate soil resistance stress isreaction coefficient kH can be considered to be frequencyindependent and expressed as a multiple of the local soilpt; = c pPp (6)kH = aE*, ( I O)¥vhere pp is the ultimate passive earth pressure and c p is ain ¥vhich a is a frequency-independent modifier.With respect to foundation scale effects on the subgrade reaction, Koseki et al. (2001) used horiz,ontal plateloading tests ¥vith plates ¥vith different sizes to demon-modifier for the three-dimensional effects of soilresistance on a pile. The value of ai can be estimatedfrom computations of the behavior of single piles subjected to monotonic loading or fully-re¥'ersed cyclicloading ¥vith a gradually Increasin*' amplitude in aninverse manner. If ¥ve derive the ultimate passive earthpressure in seismic situations from the JapaneseSpecifications for High¥vay. Bridges, ¥ve arrive at thefollo¥ving passi¥'e earth pressure coefficient KEP:cos2 cKEP = ""- fsin (c -cos¥vhereElE) sin c 2 (7)! cos 6E' )stiffness E*,strate that the - 3/4 po¥ver la¥v can only be r'evealed ¥¥'henthe subg:rade reaction coefficients are estimated in termsof a certain displacement level of the plate; n = - I can beobser¥'ed when the sub...*( rade r'eaction coefficients areestimated by referring to subgrade reactions at a certainstrain level, ¥vhich is defined by dividing plate displacement by the plate ¥vidth. Further research is required toclarify the manner in ¥vhich the loading ¥vidth dependency should be dealt ¥vith in setting the initial gradient of thebackbone curve.is the friction angle bet¥veen the pile shaft surfaceand the soil, assumed to be -6/6, ¥vhich is also employed in practice for estimating the ultimate soilBasic Hysteretic Ru!e without Loading Pattern Depend-encyresistance stress for piles subjected to cyclic loadin*', andThe hysteretic rule for' fully-re¥'erseci cyclic loading isit is used hereafter in this study. Another method ofestimating pU is to use an analytical solution based onreferred to herein as the basic hysteretic rule. The basichysteretic rule is characteriz,ed by the follo¥ving t¥voadmissible plastic flo¥vs in the area in front of an under-major features, based on the above observations ofground pile. For example, Kishida and Nakai (1979a,experimental and analytical p-y curves.1979b, 1977), Reese (1958), Reese et al. (1974), Broms1. The amplitude of soil resistance p during fully-(1964a, 1964b), and Koda et al. (2000) pro¥*ided analytical solutions vith admissible plastic fio¥v fields in thereversed cyclic loading ¥vith a fixed displacementamplitude is constant.i SOIL-PILE INTERACTION2. A reversed pl}' curve from the backbone curve179absolute ¥'alue of the displacement amplitude of point e,jy*j . The corresponding negative global control point istrends to vard a point corresponding to the absolute largest displacement !yj,**** on the backbonedefined as point g, i.e, yg=_vcl='cl .iy! = - , = -y*'***cur¥'e in the other load and displacement direction.An illustrative explanation of these t¥vo points is sho vnThe curves that connect the global control points, CiC*j,ar'e defined hereaf'ter as the external cur¥'es f'or ¥vhichm Frg 9 In Frg. 9, the p-y path runs through theorigin-a-b - c- d-a-eH, f ,g-h-e-Y- i ・ j - k-i and j have a ¥'alue of I or 2.An unloading path that departs from the last displacement re¥'ersal point and trends to¥vard p=0 is a straightl-i-m. The paired points a and c, e and g, and i and kshare the same absolute ¥'alues of p and y.The points corresponding to the absolute largest dis-linevith an unloading gradient of kFI** [. kH**,,! is providedby;placement iy: ,** on the backbone curve are hereafterreferred to as global control points. A pair of global/,- -H***1 - ( 1 1 )/*control points is located symmetrically about the origin,¥vith one point on the positive side and the other on thenegative side. The global control point on the positi¥'eAs sho¥vn by pile experiments (Fig. 2) and analyses using:the sand block-rigid body-sand block rnodel (Fig. 5), thefik o .unloading gradient can be assumed to be independent ofthe displacement level and loading history. The unload-side is ref"err'ed to as point Cj, ¥vith that on the negati¥'eing path is considered to be strongly associated ¥vith theunloading elastic modulus of soils; its gradient has theside ref'erred to as point C2. The displacement amplitudeof point Cl is y = ycl!y ",'*j , vhile that of point C*2 is y =same order of small strain level elastic modulus.yc2 lyj ***. The corresponding ¥'alues of soil resistancestress, p, are referred to as pcl and pc2, respecti¥'ely. ForUltimately, fi= I is assumed in Eq. (1 1).When a displacernent re¥'ersal occurs at a point otherthan a control point, the subsequent path trends to¥vardsexample, for the path e-f-g in Fig. 9, Iyj*** is thethe preceding displacement reversal point, as sholvn inFig. 10. Ho¥vever, ¥¥*hen strain reversal occurs numeroustimes on a path from one global control point to the otheror back to the original control point, the coding of thepcomputer pro_"._ram becomes curnbersome because theassociated past displacement reversal points must bememorized and unfolded lvhenever displacement re¥'ersaloccurs. Accordingly, additional rules are introduced, asl'iliustr'ated in Fig. I i . As sho¥vn in Fig. 1 1(a), the displace-ment reversal point on an external cur¥'e is referred to as alocal control point i, C*Li (PcLi, ycLi); the follo¥ving p-ycurve, C iCi, that trends to¥vard the preceding displacement reversal point (i.e. the global control point C 'j) isreferred to as the reference internal curve. As sho¥vn inFig. 1 1(b), Ivhen a displacement reversal occurs at a pointon the r'efer'ence internal cur¥'e CLiCj, the displacementreversal point is defined as the other local control pointFig. 9. The basic h)stcretic mechanismp[l- __ ._ ・//p!l/1J/ !i .'l'/_"":' I /,i' ; /"'/j ' l,lj'lyl'T , lllf l'liill{! ';r!l/. i_*.//ll /// / !!pp!1_!p_'1/ll (r!!p! llIf !! li/l""'/' I .'/'/l' -/f' l-y/'1l!y1i/ ' 'l lf' l'__lL__//! !J( !Frg lO. Schenlatic diagrams ofp J' behavior following a displacement reversal in the basic hysteretic ru]e (All p-y eurves start from the origi 1) SHIRATO ET AL.l 80p(a)p(b)CyC?cl'Fig. Il.C1yyclp(c)c_lBehavior of internal loops of pll' in the basic h . stcretic rulepCLj. Then, as sho¥vn in Fi**. 1 1(c), the ply loops that fol-lo¥v the displacement reversal from a locai control pointCLj al¥1'ays trend to¥vard the iocal control points CLi orCLj, dependlng on the loading direction. While p and ycontinue to respond in the area bet¥veen y=ycLI and y=ycL2, Iocal control points are not rene¥ved. For' example,the p-y path trends from the last displacement reversalpoint R to¥vard the local control point CL2 on theyreference internal cur¥'e rather than the precedin*' displacement reversal point R'.Loading Patterl7 DependencyAs sho¥vn in Fi**. 7, the path of ply from the lastdisplacement re¥'ersal point does not al¥vays trend to¥vardthe precedin*' reversal point, especially in one-sided cyclicloading. We therefore introduce a special ne¥v rule calledthe deterioration rule to account for the p-y behavior as afunction of eccentricity in cyclic loading direction.pWhen assuming different pl)' cycles fr'om the *'10balcontrol point C1, as iliustrated in Fig. 12, the re¥'ersed10ading paths from the external curve must trend to¥vardpoints that are different from the global control point Cl,depending on the degree of eccentricity in the loading:direction. This is based on the beha¥'ior of p-y cyclesymodeled in Fig. 7. Accordin_g:ly, the proposed modelintroduces the target point T] and produces a re¥'ersed10adin_path from the external curve CjZIC_2 that al¥vaystrends to¥1*ard the target point Tl, as sho¥vn in Fig. 1'_The reversed curves from the external curve comprise anunloading branch ¥vith a gradient of kH { and a strai_"..htbranch that trends to 'ard Tl from p = O. The tar*'et pointTj is defined as the intersection point of the reloading linefrom the perfectly unloaded point Z,1 of the global controlFig. 12. ProposedhT.・steretic rule lvithloading pattern dependenc¥.-(Part 1)point C and the extension line of the external cur¥'eC2Z,2C1 (i.e. the extension line of Z,2Cl).The gradlent kH* is referred to as the reference reloadIng gradient, ¥vhile the line that connects the perfectlyunloaded point Zi to the tar>'et point T{ ¥vith a referencereloading gradient is termed the reference reloading line.The reference reloading gradient is generaliz,ed as;k H* = MkHi ( 1 2)first quadrant.equation:Theref ore,lve ha¥'e thefollo¥ving:M> cek (13)2 - akFi**ure 13 sho¥vs the p-y loops pr'oduced by theproposed model for fully-re¥'ersed cyclic loading andration of p. The value of M must also be given to satisfyperfectly one-sided cyclic loading. As sho¥vn in Fig. 7(b),for perfect one-sided cyclic loadin*", the response shouldthe condition that the line of ZIC must have a cr'oss point¥vith the extension of the reference reloading cur¥'e in theinvolve a horizontal shift of the backbone curve on theone-sided cyclic loading side. When ¥ve assume that the¥vhere I f is a modifier to adjust the degree of the deterio-s SOIL-PILE INTERACTIONpppIS1BL. Y'))Fig. 13. p-y hystereses for two extreme loading patterns in theproposed modelppFig. 15. An exampie of the behavior of ply, Ivith consideration of the1'detcrioration rule, prior to reaching the llltimate soil resistancepFig. 14. Proposed h!steretic rule witil loading pot ern dependenc,'(Part 2)reference reloading g 'adient kH* has a value of the sameorder as the initial gradient of the backbone curve kH, themonotonic reloading path from point Zl increases ¥'ith aygradient equivalent to kH, reaches the backbone curve,and follows p=p**. ¥ fe therefore considered it appropriate that the value of M in Eq. (12) is taken to be the sameas the value of c k in Eq. (8). Our proposed hystereticmodel is ultimately expected to be capable of accountingfor the loading pattern dependency arising from stressdilatancy behavior' in the surrounding soil, as modeled inFigs. 7(a) and (b).pWhen an interaction spring is subjected to a fullyreversed cyclic displacement ¥¥'ith gradually increasing orstationar'y arnplitude, deterioration of p does not appearat peak points, regardless of M. We also note that forM= I .O, the loading pattern dependency is expelled.Figure 14 sho¥vs an example of setting the target pointon t.he negative side of p. The target point T2 can bedefined in the same manner as the setting of Tl, usin*' theyperfectly unloaded point Z2 from the negati¥'e globalcontrol point C*2, the gradient kH*, and the correspondingexternal curve CIZIC,.The behavior of p-y prior to reaching the ultimate soilresistance pv can also be specified by applying the aboverules. An example is illustrated in Fig. 15; notations usedin this figure are explained in the follo¥ving explanation ofspecial rules.V,rhen an internal p-y path crosses the external curveand changes the direction of travel, the global controlpoints C1 and C2 are rene¥ved, as sho¥vn in Fig. 16. InFi**. 16, the prime denotes the values or points that haveFig, 16. Updnting of giobal contro points for the case in vhich anintcmal path crosses tl]e exter}lal curve and changes its direction oftravcl lSHIRATO ET AL.182p!;J)C Bi/(b}BLTLLE/ / cL2!lZ/iZ; ¥ ¥ ll/////jl !l! .l' /;f!i*7=. 1kH' )lJ/ !// /!J// iL C..=VT c1 IC2pc(c)/BlpllllllZF J /Z2 ¥ -//T*LF,/ /l_(**L/;iHr/z.IA'!jl 7ZLsjT. /* i r!1 yl": ft /L/Fig, 18. Updating of the reference internal curve lvhen the deterioration rule is consideredCLlrf tcTt 'reversal. ¥¥rhen the reversal point is not on the backbonecurve, Cj must be set as the intersection of the backboneinter'nal cur've to the backbone curve is r'eferred to aspoint BL. As sho¥vn in Fig. 17, the internal curves thatfollo¥v a displacement re¥'ersal on a reference internalcurve CLIZ IBL trend to vard the local target points TLland TL2, in principle, depending on the traveling direction. ¥Vhen an internal pl)' path reaches point Tl_1, thefollo¥ving monotonic loading path is bound for the back-curve and a line passing through the re¥'ersal point, 1¥'ith abone curve from point TLa*radient of kH 1' One vay to include this rule into thesubroutine for' the proposed hysteretic rule in a computerp = pU.Figure 18 pro¥'ides an illustrati¥'e description of thecode is that the subroutine continues to monitor theresponse ¥vhen an internal path crosses the curve CLjZLjfollo¥ved by displacement reversal. In Fig. 18, the localFig. 17. Behavior of internal curves when the deterioration rule isconsitieredbeen rene¥ved along ¥vith the most recent displacement¥*irtual perfectly unloaded point of the coordinate of aninstantaneous displacement reversal point, z* (O, y.*).Then, the subroutine checks the relationship of y,* ¥vithrespect to the y-coordinates of the stored Z,1 (O, y,1) and Z2to point BL, follo ¥*ed bycontrol point CLl, Point Z I' and local target point TLI arerene¥ved in the same manner as the global target points;the reference internal curve CL ZLIBI is also rene¥ved.(O, y.2), in ¥vhich y,1 = -y For Ay> O follo¥vm" the lastdisplacement reversal, in ¥vhich ly is the instantaneousincrement of y, all of the points that are necessitated todefine the hysteretic r'ule, points Cj, Zj, and Tj, are up-r ITUMF,RICAL F,XAMINATION OF THF. PROPOSF,Ddated if y,* is less than y.2. In the case of zly< O after theNun7erica! Sin7u!atiol7 of Expei'ilnenta! Resu!tslast displacement reversai, all of the necessary points areupdated if y.* is larger than y,1. In addition to the updateof Ci and C2, external cur¥'es and points Z,1 and Z,2 areThe numerical model considered here pro¥*ides a simulation of the experiments ¥vith single piles that ar'ealso updated, and the target points Tl and Tl are rede-the bending resistance, considering pile yielding, isdefined by an elasto-perfectly plastic bilinear momentfined, as sho¥vn in Fig. 16.As sho¥vn in Fig. 17(a), for internal curves, a localHYSTF.RF.TIC RUl,F,described above. The piles are modeled as beams in ¥vhichtarget point T i is set on either the intersection of the local(M)-curvature ( /) relationship, as sho¥vn in Fig. 19, in¥vhich M:* is the plastic moment of the cross-section of ar'eference reloading line ZL2TLI and the r'eference interna]pile. The elastic gradient of the A(f-V/ relationship is gi¥'encurve CLIZLjBL or the extension thereof, or the intersection of the local refer'ence reloading line Z.LITL2 and theby the fiexural rigidity EI, Ivhere E is Young's modulusand I is the moment of inertia of the cross-section. Thebottom sections of the piles are connected to the hingeboundary condition, in ¥vhich the horizontal and verticalexternal curve CjZIC_2 or the extension ther'eof; thereloading branch that departs from point Zl_i is referredto as the local reference reloac.ling line, ¥vith the referencereloading gradient kH*. The point ZLj is the perfectlyunloaded point of the last displacement rever'sal pointfrom the reference internal cur¥'e CLIZL]BL or externalcur¥*e CIZ,{C . The intersection point of a referencedisplacements are fixed and the rotation is free. Theelement length is set to 120 mm (=0.38D, in ¥vhich D isthe diameter of the pile) except for sever'al elementsaround both ends of the pile.The pile is assumed to be supported by distributedSi SOIL-PIL E INTH:RACTIONMl 83(2:F*4 o. (xf O j(2 =0 l( :T' =8 o.a . 4i). ( :=0.l,, P4 o (xi*o 32002(x)zo *l(X)(( :*Case S lVExperinlemOoo2Cel OO+ ,), (z; O SO. (=0J)2--- ExpeT ment (:)ositive)Experimcut (ncgativc)pesitive}xp5rimem (ncgatiyc),loO, (z*=0 l, (2 =0 Ia , =2100200oedisplacement (mm }disp}acement mm:]Comparison of toad-displacement relationships at therig・ 20.loaclpoinultimate soil resistance stress pU (kN/m2)Fig. 19. Moment-curvature relationship for stcel pipe piles20001000Oo¥springs, ¥vith the beha¥'ior of the springs described by the¥tproposed hysteretic p-y curve; ho¥vever, the distributed2lateral springs are integrated into discrete laterai springsat nodes because of the specifications of the soft¥vare usedin the simulation. The load-displacement relationship ofeach integrated lateral spring is assigned on the basis thep-y relationship at the depth of the middle point bet¥veenthe node of interest and the neighboring deeper node, andis estimated by simply rnultiplying the lvidth and lengt.hof the beam element by p of the corresponding pl)'Nc¥Kishida and Nakai¥¥ ( 1 979!rl **l:+*¥-4¥.*¥ !):,¥i.¥ a:.p=5 . Oa:p=4 .O '¥-6r'elationship.C'alib/・ation fo/' the Backbone C*u/'ve of the p-y Re!at!onsll ipFirst, the backbone cur've ofp-y is set via cornparisonsof the results of monotonic loading calculations lvith theresults of the fully-reversed cyclic loading experiment.The experimental result of C*ase S1 vith fully-reversedcyclic loading is used to estimate the values of parameterscxp and (xk in Eqs. (6) and (8) such that the numericalsimulation pr'ovides the best rnatch vith the rneasured8Fig. 21. Comparison of the ultimate soil resistance stress pu calculatcdusing a '=4.0, c =5.0, and that obtained b) the tl eoreticalmethod of Kishida and Nakai (1979)r"both the positi¥'e- and negati¥'e-load sides are also sho vnon the figure; the ne_ ative envelope curve is sho¥vn afterreversing the signs for both load and displacement. Theload-displacement curve up to the final displacement levelresults are not sensiti¥'e to variation in c k (i.e., the initialof 166 (240 mm =0.75D).rigidity of the backbone curve), ¥vhile loads at anyThe internal friction an_ :le c' of the tested sand depositdisplacement le¥'els ¥'ary proportionally ¥vith ¥'ariation inused for estimating earth pressure coefficients is assumedto be 39', based on the results of the triaxial compression(xp (the ultimate soil resistance stress). These trendsindicate that estirnation of' the strength parameters oftest described above. The small strain shear modulus Goat each depth used for estimating kH is evaluated usingsurrounding soils go¥'erns the calculated ultimate lateralE=2(1 + v)G and assuming v=0.5 for simplicity. Thement and soil resistance become plastic. From an en-confined stress (Tgineering point of vie¥v, backbone cur¥'es of the hyperbolic function are feasible in hysteretic p-y curves, as thenumerical results are not sensiti¥'e to the inltial gradientis replacedvith the mean effectivest.ress:(J= -(1 + 2Kb)cr(. (x2) (kN/m2), (14)3¥vhere ( (, is the eft ctive overburden stress at depth x2 andKo is the coefncient of earth pressure at rest. Kb isestimated by the typical empirical equation J b=1sin c' .Figure '_O sho¥vs comparisons of ¥'arious load-displacement relationships derived f'rom different pairs of al' andcek. The envelope curves of the experimental results inresistance of a pile, as the pile is subject to lar_ :e displace-of the bi-linear p-y curves. Eventuaily, ¥ve select arelevant value of c k=0.1 in the range 10-2 to 10-1, andselect a ¥'alue of aep=4.0.Interestingly, the in¥'ersed ¥'alue of the ultimate soilresistance stress pU,vith c!tp=4.0, is very close to a theoretical solution based on assumed admissible plastic fio¥vsin front of a pile, as sho¥vn in Fig. 21 . Figure 21 comparesthe ultimate soil resistance stress pv, obtained fromEq. (6), ¥¥'ith cep=4.0 and pv Pro¥'ided by the theoretical SHIRATO ET AL.l 84soii resistance stress p (kN/m2)200200 400Ooa:p=4.0 //¥¥7//-2' l1/-,'ScO/, ¥ Kishida and Nakai1 OO¥ '3:: ,5C lO 4( 1 979):,..'//'/ 'ooC_,ase S lExperiment (positive)Experiment (negative)Simulation-6Experiment100 200(positive)Experimentdisp. (mm)-8Fig. 22. Comparisons of oad-displacemeilt relationships at tl]e loadpoint calculated usin"* vaiues of pt calculated from ap=4.0 andfrom tl]e thcoretica method of Kishida and Nakai (1979a)(ne*"ative)Fig. 24. Comparison of tlre ex'periment- antl the simulation-derivedsoil resistance p distribution lvith respect to depth (a =4.0, o!k=0.1): the numbers on the figure represent the displacement levels intcrms of dbending strainO O.OOI O.002 O 003the same displacement amplitudes, in ¥vhich the measuredvalues for positive loadin*' displacement are sho¥vn afterreversmg the signs. The calculated and measured distributions averagely are in good agreement, especially fornegative loading displacement.There is an inconsistency between the calculated ando., ,)9s-4in ¥vhich the pile behaved in the elastic range. Figures 7-3and 24 aiso sho¥v the measured values in the fir'st cycles ofCase S1SimulationExperiment (positive)-------- Experiment (negative6measured results In ter'ms of the depth at which themaximum bending strain and soil resistance stress appear. The calculated depths gradually move do¥vn¥vardfr'om shallo¥v to deep positions as the interaction springsbecome plastic. The same trend is observed in theexperimental results ¥vhen the pile was displaced in thene_ :ative direction; but notvhen the pile ¥vas displaced inthe positive direction. This inconsistency is inferred toFig. 23. Comparison of the experiment- and simulation-derivedresult from the lack of consider'ation of soil densification.bending strain distriburions lvith respect to depth (ap=4.0, otk=In the experiment, the densification effect of the sur-0,1): the numbers on the figure represemt tl]e displacement leveis inroundin*' sand ¥vith repeated loading leads to an increasein the mobilized internal friction angle of c; this in turnleads to an increase in the passive resistance of the sand.Ho¥ve¥'er, the mobilized soil resistance is considered to beterms of 6equations of Kishida and Nakai (1979a). Fi_ ure 22 sho¥vsthe relationships between horiz,ontal load and horizontaldisplacement at the loading point, ¥vhich ¥vere calculated¥vith the values of pv obtained using both methods.Experimental results are also plotted for compar'ison.The t¥vo sets of calculated results are in good agreement.We therefore consider that theoretical solutions for theultimate soil resistance stress may also ¥vork lvell.We no¥v briefiy compare the calculated and measuredsimultaneously affected by the softening effect of theevolutions of shear strain and negative dilatancy in thesoil. The loading pattern dependency of soil resistance¥vill therefore cause a much greater change in pile behavior, as stated above. Therefore, ¥¥'e ¥vill not model theeffects of soil densification on the backbone p-y curve.Fu!!y-/'eve/'sed C:yc!ic L oadingdeformation of the pile and soil resistance distributionswith depth. Figure 23 sho¥vs the calculated and measuredbending strain distributions versus depth at amplitudes ofThe calculated responses to fully-reversed cyclic loading ¥vith the proposed hysteretic rule for p..;' must bedisplacement levels of I , 3, 5, 7, and 9 . Figure 24 sho¥vsreversed cyclic loading exper'iment case S1 is simulatedwith the following t¥vo conditions: (1) Vhen the loadinpattern dependency in the hysteretic p-y curves is disre-the soil resistance stress distributlons ¥vith respect todepth at the amplitudes of displacement of 1, 3, and 5 ,independent of M in Eq. (1'_). Accordingly, a fully-is SOIL-PILIl¥TTERACTION IS5200 (a) Experunent (b)200Simulation d : (c) Simulation / ' 'IVf = I 'o 200 JVf = o_10.!csovOO // f ///"Illj!il/oO ・",¥[SimulationSimulation2020 - , --- Experiment200- --- Experhnent-._. ICQ-200- (mm)OOdisplacement' ' -200O 200displacement(mm) displacement(mm)[OFig. 25.200200OComparison of catculatecl toad-displacement curves at tlre toading point for M= l'o and lv=0,10 and ex'perimentai resuits (Case Sl)100 Expenment j=100 M I O J100 M O iO' Ail .e'o$:!c'jS')- 10[O- -10- I OO-*5,)O5050displacement y (mm) displacement y (mm) displacement y (mm)GL -o_48 mExperiment M = I .O I M = O.10**JcccL)OOJec',:;SJe),*il_ :';o-20O y20-20 O y20-20 O y20displacement(mm) displacement(mm) displacement(mm)cvGL -2 04 mFl" 26. Comparison of calculated pl)' curves f0!v= I and !V=0.ro and experimentaresutts (Case Sl)garded and the hysteretic rule is rnade a peak-oriented calculated curves ¥vith M= 1.0 and M=0.1 . We thereforerule, i.e., M= i, (2) ¥vhen the loading pattern dependency conclude that the proposed hysteretic i'ule ¥vorked as¥vas taken into account using M= ak = O. 1, i.e., kHr= k}{, expected.in Eqs. (12) and (8).The calculated load-displacement relationships at the One-sided Cyclic Loadingpoint of loading are compared in Fig. 25. For compariThe value of' M specifies the deterioration of soilson, the experimental envelope curve is shown withresistance caused by one-sided cyclic loading. Backdashed lines in Figs. 25(b) and (c). The results of thecaiculations and the experiment are almost identical;analyses with various values of M from 1.0 to 0.053 areconducted for the one-sided cyclic loading experimentdiscrepancies resulting from the difference in M are barelycase S3, in which M= cxk = O. I is considered to be the rnostdiscernible in the calculated results. Calculated andlikely value based on the phenomenological inferencedetailed above. Figure 27 sho¥vs the measured loaddisplacement curves at the point of loading and p-ycurves at depths of -0.48 and -2.04 m.measured p-y curves at depths of -0.48 and -2.04 m( - I .51 and - 6.41D) are sholvn in Fig. 26. The proposedhysteretic rule is capable of accounting for the measuredp-y loops; there is no discernable difference between theThe calculated curves are sho¥vn in Figs. 28, ,_9, 30 and 1・==*SHIRATO ET AL.l S6/ 'd4'/d //1/'j '4'1iV'I;!2flj ;A!ll!/j'JI" ';Z i OO"z t ooA+A' yi"r '*':!:'///;;i;i//;;::///:/i' " "/1/'/jl/i///j:/:jj'/j";<i;""{;' ;/:""':/:/;)oo2001 OOo300OO 3 OO200disp]acement (mm)odisplacemem (mm)['200iGJL O 48 mzzJiE 4001- aL -2'04 m !200f CJL 048mGL 2 04 m400;z..':lo0't) ' Jl f !f:;;('; i;i; i;::;'rj:/';S:}:f!( ':/,:;!V1:I OOi),)pjlei/V// // ri jf200L'/; If ;("" '//"'-" :' 41_ /)//+ ' is!i!';;;". 200'ifff'i' ' '"r ' /if //_'/:(/.:4/' 2:://_l'/!!j::C4///:"'r '"'1' ,""o)O 20 4060di:) placement( Illlll )Measured load-displacemeni curve anti p-..1' curvcs (Case S3)Fig. 27.i15ojO 20 40".O 20 4060displacementy (mm)displacement .T' ( mm )(mm)J!i".)O- f r fa" f;/j'i: i'[,.,v:*' [Odispiacement20 ,40'l/o[;':Fig. 29. Calculatcd loacl-displacement curvc and p-y curves (Case S3,M= 0.25)Case S2- Case S3'Joo:Z I OO'A4_.I .,, '/'/,,,A(/// ..;,ooo300200l OOOO 300200dlsplacement (mm)odisplacement ( mm)200LGL O 48 m20ot GL o 48 mCJL '_ 04 m :::: 400[:i:l//"400;- GL2'04 m Jl*ioot):!)/'/V1,i i)' OO/Ploof/1 /' { /1 r]l} /J' /// i;::;/':i:; ;!/';(iSi;/"/:i""= 'f /' ;/1'iAfr[ !i{, ';:='P/L"'ff(/"/V:oft)(,)+ l):. [)o iv {"''. _'oo: l!'ifi/{i *,"/'1'*!Ii<" '"* l!"'t !i'-1""*p・-・'**._, OO 20 40(mm) 60displacement . 'O 20 40displacementy (mm)Fig. 28. Calculated load-displacement curve and p-y curves (Case S3,! /= 1.0)3 1 , in ¥vhich the experimental envelope curves of Cases S2O 204060di >-placement y (nlm)"O 20 40displacemer t;* (Inm)Frg 30. Calculated load-displacement curv8 and p-y curves (Case S3,Jt/= O. 10)degraded b ., shaft yielding.(fully-reversed cyclic loading case) and S3 (one-sidedWhen M= I is used (Fig. 28), the calculated p-y curvescyclic loading case) are also sho¥vn for comparison. Notethat only the p-y curves are sho¥vn up to a displacementfollo¥v a peak-oriented rule, the calculated load isoverestimated, and the calculated load-displacememlevel of 66, as the reliability of the strain gauge data ¥vascurve is close to the fully-reversed cyclic loading result ofii SOIL-PILE IN'TERACTIO¥'Case S2shape. This difference in the shapes of the p-_v curves can- - Case S3cause the discrepancy in the experimental and calculatedresidual displacement of the pile, holvever, it is specu-t ll//'.100/.・'Jr 1//1l'lated that this does not necessarily mean that the underes-timation of the residual displacement of piles al¥vaysoccurs in dynamic analyses, because the proposed modelhas a tendency to underestimate hysteresis damping.' +!*/. !A ' J'oO I OOdisplacement (mm)Ioo'- ////i; /i/; /:ll///i:;illl !/f//:/__;)'d;j //;i'. /':f__f: oll'//f/CONCLUDING REMARKS20020ot GL 048 m lf'v 'fj { !:18Tji f']1In this study ¥'e proposed a ne v hysteretic rule for p-y300cur¥'es that can be applied to problems ¥vithin a frame¥vork of total stress computation. The proposed hysteretic rule satisfies the characteristics observed in lateral400!' lGL -2.04 m JzJ:);Ji fl':v'lO 20 6040di -placement ; (nl n)In particular, the proposed rule takes into account t.hefact that the soil ,esistance intensity of piles varies lvithdifferent loading Patterns, as soil dilatancy behavior[ ./ /';}1AA !;'1!/1l 'JlO' f"/j/'".t(r/!(tll,/i !1!11! l)cyclic loading experiments of piles and soil element tests,_ 'I' "'.1;t//J;/. c{/!if:1l !f)1 JO 20 40dispiacementy (mm)Fig. 31. Calculatcd load-displacement curve and piy curves (Case S3,M= 0.053)¥'aries ¥vith cyclic loading patterns. Although theproposed model responds to f'ully-rever'sed cyclic loading:in a similar ¥vay to the peak-oriented hysteretic rule, theoriginal deterioration rule is introduced here using atarget point; this is capable of accounting for the changeIn hysteresis ¥vith the degree of eccentricity in cyclicloading direction. The pr'oposed hysteretic rule did agood job of simulating the responses of piles embeddedin sand and subjected to either fully-re¥'ersed cyclicloading or one-sided cyclic loading at the pile top, includ-Case S'_. The three loops at each displacement le¥'el areing simulating the relationships of load versus displace-steady and indiscernible. The same tendencies are alsoment at the point of load, moment distributions in theobserved in the calculated ply cur¥'es. Even for M= O.'_5(Fi . ,_9), the calculated load is still o¥'erestimated. Thecalculated load-displacement curves at the loading pointspiles, and p-y relationships that vary ¥vith load patterns;are not sensitive to variation in M ¥vithin the range I .O topile beha¥'ior because of the ¥'ariations in cyclic loading0.25.pattern.The application of a rnodel that uses Winkler-type interaction springs to the dynamic anal),sis of foundationsM=0.lO pro¥'ides the best match f'or the calculatedresults, as sholvn in Fig. 30. Corrrpar'ison with theexperimental results sho¥vs that the proposed p-y rulepro¥'ides an excellent prediction of both the loaddisplacement curve and the pl)' cur¥'es. When M= 0.053is used (Fig. 31), the load is underestimated. Thedifference in the numerical results for M= 0.25 and O. 10 islarger than that for cases M= I .O and 0.25. These trendsindicate a possible transition zone in terms of M atapproximately M=0.10; this is likely to be a reievantvalue in expressing the loading pattern dependency. Theinferences described above, in the de¥'elopment of thene¥v hysteretic rule, are therefore supported by thenumerical results that the backbone curve of p-y isapparently shifted dependin*' on the eccentricity of10ading direction in one-sided cyclic loading, and that the¥'alue of kH, approaches kN.In terms of the relationship bet¥veen load and displacement at the point of' Ioading, the calculated accumulationconventional peak-oriented hysteretic rules ¥vill neverprovide appropriate results that sho¥v the difference incan be carried out in t vo stages. In the first stage, f'ar-fieldexcitation is computed using a relevant method such asone-dimensional nonlinear or equi¥'alent linear earthquake response analyses; these are relevant to the cyclicshear deformation rnode of the far-field soil, as is expected to be excited during earthquakes. In the second stage,the far-field ground displacement at each depth is inputinto each end of the distributed interaction springs underthe proposed hysteretic rule; this is pertinent to the cycliccompression-extension def'ormation mode of' the nearfield soil. Accordlngly, the rele¥'ant soil deformationmodes in the far- and near-fields can be taken intoaccount, respectively, ¥vhen the proposed hysteretic rulef'or p-y is used. It is also possible to perform the first andsecond stages simultaneously in a coupled system.of residual displacement vhen the load is unloaded toACKNOWLEDGMENTSzero is less than that obser¥'ed in the experiment. This isbecause, once the pl)' curves cross p= O, the cur¥'es headstraight f'or the target point on the correspondin*" side. InMaeka¥va of' the Uni¥'ersity of' Tokyo, his doctor'al studiescontrast, the actual p-y curves may ha¥'e a more curvedstudy.The first author vould like to ackno¥vledge Prof. K.supervisor', for valuable suggestions given during this 188SHIRATO ET AL.  〃18⑤y1刀ρ05iμ1η,3i玉5−3120(in、lapanese).RE、FERENCES1)Agalby,S,,Ku漁awy,E and Trautmann,C。(1992):Experime瓢aI  sτロdyofdralBedlateralandmome隈behav沁rofdri!iedshafεs  (iuringstatlcandcyclicload玉職9,E1εα1『1‘Pol∼・θ1甲1∼e∫8σκ1∼∫115伽1θ,  (TR−10223).2)Boulanger,R、W.,Curras,C,」、,Kuuer,B,L。,Wllson,D、W.and  Abg鮭ar1,A,(1999):Selsmic soll扁plle−s葛ructure lnteraction  expe錘menεs and anal)一ses,/. Gθo∫θc11. Gθo∈∼∼7、タノト“E刀9’8.,ASCE,  125(9),750−759.3) Broms,B.B.(1964a):Lateral resまsεance of p葦}es lη coねesioηless  so玉ls,ノ、So’1A4θc1∼.Foεイηゴ.Z)’v、,ASCE,90(SNI3),123−156.4)Broms,B.B、(1964b):LaIeralresistanceofpiles1nco盤esivesoi!s,/.  So’1Mθch, Fo醐4、P1、,.,ASCE,90(SM2),27−63.5)Curras,C.」.,Boωanger,R.W.,KuIter,B.L.andW道son,D,、V.  (2001):Dynam茎c experimen!s andεミrla】yses of a p員e−group−supPort−  eds葛ructure,/。(3θαθc1∼.Gεoθηv1P.Eηgrg.,ASCE,127(7),585−596.6)Fuk巖,J。,Klmura,Y,andOokos鮭i,M.(1997)=S[rerlg出andduαil−  ltycharacte醸lcsofpllefoundations,P”oc.2ηゴ∫∫の一/αραア∼  H/orκ51∼oρ oη Sε’∫117’c Dθs’911α1κノRα∼日oゾ『’oゾ」B’“’ゴ9θ5, 7セc11η1ぐα1  Aグθ1P70/r‘7’1‘ノ’〃71αノ『P罪VR1,P∼VRI,(3503),255−274.7)Fukui,J.,Klmura,Y.,Ookoshl,M.andRanno,A,(1998):Large  sca丑e exper霊磁enむal sIロdy on }10rlzonεa}re夏or圭ag forces of s圭Pgle  piles in sand, 7セ0111ゴco’ハ4θηzo1’‘7’141〃η oゾr p耳VR1,Pub…ic V㌧「orks  Research Instltuτe,(3552)(1n Japanese)、8)Gazetas,G.a簸d Dobry,R.(1984):Horizontal respo巨se of plles ln  layered sQils,ノ」(3θo’θ(r1∼、Eη913.,ASCE,1薫0(1),20−4α9〉Japan Road Assoc1a60n 〔2002): Sρθol万cαπoη∫∫oヂ β’9hwの・  B1ヅゴ9(∼∫,」P‘71’1/蓄乙S正’わ5f1マィ‘’置”響(∼5,Nlaruzen,Tokyo(in Japanese).10) Kavva(!as,へ僅.and Gaze亡as,G、(1993):K1賞emaτ1c se葦sm妻c response  and bending of free一鼓ead p員es玉n layered soi1,(3θo∫εc11η’(1μθ,43(2),  207−222.11) Kimura,Y.,Ookoshi,薮1.,Ban簸o,A.and Fuku玉,」.(1998):Experi−  me蹴sonrestoringforcecharaαeristlcsofsingleplleembeddedin  sand and s融bjecτed εo 墨aεeral }oads, Proc詞 ノ,∫∫ 5〉’〃7ρ. ∠)1’(』1i〃んv  ∠)θ5’9’h、/θ‘ho4∫o’∼B1“’‘ノ9ε∫,267−270JSCE(in.∫apa且ese).12)Klshlda,H.aRd Nakai,S.(1977):Largede餐ectionofaslnglep註e  under員orizontal LQad,Proc,9∫h ICSA4Fだ,5ρθc’α1り7Sθ5s’oη 10.  Iu【y,14,玉9フ7,Tokyo,ISS}v隻FE,127−134.B)Kishlda,H.and Nakal,S.(1979a)l A職alysls of a lateraily loaded  pipe w玉th Ωo“41巨ear subgrade reaCt玉Qn, アン’αη5, 0ゾ’パ。/./・,(28玉),  41−53(ln.}apanese)。14) K玉s員呈da, H, aR(玉 Naka玉, S、 (1979b): No駐li嶽ear玉【y of subgrade  reactionintensityanddlsplaceme煎,7甜c11’イo一κ’50,JGS,25(8),  41−53(in、lapa鷺ese)、i5)Koda,M、,Takem鷲ra,J.and Kusakabe,O.(2000):Modell鷺g arld  eva搬aε妻on of p−y cur、・es Qf s沁91e p員e恥sa鷺d,訊Gθo’θぐh、Eη9ヂ8.,  ,JSC}三,(645/11レ50),19互一207(玉“Japanese).16)Kondou,M.,Muro鷺o,Y.,Nishimura,A.and Sa鳳o,N。(1998):  Study on property of restor玉ng force of p㍊e foundaτ玉Qn−soi歪玉【1  eart盤quakereslstanIdeslgn,7hε1011∼/卿α1∼勲r∼1∼σ照舵£17gi11θθ1・至7)Koseki,」,,Kurac厩,Y.and Ogata,τ,(2001):Depeadency Qf  horizon【a丑and ve賞ical sロbgrade reacε呈oa coef澱cienεs on 墨oad1ng  wi繭,(eds.byTa【suoka,F.,Shibuya,S.andKuwano,R.),  /4ゴソα11cθ4加わo’門α’01γS惚∬7、3惚’ηTiε∫”ηgo∫Gθo〃1α’θ∼ブαな,Lisse:  Swets&Ze玉tlinger,259−264.18) KubQ,』L (1965):Experimental study of the behavior of latera践y  !oadεdpl!es,Ploo,61h10SM圧,2,27シ279.19) }〉lak玉, T、 (2002): Seism玉c perforτ嶽ance evaluation of reinforced  concre覧e p1les under grou鷺d,P11.∠).】η1ε評∫,Un蓋vers盆y ofτokyo(由  Japanese)。20)Masuda,T, ,Tatsuoka,F.,Yamada,S.a肩dS&to,丁、(1999):Stress−  strain behavlour ofsand in plane sτraln compresslon,exte脆sio“and  cychc夏oading tests,∫o〃yαn4Foi”π1‘π’01∼∫,39(5),3玉一45.21)Maζlock,H。(1970)=Correla!めPsforε1⊇edesignofla蕊era!lyloaded  P㍊es玉n soft clay,P〆oc211ゴ鴻1∼1L(2廊1∼01響7セc17.Co尺プ。,Houstonラ  ■X,USA,1,5ηづ9422)ヱvlayne,P.,Ku比awy,F.andTraUtma蹟n,C.(玉992):Experimental  studyofundrainedlateralandmomentbehav1orofdr員iedshafts  d瑳r茎ng sIaεic a簸d cyclic Ioad三ng,だ1θご〃”ぐPo、vεヂR〔∼∫(ぞ‘71℃11∫1∼3’∫’μ‘θ,  (TR−io221).23)Nogaml,τ,a簸d Novak,凝、(1980):Coe醗cients of so軽reacεlonτo  pllev重bra“on,広Gθo∼ε01L石’∼9/8.,ASCE,106(GT5〉,565づ70・24)Nogaml,T.,0【anl,」.,Konagai,K.and C駐en,狂.(1992):鼓o難一  1inear so11−P圭玉e玉nteract玉on model for dynamic Iateral mot三〇n,ノ、  G8αθ01L五ηg/1g.,ASCE,118(1),89一玉06.25〉 Novak,》1、(】9フ4):Dy澱amic s涯ffness and damp玉ng of piies,/.Cα1L  Gθo∫εごh.Eη9’響.,11(4),574−698.26)Poulos,H.apd Dav1s,E.(1980)=∫〉’1θだoμ11θ『σ!’oπ5、41解ヶ∫Z∫σ’14  0ε∫19n,Joiln V》三ley and Sons.27)Reese,L.C.(1958)=D1scuss1ononsoilmodu1騒sforlateraIlyloaded  pile,T1”αη5。パSCE,123,玉071−1074,28)Reese,L,C.(玉993):Tれe beg膿ning of p−y curves,a personal  perspeCt1ve,Pノ’oごFE万翫4肋ノ需1∼o〃01∼08塞ノ14紐9/7》∼・の?  β1・’ゴgθ5,弄01・£Y”・θ〃∼θ£vθ11’y,CryStal Clty,VA,USA,FHWA,  36−4玉。29)Reese,L.C.,Cox,W.R.and Koop,R.工),(1974):Analysis of  latera蓋叢y loa(玉ed p韮es玉n sand,Plloご.6’h〆垂11’L C喚1201ぞ】rθご11、Co’ミブ1,  Hoロsωn,TX,USA,II,4フ3−485,30)Rowe,P.W.(1962):丁11e stress−dil謎ancy relatlon for s【a“c  equ1libr沁mofanasse瓢blyofpa躍cleslnco猷acと,Ploo.Rの㌧Soご、  五〇1∼ゴoη,5θヂ1e5。4269,500づ27,31) SぬiraloシNI。(2004)=Computationa玉seismic performance assessme【箕  of a p澱e fo貸ndation subjecte(i to a severe eartれquake,PILO.  771θ訂5,Unlvers1ty ofTokyo、32) Sblraζo,八・1.,Kosek韮,、∫、,Fロk避i,」、andK萎mura,Y.(2006):Efぎec1sof  stress−dilatancyb曲avlorofsoiionloadtransfeぬysτereslsinsol!一  pilel猷eractlon,50〆’∫σ’1ゴバα〃∼ゴα”0’∼、∫(inp血εing}.33)Yoshida,王.and Yoshi鷺aka,R.(1979〉:A me由od to estlmate  modulus of 姓orizontal subgrade react玉on for a p疑e, 501’5 αηげ  」Fo∼〃漁1/0175,12(3),H7,
  • ログイン
  • タイトル
  • Ground Behavior due to Tunnel Excavation with Existing Foundation
  • 著者
  • E. Sung・H. M. Shahin・Teruo Nakai・Masaya Hinokio・Makoto Yamamoto
  • 出版
  • soils and Foundations
  • ページ
  • 189〜207
  • 発行
  • 2006/04/15
  • 文書ID
  • 20899
  • 内容
  • SOILS.へND FOUND.へTIONSVo1,46,No、2,189−207,.へpr,2006、耳apaaese Geoζecllllical SocietyGROUND BEHAVIOR DUE TO TUNNEL EXCAVATION            WITH EXISTING FOUNDATIONEuNsu SuNGb,HossA正N M.SgAHIN玉D,TERuo NAKAliii},MAsAYA H茎NoK夏olv)and MAKoTo YAMAMoToi)ABSTRACT  Two−dimensiona1(2亙))and three−dimens玉orlal(3D)model tests ofl芝unnel excavation wlth Ilearby exist量ng foαndationαre carr呈ed out to inves盛gate the星nHuence of t麺e existing foundation due to the interaction bet∼、・een groun(I and theexisting structures.Three types of foundatiolls:flat foundat呈on,group−pile foundation and pi豆ed raft are considered・2Dα11d3D負!11te element analyses using subloading∼ij model are also conducted、The deformation mech&nism anddistr呈butio}10f earth pressure during turlne玉excavatioll ill the ground with nearby foundation are found to be differelltfrQm重hose of green最eld condltion.Surface settlement trough due to tunnel excavatloll lmhe ground wi出existingfoundation does not follow the usual patte玉・n of a Gauss圭an disξributive cur∼・e,which can be observed in the case ofgreen行eld.Especially,in the case of plle foun(iatio11,∂p,由e distance between pile tip aud tunnehs an importantf&ctorforthegrounddeformationandsurねcesetdement.Forashol’ωlstanceρP,althoughthelengthofplleislong,the ground deformadon is concentrated at&place neaII毛he front pile and the rotation of foupdation becomes largeLThe maximum surface settlement in the case ofexisting foundatlon is also larger than those ln the case ofgreen field.Due to the ex宝stiRg foundadon,unsym【ne童rical distributions of earth pressure occurred at t}1e bottom of重he grounddue to tulmel excavation,both in mode1毛ests and numerical analyses.The earth pressure at the crowτ10f tu重mel in thecase of existing foundation is almost the same as t瓦ose in the case of green neld。The ardling at the s難oulder of tunnelin由e case of existing foundation,however,is much Iarger thamhose in the case of green負eld due to the dea(110adexerted on the foundation.The numerical results agree∼、7ell with the res漁s of the model tests.Keywords:const1・uctioPsequence,earthpressu1−e,nniteelementallalyses,肱foundatlon,group−pilefoandatio熱,mo(1el test,plled raft,surface settlement,tunnel excavatio【1(IGC:E2/E6/E12)excavation.The responses of’single pile and group−pi玉e1NTRODUCTlONare computecl sep&rately using εし boundary element  As a lot of stmc宅ures exist alongside the road whereanalysis.Sh&hin et al.(2004&)presented surface setde−ξun勲els are usually excavated,the iateraction of existingment and eart負press“re ln the green field conditio難structure εしnd tun董1eling should properly be considered由roug勤出etrapdoormodeltestsandnumericalanalysesuslng elastoplastlc subloading∼ij modeL They discussedduring tunnel construct玉on。 This intel’act玉on could beabout蛮he e仔ects of ground depth on the ground deform3−thought in two cases. O熱e is the inauellce of tunneltion and earth pressure a!畠ound tunnel.They also clis−constructlononexistingfoundatlons,alldtheotheristhein負uence of pile constl齢uction &nd Ioading o熱 exist呈ngcussed about t紅e inf董uence of shallow εoundation (flattunne至s.The efぞect of tunneling on亡he p圭le foundationfoundation)on ground behaviors due to tunnel圭ng,anclhas not been fully understood.conclu(1ed that surface settlement and earth pressureαre  Recently,several studies on this i飢eraction problemvery muc短岨uenced by the existing load of Hat founda−have been publis紅ed.Among them,Jacobsz et aL(2004)tion during tunnel excavatiol1.However,the e葺ects ofpublishedthestudyabout由ebehaviorofsingledrivendifferent types of foundation have not been discussed inp三1e due to重unneling.T}1ey discussed about the cha鍛ges由eirresearc熱.  This paper presents model tests and numerical analysesof p星le settleme熱t,base loa(i and sha負friction of pile in(iense sand using centrifuge test.They also presented由eto lnvestigate the mec紅&nical behaviors of tunneling ina任ectedareaforpilese毛tlementduetotunaelexcava重ion.重he ground with exlsting foundatio貧、Surface settlementsLoganathan et aL(200重)presented a closed£orm solutlonto estimate t紅e呈nduced ground n1ovements due to tunne1the existing nearby foundation,and is compare(1with thebiili旬吋and earth pressure due to tunneling are量nvestigated withGraduateStudent,Departme搬o∫αvi1Engineering,NagoyaIns芝i瓢eofTechnology,Gokiso−cho,Sllowa−ku,Nagoya466−8555Japaa・JSPS Postdoctoral Fe110、v,di疑o、Professor of Geo{edmicai礁ngllleering,Deparエmenエof Civil Engineering,dl【【o(nakai@mtiLace、煎ech、acjp).Researcll、A.ssociate,Department of Civil Engineering,ditto、餉nlemanuscriptfor【hispaperwasreceivedforreviewonJulle15,2005;apProvedonDecen}bed4,2005・Writte自discussionsonthispapershouldbes縫bmittedbeforeNovemberl,2006totheJapaneseGeo【echnlcalSocie芝y,4−38−2,Sengok娃,Bunkyo−ku,τokyol12−0011,Japan.U罫)onreques口11eclosl自gdatemaybeexこendedonemon由.189 1 . ^lSUi¥TGJ}OET AL.Laser tv e Dis lacementTransducer8cm:)ad Load;; j . !:lcm(a) Flat foundationlass of'80mluminum RodsLoad CellsT'-cmlockL pndleFig, 1. 2D trap door apparatus for group-pile fot nda{ionhresults of the tunnelin9: in :reen field condition. Theinfluences of different foundation types such as a fiatfoundation, a group-pile foundation and a piled raft onT j !0.5cm(b) CJroup-pile foundationand Piled raftthe ground deformation and earth pressure are alsoinvesti ated. As lvell as 2D model tests and numericalrig. 2.Foundationst setiin2Dmotiel tcsianalyses, 3D model test and numerical analyses areper'formed to invesrigate the effect of sequential tunnelexca¥'ation for this interaction pr'oblem. Subloading tijmodel (Nakai and Hinokio, ?004) is used in both 2D and3D numerical anal_¥'ses. CJround of I g model test of lo¥vconfining pressure is used in this study.2cm(a) Flat foundationDF,SCRIPTlor l OF MODF,L TESTSApparanls of 2D I loclel Tesrs'_D model tests are carried out to investigate the effectsof ex'isting: foundation to the g:round beha¥*ior and tlledistribution of earth pressure around tunnel due to tunnelexca¥*ation. Figure I is a '-D trap door apparatus of thegroup-pile foundation used for the purpose. Theapparatus consists of 10 brass blocks, marked ¥vith blockA to J, each of ¥vhich has a vldth of 8 cm. Three blockslvith load cells are used to measure the earth pressures inthe site of lowering block F and both side (block E andblock G). Each block is di¥'ided into four small parts tomeasure the earth pressure dlstribution on each block.Surface settlements are measured b.¥" a laser-tv. pedisplacement transducer ¥vith an accuracy of O.OI mm.Details of this apparatus can be referred to the pre¥*iouspaper (Nakai et al., 1997). Three types of foundation areused in the '_D anaiyses - the ffat foundation, the grouppile foundation and the piled raft. Figure_ sho¥vs this flatfoundation, group-pile fcn.mdation and piled raft used inthe '-D tests. The fiat foundation is made of an iron plateof 8 cm in ¥vid h and I cm in thickness. The cap of thegroup-pile foundation and the piled raft is made ofaluminum plate of 8 cm in ¥vicith and 2 cm in thickness.The pile is the plate of 5 mm thickness, ¥vhich is made ofpolyurethane vith the Young's modulus of I .06=:<105 kN/m2. If ¥ve assume the similarity ratio of l:lOO bet¥¥'eenthe mociel test and protot_¥'pe,his corresponds to the0.5cm(b) Piled raftrig. 3.Foundation usedin3D modeitestcondition that piles ¥vith a bending stiffness (EI) of6.7*105 kN.m2 and an axial sriffness (EA) of 1.4*107 kNare arranged ¥¥'ith a space of 5.5 m in prototype. The rafttouched the ground surface in the piled raft, ¥vhereasthere is a gap of 7_ cm between the raft and the groundsurface in the gr'oup-pile foundation as sho¥vn in Fig. I .The lengths of the pile (Lp) ¥'aried according to thepattern of the test (8 cm or 16 cm). The distance bet¥veenthe front pile and rear' pile is 5 cm in all the group-pilefoundations and the piled raft. To impose the existlng rTUNNELING ¥¥*lTHEXISTiN(,}pOU¥. TDATIONl'.,lBilil8cm i Io ckI ¥8cm Load Cell(a) Block-by-block excavationPhoto 1. 2D modelest for piled raft (D/B= 2.0, Dp/B= 1.0)T ral sd u c e r10ad, the same ¥¥'eight (3. 14 N/m) is placed on the centerof' the raft during tunnel exca¥'ation in every test. Photo 1sho¥vs the '_D model tests for the piled raft. By takingphotographs of the ground during the experiment ¥vith adigital camera, the trend of the deforrnation inside theround can be ¥*isualized. In the '_D model tests, aluminurn rods, ha¥'in_ ,_ diarneters of I .6 and 3.0 mm and mixedvith a ratio of 3 :2 in ¥veights, are used to simulate the soilmass. Both t.¥'pes of aluminum rods are 50 mm in length.The unit ¥veight of the aluminurn mass is '-0.4 kN/m3 atScn] Pu!kout Devicethe exper'imental stress level.(b) Sequential excavationAppa/'atus of 3D Mocle/ 'TestsFigure 4 represents 3D apparatus of t¥vo types. Figure4(a) is the apparatus for the block-by-block exca¥'ationused to investigate the change of earth pressure by tunnelexcavation. Figure 4(b) is the apparatus for the sequentialexcavation used to investigate the change of the surfacesettlement and earth pressure. Details of' these apparatuses vere described in the pre¥'ious papers (Nakaiet al., 1997; Shahin et al., '_004b). For measuring earthpressures in the block-by-block exca¥'ation, three blocks¥vith load cells are placed at the position of block F and{i;;{**of aluminum, and the Young's modulus of the pile is1 .59' <107 kN/mL. The distance bet¥veen t¥vo piles is 5 cm.Six laser-type displacement transducers are used topressure distribution. Figure 5(a) illustrates these threeblocks with load cells used in the block-bv-block excavation. In the other type of 3D tests, there are four bars atthe top on the pullout device as shown in Fig. 4(b). Thethickness of each bar is 4 mm. In the tests, four bars atas sho¥vn in Photo 3. The displacement transducers areplaced in such a vay that the displacements in threeFi_O*. 5(b).{)D tests. All data are recorded in a PC throug:h a datalogger. Fi_"._ure 3 sho¥vs the geometry of the flat foundation and the piled raft used in 3D tests. The size of' cap Is8 x 8 crn ancl 2 cm in thickness. In the 3D tests, the piledraft has f'our piles as sho¥vn in Fig. 3(b). Each pile is mademeasure the displacements and rotations of the flat foundation and the piled raft for the 3D sequential exca¥'ationmodel tests can be carried out lvlth any desired exca¥'ationsequence. Photo 2 is the apparatus used in the model test.In the pullout apparatus, there is no device for measuringthe earth pressure at the top of sliding bars. T¥vo blocksof load cells at a side of block F are set up to measure theearth pressure for the sequential excavation, as sho vn ini3D mode] tcst apparatusesbeside the tunnel as sho¥vn in Fig. 4(a). A block ¥vith loadcells is divided into eight parts so as to measure t.he earththe top are pulled out to simulate tunnel exca¥*ation.Since the sliding bars can be pulled out independently,;Fig. 4.Sur'face settlements are measured at the trans¥*ersecross section of the ground by a laser-type displacementtransducer in the same ¥vay as the 2D tests. The dead loadis applied on the center of the plate in the same lvay as thedirections (X, Y and Z) and the rotations about three axes(X, Y and Z) can be calculated by the data of thesedisplacement transducers. In the 3D model tests, aluminaballs mass, ¥vith diameters of ,.O and 3 O mm and mixed¥vlth a ratio of 1:1 in veight, are used as a soil mass. Theunit ¥ 'eight of' the mass of alumina balls is 7-2.3 kN/m3 atmodel stress le¥*el.DESCRIP'I'lON OF NUMERICAL ANALYSES ANDCONTEi 'TS OF MODEL TESTSFEIVI Mesh anc! Silnulatioll of E.rcava!ionNumerical analyses are conductecl in the same scale asthe model tests in plane strain condition ('_D) and 3Dcondition. Figure 6 sho¥vs the 2D meshes for the ground¥vhere D/B=2.0 f'or (a) flat foundation, (b) group-pile :!'SUNG ET AL.1 92Photo 3. I,aser-displacement transducers for measuring rotation offoundationPhoto 2. 3D mode] test with pul]ing out tunnelingTunne l a xisMeasurement line of(a) F]at foundationearth pressure8cmMeasurement hne ofsurface settlement(b) Group-pile foundationY(a) Block-by-block excavation(c) Piied raftTunn l axisFig. 6. ¥. Ieshes for 2D ana] .'ses (D/B= 2.0)Two blocks with load cellsMeasurement line ofearth pressure/Measurement line ofsurface settlementX jY(b) Sequential excavationFig. 5.Arrangement of blocks with loati ceils in 3D tests (plane vie,v)step. Elastoplastic joint elements (Nakai, 1985) ar'e usedat. the interface betlveen the piles and ground in the 2Danalyses. A friction angle of 6= 18' is used in the jointelements. The value of the friction angle is determinedfrom a slidin*' test of piles and mass of aluminum rods.Figure 7 sho¥vs the 3D mesh for the simulation of theblock-by-block exca¥'ation and the sequential exca¥*ation.The left-hither part in the mesh is not illustrated to sho¥vthe near piles in the ground. The front and the back face,as vell as both the side faces of the 3D mesh, are free inthe vertical direction, and the bottom face is kept fixed inall directions. The 3D block-bv-block exca¥'ation ¥vith ado¥vn¥vard displacement of d= 4 mm is imposed block byblock from A to .J, and measurement of earth pr'essure isperformed in block F and the other t¥vo blocks ¥vith loadfoundation, and (c) piled raft. Both side faces of the 2Dmesh are free in the ¥*ertical direction and fixed in thecells, as sholvn in Fig. 5(a). The numerical analyses of thelateral direction. The bottom of the mesh is fixecl in all theblock-by-block excavation are carried out by applying avertical displacement of 4 mm at the bottom nodes alongdirections. To simulate the exca¥'ation in a 2D condition,one block length (8 cm), as sholvn in Fig. 7(a). The eartha 4 mm ¥'ertical displacement is imposed at the nodalpressures in the numerical simulations are calculatedfrom the centerline elements; same as the model testssho¥vn in Fig. 5(a). In the model tests, the do¥vn¥vardpoints corresponding to the top of the lo¥vering block F inthe model tests, ¥vith an increment of O.002 mm at each rTUNNFLING ¥VITH EXISTING FOUNDATION '80cnl ::__* "';r t_:Dj ;SS;i ;- ;; si ':SS;' ; i ii;'t;f-' "' 1:'i{_ -f'Yi:;c;t ii $'*i-"; : _' '' sl : 1 "s ' "'!''f%'' ;-;*1;i:: :;;;;;::; il;,. ;;; Tl :ci' :;1' ' ]; ;:; ii: ;;;:::{,;"r;;::::; ;:;:;;s ; *;' *i'};; ' ; xs''/1;s :; ' / 'tri:("'i;; ;';: ;'*; '+ '*'-'I:"#:"- -tF;"; ' r' ;i1 ;;:T : :; ; "_ ' ; ' ; '1r;'t:f;';'i} -" ; :['-"' : -"*' ;+ ;#'Y¥ px_ *; ; s'#'* F ; s _:-'-:T - * --*.#*;* -*''ii"_"- *H1'-*f : "*-* =+** sl-_? :':'-= " .,f4_/"-f:';{i;';(a);se':;"':_ '-14'' _rl""::i;f/'rs'{';'+- ;- ': ; "ITs'#'; ';F' ':' ; _3ror o 2=04;Fig. 8.:t ;Stress-strain of aluminum rod mass35(a) Mesh for 3D analyses (D/B=2.0),_,(The lefi:-hither part is not illustrated to showbthe piles in the ground)c- observed f(r .=19.6kPa)- calculated LL/ a := 9 6kPa)i -----caiculated a *02kcpa)b-o-o 8Measurlng Line of earth pressure-o 6;,' E F H 'e$h_ 2*'6s 6 7 8 9 ( __)Triaxial testv Y y F =o:b) : ))42*rd:::4rnmBlock_; - 8dill c/i/ ; :; ; 1 ;6 7_l' f -s: #:T= * **SSSfL ' ;f r ;.SS-' f; ' . 'L/" ://' 1'so 2 (?s)#'**: " ';me1'_or: 'i"' ti' :i ::;:T ;f" '-'+ "'s_ ': ' L;S s ;- (t s:x;H ";; r ":ss:s ;** s# " ? ';+ ;" ;: *^isebs :rYed[: (1 i 1 9'6kPs-- slouls!e ( Q ' s 19 s'k?s- - - csl:ula c ( c?,jri: t' <'surface settte(nentMeasureme'--ea cuistE r (; =1q' Li:)_ - 1:a cu!ite [: o ;=e'2 ?a:.'-- 't: t;j:: : :; ;;' '- ; "ss_ ""+ ' - :#;:';;'; ' -'; ' 's; :"*'- ;T' : T ' ;tt_ '! '-*p ' ""- ;'*1;'f"+ '=t f''=:: const o erYe r T 1=19 kPZI - cl 1: const-'t;' 8 O cm# 4F 's193c)omp.-o 4(Tf const.' l+0 _,o..:"0204S (o/o)( alumina balls mass for 3D)(b) Block-by-block excavationFig. 9.Strcss*strain curves for alulnina ballsMeasuring Line of'surface settlen en(Y=0)Table 1.Parametcrs of soil matcrialI'_cm(c) Sequential excavationFlg. 7. Mesl] for 3D anal)ses and simulation of tunnei excavationN (e c at p= 98 kPa &= O kPa)O 008O 004O_30Rcs = ((TI l(T )cs(* mp l1 80pl .20aO 201300displacements of c!=4 mm are imposed sequentially bypulling out the sliding bars. In the same ¥vay, the numeri-cal analyses of 3D sequential excavation are carried outby imposing a downward displacement of d = 4 mrn to thecorresponding nodal points. The process of applying thedisplacements to the nodal points in the analyses isna balls, frorn vhich it is clear that the strength anddeformation beha¥'ior is very close to dense sand. Solidlines in the figures represent calculated cur¥'es, ¥vhich arecoincident on the lvhole with the test values representedby dots. On the other hand, dash lines represent the theo-illustr'ated in Figs. 7(b) and 7(c). Surface settlements ofretical predictions of stress-strain relation under a confin-the model tests and the numerical analyses of this sequential exca¥'atlon ar'e measured at the prescribed places asing pressure of I /100 the confining pressure used in thetests. From the stress-strain behavior of the element testssho¥¥'n in Fig. 5(b). Earth pressures in the numerical a-simulated ¥vith subloading tjj model, it is noticed that thisnalyses are calculated along the elements sho¥vn in Fig . ,5 (b) .model can describe the dependency of density and/orCoilstitutive Mode/ Used in Nun7erica! Ana/ysesconfining pressure on the stiffness, str'ength and diiatancyof soils. The model pararneters for the *'round are sho¥vnmodel (Nakai and Hinokio, '-004) isin Table 1. The same material par'ameters are used in theused to simulate the ground material. This model cananalyses for the aluminum rod mass and alumina ballsmass, vhich are independent of the ground density andThe subloadirrg' tjproperly describe the influences of intermediate principalstress, the dependence of the direction of plastic flow onthe stress paths, density and/or confining pressure on thedeformation and strength of soils. Figures 8 and 9 sho vthe results of biaxial tests ¥vith the aluminum rod massand triaxial test with the alumina balls used in 2D and 3Dmodel tests, respectively. These figures show the positiveand negative dilatancy of aluminum rod mass and alumi-the confinin*' pressure.Initia! Sti'ess of G/'oulld and the Deac! LoadThe initial stresses of a green field are assigned to themodel **round in all numerical analyses before the deadload is applied, which is accomplished by irnposing bodyforces of y = 20.4 kN/m3 for the aluminum rod mass, and 1{SIJNG ET AL194Table 2.12Contents of model tcsts antl numerical anah. sesType ofDimensionT)'pe offoundationD/B D lBexcava ionl8Z:_ClBlockexcavation2Do:Group-pile (Lp=o o 02 0.04 v/B fPig. lO.CJroup-pile2.0 1 O6)A-piledr ft (Lp=8)piledraft (L 16)O.063^OO 08 O3 ^O IlResuhs of bearing capacit .' bl_ 2D model testsO2.02 O l.O3.0 2.0Piied raftO- Flat foundation- - Group-pile (Lp=8)3.142.0foundation4Dead ioad:FlatfoundationFku)_^OBlock-byfoundalion3 O-biockexcavationPiled raf_) .O IO3.0 1 .O2.03DSequentialexcavationFlat,_.Ofourldation3.02.0 1 .O3.0 l OPiled raft2^O10' 1.0Btl FoundationC>Settlement trou)a Flat foundationDead load: 29.43 = Piled rafi(L 8)in greeu fil eldD: depth:, Def rmation region due to n{h Piled rafi(L I 6)nnel', ex vation in green fie]d co0.06 0.08 0.l0,02 o 04v/B fFig. 11.Results of bearing capacit) by 3D model testsTrap doorl,ocation of foundation in model test and numerical anaiysesFi('. 12.body forces of y = '-'-.3 kN/m3 for the alumina ball to all, Q=3 . 1 4N/cmelements under gra¥'itational condition, starting from a¥'ery small confinin>" pressure of pa= 9.8 x lO6 kPa vithIn all 2D and 3D tests and numerical analyses, the deadload representing existing structure is applied on thecenter of the foundation to investig:ate the influence ofround deformation and earthpressure around tunnel during tunnel exca¥'ation.Because of the dead load, the initial str'ess field is differentfor the bearin*' capacitv_ for the flat foundation, thegroup-pile foundation and the piled raft as sho¥vn inFig. 10, and its value is about I /,- of the residual load forthe flat foundation. In Fig. 10, the left ¥'erticai axisrepresents the applied load and the abscissa representssettlement of foundation (v) normaliz,ed ¥vith the width ofthe plate (B ) in the model tests of bearing capacity. Thebearing capacities for the group-pile foundation and thepiled raft ¥vith different pile lengths are also Indlcated inJ lJO.2L- 0.3LO.4 D/B=2.0S:)- d=1mm;!:ii1{h 4mrn ' -=j - 4mm (green fieid)0.5{l 'oJO. I)/O.7_.-0.3from those of g:reen field. The ¥'alue of the dead load inthe '-D model tests and numerical analyses is 3. 14 N/cm(average pressure= 3.92 kPa, ¥vhich is aimost equi¥'alentto the load of a t¥venty-stories building ¥vith a similarityratio of I : 100). This load is decided by the )_D model testsi: jIl i ' I lOO. lan initial ¥'oid ratio of e= 0.35.the induced stress on theitionD/B=3.00.4 (a)observedO.) 0=-16 -i2 _+ 1-4 O 4 8 12 16 20ITTTTFig. 13.x(crn)Observeti profiles of surface settlement (Flat fot ndation)Fig. 10. Although the bearing capacities of the group-pilefoundation and the piled raft are larger than those of theflat foundation, the same dead load is applied to all foundations to investigate the infiuence of different type offoundation and the length of pile. Figure 1 1 indicates the3D results of the bearing capacity for the flat foundationand the piled raft ¥vith different pile length. In all 3Dj TUNNELING ¥VI'TH EXiSTING FOUNDATIO¥. *l 95model tests and numerical analyses, the dead load is assumed to be 29.43 N (a¥'erage pressure=4.60 kPa) vlththe same reason as the '-D condition.distance bet¥veen the pile tip and the tunnel cro¥vn (Dp)are in¥'estigated ¥vith t¥vo kinds of Dp/B= i.O and ,_.O.The tip of the pile of all model tests and numerical ana-Contents oj' Mode! Tests anc! Numerica! Analyses'Table 2 sho¥vs t.he contents of the model tests andplaced in the distance of I .OB from the centerline of thetunnel, ¥vhich is ¥vithin the region of large deformatioudue to tunnel exca¥*ation as sho¥vn in Fl :. 1'_.lyses is located above the tunnel. The foundations arenumerical analyses for 7-D and 3D condltions. Threedifferent types of foundations, the flat foundation, thegr'oup-pile foundation and the piled raft are used toRESULTS AND DISCUSSIOr ITS IN 2D CONDITIONin¥*estigate the infiuence of different foundation type onthe earth pressure and ground deformation in the '_DSulf'ace Sett!eine/7ts anc! Grouncl Dejbl'mationFigure 13 sho¥vs the profiles of surface settlements ofmodel tests and numerical analyses. In the 3D tests, onlyt¥vo types of foundation, the flat foundation and the piledraf't, are considered. The influence of the overburden tothe interactions betlveen the deformation of ground andtunnel excavation are investigated ¥vith t¥vo differentdepth of ground in '_D and 3D model tests and numericalanalyses, ¥vhose ratios of dept.h (D) to ¥vidth of' blockthe model tests for applied displacements of I mm and4mm in the case of D/B='_.O and 3.0 for the flatfoundation. Figure 14 represents the computed profiles of'surface settlements corresponding to the observed results(B=8 cm) are D/B=)_.O and 3.0. In the cases of thegroup-pile f'oundation and the piled raft, the effects of the, Q=3 . 1 4N/cmO0. l0.2 ld=1mm I{ * 4mrnF: O.3L・ ) 0.4L;:Oe)D/B=2.0: - 4rnm (green field) .,jO.5 [OO. le)I0.2 LO 3O . t0.5D/B=3.0(,b) Fompute;O -16 -12 -8 -4l , I , I ,4 8 l' 16 20oL!J't!Jri('. 14.Fig. 15. Deformation zone of modei test (Flat foundation)Computed profiies of surface settlement (Flat foundation). * '-// i !,f' f! J'+ ". --*t' / r"r .! // f ' ,! /' /fS+_ , "'.+.*r*r!1!.'ifl,! r* f"r.',r I f! ;! r , t' I :!S<r ...s:'""//.!F'/1'/'//i*.'/,'!' ' ,'!'Ifr,!111!i'.' i =' ;*! /i*1l fll,1',i ' / ;= !!,rill!!//! /J-S"i," S' ,, !! J/!.':'/1!dr.' !i J ! r t,r!.'.'!li! ''ss l'L!!]( . isJ'I'l'l'f: ' ii'i': '!; , 'i*: :i ,:! 2 ,?; 1!'!ij i! !!1l/J'!1iI "!ii'rll'i":,'!!.'1'ii!"!1!ii'J''!!!i"It!iifl"'i"'*'i!J;,+.'ii'iif"J'Iii'='i ''ilJ ;!'*111 f/ li//iii'lil'll !!!! : *'= il'i'*=!j'!1ljl"'+i i! i:ii'f!111'!iiJiiJlii/"fJ*lJlril!!!? * .: 'i:1ijiiij"'!;-i;Ij :i':・・..-i;:. IS,y!'/'='' f! '1lIff.".=.iii""I'.LnL'{ il]' f*: J" ; I rir' 'D/B ao * j I:'!,jj" a()0'00 o'02 bo,04 o.06 o,08 8,io0.00 O.06l'iO.18 O.24 O.30(cm)Fio, 16.Computcd (a) displacement vectors and (b) shear strain contours of ground (Flat foundation) SUNG ET ALl 96t Q=3 ・ 1 4N/cmO -l I I i I_lr; /,l Iy f.r"*'lOlO.2O.3O 40.5D/B=2.0Dp/B= I .Opile=8cm: O 2* O 3-*. *,J O.4- d= I mmjh 4mm'-*_ ODIB=3-+1 ;ir'L t_"J //ODp/B=2.0Pile=8cml l:.J. O 5O'0304'r ' -PI';;/17h ' /- FTri !j7J :_'1L: J IO,lO.2l! * '): ----"-' f-LJ O.le'*_, r:L JI; 4 iD/B =3 . ODp/B= I .OPil e= 1 6 cml !I (a) ObSer{vedO.520 16 12 -8 -4 O 4 8 12 16 20{I x(cm)Observed profi es of surface sett]ement (Grotlp-pjle)Fig. 17., Q=3 . 1 4N/cmI 'i,-O , lLrlLriLJ11LrnLr r/ L n 'o.{o 20304D/B='_.ODplB=1 O, pfle=8cmo 5ll ..l green fleldjEF g 19. Deformation zone of Inodel test (Group"pi!e)-, O.le,s:)e,,?O.20304D/B =3 . OD plB=2 . O, pile=81 cmO.5l I , I * l0.1O^2O 304O ,5D/B=.3 ODp/B= I O, plle=1 6cm20 16Fig. 18.l (b) clomputedJ' -8 -4 O 4 8 12 16 ,_O7 x(cm)Computcd profiles of surface settlement (Group-pile)in Fig. 13. The surface settlement of green field ¥vithapplied displacement of 4 mrn is also plotted in the samegr'aph ¥vith solid lines. The position of the dead load isdepicted at the top in each figure. As sho¥vn in thesefigures, the maxnnum surface settlement ¥vrth the flatfoundation is larger than those of green field for D/B=2.0 and 3.0. The location of the maximum sur'facesettlement moves to¥vards the position ¥vhere the deadload l 'as applied (Shahin et al., 2004a). The range ofsurface settlement tr'ou :h is shorter than those of thereen field condition due to the concentration of settle-ment trou h are almost the same as the model tests.Figure 15 sho¥vs the observed movements of the modelround due to lo¥vering block F. These figures areproduced by superimposing t¥vo photos-one ¥vas takenbefore lo¥vering the block, and the other lvas afterlo¥verin*' the block by 4 mm. It is re¥'ealed in these fi*・uresthat the deformed zone in the case of D/B=2.0 and .3.0spreads from the top of the lo¥vering block to thefoundation, ¥vhich is different from the results of the*'r'een field condition. The deformed zone spr'eads o¥'er allparts of the fiat foundation. Figure 16 represents thecomputed displacement ¥'ectors and shear str'ain contoursof the numerical analyses, ¥vhich are dra¥vn ¥vith the samescale for all ground depths. The computed results of9:round mo¥'ement sho¥v the same tendency ¥vith themodel tests result sho¥vn in Fig. 15. For D/B=2.0, thedevelopment of shear strain concentrates to the left of theplate, and to the right of the plate for D/B= 3.0, ¥vhichincreases the rotation for the plate as sho¥vn in Figs. 13and 14. Ho¥vever, if all parts of the flat foundation arelocated in the large deform Ltion region as D/B= 3 .O, therotation value of the fiat foundation becomes smallercompared to those in the case of D/B=2.0, in ¥vhich thement around the flat foundation. The rotation of theleft side of the plate is more affected by ground deforma-plate is in anti-clock¥vise direction in the case of D/B=2.0, but in D/B= 3.0, the rotation occurs in the oppositetion.direction. The computed surface settlement and settle-Figures 17 and 18 sho v the profiles of surface settlements for the model tests and numerical analyses in thei FTUNNELING ¥¥*lTH EXISTING FOUNDATIONl 97l;: 'Ji J i !l ,i J, ,1:tllJD/Bs:3'o' DP/B::2'oLLi '.+*.. I! f+.. !ri r' =11'i 1' 'T_]'$i itf iiil'if?ii;{i;{"i / ;) '?/r,riii)'?ri:';:;,'! i :'{;" i fili" ? !If";"'}': L i'ij!;}i1""' l!!r';(i,t.; !'i..・*i{ ' }jji I i;';!!;;;"!"'; 'i]t' it!;; i,i*[ITllrlD/B 3'O' DP!B=1'OD!B 1.0, Dp/B=1.01 t 1 l{I ) !'i!; il lLi i !J:rdi 'llTl};f} Ti fif${i7'i i1 $iiJ.It l{ij { I"!iJ{[Ji".i ,llJ{'1 1 <;1 '!liT';:lJ'if'!i -' '!L il J'!i"'!;ttiJ;i".Ji!'f IJ F /;(;ii 'iJ il ;''!'( a)(b)/L I iL **O OO O OFig. 20.O,a4 O,a6 O.08 a.la 0.00 O,a6 dli2dli2 OO 18 O 24 O 30(cm)Computed (a) displacemeut vectors aDd (b) shea strain contours of ground (Group-pile)Q=0 , 1 4N/cmO-- r='=iff'l='14r frr rflirrp* j,-is1= 7O.Oi- i ,1'0.10.3O.Q=3 ' 1 4N/cmD/B =2 . ODp/B=1 O0.2DplB= I .Oo 0.4pile=8cm{} green4mmfiel0.51 rrD/B =,_ . OF: 0.3- - d= I mmP'Ie=8cmLrn lO. 1r!T , l}' 4mm' greenfield'*5 0.5F:)e)O0.lO.2O.3O.40.5r**'"i t ioO*if '---- 1, ..', != L J ¥ JD/B=3 . O JrlJ*-*'Lrrl/O.2O.3O.40.5Dp/B=2.0p'le= cm (a) bsee-20 -16 -12. -8 -4 O 4 8 12 16 20/D/B=3 . ODp/B='_.Oi e=8 m(b) comp te20 16 12 -8 -4 O 4 8 12 16 201 x(cm)Fig. 21. Observed profiles of surface sett!ement (Piled raft)lO.lll7 x(cm)Fig.2・-Computed profiles of surface settieme It (Piled raft)case of D/B = 2.0 and 3.0 for the group-pile foundation.These fi**ures sho¥v that the maximum surface settlementwith the group-pile foundation is larger than those ofmaximum surface settlement moves to¥vards the positiongreen field for D/B=2.0 and 3.0. The location of thedirection for D/B=2.0, ¥vhich is the sarne trend to thevher'e the dead load vas applied as the results of the ffatfoundation. The r'otation of the plate is in anti-clock¥vise 1 i.,SUNCj ET AL_l 98results of the flat foundation. In the case of D/B= 3.0,the observed amount of rotation to¥vard exca¥'ation forDp/B=2.0 is smaller than that of Dp/B= 1.0. On theother hand, despite of the different ground depth, therotation of the plate for D/B='_.O and Dp/B= 1.0 isrotation and amount of surface settlement for the piledraft are a little smaller than those of the group-pilefoundation because of the contact condition bet¥veen theplate and ground. Figure 23 represents the obser¥'eddeformation of ground. Figure 24 represents the com-similar to those of D/B= 3.0 and Dp/B= I .O. Therefore,the rotation of the plate ¥'aries on the distance (Dp)puted displacement vectors and shear strain contours forbet¥veen the pile tip and the excavation block. If both thefront and rear pile tips are located in the large defor'mation region as Dp/B = 2.0, the amount of rotation ¥vill bethe piled raft is similar to the group-pile foundation. Inthe case of the group-pile foundation and the piled raft,the piled raft. The ground deformation mechanism forsmaller. As the tip of the front pile is located ¥vithin thelarge deformation r'egion and the tip of the rear pile islocated outside the large deformation region like Dp/B=D/BF 0' DP/B 1'O"I i7"';;':*' - ' ; '-i-**"''*"*'** ';' i/ ;':1: *+: ;.;j ; : ., : - 1}:l'.'1 .O, the amount of rotation is larger than that of Dp/B ='_.O. The computed surface settlement in Fig. 18 sholvsthe same tendency for rotation to the obser¥'ed results.Fi ure 19 indicates the observed movement of the model_'..round for the gr'oup-pile foundation. As sho¥vn in thesefigures, the induced deformation of ground due to tunnelexcavation is extended to the front pile in the case of''_'; *': :*"- i*'**..:;*;::;:,; ;: :i :;*:'i(. -'E_; ' - '-' _ _ '" -_"; i 'l ;:.il-' <:-, '****': ;:::i; ;,;'*' *;*";:;;'*** /'"'*'i*j:i: ""; i" ;***** f:;*f: l**':++?: ;D/B 3'O' DplBf;s* s*;f;i _"*' _i"**-i*'-' : 1* i' ;**' ' = ::':, "*"" * "' '.i-':lL{:.' L'; ';; :;i;';:j:+:; ii' i i; ';t:':':; :; *=* " *' ._It' ';; i: :* *.{;^ _.+ * * *..".:+**'"'*;i*':" ;+'*'*.," * " ': :il'l::, ; : {・::;i ' : ::***s -'"*'+'**t ',S,;: ; { "{ j{;i *"i ' +: "-*;: " : (;'+:-""" ;;!:"' ,._ i*-'i:';i{ { t;・ -;ii !': ; **t;., i *' *;*..*'^* ; ._:* +'* . _*0_ ** ; ' '_. ;{ ' .' i :: '*'f '+*i**-' 'i -' -'..' * *_ "' '- '_: '- '..1:;i '; ii-:; ;; : !,: ' "' ';i* ' i ; ;* ' *"+T'+*;j;::; ]*;'; ; ;*;"+* ;: ' ;'t]i-''; -' '-i*" _.*_ .*i'*" *'*'i'"""* _"'- - ...'" 5 * i;. ;";;;'_'; '_'_i' i t 'i; i__i: ;{ {ii'i:'- i; #+""' ;:_' ;f; -i:+*;*_**: ::]l:{ :*i'*'r'*'+; ;{';i'-ii* < * *{';i:;' sr " ;;;::;ir r + ;'i;"'_ ;f':::;if ; l:; ;. *'{;{: i;i'='; iiij'!i"-]ll'li:::;:.' :i- ; "'"' ;--"''*+*' " +"'". +='!'+*:*'*"*' ' i' '*' !<i- ' , T ;:;-''"':*;;*f *'"-'observed 9:round deformation of Fig:. 19.i i ::__ ii*=':. ;;( .__t'i 'f'**' "i'*exca¥*ation concentrate at the front pile for Dp/B= 1.0and to the rear pile for Dp/B='_.O, the same as the_ {::i: * *;"*._'.*r ;'*pile in the case of Dp/B= '_.O. Figure '_O sho¥vs the com-the group-pile foundation. The computed displacement¥'ectors and shear strain in the 2:round due to tunnel_' :i} ' ; , ;::;: ' :' i ;':{ {::i;i.:;*"':""-*' _' '*=._ ' -f +;:::'; <::i* :{:{:'<:: *' , ::;;;;;;i);j-;:'*']'":;')1i/:puted displacement vectors and shear strain contours for*i*, i'**='*-'_''+ " *;"' "-Dp/B= I O for D/B=2.0 and 3.0. Ho¥vever, the deformation is extended to the rear pile as ¥vell as to the fr'ont."-<""'=."'.';"i:i:;1:1:,':i' i :1+*':^ *+ ;; 1 '+*;1**; *+*,,*= ;i '){;* ';+-*;*+ '** i ;; '*)'; / '-'1:**;r;;; ;i ;**' -;:': " i';Figures 21 and '_?_ sho¥¥' the surface settlement profiles*'*- "'-*--'i:-_! {**" '!:'* t";'-s-* ""'-_"'*''. ; " : ' ' {:i:'(';:"i*'*'"_:";;;"' "t ;' : ' :';'; **-*'*'-' "*'" " t '*+{of the model tests and numerical analyses for the piledraft^ The shape of the settlement curves is almost thesame as the group-pile foundation. Ho¥¥'ever, theD/B1O, Dp/rig. 23Defornla{ion zone of modei tcst (Piled raft)l.G. ** J , j ,t'J tl,"' 'f $tlifF'r!rr,'I = tiiJi' r ,,t'Iif.'.',t' 'iil!"i +,* I '; * ' '' * f'!'= ;ii!! 'iff !!/'f/r'!i"i!'F f," "'jl'r'f'F'; i ,f!? ff !t !i:,1'!.,,= j , i/ " " r " '! ((":'* * " I'I ' '' !'J j f ' f" ! J,s$ '}=":s J i f i e"'!r,',.' '...*.*, i ,jl"f!' ..; I,! r t 't,!."l* j. ;1 j.: ,, ,'i< { j iD/BF3.0, DplB '_.OD/B=3.0, DplB=2.0t ; i .=ii F jjj'.,i;I 1{-i! IiJ""' iiI I J I , r! ifj ' , JSi ifiiJ Ji L ' r,ii L i 'J J ' i ' "r I i itii liI ' i $'J t ,iii/i'//i ":' - "t'L'iF"!'fr* ' * !i!;'iti* i " Il 'ii f"" j!!!'r.'1J'! "':/ " ! i i r 'JJ'i'ii ril"f ji; f!i""Jll!'i!',',!"" i / / ' i / i /}'}'i 'f/i+ ,!;p','!1i i ff' if " r/;'r//":!ItE'I"'!;-'r f' " Ft*,,, ,, .**,**='' +'F'*'!if'*! "'**i ti ! f ! 'Y'*fri'i"!i';f;.t " """ 't :: 'l! i, t' i,,, ,, t;!j!: ':.';, I ! ! $ ' i!i *,.,. fj.j-':Ilf 'f:j,FI! {' ii""j*'i "+! iiii"'*'I j I ' f "'} '!,,i{}W:i!1.. i!!"'? t' ()ao ooFi('. 24.a.02s*^*"O 04 O 06 O.08 a la(cm)O OO 006 012 O.i8 a .?_4o .30Computed (a:) displacement vectors and (b) sl]ear strain contours of ground (Piled raft) FTU*i¥+NELli¥'GVITH EXISTli¥'G FOUN DATION'lc). ().Q=3 ' 1 4N/cm- 'd= O OOnw J3LQ=3・ 14 +/cml- - - d=0 OOmm BLl OOf' O OOmsl JT ¥ D/B 2.0eJ)5h I OOi- 4 OO2 ・_ - 4 OQ(gr en fieidj4_!' ¥ _;t ;_ _ _ _ -f- rlt¥ihb't:Y' '** h" '- '-i -iYrt¥i' _- * s/'-',i '7j . _ _fy_*,i40/ii"" * (¥.-{ 20o180¥ ,l'?' *.¥_ D/B=3'O ! io0L//j ,X ,f:f_ LL¥*"'t・_・ _ _It,- ;-c140ir t_ ':'i :-.; .;':._..'IF , --1*Y_, -1 60cl/7^j **O 16 l, 8 -4 O()4 8 1'_ 16-1 20a observed l=_ : il/ ; __ __l]60l,-'-'・- ・, v''o? C rf '- r*'_*=! .'_#loooo140DIB=3.0 I l 20Dp/B=1 .O -j iOO, ;-- Pile=DIB=' O [ii:・ ',L.1-'''*+*****', **+;t **'_**.?**vl. _160!*v*..'0i401 --- -/_ i **r;:1 (1,*!:'* r r _j40: ,. a -__*+f" J/rl" '+ ¥,-,_._. j 20= (a)obser"ed ":i! 20f *i? 20t'qo-16 -12 -8.+-4 o 48 {2 16 Ox(cm)-* ¥eD/B=3'O -;r '--_1 oe-; I oo_'. 1:" 'x'.--*, ' - - - 160v -*- : ,,.*.o6cln1 SO-' 80_-:'f -' -.- , /' -'-''1 " ! 140eQ=3 , 1 4N/cm.. J! i.- :;*! 1 1;:)i :, I]20IObserved distriba tions of earth pressure (Flat foundation)l ' "'_.."-'J' '_It rl/' //]' -] SO "._+¥x(cm)LJJJJ, l: _";1'_,_i"T .1:-..-.-.- =j.!;T..,.. ,=,,D/B=3 O J 12GDp/B=2.0 _'2 i- pile=8Cm100qi "TI .,j-' 40/, ¥Ioo¥ 1 } 80i602i- !"2G:Y2trq=;, Dp/B= I .O--j'pile=8cm* ;.Ficr. 25.D/B=2.0 1 201 [L -green r eld40:. ol[--------i40{ - 40060r +**j,!, !;- )-- O OOmm AL-?rO 05'- I OO' ++80! -fi__._._,'80Jl i,.,,: c;/ ; :f.tJJLuJFig. 26.represent the values of applied displacement used to.!8 -4 ' OObserved distributtons of earth pressure (Group-pile)' ' :_,il'-':[ 60;I f *O 16 -12_"Fi('. 27.-14020(b)COl lputed4 8 l'_ 160x(cm)Computcd distributions of earth pressure (Fiat foundntion)simulate for tunnel excavation. The dotted curves lvithblack circular marks represent the earth pressures beforethe dead loads are applied in Fig. 25, ¥vhile the ¥vhitecircular represents the earth pressures after the dead loadsare applied. For D/B=7_.O and 3.0, irrespecti¥'e of theground depth, a significant amount of load transfer fromthe cro¥vn to each side is observed due to the developmentof ground arching (Murayama and lvlatsuoka, 1971;the rotation of the foundations is in anti-clock vise direc-tion for Dp/B= l.O since the f'ront pile is significantlyAdachi et al., 1994; Shahin et al., ,_004a). This archin_",_effect is more remarkable at the shoulder of tunnel ¥vhereaffected by the def'ormation of ground. On the otherthe dead load is applied. Asymmetry in the earth pressurehand, the rotation of the foundations in the case of DplBis observed in the model test ¥vith the dead load. The=2.0 is smaller than that of Dp/B='_.O since both theresults of the numerical analyses, as sholvn in Fig. '-6,front and rear piles are located in the large deformationagree ¥vell with the results of the model tests both in shapereg:ion.and quantity.Dist/'ibution oj Eartll Pressul'e allcl Axia! Fol'ces of Pilesof the model tests and numerical analyses for the flatfoundation due to tunnel exca¥'ation. The left verticalaxis represents earth pressures normalized ¥vith initialof the model tests and numerical analyses for the grouppile foundation. The shape of observed earth pressure isthe same as the results of' the flat foundation except forD/B = 2.0 and Dp/B = I .O. Although the ground depth ofthe flat foundation and the _ roup-pile foundation is theearth pressure, and the right vertical axis represents thesame as D/B=2.0, the arching efi:'ect in the side ofvalue of earth pressure. Legends ¥vlth different markslo¥vering block F is significantly de¥'eloped due to theFi**ures '_7 and 28 sho¥v the earth pressure distributionFigures '-5 and ,_6 show the earth pressure distributions lSUNG ET AL._・ooQ=3・4N/cm- d= O OOmm*.***l 40140V OI 054 .JOO3Q=3 ・ 1 4N/cm4D/B=2.0 -; 120Dp/B=1 .O j4 aogreen fieldl OO* pile=8cm lj73r- 80*',4l=a'tD/B=2,,O I] 120OO!'---- pil 8=clmO Ijl[__.__Dp= . :l400i,!'/ j" -' *. 'L*,rh .40r C.-.'_Oj'l_20o*,o:Ol 40bLt:;"'[Dn/B=2 OPile=8cm-lIOO[J_ 180n! ---・rrf?=y'*' '+* i _l -'-';lt:< c'*_'_ ' , [ h= 60aaao{)c1rl-16 *12 -8-Ol 40ff' D/B=3.0 -:JI 120Dp/B=1.0 JFig. 30.-'4-v i'xJ * j,"I712 16ox(cm}Computed distributions of earth pressure (Piled raft)='=*prp-= J80J'*"rh*= l60*"-ocf-*'*'*'-i40Lj= 20f* ^(b)computedo !_-16 -12 -8*4 o 4O8 12 167x(cm)Q=3 ・ i 4N/crni 4_-)J 1 40' - d {} oaTn n BLH OOChnrnALJl 1 70005D/B=2 O IDp/B=1.0JIOOJ_+ I OO400'[_ - green field,/¥..Pil¥8cml802f-:60.n- L L¥HJX l__ : r¥"':' :--*¥ * --"If140120O140*:, oDfB=3.0 120 :2rnri itl't¥L1/J・J-L-f!----------h- 'l*r-1 4 4; *"Ir'!liDp/B=2.0 1100pile=8cm JJ80rl ' '!::{r'/Jfcr' ( ! :/ ' i ¥/"_¥ o !'. J40J_______1,60'a observedo-16 *12 -8 -4 OI l4existence of the front pile of the group-pile foundation.Figures 29 and 30 sho¥v the earth pressure distributions ofthe model tests and numerical analyses for the piled raft.The earth pressure distributions for the piled raft aresimilar to the results of the group-pile foundation.Fi**ures 31 and 32 sho¥v the computed axial forces ofthe front and rear pile in the case of the piled raft for eachComputed distributions of earth pressure (Group-pile)Fig. 29-4 O 4 8Pile=i 6cm I I OO)Fig. 28.40OLJi:J60Cb)computed J20- '_Oo**J"_r*rJ^fvllv :' J i!J- 40+i!1 Dp/B=2.01 1 20!'rpr]e=8cm' iloo180 ; q2 [1 40qO1;'T. D/B=3.021D/B=3'O --jl20*h- 160{ . =40!s;* ff/' ? ,lI OO_ 80-;reenrleldJl 60*+.L* **C O-0052f,* fzjl'r *"*._ -2'>20excavation step. These axial forces of pile are calculatedfrom the vertical stress of pile multiplied by the sectionalarea of pile. The left vertical axis represents the grounddepth (z) normalized ¥vith the pile length (Lp)^ Jacobszet al. (2004) has sho¥vn that for the single pile, the axialforce at the pile head remains in the same ¥'alue as thedead load, ¥vhile the axial force at the pile tip decreasedaccording: to tunnel excavation in the iarg:e deformationregion. In the present analyses, the axial forces in the pileof the group-pile foundation and the piled raft areincreased or decreased at the pile head as ¥vell as at thepile tip according to the relative location of each pile andtunnel. In the case of D /B= 1.0, the axial forces of thefront pile are remarkably decreased at all parts of the pileby tunnel excavation and increased in the rear pile. Onthe other hand, the axial forces of the front and rearpile in the case of Dp/B=2.0 have not chahged much.Figure 33 indicates the percentage of the load supportedby piles at pile head and the load supported by the groundunder the plate to the total dead load of the piled raft ateach excavation step. The percentage of the head loadfor the front pile decreases due to tunnel excavation forODp/B= l.O. On the other hand, the contact load underxrcm)the plate increases in the piled raft. Ho¥vever, in the case8 12 16Observed distributions o earth pressure (Pi ed raft)of Dp/B=2.0, the percentages of the head load andcontact load are not significantly changed since thereduction of ground stress under the piled r'aft due to rTUNNELINC ¥¥,lTH EXISTING FOUNDATIONo piie hes ' ''fA 'J[o[- DIB: !-' oJDP/B: 1 c(From pile)i04r0=201pile head:O 2 L DfB=2 O!l[ Dp/B=1 O! l' (Rear pilej04[06Ll08Lll ! Pne tiPl Pile npo jPile hea ' IiO pile head'or DP/B:D/13:;3 !o ll2'o(Front pile) l_O 4 LF:;[ f: L,N06 j ! !08[- Ie[,L)/B=3.002i.-." d=0 oo lmDplB=2 O(Rear pile)[0608-2eorl ;Tr]Piielp3ocl I Pue lp ' _ J 4- oo=- 400+., *O pile head' 'oPileheac o[I)/B30204-04-i i'B.c6[06-'.*=lf DfB=3 ODplB=1pile)O[= (RFear02r r)p/13;;1 oL (Frontpue)[l I plle tlp I { l.il.'*jP =08Lfl I Pile np [ocFig. 31. Cornputed axial forces of pile according to excavation (Frontpile)tunnel excavation is uniforrn. ¥Ve applied the same deadload (3. 14 N/m) as the existing building load in every test2Fig. 32.Computed axiai forces of pile according to excavation (Rearpile)these figures ¥¥'ith solid line. The settlement trough rsasymmetric about tunnel axis, and the amount of settle-to investigate the influence of foundation types. This loadis roughly one third of the bear'ing capacity of the fiatment is larger than the results of the green field conditionfoundation, so it is rather smaller than the bearin*'settlement trough is also smaller than those of the greenin the same ¥vay as the 2D results. The width of thecapacity of' the piled raft. As a result, the initial percent-field since the surface settlement concentrates at theage of load supported by the raft is smaller than that to bevicinity of the fiat foundation. After the excavation frontusually known.passes the flat foundation, surface settlement increasesRESUL1'S AND DISCUSSIONS IN 3D CONDITIONIlAxial forces of'pile (N)Axial forces ofpile CN)much rnore si**nificantly when compared to the *"reen fielddue to the influence of the dead load. The rotation of theflat foundation f'or D/B= 2.0 and 3.0 occurs as the sameSurf'ace Sett!enzent and Rotation of FootingFigures 34 and 35 sho¥v the profiles of surface settlements of the model tests and numerical analyses for thesequential excavation in the case of D/B = 2.0 and 3 .O fortrend in the fiat foundation of the '-D tests. Thethe flat foundation. In these figures, Iegends ¥vith differ-results. The elastoplastic joint elements are not im-ent marks represent the position of the excavation frontplemented in 3D pro*・ram used in the analysis. Therefore,in the tunnel excavation direction, ¥vhere the value of zerorepresents the excavation f'ront reaches the center of thethe computed settlement of the foundation is smallerthan the observed ones, because the slippage bet¥veenfoundation and ground ¥vas not considered.Figures 36 and 37 sho v the profiles of the surfaceflat foundation. The surface settlement at the end ofexcavation for the green field condition is also plotted incomputed trough of surface settlements is the same shapeas the obser¥'ed one shown in Fig. 34. The magnitude ofsurface settlement, however', is smaller than the observed SUNG ET202AL_1 OOQ='_9・4JN80O60O. 140.2O.3*"_ O.4':0.51o 100I:s6e);4"- -20cmH O4{h 28D/13 =7_ . O- greenfileld(2 S )o**o 80'-Excavation rrontOO.1O,?_20.3D/B=3.00,40.5es:s: 100S-20 -16 -12 -8,-4 o 448 l'_7:6 ,_Ox(cm)Computed profi!es of surface settlement (Fiat foundation)Fia. 35.Q=29 4:)Nd (!11 Tl)Orig. 33" Li*'-/:,,:L S*i::___._:i;::.S fL _ _F/:O. lChange of supporting ratio of piled raft due to tunnelO.2 Excevat{onFrontexcavation agninst imposed displscementO.3- -20cm- r '4D/B=2.0O . 4 [ Ii: oj'0.30.4)0.5-+^ :_ _t = *,*=. -・'/i"-0.1O.?__* ' P L _i_t-,1Excavation Frout- -2acm l=-- *4:* .+ '111--- )S4- O D/B:ss2.00.11}-.--'i 7- - ? ;' Ft:;'1A-l !'' "0.20.30.4O.5L;( __::"-/, ' _: '_ '* "'i:..c '*. ; ;',* * : *Jo.o.2o.3o.4O 0.5. J-(irD/B ='.sOI , I DplB=2.0i ,t_ j -i_;T___ ": . I_t:TTD/B=3.0; x(cm)i i 4*_*'*+'*'_..;1 ,i * i# ,:O. lO.2O.30.4D/B=3 . OD p/B= I . o05-20-'_O -16 -12 -8 -4 O 4 8 1'_ 16 _'OFi('. .34^een fleld(2s )O; '__- -' LTT!i-Vo, i,:Teenfield(2S)t)-oeJl)Ol ,D /B, l'O.5-1--= 18Q=29.43NOT' .**:'; ;i *+1 'Fio. 36.-16 - 2-8*4 o 4{ ll8 12 16 ,_OObserved profiles of surface sc tle ,]cnx(cm)(Piled r2ft)Observecl profiles of surface settlen]ent (Flat foundation)settlements of the model tests and numer'ical analyses forthe sequential exca¥'ation in the case of D/B = ,_.O and 3.0for the piled raft. The shape of the surface settlementtrough in the final exca¥'ation step is similar to the 7-Dresults of the piled raft. In the case of Dp/B= 1.0, therotation of the plate is to¥varcis the excavation ¥vhentunneling is finished. On the other hand, in the case ofDp/B='_.O, the rotation of the plate is in the oppositedirecrion to excavation as the result of the piled raft in 2Dmodel testFigure 38 indicates theface settlement* vhen thecenter of g:round for theposition of the foundationcomputed contour's of the sure¥.'cavation fr'ont reaches theg:reen field condition and thefor the fiat foundation and thepiled raft. Figure 39 sho¥vs the computed contours ofsurface settlement for the excavation front reaches atthe end of exca¥'ation. The foundation outline is alsoindicated in these figures. In the green field condition, thesettlements at the end of tunnel excavation are symmetricalong the tunnel axis. In the cases of the flat foundationand the piled raft, the settlement concentrates around thefoundation, and the settlement trough is asymmetricalong the tunnel axis. Especially, before the tunnel frontreaches the foundation, the surface settlement troughforms asymmetrical in shape, and the settlement concentrates around the foundation.Figures 40 and 41 indicate the rotation an_._"le of the flatfoundation and the piled raft for the 3D model tests andnumerical analyses of sequential excavation in the caseof D/B= '_.O for the flat foundation and D/B= ,_.O and 『203TUNNEUNG WITH EXISTING FOUND、へTION墨Q軍29・43N・即2、器 00.10、20.30.40.5εxoa、’aび0口Froa【くト ロユぽじみ DIB;2.O‡言芯西2141[EDp町ILg一   9鷲9耶恥麺‘2sナo﹁⋮1レ01箏 0卜25(cm)1毒蔓0・1糞o・2.0(a)Green負eld(Z)昭讐3。0)至0、3葛0.4DIB=3.OD辞/3撚2・o傘 0IO5㏄0.5心0.19095.G                『0.20.3DIB=3.0 コ。3gl        EPP/3篇玉・P0.4^ 0199綴0.58 12 16 20一20−16一里2 −8  −4  0  4             響f「io 37. l      x(cm)CompuIed prof圭les of surface se吐盤eme蝋(P難ed rafI)(cm)(b)Flat fo昼ndat量on(1)昭茸3.0)        論Direαi。n。f         ltu照elexcavation}         Iou烈dation outlinまa.765’o麟.e6鞠301δ2ヨβ哨轡昏ξ了(cm)(c)Pi豆ed raR(1)昭=3.0,五)〆B謀LO)(a)Green field(Z)昭嵩3.0)Fi9.39.Pl9龍eviewofcomputedconこoursandvぬesofsurface傘   seIdemen亘(when excava甑on童s簸nished)Z)P/β繍歪。O for tke pi正ed raft. Posit重ve va重ue of therotatiou ang正e about tunnel axis (Y) represents therotation tow&rd the tunnelεしs shown重n Fig。40.Positivevalue of the rotation angle &bout lateral axis (X)represents the rotation hl the opPosite d圭rection of tunnela(董vance. Before tunnel excav段tion advances to  t圭lepos圭tioll of the foundation, the rotation of ξhe plate(b)Flatfoundation(.0/8漏3.0)about X−axis beglns toward the excavation front,andwhen the excavation frollt passes the posi重ion of thefoundation,the foundation returns to its illitiεd state.On the o重her hand,when the excavation front 星ies inη.一〇between about−10cm to10cm,from the center of騨0,228foundatio11, the rotation of the plate abouξ Y−axis‘gr&dually increases.Aftel’then the lncrease of rotationoabout Y−axis ceases and mainta圭ns its state u簸ti星tunnelexcavation is nnished.The amouPt of rotation for thepiledraftissmallerthanthoseoaheHatfoundationdue一〇(cm)(c)Piledra負(D昭嵩3.0,Dβ皿LO)to the sti飾ess of the pile.Tlle computed results describethe tendency that the rotations 段bout Y ax圭s in the flatfoundation are larger thεしn those in the p圭led raft,thoughFig.38. Plgne v量ew of compu[ed con重ours 践nd v&lues of surface由ecomputedrotationsaboutX段xlsinbothfoundations   se載demen叢(whenexc貸、『aIionfromreaches旦tfound蹴ion)are underestimated, SUNG ET204AL.Q=29.43Nl .5[ := ',llilfield resut :JDIB 'Or[ 1 - rl _] /f ri __: y80Jt60fo.5leQO O- (AL)O *"hF7:lc 1 520 ,* - initial(BL)Lowerin : BlocD/B:ss'3 . OeoO.540-r*Oo::1'120--J1_ GJIlOO!gol[- - -]l ---O 5oo-O.)_20 -16-1 2!-t'rl:[ II nl /i- I60-1 40[ 201-4 O 4 1 2 1 6 20o8I-8 -4 O 4 8 12 16 y(cm)20--,l; x(em)Position of excavation frontObserved distributtons of earth pressure (Flat foundation)Fier. 42.Observed rotation an"*le of flat foundation and piled raftFig. 40.Q='_9'43Nl .57-D/B=2.0green fie}d resul---FJl/nitiai(AL,)Lolverin B}oc/=E-{HJG.boeoOeD/B=3 . O,o60401o.5:::8020O,q,*eei 20 1;$$1 OOeo**_el -0.5::: 1.580601:40120O.5oO8oO)20 16 -12 -8 -4 O 4 8 12 16 20Position ofexcavation fronty(cm)126ul T x. '(cm)Fio. 43.Computed distributions of earth pressure (Flat foundation)Frg 41. Computed rotation an"*le of ffat foundation and pileti raftDistribution of Eartll Pressu/ e Considel'ing Sequentia!ExcavationFigures 42 and 43 sho¥v the obser¥'ed and calculateddistributions of earth pressure along the line that isperpendicular to the tunnel axis of the model test, assho¥vn in Fig. 5(a), in the case of flat foundation. In theseflgures, the results of block F and block .J in green fieldcondition are also indicated to check the influence of thedead load. As sho vn in these figures, Ivhen block E. islo¥vered, earth pressure above block F is increased sincethe arching of 3D sequential excavation is developed inthe lon9:itudinal direction as ¥vell as in the transversedirection of the tunnel axis. When block F is lo¥vered inthe case of D/B= 2.0, the values of earth pressure abovelo¥vering block F are similar to the results of the greenfield condition. O_ n the other hand, the amounts of earthpressure at the side, ¥vhere the flat foundation is located,are lar*'er than those of the green field due to the deadload. After excavation front passes block F, the earthpressure in the transverse dir'ection of the tunnel axisincreases since the longitudinal arching is disappeared.On the other hand, in the case of D/B=3.0, the earthpressure distributions are similar to the results of the_ *reen field condition since the influence of the dead load FTUNNELiNG ¥¥*ITH EXISTING FOUN. 'DATION205,Q=29.43N" /B=2.02[_Dl ! n1 --- ; i:(AL)Lowerin Blockf[ :fiJ 40N{ 80 q,.b*'10D/B=3 . O12r Dp/B=1 O' //1:¥,n t: A/*o[ 8;;sF60l 40' E Ir :[ :lo-4 O 4 1 _7 1 6 20is less significant due to the far distance betlveen thetunnel and foundation. Computed r'esults in Fig. 43 sho vthe same trend and quantity as the observed results.Figures 44 and 45 sholv the change of earth pressure forthe piled raft. In the case of D/B= '_.O and Dp/B= 1.0,earth pressures ar'e similar to the results of the fiatfoundation. Earth pressures in the case of D/B = 3.0 andDp/B=2.0 are also similar to the fiat foundation ofD/B=3.0. However, in the case of D/B=3.0 andopoints from the tunnel axis are ,5.5 cm (Pl) and 7.5 cm(P2), respectively. The left vertical axis represents earthpressure normalized ¥vith initial earth pressure. Abscissain the figure represents the location of the excavationfr'ont from the center of the flat foundation. In theseO8IIPig. 457 x(cm}Computed distributions of earth pressure (Piled raft): 70i 60l_8 - DIB=2_Ol.6. f_14.l_'_ Ll *-- 'O,8 LO 6* 2l 8 ! D/B=3,0! 40・ "・ 30*From Qlnnel a:xis)pl 5 5cm- -p2,75cm-Pl,S Scm(_ reenj :eld)- - - P2,7 jcm(grcen e d)14and the piled raft (Fig. 5(b)). The lateral distances of the20-4 O 4 1 ' 1 6 20Dp/B= '_.O due to the transfer of the dead load throughat t¥vo outside points located below the fiat foundation40E l1,6Figures 46 to 49 indicate the histories of earth pressure/ -/lDp/B=1.0, earth pressure increases more compare tothe piles.80E /{IIIl l x(cm)44. Observed distributions of earth pressure (Piled raft)lOO6020JL120D p f ;= I .O18F'ig.OD/B=3.02' roo- ,.*40o120n1r20180f- crll 60-lJ 20*'. ;l looJrt ' //11_ _ _iJ 120100l80*1[l f' *60 eelecH "DIB=3 .ODp/ =2 .O,20I-4-EcoOFF 1 120-G{hiinitial(ALLowering Bloo/lf D/B =3 . OO6040/-4'-iniuai(BL) J 202 = IBDp =_.O[: l[_ : /---FJ------LI [---i--80green field resuli , -l I[Dp/B= I .O* 1*f 6040E[I - .lr-[- l -I /Ol iD/B=2.02l pD IIB=_O[r --I,Q=29.43N____ _r_ee_nJF eld resui I gol OO*・ *"9080701 2l=O 8 i. J-601 50- 40JO.624 ,O 16 12-8 -4 O 4 8 l' 16 20 '4Position of'excavation front y(cm)Fig. 46. Observed histories of eartt] pressure at two points, P& P2,s lown in F'ig, 5(b) (Flat fountlation)figures, the results of the green field are also indicated tocheck the influence of the foundation. The arching effectfor the 3D green field condition has already been discussed in a previous paper (Shahin et al., 2004b). As itcan be seen from these figures, earth pressure increasesmore for the foundation compare to the results of thegreen field condltion. In the case of D/B=2.0, earthpressure at both points si**nificantly increases due to thedead load. In the case of' D/B = 3.0 and DplB = I .O, earthpressures for the flat foundation and the piled raft arealso larger than the results of the green field condition.In case of D/B= 3.0 and Dp/B= '_.O, earth pressure for −206SUNG ET.へL.  2  1.4601.41。2 1     一一一の一噂一一一r昌40   至這・2lFr。mI㎜el砥ls 30,90.6 2           I           I rom艶nnel axis30   I     l  l  I 一びP1,5.5cm         I  p1、55cml9鯉面eld1,8DIB 3.0   :榊榊鷲・75cm(g鱒側1.6  1.4 1            70 1         ___一一一一 6090           }80_         1L   i醤       喝㍉て需需r闘”げ  ’           5070,9 簿60唱r二__一_一  韓50ε 一 一 一 一 榊 嘗、  夏    心 飛             冊  需−1   鵡i OO DpB=2・O  l1.2L1,20、8401 「 l     l  l ∫      貿甲1−PL55醜τeen㈱IOO讐:謄”P2ラ75cm(grε謡elq − 1            80董一一騨一一一一一轍 −0−P1.55cm     簿 1            901.450_ ___一一一一一一一一 鞘やD/B=3.016」                       10.8      四一              〇     アヲロ      ヨ じ     あ  1、8           E 1嚇イ』}  一一一一一一一一一一一一一一曽→ 0.8曝0,660 Dp/B=LO睾撃    、        璃  1.2 DIB繍2.0     [正.6             」5070  l  l  I’I  l     I』l  I     l1.8DIB=2,0  1,6 270  l  l  1.8           も欄  0,8 1            4040           I0.60.6一24轍20麟16−12−8 −4  0  4  8 葉2 16 20 24 2      Posltlon ofexcavation倉ont  y(cm)正.8隻.6Fig.47。CompuIedhisloriesofear{hpressurea“wopoin重s,pl&P2,   showninFig。5(b)(Fi禽げoundalio鑓)   I  I I l ! I I    1100DIB鷹3,0工)p/B識LO1.4      90笹尋8・701、260 1  ___一一一一    蠕篇て一一   π曽鴨齢  ”一榊鴨冊曹欝}需智騰 270i.81.6rDp/B篇LO  I一一一一                      一 ,1.2           :,一一“一一一一一i0「80.6紅綱轍一モー一} 一__L〆  _一一一一一一一一‘         蜘榊一ヂ           『           『            Frommnnel axis1一601、Blガ筐壁幽1840      Pos量tion ofexcavatlon fヤont  y(cm)Fig。49。ComputedhisIorlesofear症hpressure段1Iwopoi臓重s.P丘&Pユ・   shown i罰Fig。5(b)(Pi韮ed田f{)100Dp/B竺2.0   :901.4          :          :80          1          6          1                 )70          1            岬榊_瀦  ,960簿  一 、壱 1    、、騨一、  じ需二騰”………『ゴ  奏50℃0.8             ノ         、          ㌧    ε40           箪0.6   i  l  l  l  l      l  l  l      l 2were carried out to investigate the interaction problembetween exist宝n黛foulldation and ground due to shallowtunnel excavation.The 宝n且uences of foundation type,plle Iength and overburden on settlement and e&rth pres−sure of the ground were investigated.From the results ofthemodeltestsalldnumerlcalanalyses,thefollowil19poillts can be concluded:  l    E i l l I i      i王.8            E(1) D/B−3。0   1     1 1001、』…潮ンー0.8The m段ximum surface settlement due to tunnel ex−cav&tion hl the case of existing foundation is Iarger                      」J gO腿o.、1蜘5040301.6§,i28           10,6           11  1111一24−20−16卿12−8 −4  0  4  8 董2 16 20 24501←P15.5cm 2             、、占口’0,D/B=20 L}/b_.U           Ithan those in the case of黛reell負eld.丁董le玉ocation of80the maximum surface settlement moves towar(i the70posit圭on ofthe foundation.The width ofthe surface60settlement trou黛h is smaller th&n those of green盒elddue to the co熱centration of surface settlementaroun(ifQundatlon,50400、6(2)一24−20−16−12−8 −4  0  4  8 12 16 20 24The deformation of ground in the c&se of existingpile foundat呈oll depends on the distance beτween      Posltion ofexcavation f}ont  y(cm)tuanel and p至le tip,and the dept封of ground as weILln the situαt茎on w}1ere o鍛茎y the front pile locates inFig,48.Observedhis重oriesofe韻hpress腿rea{twopol猷s、pl&P2,the large deformatlon reglon,the deformation   s恥own in Fig。5(b)(Pi嚢e{量ra貴)extencls to the aτea of the front pile o∫the group−pilefoundation and the piled raft.On the other hand,when all piles玉ocate玉n t}le large deformation reg玉oa,the aat foundation and piled raft are almost same as thethe deformation of ground exten(is to all P量les,andresults of the green負eld.the  magnitude  of  the  differential sett璽ementdlecreases.C(:)NCLUSI()NSで〉lode茎tests alld elastoplastic 燕鷺圭te element analyses(3)In2D cond三tion,the rotation angle of flat founda−t量on量s smal重er thall those of the group−pile founda一 r,TUNNELIN'G ¥vITH EXISTIN Gobser¥*ed. Ho¥ve ・*er, zhe results of' the '-D conditionf'oundation lies in the lar_"..e deformation region. Theare qualitatively the same as those of the 3D condl-rotation of the group-pile f'oundation and piled raft,ho¥ve¥'er, ¥'aried ¥vith the position of the front and(4)(5)rear pile.The arching in the shoulder of tunnel developed in aprocess are taken into consideration, is a po¥verful toolmuch larg:er area due to the dead load exerted onfor the prediction of ground mo¥'ements and earthexisting foundation than those of green field.The axial foFces of' piles increase or decrease accord-pressure in tunneling,of foundation is located in the laFg:e deformationregion, the axial forces of the f'ront pile decreasesremarkably, ¥vhereas the axial forces of the rear pileincreases. On the other hand, if both front and rearpiles locate in the large deformation region, the axialforces of both piles may be almost the same.Rotation occurs not only in the longitudinal direction (Y-axis) of the tunnel but also in the transversedirection (X-axis) under 3D condition. The rotationabout X-axis ¥'anishes ¥vhen tunnel exca¥'ation iscompleted. Therefore, 3D analysis is required forproper prediction of the rotation of' foundation.(7)(8)Arching: is f'ormed in both trans¥'erse and long:itudinal direction of the tunnel axis due to tunnel excavation, e¥*en in the ground nearb.v existing foundation,¥vhich emphasizes the necessity of the 3D analysis inpredicting earth pressure of the ground precisely.It is re¥*ealed that joint elements betlveen the pilesand ground play an important r'ole on surfacesettlement durin tunnel excavation ¥vith existin9:foundation. In the present study, the 3D numericalanalyses underestimates surf'ace settlements lvherejoint elements ¥vere not consldered in bet¥veen thepiles and ground. Therefore, joint elements shouldproperly be considered in 3D numerical analyses forproper prediction of surface settlement.(9)'T=he g:round mo¥'ement is ¥'isualized in the '_D modeltests, f'rom ¥vhich the mechanism of the grounddef'ormation due to tunnel exca¥'ation can bedemonstrated. Ho¥vever, the construction sequencesare not employed in the 2D model tests ¥vhichemphasize the necessity of 3D analysis. It is alsoimportant to use the same constitutive model both in2D and 3D anal_vses. In the present research, 3Deft ct in surface settlement and earth pressure isi;!tion.Finite element analysis, in vhich elastoplastic beha¥'iorof' soil, initial stress condition in ground and constructioning to the location of each pile. If' only the front pile(6)POUNDATION 2a7tion and piled raft since the location of the flatACKNO ¥rLEDGEMENTS'The authors are very much grateful to Takashi Sadaand Yusuke Tabata of Nagoya Institute of Technology,for their contributions in the model tests. The authors¥vould like to thank Prof. Feng Zhang of NagoyaInstitute of Technology and Prof. Mai'cio Muniz deFarias of' Uni¥'ersity of Brasilia f'or their ¥'aluablesuggestions and comments regarding this research.REFERENCES) Adachi, T.. TamuFa. T., Kimura,,1. and Aramaki, S_ (1994): Ear hpressure dislribution in trap door tests. Proc29!11 Jl;n ,Vfu. Conf.Sj fFF, 3, 1989-1992 (in Japanese).2) Jacobsz, S. V,, Standing. J R , lair. R. J , Hagi¥vara, T. andSugiiama. T, (2004): Centrifuge modeling of tunneling near drivenpiles, Soi!s rlnc! Fbtuic!ations, 44(1), 49-563) Loganathan, N.. Poulos, H_ G. and Xu. KJ. (2001): Graund andpile-group responses due to tunneling. Soi!s clnc! Fbluic!(7rions, 41( ),57674)lurayama, S. and h,Iatsuoka. H, (1971): Harth pressure on Lunnelsir saFld} ground, Proc. JSCE, (lS7), 95-108 (in Japanese).5) Nakai, T. (1985): Finite element computarions for active and passiveeanh pressure problems of retainl g wall, Soils clncl Founc!(1!ions,25(3), 98-1 i2.6) Nakai. T. and Hinokio, _ (2004): A simple elas oplastic modeltor normaily and overconsolidated soils ¥vith unified materi llparameters, Soi!s anc! FounrJations, 44(2), 53-70.7) Nakai, T.. Xu, L. and Yamazaki. H (1997): 3D and 2D model tes sand numerical analyses of set lements and earlh pressure dLle totunnel excavation. Soi!s anc! Fb ulda!ions, 37(3), 3 l428) Shahin, H. i . Nakai, T., Hinokio, i., Kurimoto, T and Sada, T_(2004a): Influence of surface ioads and construction sequence onground response due to tunneling, Soi!s anc! Founc]a!ions, 44(2),7 1 849) Shahin, H. N'I., .Nakai. T.. Hinokio. 'I. and Yamaguchi, 'I.(2004b): 3D effects on earth pressure and displacemems duringtum el e¥'cavarions. Soi!s anc! Fbunc!a!ions, 44(5), 37-50
  • ログイン
  • タイトル
  • Main Factors Governing Residual Effective Stress for Cohesive Soils Sampled by Tube Sampling
  • 著者
  • Hiroyuki Tanaka・Masanori Tanaka
  • 出版
  • soils and Foundations
  • ページ
  • 209〜219
  • 発行
  • 2006/04/15
  • 文書ID
  • 20900
  • 内容
  • SOILS AND FOUNDATIO 'SVol 46,No1)209-2 1 9,Apr. 2006Japanese Geoiechnical Societ}MAIN FACTORS GOVERNING RESIDUAL EFFECTIVE STRESS FORCOHESIVE SOILS SAMPLED BY TUBE SAMPLINGH ROYUKI TAN. 'AKAi) and MASAN_ ORI TA¥_ ,AKAii)ABSTRACTThe residual effective stress (p ) ¥vas measured for various clayey soils collected from ¥'arious parts of thevorld,includin*' Japan. All samples studied in this paper vere retrieved by the sarne sampling method, i.e., using theJapanese standard sampler. Ho¥vever, measured pf/(T(.., ¥vhere (7'(, is the in situ effective overburden pressure,considerably ¥'aried for different sites as ¥veli as ¥vith depth. This paper examines main factors go¥'erning the pf ¥'alue,focusing on location of the sample in the sampling tube; transportation of' the soil sarnples; time duration bet¥veenretrieval of the sample and extrusion of the sample f'rom the samplin*' tube; overconsolidatiou ratio (OCR); claycontent and plasticity index (Ip). In addition, the p values are correlated to the volume change generated ¥vhen the insitu a(.. is applied in the oedometer test, vhich is extensively used for assessrnent of the sarnple quality. The largestmeasured value of pf ¥vas f'ound at one third of the sample length from the cutting edge of the sampling tube. Theeft cts of the transportation and the time duration from the sampling to the extrusion of the sample are not prominentfor the pf value. Any clear relations bet¥veen pf and ie/e* are not found, where Ae and e. are the void ratio changecaused by applying (7(.. and the initial ¥'oid ratio, respectively. Among f'actors examined in this paper, OCR is the most{effective factor: i.e. , as OC*R increases, p /(r(.. ratio increases for every studied site. Ho¥ve¥'er, ¥vhen compared at differ-ent sites, the pf/(7(,. ratio at the same OC R is considerably different. In spite of some exceptions, thel'e exists a tendencythat p lcr(,. ratio increases lvith the increase in the clay content as**vell as ll"Key lvords: clay content, cohesive soils, disturbance, overconsolidation ratio, plasticity index, residual effective stress,sample quality, sampling (IGC: C6)!conditions, ¥vhich is sometimes called "perfect sam-(for example, Andersen and Kolstad, 1979; Lunne et al.,1997). Furthermore, in Japan, since the undrained shearstrength for the stability analysis has been traditionallymeasured by the unconfined compression test, the valueof pf remaining in the specimen is quite important toevaluate the test results, i.e., unconfined cornpressionplin*"', this ne*'ati¥'e pressure may be equal to the meanstrength (q**).confining stress, assuming that the material behavesFactors governing the pf value have long been studiedby many researchers (for example, Okumura, 1971), andit has been identified that the sampling method has aIINTRODUCTIONWhen clayey soil is retrieved from a certain depth, partof the in situ confining effective stresses remains in theform of negative pressure in the sampled soil. In idealielastically (p , = ((T,',. +2(rl'**)/3, ¥vhere a(.* and crf,* are thein situ vertical and horizontal effective stresses, respec-Iftively). Ho¥vever, the actually measured ¥'alue of thenegative pressure (hereafter, it is called "ResidualEffective Str'ess (pf)") is some¥vhat smaller than thep , due to the release of the confining stresses andquite an important influence on the measured ¥'alue of pfdisturbances caused in the handling processes such as soilsampling, extrusion of the soil sample from the samplingtube, and the prepar'ation of a specirnen for' the measurement of pf, including storage of' soil sample. It has long(Tanaka, '-OOO). As can be seen, the amount of pf forsamples collected by the Shelby tube and ELEIOO samplers is only half of that from the Japanese standardsampler (JPN in the figure) or the Sherbrooke sampler.been thought that the amount of pf is one of the goodEven though every geotechnical engineer no ¥'adaysrecognizes the irnportance of sampling for deriving(Hight et al., 199'_; Tanaka, 2000). Figure I sho vs diff rence in the pf values of samples collected by six differenttypes of samplers that are extensively used in the ¥vorldindicators for evaluatin*' the sample quality, in additiongeotechnical parameters, ther'e are no sampling rnethodsas an international standard. C onsequently, dift rentsarnplin_ : methods have brought difficulties in comparingto other indicators such as volumetric strain caused by}the recompression process ¥vhere the specimen isreconsolidated under the in situ efi cti¥'e stress conditionsIi*}*IAssociated Professor, Hokkaido University, iapan (tanaka@eng,hokudai.ac jp) (formerly Port and Airport Research Institute)Port and Airport Research Institu e, Japan.The manuscript for thls paper ¥vas received for review on ,Iarch 25, 2005; approved on January 13, 20061¥rritten discussions on his paper should be submitted before November l, 2006 to the Japanese Geolechnical Society, 4-3S-2, Sen_a*oku,Bunkyo-ku, Tokyo I 12-001 l, Japan. Upon request the closing da e may be ex ended one month.'{;209{i +-H. TANAKA210ANDthe soil par'ameters including pf value for soils of differentcountries or reg:ions.A geotechnical group at Port and Airport ResearchInstitute (PARI) (formerly Port and Harbour ResearchInstitute, PHRI) has carried out the site investigationusing the Japanese Standard sampler at various sites inthe world, including .Japan. They have revealed that the¥'alue of pf is considerably different at different sites, eventhough the soil samples ¥ver'e retrieved using the samesamplin_method. In this paper, based on these ex-perimental results, main factors governing the pf value ofcohesive soilvill be discussed.oe JpN>( She I bycb¥ ' NG 1 54n_ ri e A ELEIOO5xe¥ e(> Sherbrooke.LAVAL'>- ¥¥eE ;cc), '¥+JQoe e'.., c)c::LSOIL SAMPLF.S AND TF,STING METHODSan7p!es Used in Tl7is Stuc/yNaturally Deposited SoilsA11 naturally deposited soils ¥vere retrieved by theJapanese standard sampler, ¥vhich is a thin ¥vall fixedpiston sampler. CJeometry of the sampling tube is 75 mmin inside diameter, 1.5 mm in ¥vall thickness, 6 degree intapered angle and 1,000 mm in length (length of the soilsample is 800 mm) (JGS1221'_003). All samplings ¥verecarried out under the super¥'ision of the authors. Exceptfor tests noted in this paper', all the soil samples ¥veretransported to PARI's laboratory by private transportservices. The sampling tubes containing samples ¥verestored in a wood box and ¥vere rapped ¥vith rubber foam.Main geotechnical pr'operties of the soil samples usedin this study are indicated in Table l. For more detailedproperties including their soil profiles, readers may referto literatures listed in Table 1.reconstituted conditions. T¥vo kinds of soils ¥vere usedo' L ¥.¥¥, 16al15TANAKAReconstituted SamplesTo identify factors ¥vhich affect the p value, somenatural and artificial samples ¥vere prepared in the-'>* ¥_c 10*i.for studying the influence of OCR: Honmoku andSingapore clays. Their properties are sho¥vn in Table 2.'¥ voThese clays ¥vere thoroughly remolded by a mixer at aAriake Clay¥..¥vater content of about '_ times of liquid limit (lvL).To study the influence of ¥'ariation of clay content and10Residual Effective Stress, p' , (kPa)lp on the pf value, artificially mix'ed soils ¥vere also used,¥vhose properties are indicated in Table 3. The mixingratio in the table is defined by the dry ¥veight. T¥vo kindsrig. l. Residual effective stress for samples retrieved b .・ varioussarnples (after Tanaka. 2000)of mixing materials ¥vere prepared: Toyoura sand andDiatomite soil. Toyoura sand is ¥vell-knolvn sandmaterial in Japan, ¥vhich is extensively used by thelable 1.Geoteehnica] properties of sampled soilsSite Country Depth (m) LL(' ) pL ( ) = IpAFiake IJapan- I S 55 1 78 276_? _)8- I 1 6Bothkennar UK3-16 55-7722-32 3245lLoviseville Canada698 212346-5766-8222-24Sm"apore S n"apore 628 42-579- 1 9Itn (o!/o) CF* (o/b) sL = = = (kPa) OC_R= = * References42-200 60-70 8-40 1 O 1 6 Tanaka H ()OOO)5168 4050 2060 1 .7-3 6 Tanaka, H. ( OOO)64-76 )55-70) 3 845-66Tanaka, H. et al. (200lc)6- 78090 : l.ll 6 Tanaka, H. et al^ ( OOla)50-60 17-904181 50-70 2040 .1- 7 Tanaka. H et al (200 a)30-43 50-70 20-40 1 :) ) ) Tar aka H ()OOO)Yamashita .Japan 20-38 92-126 3855 5477j4 100 6070 130160 1 / ) 7 Tanaka H el al ()OOlc)Ban rkok Thaiiand 6-1 7 46-101 19 '7 41735l9 1 34-48Drammen Nor¥vay8-225473PusarKorea6--)2 22--7630-47CF Cla¥ Fra'mem (<)tm) s**' *lea:sured by Vane, OCR>Table 2.NameSand ('!.・ )46-65 50-70 22-38 1 0-3 ,p* is measured by CRS tesTanaka. H¥vith a strain rate of 3 3 xOet al. (200lb)s 1Geotechnical properties of reconstituted samples for studying the effect of OCRSilt>5 tun ( ・b)C_lay< 5 Ltm (e., )It'l (O. )tt'P (')' )I PHonmoku444^50.5ll83979Singapore22s . 870.08531*4 rRESIDUAL EFFECTI¥,E STRESS FOR COHESI¥,E SOILS'Table 3.Geotechnical properties of aFtif'lcta:1 nlixture for studying fhe effect of cla) conteni and plastlcit) indexNameKaolinKaolinBan kokBang:kok211Sand ( + )Silt > 5m ( ,t))Cla¥'< 5 Lun (o, )T"[ (qo)}t'P ( ・ )I.SandK:S = 100:Oo208069353475:25251516054302450:50so10403823l, 5:757551._oO: I OOl OOooDiatomiteK:D = 100:Oo208069353475:25o386283483550:50o5446lOl683325:75o63371 128824O: I OOl772・2177892663SandB:S = ICO:O*)-75:2526165863204350:5051103945162925:7575)1929o: i ool OOooB:D = 100:O22177266375:2533364426750:50,-4454Diatomite25:75l5346O:1COl772,89l 09geotechnical researchers to study behavior of granularmaterial. Their properties can be easily found in manypressure of 98 kPa in an acryl cylinder ha¥'ing 8 cm innerliteratures. On the other hand, the diatornite soil may becylinder, the silicon grease ¥vas applied inside the wall of"rather an unfamiliar material for rnost r'esearchers.In this study, the diatomite soil vas collected fromHiruzenbara, Okayama Pr'efecture in Japan. It is veryrich in diatoms, a kind of micro fossils, ¥vhich may bethe cylinder. In addition, the initial height of the recon-classified as silt according to its particle size, but it hasdiameter'. To reduce friction bet¥veen soil and thestituted samples ¥vas determined in such a ¥vay as tobecome its final height 12 cm af'ter consolidation, so thatthe effect of friction bet¥veen the inside vall and the soilsarnple ¥vould be minimum at time of the extrusion of thecapacity of holding a lar*"e amount of ¥vater because of itssample from the cylinder. After ensuring the end ofporous nature. Interestin*' behaviors of diatomite soilsmixed ¥vlth various soils have been reported by Tanakaconsolidation by so-called ,3t rnethod, the sample ¥vasextruded from the cylinder. Then the sample ¥vas(2002); Shilvakoti et al. (2002); Tanaka et ai. (7_003). Twolvrapped by plastic film and sealed by ¥vax ¥¥'ith pine resin.types of clayey soils lvere used as base material for theTheymixing: Bangkok clay and Kaolin. The properties ofuntil testing.vere then stored in a temperature controlled roomthese clays as ¥vell as mixed soils are indicated in Table 3.In case of sand mixed soils, plasticity index' (I ) decreases¥vith an increase in the sand content. Other¥vise, fordiatom mixed soils, their lp does not chan* e and T,,Lincreases even though clay content is decreased. As aresult of increase in content of diatomite soil, acti¥'ity ofthe mixed soil apparently increases.Reconstituted samplesivere consolidated under aMeasu/'elnent of Residua! Effective StressThe apparatus f'or measuring the pf value is schematically illustrated in F=ig. 2. The specimen ¥vas placed on acerarnic disc lvith an air entry value of ,_OO kPa. Theceramic disc ¥vas boiled, prior to measurement for fullsaturation. The ceramic disc ¥'as placed on the pedestaland de-aired ¥vater vas supplied through the flushing line H.21-7TANAKA AND M. TANAKACeramic Dlsc/ 200 kPa)rrT41 AirEntryValue =(p .) could likely be more reasonable normalizationparameter to consider the pf vaiues. Ho¥vever, since the insitu horiz,ontal effective str'ess (cTl'*') is a very difficult valuefor measurement and in this study it has not beenFiushing Piezometerr r LiiLLLJFig. 2. The appa atus used for measuring the residual effective stressmeasured, the pf vaiue ¥vill be interpreted in the normaliz,ed form of pf/cT(. .As mentioned earlier, although ail samples wereretrieved by the same sampling method, i.e., using theJapanese standard sampler', considerably large variationin p /a(.. values are observed for various soil deposits andin this stud¥.at various depths; for example, the p; value for Drammenclay at some depths is as lo¥v as nearly zero ¥vhile p'lGr', ' ***for Ban*"kok clay at depths betlveen 12 and 18 m exceeds0.35. It is also interesting to note that even at the samesite, the p;/(T(.. varies vith depth and there is no consistency in its trend. For example, pflcT(.* at Yamashitao<! (K!. <10E'20! o!Vo :_ - l'e<! d o __ j_ r .)pl5e' oe._< oo=. ?eo vv O-¥¥aO<;o ' =-_3v= -- rov _ _ Ariakev,¥¥ evv Bothkennar-B3- Lou'seville>>vvSingapore*¥¥/ r>v -(PBangkok3040/ <1 Drammen>>1/ >_ o Yamash'tapusanoo olo 2 o 3 0.4 0.5p' IcFig. 3. pf/(T(,, ratio for sampi8s Fetrieved from various sitcs, using theJapanese stantiard san]pierand Sin_9:apore sites is nearly constant lvith depth. On theother hand, at Pusan and Louiseville sites it generallydecreases ¥vith an increase in depth, ¥vhile at Bangkok siteit increases ¥vith an increase in depth. It might beanticipated that at _g:reater depths, the value of pf may besmaller because of difficulty in sampling. Ho¥vever, sucha tendency cannot be obser'¥'ed in Fig. 3. Indeed, Tanakaet ai. (2002) have reported that no reduction in thesampling quality had been obser'ved for the Pleistoceneclay in the Osaka bay, ¥vhere the sampling lvas carr'ied outas deep as 400 m. The purpose of this paper is to identifymain factors gover'ning the p; value in naturally depositedsoil samples.Location of San7p!e in the Sanlp!il7*" TubeStress conditions of the sampled soil in the samplingtube are rather complicated but are very impor'tant tointerpret the pf values. ¥ fhen the soil is sampled anduntil the excess lvater 1 *as coming out from the ceramic¥vithdra¥vn from the ground, the vertical total stress ondisc. Then, the excess lvater vas ¥viped off by filter paper.the soil sample (a,)The capacity of piezometer for measuring negative pore¥vater pressure is - iOO kPa. Since the p; ¥vas measuredpr'ior to the unconfined compression test, the testedspecimen ¥vas trimmed by a ¥vire sa¥v into the siz,e of35 mm in diameter and 80 mm in height. After placingthe specimen on the ceramic disc, the measured negativepore ¥vater pressure ¥¥"as gradually reduced (absoluterelease of the ¥'ertical stress. Ho¥ve¥'er, due to the frictionvalue increased) and became constant: this constant value¥vas defined as the p acting in the specimen. The timeconsiderably vary ¥vith the location of the sample along:the tube. Some excess pore ¥vater pressure (u) exists in thesoil sample due to the ¥'olume chan*'e caused by pushingof the sampling tube as ¥vell as the friction acting on theduration of attaining this constant value varied indifferent specimens, but usually less than one hour forvould ideally become zero due tobetween the inside ¥vall of the samplin*' tube and the soilsample, a fraction of the (Tkeeps acting on the samplein the vertical direction. Also, some amounts of thehorizontal str'ess (al') is acting on the sample, because thesoil sample is confined by the sampling tube. Thesestresses may not be uniformly distributed, insteadvas measured under atmosphere condi-inside ¥vall of the sampling tube. And this u is alsoconsidered to vary along the tube immediately after thetions because its ¥ralue is smaller than the ca¥'it}., pressure,soil is sampled. Ho¥ve¥'er, such ¥'ariation in ¥'alue of u getsi.e., 100kPa. Ho¥vever, to avoid the specimen fromredistributed and equilibrated to a cer'tain ¥'alue after acertain time. In fact, Tanaka et al. (,_OO'_) ha¥'e measuredthe dissipation time of the negative pore ¥vater pressure inmost soils. In this study, the negative pore ¥vater pressureof all samplesgetting dry during the measurement, the specimen ¥vascovered by an acryl box.a specimen by submerging one side of the specimen intoTEST RESULTSFigure 3 sho¥vs measured pf values for sampled soil at¥'arious sites, normalized by (T(.. and plotted against thedepth. In stead of (r(.., the mean in situ confinin_._" stressthe free ¥vater. They revealed that for the specimen ¥vith5 cm hei**ht, it took 3 hours for the disappear'ance of thenegative pore lvater pressure. In usual in¥*estigationspr'actices, samples are extruded se¥'eral days after the soilsamples are recovered at the site. Therefore, it is likely_. r,RESIDUAL EFFEC'TIVE STRESS FOR C OHESIVE SOILS(JGSO1'_-2000): one is the extrusion from the cutting:' - - on $ite {em) l' i l 1F:70h80esrit::eOemk}(7m) 'edge (in ¥vhich the moving direction of the soil sample is' ' ¥¥¥t;'A After¥,3 '/eek"S (7m);Wopposite to that in sampling) and another rnethod is: r¥. l¥opposite to the abo¥'e method (the direction is the same incE(,)sO40the process of both sampling and extrusion). In thls¥. o' :1:;fAriake clayinvestigation, the sampling tube ¥vas vertically installedlvith the cutting edge on top, and the sample ¥vas extrudedf'rom t.he cutting edge, i.e., the former ¥vay...r¥¥: '・ !uJ30;l' ¥i¥'Tii//.. /oIn the process of not only samplin*', but also the'f 'ico20ao OOllelOextrusion, the soil in the tube rnay suffer' from shearingOos o I o O 1 5 O 20 o.25p .! Q'21nS rTlp!ing ExtrusionFig. 4. Distribution of p;/a(*, in the sampling tube: Ihe measurementwas carried out immediatel) and 3 weeks afteF samplingaction on the inside vall and the sampled soil, Ivhichcauses reduction of p . As far from the top edge of thesampler, the intensity of the shearin*' is much more significant because of lon_".. travel distance during both thesamplin*' and extrusion. Therefore, the pf distribution inFig. 4 is quite understandable. Also ¥ve have to considerthe influence of disturbance caused by making the borehole. Closer to the bottom of' the borehole (upper part ofthat the value of u in the sampling tube gets redistributedand attains a uniform value throughout the sample in thesampling tube. In the process of' the redistribution of u inthe sample, both ci* and (7. may change. When the sarnpleis extruded from the sampling t.ube at laboratory, thesample suffers frorn shearing and excess pore ¥vater (ne*'ative or positi¥'e) pressure may get generated in the sample.Ho¥ve¥'er, ¥vhen the soil is exposed to t.he atmosphere,both total a+ and ah become zero and the resultantFig. 4), the soil may be disturbed by drilling. The reasonf'or the largest p value at the cutting edge of the samplevhich lvas extruded at the site Is most likely due to thecreation of vacuum at the vithdra¥val of the samplingtube. This large negative pore ¥vater pressure rnay havebeen redistributed, ¥vhen the sarnple ¥vas extruded three¥veeks after the sampling.It is kno¥vn f'rom previous researches (for exampie, seeMatsumoto et al., 1969) that the unconfined compressionnegative pore lvater pressure, i.e., pf is finally generatedin the specirnen.In order to investigate the effect of redistribution of ustrength (q ) is ver'y sensiti¥'e to the location of' 80 cm longon the final pf value, the pf value vas measured at t vodifferent stages: immediately after sampling at the siteand 3 ¥veeks after sampling, ¥vhen the value of u in thesample is considered to have been redistributed to reachan equilibrium condition. In case of the test on site, thesample ¥vas extruded from the sampling tube immediatelyafter vithdra¥ving it fr'om the ground. The soil samplesyields the highest strength. This fact is supported by thefindings from this investigation that the large pf ¥'alue isobserved around this location. For practical applicationfrom this investigation, It is recommended that specimenssample in the sarnpling tube: the sample at the position ofone third of the sampling tube from the cutting ed_"..efor laboratory test, especially to measure mechanicalproperties, should be obtained from this location. In this¥¥'rapped ¥ 'ith thin plastic film and ¥vax lvith pine resin.investigation, all samples for measurement of pf ¥verelocated at one third distance of the sampling tube fromthe top edge. Therefore, the variation of pf caused byThe time required for the above procedure was less thandifferent location in the sampling tube is eliminated.¥vere cut into 10crn length by a ¥vire sa¥v andvere0.5hours, ¥vhich may be considered short enough toprevent the r'edistribution and attend ne¥v equilibration ofu. These samples ¥'ere transported to the laborat.ory andthe pf vaiue ¥vas measured for these samples. On the otherhand, other samples ¥vere kept in the sampling tube for 3¥veeks, and then the sample was extruded at the laboratory and prepared by the same manner as that on site.Figure 4 sho¥vs the distribution of pf value at differentlocations of the sampling tube. As can be seen from thefigure, the distribution of the pf is nearly the same for t¥vodifferent timings of extrusion, except near the cutt.ingedge. This fact demonstrates that the distribution of thepf value is not mainly determined in the process of samplin*', rather it is created in the extrusion process of thesample. For interpretation of the pf distribution indicatedEffect oj' Duration of Stol'age a/7d Transpol'tationThe tirne required for transportation from the sampling site to the laboratory of PARI considerably variedfor different sites. For example, ¥vhen the site is located inJapan, this tlme may be less than I ¥veek; vhile in case ofoverseas, it sometimes took more than I month to obtainthe samples at the laboratory. Also, ¥ve have to considermethods for transport of samples. Although all transported samples lvere kept in the sampling tube, chancesof encountering disturbance on the samples are muchhigher in case of overseas than those in Japan.At the Ariake site, the infiuence of duration bet¥veensampling and measuring the pf value was studied. Asalready sho vn, it ¥ 'as found that the pf value isin Fig. 4, the extrusion method employed in this investi-considerably aftbcted by the location in the samplin_O*.g:ation should be noted. T¥vo methods for extrusionfrom the sampling tube are allo¥ved by the JGS as astandard for preparation of undisturbed soil samplestube. Therefore, soil sample was extruded immediatelyafter sampling, and cut into 10 cm long blocks and oneblockvas furthermore cut into four pieces on the site as lH. TANAKA AND214¥. l. TAN_ AKAoc 7.5ClllO siteol 5 ocmi:O Oco5-1o At PARI Laboratory8l K)r)oooocoo; ) OcmoBothkennar Clay15oFig. 5. Cutting specimens foF stutiying on infiueuce of time durationbetlveen sampling and measurement of tt]e residual effective stress20o51015o202530p (kPa)o 20Fig. 7. Effect of transportation on the residua! efrective stress: Theva]ue of "on site pf" Ivas measured on the samples extruded at theROsl5/site: "At PARi Laboratory" samples were transpoFtcd frorn theuBothkennar site to the laborator)' at PARIthe infiuence caused by the transportation on the pf issmall enough not to be detected in such an invesrigation.:: a loco a5This conclusion is also confirmed by the fact that the q**value measured in the site and at PARI's laboratory ¥vasAriake Olay Sample Depth is 8ma ooc200 40a saothe same in both sites of Drammen and Bothkennar (seeTanaka, 2000).80GEf psed Time (day)Flg 6. Chamge in tl]e Fesidual effective stress during storageThe reasons for occurrences of different pf ¥'alues atvarious sites shown in Fi . 3 are examined from ¥'ie vpoints of human factors, such as sampling quality differences caused by different locations in the sampling tube,timin_ for measurement of the pf value and influence ofsample transportation methods. It is found that theirsho¥vn in Fig. 5. The pf value for a piece of the block ¥vasmeasured at the site just after the cutting. Other 3 pieces¥vere ¥vrapped bv_ thin plastic film and ¥vax ¥vith pine resinin the same way as done in the usual investigation. Thesesamples ¥vere stored in the room ¥vith constant temperature ('-O'C), and the measurement vas carrled out untii aslong as ?- years after the sampling. Test result is sho¥vn inFig. 6. It may be concluded that the pf value is constantfor considerably long time period, provided that thesample is properly sealed and ¥vell stored. Ho¥vever,effects are ¥'ery small and can be practically ignored onthe measurement of the pf, so that the difference in thepf/a(.. at different sites should be caused by their'-eotechnical properties. In the follo¥ving sections,various factors influencin>・ the pf values ¥vill be evaluated.Effect of OCRAs sho¥1*n in Table l, the overconsolidation ratio(OCR) ¥'aries¥'ith sites. It is knolvn that ¥vith an increaseHight et al. (1997_), ¥vho carried out the similar' test forin OCR, the coefficient of earth pressure at rest (A'.)increases so that the mean stress (p ,) also increases,Bothkennar clay, have reported that the pf changes ¥vithalthough the value of (T(.. remains unchanged. Thus, it istime. At present, the reasons for' such a different conclusion are not identified.The effect of transportation on the p; value lvas studiedon the Bothkennar clay. Test results are sho¥vn in Fig. 7.anticipated that the pf/(7(.Some samples vere extruded at the site of Bothkennarand the pf was measured by the same apparatus as sho¥vnin Fig:. 2, ¥vhich lvas brought from .Tapan. Other samples¥vere transported by a commercial cargo ser¥'ice in a similar ¥vay as pre¥'iously described. Though some scarter inmeasured pat both the Bothkennar site and theiaborator_v of PARI can be recognized, any prominentdiff rence cannot be seen in the p; values measuredbetween these places. Therefore, it may be conclucleci that¥'alue for heavily over'consoli-dated soils is larger than that for slightly o¥'erconsoli-dated soils. Using reconstituted soils, namely, Honmokuand Singapore clays, ¥vhose properties are indicated inTable 2, the influence of OCR on pj/(T(,. ¥vas examined.Ho vever, it should be noted that the reconstituted soil inthis study cannot be simulated in the sampling process,¥vhich means only the effect of the extrusion on the pfvalue can be examined.The maximum consolidation pressure (p,'*,**) in reconstituted mechanically overconsolidated soil ¥vas determined in order to obtain the final pressur'e of 49 kPa (p(*)after s¥velling. For example, in case that the OCR is '-.O, RESIDUAL EFFECTl¥,E STRESS FOR COHESI¥,E SOILS21530Ariakec5P fc" e(K::KK OSe-$ : OCR["';::Bothkenn rLou sev'lie ;25, v singa ore i042JO¥*CF()eIo03O BangkokeBangkoki:Singapore<! Dram Tlen i> Yamashitaee/ :///:>/1 1: ,;o Yea;ashitai o pusanCl o 5¥*¥I OT']LLileas02p'r/cF _Qo singapore clayooOo2468BothkennarAri a ke e /'/ tv/ , I v':Louiseviiiepusanc; flY"-"(lr'vt vvl /r; <1 -;/6:r<amr enoocR);>Oo2053a35OCRFig. 8. Relation between OCR and p /( (,, for reconstituted samples:p ,/a(,, was calculated using equation (1), assuming a = 0.5 and 1.0Fig. 9. p:/( (,, Yalues shown in Fig. 3 are replottcd agninst OCRthe sample is consolidated to p *. of 98 kPa. Afterconfirming the end of primary consolidation under thetest result that ¥vhen a highly o¥'erconsolldated soil sarn-p ,* , the sample ¥vas s velled under the p(* to 49 kPa.ple is disturbed severely, the sample presents greaterAfter the completion of the s velling, the soil ¥vasundrained shear stren th in the unconsolldated un-extruded from the cylinder.drained (UU) triaxial test than the high quality sampleThe test result is sho vn in Fig. 8, plotted in the form of'p lcT(.. It can be seen in Fig. 8 that the pf/(r(, ratiodoes, due to the positi¥'e dilatancy.Figure 9 plots pf/(7(. to OCR for naturally depositedincreases ¥vith an increase in OCR. In Fig. 8, the p,'**lcr(.soils sho¥vn in Fig. 3. It can be seen that the p;/(7(ratio is also plotted, ¥vhere p*'** is the rnean stress befcu"eeach site increases ¥vith OCR. As mentioned earlier inextrusion of the sample from the cylinder. From se¥'eralFig. 3, there is no consistent trend of the pf/(y(. ratio forprevious studies, it is kno¥vn that the K! f'or o¥*erconsolidated clay is related to OC R ¥vith the f'ollo¥ving equationeach site to the depth. This may be mainly due to the(see, for example, Leroueil and Hight, 2003).K. = K ***OCR* ( I )¥vhere, K.*** is K at normally consolidated (NC) state, and(x is a constant. The ¥'alue of' ( has been experimentallyfordifl:erent OC*R to the depth. That is, for a restricted slte,the ratio of' p /a(, can be ccu'related ¥vith OCR.Ho¥vever, it can be easily recognized that even at thesame OC*R, the pf/(T(. ratio is considerably different f'ordifferent sites. For example, at OC R= 1.5, pflcT(* at theBangkok site is about O.38, ¥vhile that at the Drammenobtained and expressed by several researchers, forexample, the a=sin c' that is proposed by Mayne andsite it is as small as O.04. Also it can be seen that theKulha vy (1982),exarnple, pflcT(,. at t.he Yarnashita site is not so sensitive tovhere c' is the internal effecti¥'e frictionincrement of pf/(T(.to OCR ¥'aries at dift rent sites. Forangle. On the other hand, Hamouche et al. (1995) hasOCR as that at the Bangkok site. These trends indicatereported that the ai is I .O for eastern Canadian clays. Thethat the pf/(T(.(71'* in the p * vaiue in Fig. 8should also be other important f'actors go¥'erning the pvalue.vas calculated from theEq. (1). In this calculation, K * is assumed to be O.5, butthe ( is O.) (c' =30') and 1.0. It is interesting to notethat, pf/(T(., increases ¥vith increase in OCR, but rllis ratiocannot be onl_v related to OC.R, but thereEffect OJ Cla.y Contentsis gradually close top */( '.and exceeds p,*,/(T (a O ))Another considerable factor may be the component of'at large OCR. This means that at high OCR, thesoil particles. The reason for retaining high negati¥'e poreconfining stresses before the extrusion can easily be keptin the form of' the pf, even though it is subjected todeformation due to the extrusion from the reconsolida-pressure in the specimen is the capillary force acting onthe water existing in the pores. Classical physics teachesus that finer the capillar}' tube, stronger the capillarytion cylinder. Although the value of pf is not theoreticallylarger than p ,, it is obser¥'ed that the ratio of pf/p(..force. Therefore, it is easily anticipated that for the soilconsisting of fine particles, the dlameter of pores may beexceeds p ,/o'(, at large OCR. This is probably due to thesmall enough to keep high pf. Using soils artificiallymixed ¥vith granular materials such as Toyoura sand or+'generation of the ne*'ative pore ¥vater pressure, ¥vhen thesoil sample is extruded fr'om the cylinder. When the soil ishighl), overconsolidated, the negative pore ¥vater pressureis generated by shearing because of the positive dilatancy.It is inferred that as a result, a sample ¥vith the hi**h OC Rsho¥vs large pf value. Hight (1992) has sho¥vn a similarDiatomite soils, the influence of soil particle componenton t.he pf value ¥vas in¥'estigated. The result is sho¥vn inFrg. 10. The stress conditions for Fig. 10 asvell as Fig. 1'_are normally consolidated I e cr' =cr(*. As expect.ed,' ' .' "***¥vith a decrease in clay content (¥vhere the diameter of TANAKAH_ TANAKA AND *¥,i.21605os1 --v//fBangkok+sand0403(iL////// ' /¥i:0203Kaolin+sand1:o/m'¥//f/1/i!7 /d ;ee4.// B ngkok+Dlatome¥c/ /;///.(//e¥/ /02olOKaok n+Diatoml'c)oooO60 8020 4O302010OoO40Olay Content (<5L*m)eo5070IFig. 10. Relation betlveen pfl,1(,, and cla¥. content for reconstitutedReiation between p,/(T(<, and lp for reeonstituted samplesFig. 12.samplesC'o, O04Ariake eO Bothkennar _ QQoe (' aOvvvl>a o?1 t:"O1<<;<:OO< <Io;r :80Rn>Q'>>RH >>R<S8 <io osd<S v::oo e _( <!<O Pusan <O 20 40 60eve oO -r vv:;s oDrammenYamashita(,_a' 028angkoko Pvsane _ eo rl r*BothkennarLou'sevilleSinga poreovv¥v,,O Bangkok<1 Drammen <> YamashitaoovvaFE 1 Louisevilie voo03V Singapore c'ee 'e,,1e<!Q030204c,(,e:Ariake505l20O4060801 OOIoOOlay Oontent (<5t m, s )Fi(;. 13. Relation bet veen pf/(T(,, and I for naturall)Fig. 11. Relation between pflcT(,, and cladeposited sampiescontent for naturalllclay particles is defined as smaller than 5 um), there is adepositedsampleswith diatomite, Ip does not decrease ¥vith diatomitecontent, unlike the sand mixtures. Ho¥vever, ¥vithtrend that pf/(T,i, decreases, although these values areincrease in diatomite content, pf/(7(, decreases, so thatsome vhat different for mixing: ratios and mixedpf/( (,, apparently decreases at constant lp.On the other hand, pf/(T( for naturally deposited soilsmaterials.For naturally deposited soils, the relationships bet¥veenpf/(T(.. and clay content are plottcd in Fig. 1 1. There arecan be relatively ¥¥'ell correlated ¥vith lp, as sho¥vn inFig. 13. Although the Bangkok clay and Louiseville claymuch scatters in this r'elation. Especially for Drammenhave the same lp, the difference in pf/(T('. for these clays isclay, its clay content is much larger than that ofBothkennar. Ho¥vever, pflcr(, value for Drammen isas much as 0.3. Ho¥vever, the reason for lo¥v pf/(7(.* forDrammen clay can be explained by this correlation.considerably smaller than that of Bothkennar. It cannotbe said from this figure, that the pflcT(.ratio for naturallydeposited soils cannot be simply related ¥vith clay con-Effect of Void IndexIt may be anticipated that soil ¥vith hi_9:h structure easilyInstead of particle's component, pflcr(,* value for arti-loses the pf due to disturbance. In this study, easiness ofthe disturbance ¥vill be expressed in terms of ¥'oid index(1.), ¥vhich is proposed by Bur'land (1990). In thls concept, the relation bet¥veen I* and consolidation pressureficially mixed soils is related to plasticity index (Ip) in(p') for reconstituted samples can be expressed by aFig. 1'_. In this figure, only one point of Bangkok+sin_ le line, i.e., the Intrinsic Compression l,ine (ICL).Diatom is indicated because ¥vhen the diatomite ratio ismore than 500/0, the arterberg limits cannot be measured(see Table 3). As already mentioned, in case of soil mixedHereafter, J,, for the in situ conditions for naturallytent.Effect of P!asticity Incle, 'deposited soils is denoted by I *. The index of I, for allsoils studied in this paper is plotted in Fig. 14 together 217RESIDUAL EFFEC TIVE STRESS FOR COHES1¥"E SOILS30eHPooric]e ^" : - e l' l_ v<:oo-p :e10s: 10: ' s;<iLouisevil e:<1<]K; >sOCI 058oLh OOeeve*EQ,¥Q,> Yamashita,O 04O Pusan<lGood to Fair,(,< e <o%('QPvsanccording to Lunne ( 997)'s Criteie cGood to Fair)>+.Qre:<!> :>e1giExceient-1 oYam shita>osvra(,,O G2BangkokDrarnmenov'vIntrinsic Compression LineO 5O Bangkok< Drammenvv <:_<O 06v Singapo e< < Q'o1 ry'Is(:>LouisevilieSinga porevvO OBAr'akeQ Bothkennareeec:)BothkennareH20ooAr'akeO 10,!25Excelento oe10101 oOo1 oo25215303soCRc' (kPa)Fig. 14. Void index for naturall) deposited samples in this stud)Fig. 16. Sa}nple qualit) for samples using in this stud), assessed byLunne et al. (1997)'s critcria1 Ari kee Both enna0505Louisevilleov Singapore,oAe;)O Bangkok<i Drammene04i> Yamashite03o pusan've q,v r v80cS :702vo e**oCO1<v<< s<<<: (OO-o1e03>.gce:}<:0.2ecP*<:oe O 5 2.0 2 5-O 5OOOo10O oO! .l. lCLo 32o Oo 06o 08O 10e/e.Fig. 15. Relation between p,/c(,, and structure defined b_v the distanceFig. l?.from the ICL IineRelation between p /(;(<, andje/e,,with the relation lvith the ICL and p'. According toof sample quality. Figure 16 shows the Ae/eBurland's idea, as the difference bet¥veen I . and I+ corre-soils, I 'here e. is the initial void ratio and Ae is the changesponding to the ICL at the same (7(.., which is hereafterin the void ratio due to the application of the a(... The Aedenoted as (I .-1 ( cL)), is larger, the soil possesses highstr'ucture. As can be seen, it may be said that the Ariakeclay presents the highest structure, ¥vhile some points forBangkok clay sho¥v relati¥'ely lo¥v structure of the investi-was measured by the Constant Rate of Strain (CRS)*'ated soils.Figure 15 sho ¥'s the relation of the pf/(T(.. and thethe criteria for sampling quality using the Ae/e. correlated ¥vith OCR. According to their criteria, most sam-(I *-1 ( cL}) for the studied soils. The soils indicatin*" hi_ :hples with OCR greater than 2.0 are ranked as "excel-(1+.-1+( )cL)are easily disturbed and as a lesult rt rslent" , ¥vhile samples with OCR Iess than 2.0 ar'e classifiedprobable that these soils easily lose the pf during thesampling and the extruding processes. As can be seen inFig. 15, most soils are distributed following the aboveas "good to farr" and some samples are "poor ,', eventhough all sarnples ¥ver'e retrieved employing the sametendency. However, Ariake and Drarnmen clays are farAs already mentioned earlier, volume change causedIf the gradient of the e-lo*'p relation at the overconsolidated (OC ) state (sometime it is called "C*") is thesame, then soil lvith pf close to the cT('. should indicatesmall Ae. Fi**ure 17 sho¥vs the relation bet¥veen Ae/e. andpf/(T(' . In spite of the above anticipation, no distin*"uished relation bet¥veen these indices is recognized.Figure 18 replots the relarion in Fig. 17 by grouping databy application of the in situ stresses is used as an indicatorwith the same OCR. From these figures, it may befrom this tendency: i.e., in spite of large (I .-1 (lcL)),pf/(r(,. for Ariake clay is not high, ¥¥'hile Drammen, whose(1,.-1,(icL)) is not large, sho ¥'s quite small p'l(7'**.C*orre!ation vvit/1 Another Index Indicating Disturbancefor all testedoedometer test (the diarneter of the specimen is 60 mmand the initial height is 20 mrn)l0-6 svith a strain rate of 3.3 xI (0.020/0 /min). Lunne et al. (1997) have proposedsampling method. H. TANAKA AND ,1. TANAKA_ 182-1 5A 1 5-2 OO 2 0-3 O04OCR. Ho¥vever, if the site is different the p'lcr'cannot be simply related to OCR alone.5) Clay content and plasticity index (Ip): There exists a, ' *.etendency that the p /(T(. increases ¥vith an increase in03¥pf ¥'alue, distinguished relation can be observed ¥vithOCR: i.e., the pf/(7(. increases ¥vith an increase inOOR I O1 205clay content or lp. Ho¥vever', it is also true that thereE!eO 4 ; e :02eoOC'Eare some exceptions.6) Aele.: As an assessment of the sample quality, Aele.is extensi¥'ely used, ¥vhere Ae is the chan_ e in the ¥*oidQEratio when the specimen is consolidated Lmder the cr(.cOCO1and e. is the initial ¥'oid ratio. Although it vasanticipated that there is some correlation bet¥veenRoao.oooa0204o a6 a a8 opf/(T('. and Zle/e., no clear relation bet¥veen them ¥vasOe,' eidentified.From these summaries, it is found that the pf value,Fig. 18. Replotring relations in Fig. 17, with OCRconcluded that as far as the sample quality is _g:uaranteedlvith a certain level, i.e., unless the sample is extremelydisturbed, there is no clear r'elalion bet¥¥'een Ae/e. andpf/(T(... This conclusion suggests that !le caused by thetestlng process, for example, setting the specimen into theoedometer ring, is much larger than that yielded from thepressure of pf to cT(*.¥vhich is a very important value for' assessment of theunconfined compression str'ength (q,,), is affected by notonly sampling method but various other factors based onsoil properties. These factors are complicatedly relatedeach other.REFF.RENCF.Si) Andersen. A and Kolslad. P (1979): The NGI 54 mm sampler forundisturbed sampling of clays anci representati¥'e sampling ofcoarser materials, Proc. In! S_T*nlp. Soi! Samplil7g, Singapore,l 3-2C_ON CLLi SIONSThis paper in¥'estigates the main factors governingresidual effective stress of cohesive samples retrie¥*edfrom ¥'arious parts of the ¥vor'Id. E¥'en though all samples¥vere recovered by the same sampling method using theJapanese standard sampler, the measured residualeffective stress (pf) Ivas considerably differ'ent at differentsites and depths. To e]iminate the influence of the in situeffective o¥'erburden pressure ((T(. ), the pf ¥'alue lvasevaluated in the form of pf/(T(, . The paper exammed themain factors governing the pf/(T( ratio. The interestingfindings in this study are as follo¥vs:l) Location of sample in the sampling tube: The sample10cated in the one third of the sample length from thecutting edge provides the highest pf. Also, the valueof pf has been determined in the pr'ocess of theextrusion from the sampling tube asvell as the sam-plin_,*".?_) Time duration from the sampling to the measurement of the p : The value of pf ¥vas measured as lon_",_as t¥vo years after the sampling, including immediatemeasurement on the site. It is found that if thesample lvas properly stored, the infiuence of theduration on the pf ¥'alue is not prominent.3) Tr'ansportation of the sample: All samples ¥vere2) Burland. J. B^ (1990): On the compressibility and shear s rength ofnalural clays, Georechnique, 40, 329-3473) Hamouche. K., Lerouiel. S.. Roy, I a;nd Luleneg er. A J. ( 995):In situ e¥*aluation of Ko in eastem C_anacia cla¥'s. Ccln Geo!ech J..32, 67/ -688.4) Hight, D. ¥V. (1992): A revie¥v of sampling effec s in: Cia¥'s andsands, Proc Inr. Conf. Offsllore Fotulda!ioils (7nd S!te fnl'es!igation, Socie y for Underwater Technology, Kle¥ver* 1 15l465) Higllt, D ¥¥,* , Boese. R., Bu cher, A. P , C*la¥*ton. C. R,, 1. andSmi h, P R ( 992): Dis urbance of the Bothkennar clay prior tolaboratory esting. Geo!echnique, 42, i99-2176) Leroueil, S. and Higl t. D. (2003): Beha¥'iour anci properties ofnatural soils and soft rocks. Characrei'i alion anr! Eilgineer'ingProperties oJ Nalura! Soils, 29-254.7) Lunne, T^, Berre, T and Strandvik. S (1997): Sample disturbanceeffects in soft lo¥v plas ic Nor¥vegian clay. Pr'oc. Conf RecenrDeve!opn,e,lis in Soi! (7ncl Pal'e,neni , lecha;1ics. Rio dr Janeiro.81-l02.8) N,Iatsumoto, K.. Horie, Y. and Okumtlra, T(1969): Studies onboring and sampling of saturaled alluvial clays (4th report)= Reporrof t/1e Po,-! anc! Ha/'boul- Research Inslin!re, 8(2) (in Japanese)9) ¥, iayne. P. ¥¥r, and Kulha¥vy, F. H_ (1982): K )-OCR relationshipsin soil, J. Geotech En*"/ ;" Dil' . ASC_E, 108(CT6), S51-872.lO) Okumura, T (1971): The ¥'aria ion of mechanical properlies of clavsamples depending on its degree of dislurban:ce. Proc. Specia!,_1'Session, Qua!it_T* in Soi! Samp!ing, 4lh Asia,1 Reg!ona! Colrf_S1'1FE, 78-81^l ) Shilvakoti. D. R., Tanaka, H.. Tanaka, ¥,1. and Locat, J. ( 002):Influences of diatom microfossils on engineering properties of soil,Soi!s alld Fou,Idaiioils. 42(3), ll7.2) Tanaka, H. (2000): Sample qualily of cohesi¥'e soils: Iessons fromtransported keepin*' them in the sampling tube after'these sampling tubes lvere placed in a ¥vooden box,lhree shes, Ariake, Bothkennar and Drammen. Soi!s and Founc!a-¥vrapped lvith rubber sponge. In these conditions,sample disturbance effects during transportation can13) Tanaka. H. (l002): Re-examinatian of esiablished relationsbet¥veen index properties and soil parameters. Proc r_oasra!be i_"._nored even from the overseas.4) OCR: Among ¥'arious soil properties influencing ther!ons, 40(4), 57-74,Geo!ech. Eng, Pr(7ciice, (2), 3-2514) Tanaka, H. Local, J , Shibuya, S . Tan. T. S aod Shi¥vakoti. D R(2001 a): Characterizalion of SingapoFe, Bangkok and Ariake clays, 蔭RESIDU.AL蓑…FFECTIVE STRESS FOR CO賛ESIVE SOH−S  Cα’∼,Gθo∼θch.ノ、,38,378−400、15)Tanaka,}{、,Mishlma,O,,Tanaka,M.,Park,S.Z、,Jeong,G.H、  and Locaし,J.(2001b):Charaαeriza{ion o∫I Yangsan clay,Pusan,  Korea,So’Z∫α1∼4Fo朋ゴo”01∼5,41(2),89−104.16) Tanaka, 9., Slliwakoti, D、R、, で〉11shima, O、, ∼Va【abe, Y. alld21917)Tanaka,H,,Riξoh,F.andomukal,N、(2002):Qua翫yo∫samples  retrieved fronl great depth alld i[s inβuence on consolida嫉on  proper【ies,Co’∼、Gθo’(∼ぐh、ノ.,39,1288一王301.18)Tanaka,冠、,Shiw&koti,D、R.,Omukai,N、,Ritoh,F,Loca乳,J、  and τanaka,M、(2003):Pore size distribution of c玉ayey soi!s  Ta自aka,M (2001c):ComparisQn of mechanical behavior of two  measured by mercury intrusion poros1metry and its relat三〇n Io  overconsolidated da》・s=Ya貰1asbiξa and Louisev註1e days,50’15αn4  hydraulic conducti、「ity,50〃5α〃‘ゴ君o∼〃∼ゴθ∫’o∼∼∫,43(6),63−73  FOε’刀面”o’∼∫,41(4),73−87、
  • ログイン
  • タイトル
  • Undrained Shear Strength of Cement-Treated Soils
  • 著者
  • Kiyonobu Kasama・Kouki Zen・Kiyoharu Iwataki
  • 出版
  • soils and Foundations
  • ページ
  • 221〜232
  • 発行
  • 2006/04/15
  • 文書ID
  • 20901
  • 内容
  • SOILS AND FOUNDATIONS¥rol^ 46, No_ 2, 221232, Apr '006JaF)anese Geolechnical SocielyUNDRAINED SHEAR STRENGTH OF CEMENT-TREATED SOILSK YONOBU KASA ,iAi} KOUKZE 'i) and KI¥'OHARUI¥¥'ATA (lii)ABSTRACTIn order to evaluate the eft cts of cementation on the mechanical properties of cement-treated soil, a series ofisotropic consolidation and undrained triaxial compression shear tests ¥vere performed for cement-treated specirnensof Ariake clay, Akita sand, Rokko Masado and Toyoura sand. This paper evaluates factors aft cting the shear strengthof these cement-treated soils. The follo¥ving conclusions are obtained: l) C*ement-treated soil has a normally consolidated line in e-In p' space vhich depends on the mixing cement content. The consolidation yield stress, p;・, of cementtreated soil increases ¥vith increasing cement content and initial specirnen density. 2) C hanges in cohesive strength dueto cement-treatment can be r'epresented by a tensile effecti¥'e stress, pf・ Strength properties can then be normalized bythe augmented consolidation stress, (p +pf)・ 3) The shear strength properties of quasi-o¥'erconsolidated clay can berepresented by the yield stress ratio, R = (p(. + pf)1(p + pf). 4) The undrained shear' strength of cement-treated soiis canbe represented as a po¥ver la v relation of the yield stress ratio, R, and the augmented consolidation stress.Ke¥.' ,vords: cement stabilization, consolidation, o¥'erconsolidation ratio, shear strength, undrained shear, yield stress(IGC: D6/DIO)through unconfined compression tests lvhich are f'avoredINTRODUCTIONdue to their simplicity and ability to represent propertiesat lo¥v confining pressures. In recent years, the strength-Conventional cement stabilization such as deep mixingmethod (Terashi and Tanaka, 1981) has been used mainlyfor improving the bearing capacity of soft ground, lrlrecent years, ne¥v cement stabilization methods ha¥'e beendeveloped in order to make effecti¥'e use of r'ecycledtics (Shibuya et al., i99'_) have been reported using triaxi-geomaterials 1¥'ith appropriate considerations of en¥'iron-al compression tests. These prior studies have clarified themental impacts. For example, dredged soils mlxed ¥vithmechanical properties of cement-treated soils.dependency on confining pressure (Yajima et al., 1997;Ue et al., 1997), dynamic shear properties (Ito et al.,1994; Yamamoto et al., 1996) and lo¥v strain characteris-videly used as fills in many reclamationVarious indexes ha¥'e been used to represent theprojects in Japan (e.g. Tang et al., 2001). Sandy soilsmixed with cement by the pre-mixing method (Zen et al.,1992) have been used in reclarnation works in Japan tomitigate against liquefaction in reclarnated land. Fromthese backgrounds, a number of types of cement-treatedstrength of cement-treated soil as sholvn in Table I . Theseinclude: a) the ¥veight r'atio of cement to dried soil (calledsoil (namely cement rnixed soil) have been created for thepurpose of increasing the bearing capacity of softmix design for cement-treated ground. In addition,Yajima et al. (1996) have characterized the failureground, reducing the unit ¥veight of fills, Iiquefactioncriterion for light-¥veight cement-treated soil using voidcountermeasures and recycling of construction surplusratio and unconfined compression strength, whilesoil.Minamiga¥va et al. (1987), Miura et al. (2001) and Suzuki(1990) evaluated strength using the conventional vatercement ratio (i.e. weight ratio of ¥vater to cement) similarto the concrete. Omine et al. (1998) proposed a two-phasemixture model for predicting the stress-strain relationshipcement have been"cement content") and b) cement ¥veight per ¥vet soilvolume of i m3 (called "cement amount"). Cementcontent and cement amount are widely used in practicalIn general, the primary factors affecting the shearstrength of cement-treated soil, are the amount of mixedcement (e.g. Terashi et al., 1980; Matsuo and Nishida,1969), types of cement (e.g. Kurihara et al., 1994; Kamonand Katsumi, 1999), physico-chemical properties of theof' cement-treated soil based on the consideration ofnature soil (e.g. Kuboi and Nishida, 1999; Okabayashistress distribution and strain ene 'gy. Tang et al, (2001)et al., 1999), curing conditions (e.g. lvlino¥va et al., 1998;evaluated the variation of unconfined compressionMishima et ai., 1995), and, specimen size (e.g. Hayashistrength of in-site cement-treated soil due to the changeet al., 1997), etc. These factors have been examinedof ¥vater content and cement amount.i:]}Kyushu University, Japan (kasama ・civil_kyushu-u.ac_jp),Iinistry of Land Infrastructure and Transport, Kyushu Regionai Developmeni Bureau, JapanThe manuscript for this paper was recelved for revie¥v on June 16, 2005; approved on January 13, 2006¥ 'ritten discussions on this paper should be su:bmi ied bcfore November l, 2006 to ihe Japancse Geotechnical Socie y, 4-3S-2, Sengoku,Bunk.vo-ku, 'Tokyo 1 12-0011, Japan_ Upon request the closing date may be extended one month.221i.. KASA¥. IA ET AL.222Table 1.TargetIndexes for describin" the stren"th of cement-treated soilsFunctionsiren thc! , = f (.C!LC_ommenlsInciex,, )Ref erenceA**( ・h): ¥veigl t ralio of cemem toFor cement-treated sandydried soil (cemem con em)soil by the pre-mixing(1990) andZ,en et al( 1 992)methodSu' SdC_( /m ): cemenq,, = f (Oqu¥veighper ¥vet soilFor cemcnt-treated soflcohesi¥'e clay by the deepmixin_(;* methodTerashi et al. (1980);Terashi and Tanaka(198 ),tf: inclination of failure line in p'-qFor li ht-¥vei :ht-cemen -Yajima et al.s pacetrealed soilA: imercept of failure line in p'-qItif: 2 65O.59espaceA: O^92q**volume of I m; (cement amoun ))s 2sd = *tJp 'A( 1 996)p': mean effeclive stress: ir itiai void ratio of specimenqu' suc! ' s* =f(It't /A,,) or f(A,, /Ts't})It'{)/A,,: ¥veigh ratio of cement towa er (¥va er-cenlent ratio)For cement-treated cohesivesoil with hi h'ater comem,Iinami9:a va et al. (19S7)It'o: initial ¥vater contentc/F"qu of in-si utreated soil_ (b l)fFor in-situ cement-lreatedsoil considerin the mixinb: s ress distribution parameterlb_ft/cj,, l *roi( -f.)lqqutFI"f*: impro¥'e rario* I *(x: a coef cient represen ilrgheslrain raiio of in-situ lreated soil andpareci in laboratoryori :inal soil at a half of maximumc/**i'}: qslresscJ 'K(Cq,,K: strenCo)For cemenl-treated dred edh c.oef ciemr_: cement amoumCj : minimum cement amountjc. tt:7106 ; IOmine elal_(1998)levelqu i-): qu of treated soil preof in-situ soil'Iiura et al. (2001);Suzuki (1990);Taneal(2001)cohesi¥'e soilloincrease cl*squ :unconfined compression strength,st' :ur drained shear s rength, sd: drained shear slrenIn some applications such as cement-treatment of clayhTable 2.Ph _'sica] propeFties of tcst soilsof very high ¥vater' (i.e. abo¥'e the liquid limit) and lolvliquefaction, Iarge reduction of ¥'oid ratio (dry density)can occur due to self-¥vei_"*ht. In these cases, the effects ofthe streng:th of cement-treated soil.2 . 6092 6822 6202 640ein ,tl 300l .OI )_o.977em 1o_793o.479o 605o 171O. 1 75In order to control the strength properties of cementtreated soil and make g:ood use of cement-treated soil for'D* ) (mm)futur'e practical application in various fields of geotechni-Ucal engineering, this paper considers the role of consolidation on the strength of cement-treated soil based on aseries of isotropic compaction and undrained triaxialcompression shear tests carried out on cement-treatedspecimens of Ariake clay, Akita sand, Rokko l¥,Iasadoand Toyoura sand.SAMPLF. PRF.PARATION AND TEST PROCEDUREUt:ToyouraAkita sandp. (h'lg/m3)overburden pressure need to be considereci in estimatingRokkoAriake cla¥'cement-content treatment of sand for prevention ofo.0034l .41r* (?,t)96T"I_ (?・, )86 5In51^3u n i f o rmi t ¥'l .3l¥,1asado4019^29,00sandl .52ocoef lcient, F*: fine contentof cement as the solid part of soil. The sample for Ariake"Akita sand") 3) decomposed granite from southclay was prepared according to .JGS-08,_1 (.JapaneseGeotechnical Society, '-OOO). The specimen size forcement-treated Ariake clay ¥vas 35 mm in diameter and70 mm in height. In preparing cement-tr'eated AriakeSuzurandai area in Hyogo Prefecture (called "Rokkoclay, Ariake clay ¥vas mixed ¥vith cement together ¥vithMasado"), and 4) Toyoura sand. The physical propertiesadding ¥vater to keep the slurry a target initial water con-of these materials are in Table 2.Ariake clay ¥vas prepar'ed ¥vith cement contents of 50/0,70/0 and 100/0 and initial ¥vater contents of i.5tvL, 1.7tvl ,tent. The slurry ¥vas gently poured into the mold (35 mmin diameter and 70 mm in height) ¥vith a spoon in threelayers. In order to remo¥'e air bubbles in the slurry, the'_.Olt*L and '_.5Tt*L. It is noted that the ¥vater content ofmold ¥vith the slurrycement-treated soilpart of the mold ¥vith a hammer for ever'y la_ver. AfterThe four test materials in this study are 1) Ariake clay)-) sand from Akita Por't in Akita Prefecture (called"as calculated by treating the lveightvas thoroughly tapped in the lo¥¥'er TUNDRAINED STRENGTH OF CE¥. IENT-TREATED SOILSExperiFneutal conditions for Ariake claTable 3.Cememconlent9?r]_*d-VStandard of samplepreparationInitial wa ercontem t'(Averag:e ¥taler comemafter 27 davs curin :Uotrea edConsolidation pressurep (kPa)77 8(]+83_3 O*O!l oJGS 0812-2000173 /o (2 Olt'L_)30,98, 196, 29499 . i '.50/IOl 19 O +.130?'(1 _5}t'L_)l 1 2.09・・98, 196, 2945 O!+!O7 o!!a1730, (2.0 t'L )l 66. 7q,'130 ' ) (1.l 17.30, _1t )1739・(2 O t )l 62 .Oo,b216 +(2.5T,'1 )206.4q+147?(1 7T+ )98, 196, 294JGS 0821-20001 o ,t47, 98, 196, 29428 .4qt)i73?t) (2.0Tt'l )l 5 7'l6?・ (2 5Ts'l )201 .80'(3,b93, 196, 294Table 4.Experimental condrtions for sand) soi!sUrldrained lrlaxialIsotro piccompression shear tesconsolidation testBase malerial(cenlent content)niliallrlitial rekuiverelative densitydensity*Corlsolidalionpressure p (kPa)_ 460.(j9, 35, 79 'Akita sand (5.5 , )44'.49, 98, 196730.,(}20,,・Rokko lviasado (5.5)26, 41,49, 98, 19668's5 ・b98, 19692 tToyoura sand (5.50./o)20, 31,270*63. 95 O!!O98, 196, 294*No e: These ¥'alues are defined from e.1 " em;"' of 'untreated' soils and do not account for presence of cement and celrlent reacrionfillin*' the mold with the slurry, the top of the mold ¥vasgently poured into the mold (50 mm in diameter and 125sealed ¥vith high polymer film to prevent any change inAriake clay ¥vere prepared by mixing ¥vith de-aired lvatermm in height) filled lvith de-aired water vith a spoon inthree layers. In order to adjust the density of the specimen, the mold ¥vas then vibrated and tapped at the lo¥verto _g:i¥'e the slurry an initial ¥vater content, w=2}vL, fol-part ¥vith a hammer. After filling the mold lvith thelo¥ved by one-dirnensional consolidation to a preconsolidation pressure of' 49 kPa. Table 3 summarizes the vatermixture, the top of' the mold ¥vas sealed ¥vith high poly-¥vater content. Samples of untreated 'reconstituted'contents of the cement-treated Ariake clay specimensafter 27 days of curing, immediately prior to testing.mer film to prevent change in ¥vater' content. The testconditions for cement-treated sandy soils are sho¥vn inWater contents ha¥'e slight differences bet veen the initialTable 4. It should be noted that the results of' undrainedtriax'ial compression shear tests for cernent-treated Akitaand 27th day due to chemical reactions bet¥veen thesand and Rokko lvlasado were originally reported by Zencement andet al. (i990).vater. It should be noted that test results forthe untreated clay and samples vith the cement contentslo/o and 30/0 ¥vere orlginally reported by Kasarna et al.(2000) .All of the test specimens ¥vere cured for ?-7 days in ahumid r'oom under atmospheric pressure at a ternper'atureof 20 d: 3'C. After curing, a series of isotropic consolida-Cement content of 5,50/0 ¥vas useci for the tests ontion and undrained triaxial compression shear tests ¥vereAkita sand, Rokko Masado and Toyoura sand. Theseperf'orrned according to JGS-05・_3 (JGS, '_OOO). The topand bottorn of the specimen vere lubricated by a rubbermembrane sheet smeared vith a silicone grease. Drainage¥¥'as allolved at the top surf'ace of' the specimen. Thetests used specimens 50 mm in diameter and 100 mrn inhei_ ht. These cement-treated sandy specirnens ¥vith theinitial ¥vater content of 5.50/0 ¥vas thoroughly mixed andJ KASA ,1A ET AL.224excess pore¥'ater' pressure ¥vas measured at the bottom ofthe specimen. For cement-treated Ariake clay, a slittedfilter ¥vas placed around the cylindrical surface of thespecimen to facilitate radical drainage of the specimen.External dial gage and double tube burette ¥vere used tomeasure the axial displacement and the volume change ofthe specimen ¥vith an accuracy of 0.10/0 of the initialheight and ¥'olume respectively. In order to raise thedegree of saturation of specimen in the isotropic consolidation test and triaxial compression test, a partial ¥'acu-6Lcement content: 501; : consolidationV ¥'ield stress)LT'o;:2.0lt'Lo 4c1^ It'L!---'_ LJ [ -:,>3 Lli i_um of 60 kPa ¥vas applied in the specimen set-up process,to dra¥v out pore air keeping an initial effective confinin_a._pressure constant, and then de-aireduntl'eatedvater ¥vas dro¥vnin_O._Ariake c]ayfrom the bottom of the specimen. A back pressure oflO¥l O*Mean effective stress (kPa)(a) cement COntellt 50/0consolidation step ¥vas terminated after confirming the6completion of consolidation by using the 3t-method(.JGS, 1999) together' ¥vith the measurements of excess¥vater pressure at the base of each specimen. The speci-li :::2- . 5cement content: 70li;IO_¥n,)mens ¥vere then sheared at an axial compression ratetively.c cl 02lB-values in excess of O.95 in all cases. Isotropic consolidation test was carried out usin incremental loadin9:. Each0.10/0/min for the cement-treated sandy soils, respec-cl198 kPa vas maintained for 24 hours before starting theconsolidation pr'ocess. These careful treatments ¥vereused to ensure full saturation of specimen, and achievedof O.050/0/min for cement-treated Ariake clay and1.consolidationv yield stress2 . OTt 'L'*4 L:;>31 .5'LL''"_'_'- H] ]-'-L r1___. ・ ¥¥ISOTROPIC CONSOLIDATION7_LThe e-Inp' relationships of cement-treated and untreated Ariake clay are sho¥vn in Fig. 1. The consolida-Ariake c]aytion data indicate lvell defined yield points, p(, for each ofiO(b) cement content 7010content). This consolidation characteristic is similar tothat of cohesive soil^ Yajima et al. (1997) had reportedTi, =2 . 5Tt*L cement content:6similar experimental results for cement-treatedl Oo/o0 ) > Oln lYoneyama clay. In addition, the compression index ofvith cement5content ¥vhile the reference void r'atio, eo, ¥vhich is thevoid ratio at unit p', increases ¥vith the cement content.dated (p' ; p(,). The pre-yield region is referred to "quasi-i 03Mean effective stress (kPa)can be obtained irrespective of the initial void ratio (¥vaterBased on these observations, strength properties ofccI 02:content, an unique Normal Consolidation Line (NCL)cement-treated soil should be evaluated in t¥vo rejoins,pre-yield (p'<p(,) and post-yield or normally-consoli-clthe cement-treated specimens although none ¥vere subjected to compaction or preloading. For a given cementthe NCL, ).=(-c/e/dlnp'), increasesuntreated2 . o i 'L;f l"4l/ i> 1:)consolidationo¥'erconsolidation" in this paper as there is no mechanical pre-consolidation of specimens.yield stress,untreatedAriake ciav cThe relationship bet¥veen p(. (obtained by theCasagrande method) and cement content is sho¥vn inc clFi :. '_. It is observed that the yield stress increases ex-lOponentially with the cement content, but decreases ¥vith1l 02l O*Mean effective stress (k'Pa)increasin9: initial ¥vater content.(c) cement content I OoloThe e-Inp' relationships of untreated and cementtreated sandy soils (i.e. Akita sand, Rokko Masado andToyoura sand) are sho¥vn in Fig. 3. The e-In p' cur¥'es ofcement-treated Akita sand and Rokko lvlasado underg:oFig. 1.lsotropic consolidation of cement-treated Ariake cla) 225UNDRAINED STR叢NGτH OF CEM狂NT−TRE.へTED SOILSβ1000「萩□1.39%(〉一〈〉_“         ∼             ▽1.2=・80036%△r一一ム1り1塁1盟鎚…42%匂531639む℃●ろ>79%0.9ぴ      ヨ墨  184%ム蜂200ε                 L」,I0.8Initi)     /1ce口1ent COP電ent5.59も0,7                ]    O                       CementcoI簗t瞭(%)  l      l                 つ             ら       10   10−   1び104          Mean eff奄ctive stress(kPa)Fi62. E研eα of i鶏ltial w滅e『con琶en重9nd cemen重conten芝on   consolid澱重ion yield sI『ess oεcemenHreaIed A『iake ciay(a) Akita sand,ceme膿co殖ent5.5%O.927%   (               壷swell changes ln void ratio compared to their untreated0,8counterparts,andεしlso show welレdef}ne(墨co烈solidat隻onyiekl stresses.As a result,the stress state of cement−treat−ed Akita sand alld Rokko Masado,call also be divided圭nto noITna星consol玉dat圭on and (1uasi−overconsolidationregions.The cemen書一treated Toyoura sand shows similarbehaviors at low initial relative de勲s呈ty,と)ut call not be幾施\ξ、Q0.7心o弱0.6で’δd玉stinguishe(1from由e u虚rea芝ed specimens at mediall>and high initial relative〔iensity.王n(1ense cases,consoli−0.5iilitial     七datlonyieldingofcement一重rea重edsandysollisrelatedtothe degradation of the cementation and to the breakageO,4relative densltv0.31トce陰1ellt content:5.59もof soi至particles(Bee獄and Jefferies,1985;Pestana andWhittle,19951Miura et&1.,1984).In comparison,thel    l l1[目目   1 11iしEI  l l l IlO   1α   10蘭soil particles of Rokko Masado and Akl芝a sand are quite104    Meaηeffectivestress(kPa)frag琵e, and cementa重iorl is effective i勲 preventing thebreakage of soil par重icle at 正ow pressures. Yiel(玉ing(b)Rokko Masado,cement colltent55%(Figs. 3(a) ancl (b)) is then probably assoc玉ated withdegradatiOnOfCementatiOn.171  Figure4summarized the relationships between inl亡ial   、20%(     ire13tivedensityandconsolidationyields重!lessμillorder              ▽∼Vhite:treated   、Blacklu煎ea芝cdI 1−to evaluε瑳e the c紅ange of consolid&tion yielcl圭ng stresswith the variation of(1ensity for these sandy so玉互s.Thengure also includes d3ta of Do9ラs Bay sand and Quiousand(Kwag e{al.,重9991Yasufuku e重al.,1999)whichεしretypical crusぬab至e carbonate sands.It can be seen that theconso豆kiation yield stress of ce1nenレtreated and untreat−ed Sandy SOllS have very Slmllar relatiOl1S Of relativedensityin spite ofthe difference ofso量1簸1ineraiogy except心oG℃一δ1旗、_幡、す飛、>for medium dense and dellseToyoura s&nd。Moreoverりρlof cemenトtreated sαndy soil cαn be formulated by aa ce臓lent conteut=5、5%exponelltial function of relative dens圭ty.The relationship0,4between lnitial relative denslty andμof tlle three cement−1目一i  l l日I IFl  1  目IllI        ラ       3       410   10闇  10   10treated s&ndy soils witb cement coatent of5.5%was very   MeaI}ef艶ctive stress(kPa)slmllar to that of the two mtreated carboaa重e sands(Do9’s Bay sand an(I Quiou sand).It is part星y because(c)Toyourasand,ceme!1tcontel曳t55%carbonate generated by the chemical reactioll of cementand water has the same chemical composltion of car−bonate sand.  Figure5showsthevalueo朗(;1一κ/λ)asaf’u正1ctionFig,3.粟soIropiccompressioubehaviorofceme賦¢確re謎亘edsandyso藍ls 26KASA+¥・IA ET ALlO-UNDRAINED SHF.AR STRENGTHrT T T l T1: [ ! '¥J'O: r? oura s nd solid' un reated: ' r ijl'Rokko !1asatio IEffecth!e Stress Pat/1 and Int,'oduction of Cen7entation:i A: A ita sand j ///O; L ' s i/ ';': 104,Parcanetel 'Figure 6 shows the effective stress paths for undr'ainedcemt nt contc t' _ ')o/Oshearin_ z of cement-treated Ariake clay specimens consolidated into normally consolidated stress range (p' >7:'" Q,uiou s nd; r>1lO'Lfter =asuftlku et a(' 1 999 )) j ;rltnlrp(.). The critical state line of untreated Ariake clay (M=1 .49) ¥vas also sho¥vn by a dotted straight line. The resultssho¥v the effective stress paths for the same cement con-tent and confining pressure seemed to be identicalirrespective of the initial ¥vater content (¥vith the exception of tests of p' = 300 kPa, Fig. 6(a)). The peak deviatorDo : s Ba ::'aJldv:'af ef Kwaetl { 1 999v'__El02 o1 [ !_--f20 40 60 80-20stress, q*.., occurs at an obliquity (q/p') substantiallylOOInitiai re]ati¥'e density ( ,b)higher than the critical state line of untreated clay. Thedifference bet¥veen these "failure" states and the criticalstate line of untreated Ariake clay increases with increas-ing cement content.Fig. 4.C_onsoiidsrion yield stress and initial relative densit .'Fi**ure 7 summariz,es the failure states (pf, qf= q***) fornormally consolidated cement-treated Ariake clay. It canbe seen that the failure envelope for a given cementcontent is parallel or slightly steeper than the CSL ofuntreated Ariake clay in confining pressure up to 2941,lI¥,O.9 L'o( 4vA viVOpreviously found that the inclination of failure envelopewas almost the same or sli_._."htly larger than that of untreated soil, and the up-ward shift of fai]ure envelopeACo¥. 0.8 LkPa. Clough et al. (1981) and Zen et al. (1990) haveOAriake clayC:]Akita sandARokko MasadoKaolin after Ue et al. (1997)hand, Lambe (1960) and Ue et al. (1997) had presenteddata ¥vhich sho¥ ' that both of cohesion intercept andTokyo bay c,layinclination of failure en¥'elope increasedafter CDM association ( 1 99 1 )cement content. Therefore, it ¥vas commonly expectedToyoura sand0.7 LO.6 ioVA5 10 i520Cement content (o/o)Fi'lF.vascaused by the cementation effect based on the results oftriaxial compression shear tests for cement-treated sandysoils. Gou¥'enot (1997) also r'eported same experimentaloutcome for chemically grouted sandy. soil. On the other5. EfTect of cement content on compression properties st pre- andpost-¥.' ietding¥'ith increasin_,_that the apparent cohesion of cement-treated soil ¥vouldincrease ¥vith cement content, ho¥vever, it ¥vas unclearhow the frictional stren**thvould be affected by cementa-tion. One possible reason for different effccts on thefrictional stren*'th is that cement mixin>' not only generates cementation bet¥veen soil particles, but also suppliesfine grains to the untreated soil through very fine particlesof cement content in order to evaluate the chan*'e ofin the cement agent. To a first approximation, the currentstiffness at the yield stress. It is noted that ;, and /C are thesug*'ests that cement-mixing affects only to apparentinclinatlon of e-Inp' in normal consolidation and quasi-overconsolidation regions, respectively. The data arecompared ¥ *ith prior studies for cement-treated Kaolinand Tokyo Bay clay (Ue et al., 1997; CDM associarion,cohesion of undrained shear stren9:th.Assuming that failure envelope of cement-treated soilin p'-q space is parallel to that of untreated soil, theapparent cohesion can be characterized by an equivalent1991). It ¥ !'as observed that A gradually incr'eases to¥vardtensile stress, p'*, in Fig. 8. It can be seen that p* increases1.0 in proportion to cement content, ¥vhich confirms thelinearly with cement content for the ex'periment r'eportedhigh stiffness of cement-treated soil for stresses p' <p(. Inin this study. Schnaid et al. (2001) proposed a similarapproach for interpreting drained triaxial compressionsshear tests on cement-treated sandy soiLaddition for specimens ¥ 'ith same cement content, the Aof cement-treated clayey soils ¥¥*as slightly larger thanthose of cement-treated sandy soils used in this study.The effecti¥'e stress paths for undrained shearing in thequasi-overconsolidation region are sholvn in Fig. 9. Thedata are normalized by the consolidation yield stress p+'..It can be seen that the initial effective stress paths forq/p' = O.'_O.4 are approximately vertical suggesting thatthe soil can be treated as an elastic material (no shear- 227UNDR.へIN琵D STRENGT村OF CEMENT−TRE、へTED SO至LS350400      C.S L      ・fun毛reateds・臼一11・, ノ  ini樋al /  、∼ater con{Cm 350                         /                    ×※響◇斗◇/        1、5WL               /ぐ美300  一 一 2。0料・L.)o 250ののののqりoの\       〆   〉       ブ   /150_    / \、\      〆’、 /鳴     /、、/   \>の100一 ,〆ノζ、    \      ■♂璽響Q 200’ヌ150澱\’                  ・ 一【二]一 一 39/b\・○v    !/50_ノ/   \      へ   〆/           ず          ぞ          ゲo 100     /.3〆      一⇔一5%50  /            十7%        ぐ)=装ntreated      ×    109も   1/    /O          O      50     100     150    2000   50  100  150  200  250  300  350Mean effective stressρat failu…『e state(kPa)    Meanef£ectivestressρ(kPa)(a)Ariake clay.cemellt colltellt5%Fi9・7。Failure s重旦te in難orm旦Ily consolidated s垂a重e350     C100,S.L              /  initialll欝◇\』讐 1(じ3謡802劉/詠、/\冬1111,◇/\、\\\  \    i、/  \    i〆   1   0      …l  撃60             「            ρ一グr邸詔○ε40瓢、!   ∼2二20Φり○   1100   50  100  150  200  250  300  350     Meanef罫ectivestressρ(kPa)        Cemelltconte鞭t(%)(b)Ariake cl&y.cement con重ellt7%Fig.8。350Cemem撮ionparame{erPlandcemen載co服te甑 C.S,L of            /300 しmtre&tedsoil 込 /      \』     / initialwater謁250腎200のの㊤Qりo讐ll\,/ 2・Minduced pore pressures).丁簸e e貸ecξive strength envelopeis bounded by the tension cuトoff and by the response ofthe normally consolldated samples.丁紅e characteristicsare very similar to undraiIled shear behavior of mec熱ani−     ,      /ca1豆y overconso豆idated cohesive soil spec玉lnens.    /     /150    !    ■   //   \o 100>(  ノ   ノ                  え,’//     \、//       /                      ヒ50Unゴ1πinθ4 Shθαノ’S!1ぐθng∼h iη 71θ17η5「oゾ Co1ψningPノ’ε55〃1’θ〆/                          i0 〆 l   I     10   50  100  150  200  250  300  350     Mean ef旧ectlve stressρ(kPa)  Figure IO shows the relationship between tぬe u盤一dralnedshearstrength5、(;σロ,、、、/2)ofcemenレtreatedAriake clay and the consohdation pressure,1プ。,ln thenormallyandquasi−overconsolidatedreglo11s。丁紅efailureenvelope in normally consolidated region is esξimated byconnecting the fa録ure s亡ate量n雛1e conso1圭dation pressuresρ≦>ρチ.τhe undra呈ned shear strength ratio concludes,(c)Ariake clay,cemellt colltent lO%5u/ρ5凱o.4−o.45.丁致eu曲ainedstrengthofquasi−over−consolidated specimens has a much Iess increase byμ.Fi9。6、E∬ec{ives吐ressp甜lsfornormal畳yconsolid9亘edspecimensofTheI畠e is a signif}cant infiuence of the initiαI water content  Ariakec皿犠y(P‘>ρ1)oll the undrεtined shear量ng to quasi−overconsolidated _28KASAl¥・IA ET AL200l '4v : consolidation yic]d stresscement content: 50/01 j .2・ '/'_':1¥--- - l'5lt'Ltcns on cut/ ' f' _ 2.0Tt'l-I !rlf¥¥¥( l -1 50-efJr /! "' ¥i ! '1' ¥.*¥s":;p. i OOL1;I r!f ' ¥¥e)S'r f! "¥¥.ntent: 59・)l;: 50L--Ar- l'5Tt'L"**c ., ;'・iSOol OOPI P¥1400300(a) cernent content 50/0300f cement content: 702 =200Consolidation pressure p t (kPa)(a) Ariake clay, cement content 50/01 '4Hll- 2'o t'L2 (ht'r !!--- l.)ite slon cut!l - ¥¥!vL- 2.0Tt;L'-'5Tt'i : cor;sol]dai one]d ng tressv ___ 7ic:s 250' !_ C>r' 1/ > . ' 1'i :: l¥!' iI* t'L:; 200rbO'8!- !: '・' .(/ ?: 'l ,j ' _O_6f '. ¥cs'. '.' s2 olt/ ceme i1501_I/ -O- I 51$'L; 100O 4 ' !',*!'f v'c; 50=0'2,-.} !! ':DO'. i+[_. - - 2 Olt'L' 2 5lsi Hll- 2 )ItLOO O.2 O.4 O.6 O.8 i I ' 1 4op/p,lOO 200 300 400 500i '4cement content: I Oo/o!;_tens ion cut!¥¥¥ !r(l :¥::¥¥__ l'7TvL2.0Tt'L'_'5lspL:/(' ' , ¥ [j : Il' I ¥ f!i/':i:I}/;':4¥¥¥l}!' !:: : I :i lOO O.2 O.4 O.6 O.8 1 1 ' 1 4p! p+(C) Ariake clay. celrlent COntellt I OoloFi('./ ' I i:consoi datlon leld suess/:;::400 I. _. _ .' -f7l ' i 7lt'L-c/'['i'J OO L9. 'orma ized efrective stress paths in quasi overconsolidatedAriake ela)j 2 ols'LI - ":)rc;O'6 - ../1 1¥¥i .¥・"(b) cement Content 7010eJ)O'8r h600Confining pressure p* (kPa)(b) Ariake clay, cement content 7010l'2conteElt: 7P'b) 200 ' cementcontentl loo!b: *1v>I _-/;t:f!;: HCv 1 5lt 'S12'oll'L7}t'1:IL ' ' L 2'5ls'I -'-'- _l ' lOO ;200 400 600 800Consolidation pressure p(kPa)C Cenlellt()COntellt I OoloFig. 10. Effect of consolidstion pressure and initiawater content onundrained shear stren('*th of cement-treated Ariake cla ' 駆…229UNDR.AINED STRENGT}{OF CEMENT−TR狂、へTED SOILS1400−R隻、kkし、m乙tsad〔、・一一   ccm¢縦G。ntCn【55%     Ak聡》and=㊥△圖i200− 1』・》・u㈱耳1dぐ  rc[乙L【剛じ蝋》.〔)2・奪…L ㌦\?8謄一・メ〉27し!ioOO許瓢800←  、盆、.  1 團600L  I    一。     68ul     440.ム『         .・A『400』 A一.○・ 20。(』2001一  .O. ・σ  1。1 臨.・一」  0     50    …00   】50   200   250   300   350(・ns・lida芝i・11prcssu即[“(kPal      untreated  圭)    O   …   1   2茎3   46Vold ratioεFl9.1L Undr蜘edshe貸rstreng疑handconsoiidationpressurefor  sandy so蓑s(賃f芝er Ze醜e正aL,1990)Fl9,12,服ectonvoidratioonundrainedshears重rengthofcemen重一   〔re盆吐ed Ari旦ke ci謎yspecimens.From these experimental nndings,it cau beemphasized that the undralned sぬear streng由of卜cement−treated Ariake clay should be evaluated as a function6        Ariake clay λof connning Pressure toget紅er with cement contα1t。Tatsuoka and Kobayashi (1983) have a正so reportedslmilar experimental results from undrained triaxialcompression tests on cement−treated Tokyo Bay clay(cement content=8−20%).Their experiments correspondto quasi−overconsolidate(l state.Terashi et a至.(1980)andYajima et al.(重997)have presented重he fallure eavelopeofcement−treatedKawasakiclay(ceme飢conteat;10%cement−treated Akit&sand,Rokko Masado an(i Toyourasand. It was observed that 5u slightly increased withconfining Pressure and the failure envelope shifted to▽×  10%  i箋 1屈  1蝋0 r一⇔㌧o▽79・o命         命2 × × ▽㊥命 5%> )…’esultsshownlnFig。iO.draiaedshearstreng由5、、andconsolidationpressurefor celηent cOnten口逐04…  ⇔curve as毛he function of con且ning Pressure is similar to  Flgure1至summarizes the relationship between un−▽・二 iand the wi冨200%)and Yoneyama clay(cement amount瓢100kg/m3and wi=75%)which formed a bilinear        ×)め   ハX。救門         一一\   、                命1%C,S.L ofuΩtreated soiI         l   60   80 100        200    300Mean effective stressグat f包ilure(kPa)Fig,13. Criticai s芝飢e a賎d fa韮iure liほes of cemen琶reaIed Ari且ke clayupPer−si(ie with inclleasing re豆at重ve density。 1t can beemphaslzed重hat the stress states of cemen琶reated Akita The relatlonship between void ratio and mean e任ectivesand,Rokko Masado an(1Toyoura sand were consldered stress at fai夏ure is shown in Fig.13alld compared withto bein the quasi−overconso正idation region because oftheprior obseIlvation thatρ1くρ1.Matsuoka and SuH(i993)results for the critical state Iiae of untreated Ariake c正ay.It can be observed that the f&ilure state of cement−treatedsho、ve〔i sinli玉ar experimental results for tests on cement−soil呈nθ一1nメプ space was located above the crit圭cal statetreatedToyouras&nd(cementcontent;5.3一至4.3%)atline of un亡reated clay and the difference between cemenトconsolidation pressures up to10八厘Pa.treated soil an(l untreated soil increases with the cementUn41・αinθ4Shθσ1“S〃−θng∼h in Tθ1−1ns{ゾ『Voi4R躍io(Fig.至3),lncreases with cement content alld the failure Figure 12 illustrates the e狂ects oll cenlentadon onstate lines apPear to converge towards the untreated CSLundrεしined strengt紅as a ful1αion of the initial void ratioas meaH e登ectlve stress increases.However,cementcontent.Moreover,t簸e slope of the fa重正ul卜e state lines,λ,for Arlake clay.Although there ls some scatter imhe testresults,especiallyathighcementcontent(7%,10%),content圭s a key parameter contro豆1重ng undra圭ned shearstrengt紅a重玉ow collsol重dation pressure.there is a welレdenned correlation show圭ng that5、 in−creases witぬdecreasiag vold ratio.Whe鷺combined withtheexperiment&10bservationsin婁heprevioussectloP,山eundrained strength−dependency of cement−treated soi茎PRIMAL FACTOR O賊丁}{E STRENGTH This sectionconsideIls asimpli丘e(1appro&c勤to describec豆ay shou玉d be evaluated as a funcξion of void ratio,con愚婁he shear strengt紅of cemenトtreated soils.Prior resukssolidaξion pressure and cement content.illtroduced a reference tensile stress, 」ρ1 (Fig. 8), to 230K、へSANIA BT AL.200         ○ し1ntreated寿  塩1%   %メ    100−      H      p cernel笠cOlltentl      ←5%一10%for Ariake clay語一扁r5、5%f。rsandys。ils    パイマ。 ヒ    彰  1.騙L口日選151レあ_⑯奎雛・…i           雄044       「    00,100     200     300     4001V C田shab1¢cabOIlatesa畜1d  a負¢rl椥19註ndAirey(1993)     10  璽100Yieid stress ratio1∼    ρヴ(kPa)     c   rFig,15.∫、、/ψ1+ρ1)置ndyieldsIressr窪重ioFig.14, Re段rranged Fig、7by changing the horizonIai寂xis of(ρ二+ρ二)chaτacter量ze changes in cohesion of the soils due to thecement conte熟t.Figure l4 replots the llndraiaed shearstrength of normally consolldated cemenトtreated Ariakeclay as a function ofρ‘+ρ‘.The results show a linearsoil can a歪so be expressed by the same exponentia1至unctlon and paτameter(α=0.44,η瓢0.92)as cemenトtreated Ariake clay.These results suggest that theequivalent undralned shear strength ratio of cement−relation independent of cement content an(i initial watertreate(i soils can be estimated indepeudent of the soilcontentltype,lnitlal density(water content)and cement contenLThe prior results from isotropic consolid&tion connrm               5u             一    二α籍0.44          (1)             !フ1+μwhereα玉s the eq凛ivalent undlnaine(蓬streng1h ra!圭o of nor−that yielding of cement−treated sandy soils can be deter−mined as&function of the cement content and initiaIrelative denslty.Accordingly,lt ls supposed for one ofmally consolidated Ariake clay and(μ+μ)is referred tothe reason t熱at soi三type was not圭nfiueatial in determin−as‘‘cementation−enhanced”consolidation pressure. The undrained shear strength property in quasi−over−soils, although the correlations of consolidation yieldingyieldstressratioforatleastcement一τreatedsandyco烈solldated regio勤was examined by replotting theequlvalentundτainedstrengthratio,5u/(μ+!フ1)asastress of cement−treated sandy soil and clεしyey so量I shouldfunction o{the yield stress ratio,1∼,de負ne(i by: Fτom prev玉ous consideτations about Fig. 15, it is                  μ+μ               1∼駄一漏一一一一一                (2)                  ρ1+μThis yie1(i stress ratio is equiv&lent to t圭le overconsol量da−tionratioforcaseswhereμ=0、becla茄edlnfllturestudy,concludedt鼓atcementatlonparameterμandyieldstressrat量01∼are primal factors inf達uencing the un(ira量ned shearstrength of cement−treated so茎ls.1n other words,メフl and Rare parameters that cbaracterize the e貸ect of cementatlonand量nit量al soi1(iensity on the un(iτaine(i shear strength. Figure15shows that5u/(ρ‘+ρ1)is stronglyτelate(i toEspecia歪1y, the yield stress rat茎o R was important inthe magnitude of R irrespective of由e lnitlal watercontent and cement content.Moreover,the relationshiptreated soil at low range of conso1圭dat圭on pressures andbetween5し耳/(メフ‘+ρ1)an(i1∼ can be formulated using anestimat圭ng the undrained shear strength of in−site cement−mixing at high cement content、expone鷺tial function(similar to the SHANSEP relation,L&dd and Foott,1974):               SuCONCLUS正0藤S             一一一一一=α×1∼n      (3) Paperhasevε邊uatedthefactorsin釘uencingthestrengthwhere11漏0.92is t封e gradient of the functio聡plo貰e(i in lnof cemenレtreated so1ls based on results of lsotroplcconsolldation and undralned triaxial sheaτcompression              ρ1+μ5し、/(μ+」ρ1)一ln R space.覧quation(3)is very similar to thetests.The follow玉n黛conclus量ons are obtained:Undrained Stren9重h eVaIUatiOn fOr CO封eS董Ve SOII in OVer_ (1)Cement−tre&tedsoil勤asanomaHyconsolid&tedconsolidatedregion(∼litachlandKitago,圭976;Marthy line inθ一la1プspace that is dependent on the cementcontent similar to the conso圭idation characteristics ofetα1.,1982;Mayne,1980). 丁盤e test reslllts ofAklta,Rokko Masado,Toyoura andcohesive soi璽. The consol圭dat1on yield stress increasescrushable carbonate sand(Hu&ng and Airey,1993)arewiththeincreaslngcemen毛conte熱tandformationdensityalso plotted ln Flg.15by assumingμ=0.Tlle relation−ship between∫し、/(ρ絆グ)and R for cemenトtreated sandyof the specimen. (2)The undrained shear strengt}10f cement−treated so宝1』 rUNDRAINED STRENG'TH OF CEhIENT-TREATED SOiLScan be evaluated by dl¥'iding the stress state into normallyconsolidated and quasi-overconsolidated regions irrespecti¥'e of the difference of soil type. The undrainedshear strength slightly increases ¥vith the increasing con-fining pressure under the quasi-o¥'erconsolidation andthen increases at constant rate in the normally consolidated state.(3) A ne¥v parameter pf (Fig. 8) representing the cementation effect is proposed to repi'esent the cohesi¥'e compo-nent of' shear strength. Yield stress ratio, R=(p(,+pf)/(p.(+pf), represents quasi-o¥'erconsolidated stress state.The equi¥'alent undrained shear strength ratio of cernenttreated soils, s,,/(p.'.+pf), can be evaluated as an uniquefunction of' the yield stress ratio, R.(4) The general relation for undrained shear strength,Eq. (3), appears to represent the behavior for a ¥viderange of cement-treated soils¥'ith plotted parameters (xand ll.2314) G ouYenot. D (1997): S ate of rl e art in European grouting Lechnologies, P/'ocl 2nc! In!^ Con_f. Grouncf !,nprol'emenf Geosystems, 2S338 O5) Hayashi, N . Ochial, H_. Yasufuku, N. and Omine, K_ (1997): As udy on evaluation of size effect of cemem*treated soil. Proc'. 32nc!Jpn¥;at_ Con.f. JSS* ・fFE, 2409-2410 (in Japanese)_6) Huang. J. H. and Airey. D. ¥1', (1993): Eft ci of cemem and densityon an arrificiallv cemenled sand, Pi'oc. Ist In! Con.f Harc! Soilsc!nc! Sof! Rocks, 553560.7) Iro, T., iori, Y. and Asada. A. (1994): Evalualion of resislance toliquefaciion caused by earthquakes in sandy soil sta:bilized lvithquick-Iime consolidaied briqueue piles, Soi!s anr! Fbunc!ari0,Is,34(1), 3340.S) Kamon, 'l and Kaisumi, T (i999): Engineering properties of soilstabiHzed by ferrum lime and used for the application of road base,Soi!s anc! Fbunc!arions, 39(1), 3 1-419) Kasama. K , Ochiai, H. and Yasufuku, ¥lT. (2COO): On the sLressslrain behaviour of ligl tly cememed clay based on an extendedcritical slaie concept, Soi!s anc! Fbunc!ations, 40(5), 37-47.lO) Kuboi. Y, and Nishida. K_ (1999): Experimental studies on cementstabilization of soh clay utilizing ¥vaste rock powder as a supplemental malerial, JSC E, (631 /lll-48), 1-12 (in iapanese).l l) Kurihara. H., Kikuchi, K. and Fukaza¥va, E. (1994): Experimentalstudy on durabili y of artificial soft rock. JSCE, (486/¥'1-22), 8594(in Japanese).ACKNOWLEDGEMEiNTA _",_rateful ackno¥vledgement is made to ProfessorHidetoshi Ochiai, Associate Professor Guangqi Chen,Associate Professor Noriyuki Yasufuku and AssociateProfessor Kiyoshi Omine of Kyushu University f'or theirhelpful ad¥'ice and encouragernent. The authors also ¥vishto express their thanks to lvlr. Michio Nakashima, Technical Officer of Kyushu Uni¥'ersity for his expert support12) K¥vag, J. I., Ochiai. H. and Yasufuku, N (1999): Yielding stresscharacterlstics of carbonate sand in relatio! o individual particlefragmentation s rength. Proc. 2llcl In!. Conf・ Ell*"rg. C a!c(1reousSedimen!s, 79S6.13) Ladd, C' C and Foot , R (i974): 'e¥Y design procedure forstabilil) of soft clays, J. Ceo[ech^ Eng/'g.. ASCE, 100(7), 763-7S614) Lambe, T ¥¥*. (1960): A mechanistic picture of shear strength incla} , Res. C 'on.f. Shea/' Sri'engrh ofCohesil'e Soi!s, ASC'E, 555-580.in the laboratory experiments. The authors extend theirappreciation to Professor Andre¥v J, ¥Vhittle of lvlas-15) ¥Ialsuo, S. and Nishida. K. (1969): The properties of decomposedgranile soils and heir influence on portland cement slabllization,Soi!s anc! Founcla!ionj, 9(2), 3543sachusetts Instltute of Technology for his help in editin_"*,16) ¥. Iatsuoka, H. and Sun. D (1993): A constiLulive la¥v for frictionalthe paper.and cohesi¥'e materials, JSC E, (463/III-22), 163l72 (in Japanese)17) . iayne. P ¥V, (1980): Cam-clay prediction of undrained s rengrh,NO'TATIONASCE, 106(GTI 1), 1219l242p' :mean eff ctive stressIS) ¥. ,Iinamiga¥Ya. K., Honda. K , Seriu, h,1., Kudou, T, and Kanaoka,i. ( 987): Efl cti¥'e factor in unconfined compressive streng h ofsoiH-cemnt, Proc. 32nc! Jpn. 1¥,'a!. Co,rf: JSS!V!FE, 1925-1926 (inJapanese)_!)( :consolidation yield s ress19) ¥・Iinowa, T.. Ishizaki, H_, Sakamaki, K . Suzuki, S. and Takakura,q:!1 :AInclination of normally consolidaied line in e-In p' spacecolumn improved by soil stabilizer - esting Fesults on the se¥'enyears lapse after improving York-, Proc 33rd Jpn. !Vat. ConfJSS, fFE, 2203-2204 (in Japanese),inclination of quasi-o¥'erconsolidated line in e-Inp'(! :spacepeak devialor stress for the undrained triaxial compres-p::(1995): LongJpn ,'Vcl!. Conf. JSS!YIFE, 2203-2204 (in iapanese)undrained shear strength (sconf ning pressure at undrained triaxial compression testinclinarion of failure envelop in normal consolidation inR:n :pf) :20) ¥1ishima, N , ¥,lorimoio, Y , Imayoshi, H. and Kobayashi. H_cementation parametersu'+ pf:(1998): A s udy on long term character of the soil-cemenlslon tesp :( :s,,/( p;= l-,(/;-K:r ln'Lpdeviator stress= cf*・ il /2)erm streng h properties of stabilized soils, P/'oc. 30th21) ・litachi. T and Kitago, S^ (1976): Change in undrained shearsirength characterisrics of saturaLed remould clay due to s'ellingSo!!s anc! Fblmc!ations, 16(1 ), 4558p'-q spacecementarion-enhanced consolidation pressure22) i¥,liura, N., (urala. H. and Yasufuku, N. (1984): Stress-straincharacieristics of sand in a particle-crushing region, Soils anc!yield stress ra ionFbllnc!arions, 24(1 ), 77-89.23) lvliL ra, N . Horpibulsuk. S and .¥agaraj, T S (2001): IEngineerin_"..gradient in In ju/(p+pf)-In R spaceundrained shear streng h ratioREFERENCESl) Been, K_ and Jefferies,_ ,1_ G. (1985): A state parameter for sands,(;*eotechllique, 35(2), 99-1 122) C 'D¥* i Associalion (199 ): Ce,nen[ Deep ! ifj. !ng Itletlloc! -Desi* !1anr! Constri!cti0,1 !V!anua!-, 8 -97 (in Japanese).3) Clough, G. ¥¥* , Si ar, N , Bachus, R C and Rad, i¥'. Sbchavior of cement stabilized clay at higll ¥vater con ent, Soi!s andFbunc!a[ions, 41(5), 33-45,24) iurth}, ¥. i. K.. Sridharan. A and Nagaraj, T S (1982): Predicion of undrained strengih of overconsolidated clays, Soils anc!Fbunc!a!ions, 22(1), 78-8125) Okabayashi. S . Tasaka, Y. and l¥,Iaruya, E. (1999): 'The strengthde¥'elopmen of cemen -stabilized soil (pan 3), Proc. 34th Jpn !Vc!t.Conf JSS*t・!FE, 849-850 (in Japanese)26) Omine, K_. Ochiai. H, and Yoshida, N. (1998): Esiimaiion of(1981):Cememed sands under staric loading, J. Ceorech. Eng/ "・, ASCE,l07(6), 799817,in-situ strength of cement-treated soils based on a l vo- )hasemixture model. Soi!s anc! Fbunc!cYtions, 38(4). 1729.27) Pestana. J ,1. and ¥Vhiltle. A_ i (1995): (] 'ompression model fbr 232KASAMA ET AL.  cohesioniesssolls,Gθo’θ‘h吻τ’θ,45(4},611−63L28)Schnaid,B.F、,Prietto,P, Di へ’L and Co臓soli,N・C・(2001):  15.36)Ue,S、,F謎jlwara,H.,Takeuchi,J,,Fukuda,Y、,Sakal,丁.and  Charac【erizatioほofcemenτedsandimrlaxlalcompression,ノ.  Yanag漁ra,K.(1997):Mechanicalpropertiesofstab澁zedkaoll鷺  Gεo’ech.Eπviノη11.Eηg∼g.,ASCE,127(10),857−868.  claybycemenロypesolid盗er,/SC五,(582/11王一41),217−228(ln29)S蝕ibuya,S.,Tatsuoka,F.,Teacぬavorasinskun,S、,1くong,X.」.,  .lapanese),  Abe,F.,Klm,Y.and Park,C.(1992):Elastlc deformation proper−37)Yajima,J.,Maruo,S.and Ogawa,S.(1996)=Relation between  ties of geoma【erials,So’Z∫αη4Fα!η磁f’o’∼5,32(3〉,26−46.  fa1lure cr1terion unconf}ned compressive stre旨gth and inまt玉al vo1d30〉 Suzuk玉,K.(1990):Bst玉mation of compressive streng−h of ceme降【  trea吃edcoalashes,P/oぐ、25’hゆπ.Nα’.Co17弥1∬M圧,1945−1946  (玉n Japanese).3玉)Tang,Y.X.,Miyazaki,Y.andτsuc匿da,T.(2001):Practices of  reロseddredgi駐gbycemε照reatment,30’Z5σ11ゴFα〃1伽’0115,4雲(5),  129−143.  ratlo ofl1911t−weig姓t soll,∬Cだ,(544/11王一37),251−257(ln  Japanese).38)Yajlma,.L,Nagaoka,T.andTanizakl,S.(1997)IMecha葺ical  properties and fallure c旗erlon of職ormaily and overcoPsohdated  cemenトtreatedsoil,ノSCE,(561/11B8),205−214(inJapanese)・39〉Yamamoto,T,,Yamauc臆,T.and Horibuc厩,K.(1996):E猛ect of32)TaIsuoka,F,andKobayashl,A、(1983)汀rlaxlals[rengthcharacter−  isIlcofcemenureatedsoftclay,P1「o‘.8!12E1〃甲、Co厚.SM圧,1,  gralnsizecharacteris1icso朗1eeπectlve蹴essofce獄en面xingagem  421−426、  (541/11三一35),133−146(in、ヌapa跳ese)。  met難od for preventing1玉quefactio職of sand to silt depos註s,/SC万,33)Terash1,M,and Tanaka,H.(1981):Grou鷺d lmproved by deep40)Yasufuku,N.an(i Kwag,、」.M。(1999):Sign澁cance of soll part1cle  mixing method,P1『oc/0’h ICSMπだ,3,777−780.  fragmenta【ionstrengt簸relatedtosoiicrushabillty,P’門oc.1”∼34)■erashi,M ,Tanaka,H.,Mitsumoto,丁、,Nlidome,Y.and  Honma,S, (王980):Fu駿damental proper【ies of llme and cemenト  ,45’αηRθ9’oησ1Coη∫Sル∫G万,1,53−54.  treated so玉1s(2賢d report),R(∼poπo∫f1∼εPo1甲fα1∼ゴ∫ノ‘71カoε〃●R〔∼5θσ1℃1∼  tioR cめaracter1stlcs of cemenトtrea【ed sands used for premixlng  1〃∫1i∫〃∫θ,19(1),35−57(玉n Japanese).  meIhod,ノ吻01”ρ/’11θPoπα’14飾めα’ヂRθ5θ01「 ‘1∼∫175伽’θ,29(2),35) τ員e Japanese Geotechn圭cal Societ》f(1999):S’α〃4αrゴ50∫ノαραηe∫θ  85−118(i鷺,∫apanese)、  Gεo’θご1111’ごα1Soc忽ζγ,ブ01・乙θZ)oro’oり,Slrθαヂ刀ε5∼‘E’191∫51∼vθ13’onノ,41)Zen,K.,Yamazakl,}{.andSato,Y.(1990):Strengdla鷲ddeforma−42)Zen,K.,Yamazakl,H.,Yoshizawa,H.and Mori,K.(藍992)l  Developmentofpremlx1ngmeIhodagalnstliq賎efaction,辞ocgrh  〆1,∫’αη1∼θ910’1α1Co双〆l SAグノ7E,1,46豆一464.
  • ログイン
  • タイトル
  • Cyclic Resistance of Clean Sand Improved by Silicate-Based Permeation Grouting
  • 著者
  • Yoshimichi Tsukamoto・Kenji Ishihara・Keitaro Umeda・Tadao Enomoto
  • 出版
  • soils and Foundations
  • ページ
  • 233〜245
  • 発行
  • 2006/04/15
  • 文書ID
  • 20902
  • 内容
  • ,FSOILS AND FOUNDATIONSVol. 46, No., ,233245 Apr, 2006Japanese Geotechnical SocietyCYCLIC RESISTANCE OF CLEAN SAND IMPROVED BYSILICATE-BASED PERMEATION GROUTINGYOSHI ,11CHI TSUKA¥. ioToi), K Nil ISHIHARAii), KEITARO UIEDAiii) and TADAO ENOh・IoToi+)ABS1'RACTThe cyclic resistance of clean sand improved by silicate-based permeation grouting is examined based on laboratorytriaxial tests. Specimens were prepared by the methods of sedimentation and vet tampin*'. In the former method, drysand ¥vas poured into the silicate-based solution. In the latter method, grouting lvas conducted by permeating silicatebased solution through ¥vet-tamped nearly-dry specimens as ¥vell as through ¥vet-tamped nearly-saturated sand specimens. The overburden stress ¥vas applied on some of the grouted saturated sand specimens during a curing periodtypically of one month. For all of the specimens prepared in different ¥vays as above, the non-destructi¥'e measurements¥vere first conducted of velocities of P-¥vave and S-wave propagation through the samples prepared under varyingB-values. The aim of these measurements ¥vas to examine whether the small-strain properties of grouted sand specimens ¥ 'ith **elled soil fabrics can be evaluated in the general framelvork of the theory of poro-elasticity. The undrainedcyclic triaxial test lvas then conducted on each of the specimens. The influence of **routing on dry sand and saturatedsand, and the effects of sustained application of an overburden stress during the curing period lvere ex'amined dra¥vingattention to the inner structure of **routed sand and its effects on the cyclic resistance.Key w'ords: cyclic resistance, permeation grouting, sand (IGC:D61D7)Outer tubeINTRODUCTIONInner tubePermeation grouting has been developed for improvin_O,_sand deposits as one of countermeasures against soilliquefaction during earthquakes. Among various otheralter'natives, this technique has an advantage in that it canbe implemented e¥'en under difficult site conditions suchas reinforcing loose sand deposits under existing airportrunways or' around existing bridge piers. Application ofthis technique is also being explored in order to improveloose sands underneath storagae tanks so as to make themmore resistant to liquefaction. In the permeation grouting, the specially prescribed silicate-based (SiOl) solutionis used, which is produced by extracting alkali substances(a) Insert an outer tubefrom water glass and therefor'e environmentaily harm-(b) Insert an inner tube and pull out the outer tubc.(c) Inflate sleeve packers from bottom.(d) Inject siiicate-based solution ftom strainers.less. This solution is initially permeable enough to travelthrough soil aggregates and to gradually solidify intogel-like formation.In the typical practice of permeation grouting used inJapan, a main tube consisting of inner and outer tubes,equipped with packers and strainers, is lo 'ered into thebored hole, as sho¥vn in Fig. 1. The strainers are fixed atthree locations each about 15 cm apart along the surfaceof the main tube for ejecting the silicate solution throughthe inner tube encased in the main tube. In bet l*een thei]jiiiii ,Fig. 1.Operational procedure of permeation groutingtwo nerghbormg stlamels rubber sleeved packers arefixed ¥vhich can be expanded radically by pressurized¥vater from the outer tube encased in the main tube. In itsoperation the rubber' sleeves are frst infiated ¥vithin theborehole by sending pressurized vater' through the outerAssociate Professor, Department of Civil Engineering, Tokyo Urliversity of Science. Japan (.vtsoil@,rs noda.lus.ac.jp).Professor, dit o.Kyo¥va Exeo C*orporation, Japan (f'ormerly Graduate Studem, ditto)Graduate Student, ditto.'The manuscript for this paper was received for revie¥ ' on September 8, 2005; approved on january 13, 2006.¥Vrirten discussions on this paper should be submitted before November 1, '_006 to the Japanese Geolechnicai Society 4-38*2, Sengoku,Bunky0-ku, Tokyo I 12-001 l, Japan. Upon request the closing date may be extended one mouth.233 9 34TSUKA ,lOTO ET ALtube, thereby isolating the soil mass to be grouted. Then,the silicate solution is ejected through the inner tube topush silicate liquid out¥vards from the strainers into thesurrounding sand deposits. By this operation, the silicateclean fine sand ¥vith no fines, and its specific gravity isliquid is forced to permeate into a spherical soil z,one ¥vithprepar'ation employed ¥vere as follo¥vs.its centre located at the mouth of the strainer. After acurin_._" period of about one month, a chain of sphere-Tes! Series OG*=2.66 ¥vith a mean particle diameter D50=0.18 mm.The maximum and minimum void ratios are e ,** =0.973and e ,i =0.607, respectively. The methods of sampleshaped solidified zone is formed in the ¥'ertical alignmentIn the test series O, samples ¥vere saturated ¥vith ¥vaterfor one borehole. By carrying out the same operation atother neighboring boreholes, the tar_",_eted zone of loosesand deposits can be grouted and solidified. This methodis conceived as applicable to layers of sandy soils ¥vithfines content less than 350/0 Iocated do¥vn as dee as 20metres belo¥v the g:round surface.There are past studies addressing chemical **routing in.Japan (iMori et al., 1989; Hatanaka et al., ,_OO._ andothers), which has been used ¥videly to cr'eate impermeable layers ¥vithin grounds and to stabiliz,e soil massesalone and silicate-based solution ¥¥'as not used. Sandspecimens ¥vere prepared by the method of air pluviationduring: underg:round exca¥'ation. lvlori and Tamura(1986) addressed the permeability of sands stabiliz,ed by.¥vater glass-based chemical grouting, ¥vhich ¥vas sho¥vn to(A.P.), wet tamping (W.T.) and water sedimentation(W.S.). They ¥vere then nearly saturated by circulating:water and by controlling the back pressure for varyin*"levels of the B-value and tested in the triaxial testapparatus.Test Series AIn this test series, dry sand was poured into the mouldfilled ¥vith liquid of the silicate-based solution and madeto sediment to form specimens for triaxial testing. The¥vater used contained 60/0 Iiquid silicate. The silicatebe changing ¥vith shear' deformation and ¥vas examinedcontent of 60/0 ¥vas the minimum to form specimens¥vith reference to dilatancy characteristics of sands. Kagacompetent enough to stand by self-1veight. If the silicateand Yonekura (1991) examined extensi¥'ely the unconfined compression strength of chemically grouted sandscontent is belo v 60/0, it ¥¥'as not possible to form a speci-men. With this method of sedimentation, it ¥vas difncultto precisely attain targeted values of the relative densitv. .with a ¥vide range of properties.In order to clarify basic physical characteristics of thesand solidified as above, a series of laboratory tests ¥vereThus, specimens lvere prepared so as to attain a givenrange of the relative density, as accordingly sho¥¥'n inconducted on silicate-grouted sand specimens preparedTable 1. No overbur'den stress ¥vas applied on the speci-b)* different methods ¥vith or ¥vithout surchar__."e bein_"._mens during the curing period of about one month.applied during the curing period. In the labor'atorytriaxial samples, the small strain ¥vas applied fir'st in non-Test Series Bdestructive manner to monitor shear and compressiveIn this test series, dr'y sand ¥vas first mixed to attain 50/0¥vave ¥,elocities, and then cyclic shear stress ¥vas applied toies ¥vill be described in the following pa>・es of this paper.water content. The ¥vater used here contained 60/0 iiquidsilicate. The moist sand thus prepared ¥vas tamped uniformly in several la_vers ¥vithin the mould to achieve theMATERIAL AND SAMPLF, PREPARATIONthe mould the specimen ¥vas made to stand by vacuum inthe triaxial cell, and then, the silicate solution ¥vasevaluate the cyclic resistance. The outcome of these stud-relative densities of D* = '_5, 40 and 550/0. After removingFour methods ¥vere employed in this study for prepar-circulated through the specimen and it ¥vas left at rest foring specimens for the cyclic triaxial tests, as illustrated inbelng cured for a per'iod of about one month. With thisFig. 2. The variation as such in the method of samplepr'ocedure, it ¥vas found that approximately 750/0 of ¥'oidpreparations ¥vas intended to examine lvhether the cyclicin volume ¥vas saturated ¥vith the liquid of silicateresistance of solidified sand is affected by the characteris-solution ¥¥'ith the remainin*"_50/0 occupied by pore air.The characteristics of the test series ¥vere that the speci-tics of fabric structures produced by different methods ofpreparation. It lvas also intended to clarify vhether thesustained application of an overburden stress during thecuring period may or' may not exert any influence on themode of formation of geiled substances (gel-formation)mens ¥vere permeated by the silicate solution throughnearly dry sand, but vith no overburden pressure appliedduring the curing period as accordingl_v indicated inTab]e l.within soil fabrics and further on the cyclic resistance ofsoils.The groutin_ material used in the present study is thespecially prescribed siiicate-based (Si02) solution, vhichis produced by extracting alkali substances from ¥vater:lass using electro-osmosis. The typical gel time of thismaterial is about a couple of ten minutes to hour's.rn all series of the tests, soil specimens 60 mm indiameter and 1'_O mm in height ¥vere prepared usin_"..Toyoura sand. This sand is classified as poorly gr'adedTest Series CIn this test series, moist sand ¥vith about 50/0 ¥vatercontent vithout silicate lvas tamped uniforml)r jn severallayers ¥vithin the mould, and deaired ¥vater ¥vas circulatedthrough the specimens to make them almost tully saturated ¥vith water. Similar to the test series B, the specimens ¥vith the relative densities of D*=_ 50/0 and 400/0¥vere prepared. From this sta_"._e on¥vards, the silicatesolution of 60/0 ¥¥'as permeated throu_2:h the specimenThe r,PER iEATIO*¥ GROUTiNG O ' SAi¥Dd) ')-,)ov,. .Pe ne {onef C02Preparation ofsoil specimensfelio 'ed'・t1::.;.: ・lbde-aired'・ ・,*ater,.tt ・ ,ttt,,,,t{{,,,Permeation ofsilicate solutiont,'Curing period: '.Itt' ; 1 1 -.t',.Soil specimenswith voids occupied bygel-formationSoil specimensSoil specimens1; 'ith voids occupied bywith voids almost fuil)saturated withel*fomlationgel-forrnationlvhich is irnperfectlywithirl which ai bub iesare constrained(b)Test series B(a)Test series AFig. 2.saturated ¥1'ithreepore-water(c)Test series C(d)TeSt series DPreparation of soil specimens and tcst seriesspecimen vas then cured for a period of about one month¥vithout applyin_ the surcharge. The particular feature ofthis test series ¥vas that the specimens vere permeatedgrouting through ¥vater-saturated sand. The intention ofwith deaired ¥vater first for full saturation, and ¥vere thengrouted or permeated by the silicate solution, as indicatedthe sustained application of an overburden stress of 98kPa over the one-month period of cur'ing.in Table 1. No o¥'erburden pressure ¥vas applied during:the curing period.this test series ¥vas to see the effect of overburden duringthe curing per'iod. Thus, the specirnens ¥vel"e subjected toTESTING PROCEDURESThe testing scheme vas the same as that previouslyTest Series DIn this test series, the specimens ¥vere pr'epared In theused and described in detail by Tsukamoto et al. (200'_),same manner as in the test series C by employing theNakaza va et al. (,-004) and Ishihar'a and Tsukamoto TSUKA ,IOTO ET AL.236Table 1.Teslserieslest series and summary of response during measuFements of B-value, P-1vave and S-wave velocitiesSilicate-based:routin conductedD* ( /a) PreparationmelhodOverburdenstressRange ofobser¥'edB-value Type of Vp-responseNumber ofspecimens40 & 60 A.P.O 25 & 40No¥V . T:roulin :NoO- l OType-A: I/p changesvith B-¥*alue40 & 55 ¥V s20 - 306A 35 - 45 Sedimentation into silica e solutionNo0.1 - O 8Type-B: V* s aysaround 1650 m/s50 - 6067) 5BGrou ino_ throughdry sand40Noo - o.45Type-A5540?_5loNo¥V_T.Type-B9Groutin : hFou9:h¥vater*saturatedO - 1.0sandD66) 5C5Type*A3Type-B3Type-AlType-B)Yes40(1) n ihe preparation method, A P , ¥¥T T_ and ¥V.S. siand for air pluviation, ¥vel tamping and ¥vater sedimemation, respecli¥'ely.(2) OYerburdeu stress indicates ¥vhether the sustained application of an o¥'erbuFden stress lvas made during the curing period(3) Type of Vp-response: Type-A corresponds to the response ¥vhere the velocity of P-¥vave changes vith the B-¥'a}ue in accordance ¥vith he iheoretical expression (2). Type-B corresponds to the response lvhere the velocity of P-¥vave s ays constant around V,.= 1650 n/s irrespective of theB-¥*alue(2004). Figure 3 sho¥vs the cross section of a soil sample60 mm in diameter and 120 mm in hei..,*crht, ¥vhich is placedbet¥veen the cap and peciestal equipped ¥vith the tr'ans-ducers for P-¥vave and S-¥va¥'e velocity measurements.The transducer ¥vorking as a source of P- vave generationis a piezo-electrically driven bolt-clamped ceramic transducer, ¥vhereas the transducer for a receiver of P-¥vavepropa_g:ation is a piez,o-electric accelerometer, vhich isthe highest ¥'alue achievable in the specimen. Thespecimen ¥vas then put under the p-constant condition.In other ¥vords, the lateral stress ¥vas reduced while theaxiai stress ¥vas increased and ¥*ice versa in the cyclicphase of load application to maintain the change in themean principal stress equal to zero, zlp = (A(7i +2A(73)/3=0. The cyclic stress ratio, (7d/(,-(7 ), ¥vas applied un-drained ¥vith a frequency of 0.1 Hz. The p-constantattached to a dummy metal block plug:ged into the-condition ¥vas considered important particularly in thepedestal. A pair of bender elements lvas installed at theloading tests on par'tly saturated samples. This aspect ¥vascap and pedestal as a source and a receiver of S-¥vaveaddressed in the previous paper by Tsukamoto et al.propagation, respectively, tr'a¥'elling throu ( h the triax'ial(2002) .specimen. The velocities of P-¥vave and S-¥vave propagation could be obtained by monitoring the difference intime bet¥veen dispatch and arrival.The tests consisted of t¥vo phases, ¥ *hich are the non-MF.ASUREMF,NT Ar ITD ANALYSIS OF VI' AND V,The velocities of P-wave and S-¥vave propagation ¥veredestructive test monitoring the velocities of shear ¥vave V*and longitudinai ¥vave Vp, and the destructive test ¥vherecyclic stress ¥vas applied undrained leading to large shearstrains near failure. After finishing the series of proce-measured under varying B-values. This testing methoddures of sample preparation, the specimen ¥vas firstson's ratio v at small strains are one of the combinationsfor determining the elastic material properties. The shearmodulus G can be calculated solely from the velocity ofisotropically consolidated to a given confining stress (7 ,and the B-value ¥vas measured. The non-destructive test¥vas then conducted monitoring the ¥'elocities of shear¥vave V* and longituciinal ¥vave Vp under given B-values.Herein, in order to achieve a tar_cr,_eted B-value, the backpressure ¥vas controlled lvhile the effective confining stress¥vas maintained constant. In the undr'ained cyclic triaxialtest conducted after vards, the B-¥'alue ¥vas increased to¥vas found to be useful in Infer'r'in : the basic small-str'ainproperties of soils, based on the theory of poro-elasticity.The t¥vo parameters of the shear modulus G. and Pois-S-¥vave propagation ¥vith the simple expression of G.=pV , and has been found to be independent of the Bvalue as sho¥vn in the follo¥vin section. On the otherhand, based on the theory of vave propagation through aporous medium by Kokusho ('-OOO) and Tsukamoto et al.(2002), the o¥'erall Poisson's ratio v is expressed as +'T3 -OC}(1 - 2vb)(1- - - -B)'-¥-..-}p. ..'}237PER ,lEATION GROUTliNO* OiN SAND2000Piezo-electrlc'flceramic transducer' 111Test series O11Oj lA ir pluviation!;*";;; ;;,14,"*--***;x/.+*'・ ;... *!・*・SToyoura sand (Df300/0, , a'(f98kPa)* ; i;・iSBender・*element! Vs;*,/, *t'l// ll'*Porous dlsk4*,; * *・ ・・ !* ;; ,;,, ;, ;**-' f"/ ll/ llJ / llll/dL '2( l+vb)JV 2=V [s 3 3(1-2vb)(1-B)p1000・;';II+l'J,,/'flhj// Ilimembrane. ;*#,.,.O Vp1500RubberIII/ /vb=0 '4 vb=0 ';5ll/ //// ll__-¥._.・ --_ /_//>- r I_ -¥500 -_ -JJ_C- -I)C Cr¥l : L (a)¥l" Samp leVs=2 1 2 (m/s )Orfuri+L'o0.5Bendere.1ernent1 -"Vb;sO'4 vbs;O'35 -*pedestalc1i' "";' is/cl)";^ "';" "?="';'"*s." '*i* '= "'*"'=' "'i'i ' '';;:'"' '* "' '* *; -"""'--"i""*"''';" 'i ; i :;';;**("f**"': 'SS;:ie ji< i" ';i:/!1g .///*_/ ///o.31v)OC 1'; ,: !#;s/////O; ' ' !; ; /(/ 1 = S';; ""' ;";'O._'O_ :lr _' i.....SSS; "'F'* / / />O**+ , ;, ".,*... ・**,'・・ *./////¥,_¥//___'_ e_tf/r'//////LL, t0.4・*,+ ・ *,0.2 O.40.6 O.8 lB-valueo¥JL 3i3 ¥ b ( I *2¥b)Bv;::3 - ( I -2v )Bo.2* ^ S.cee zeo-l/' IP'ectnc/ rec e eriV)>O0.1(b)Fig. 3. Cross sections of the cap, samples and pedestal , rth trans-ducers (after Tsukanloto et al., 2002)oofunction of the B-value in the follolving form,Test series OAir pluviationT'oyoura sand (Df30"/., , cr'(f98kPa)O.2 O.40.6 0.8 1B-value( Vp/ V=)2 - ,_ 3vb + (1 - 2vb)Bv=2(Vp/V,)2-2= 3 -(1 -2vb)B ' (1)where vb is the skeleton Poisson's ratio pertaining to theFrg. 4. Results of non-destructive tests in test series O (D*= 30?・ ): (a)Vp & V, aga nst B aiue and (b) oTerall Porsson s rat o agalnstB-valueskeleton deformation of the medium. Note that theoverall Poisson's ratio, v, is associated ¥vith the deforma-lar set of data f'or t.he relati¥'e density of D,=400/0 istion of the skeleton and pore ¥vater. In other ¥vords, vpertains to the undrained deformation. It is also possiblesho¥vn in Figs. 5(a) and (b). These test results ¥vereto infer the value of vby inspecting the relation of Vp/ V,(200,_). ¥Vhen the soil is f'ully saturated, the velocity ofagainst the B-value obtained from the series of tests ¥vithreference to the expression as follo¥vs,P-¥vave takes a value slightly greater than the value forthe compressional ¥va¥'e through water. By examining theplots in Figs. 4 and 5 Ivith reference to the theoreticalrelationship of Eq. (2), it may be reasonable to assumethat the skeleton Poisson's ratio v takes a value of about+ v )( V 4/_7(1V,)13 (2)In ¥vhat follows, the relation of Vp and V= against theB-¥'alue is adopted in pr'esenting the data of the tests.Figure 4(a) shows the results of the test series O in termsof the plot of the B-value against the velocities of P-¥va¥'eand S- vave propagation, Vp and V.*, for the case of nongrouted Toyoura sand ¥vith the relative density of D*=presented in the previous paper by Tsukamoto et al.0.35 for Toyoura sand.Fi**ures 6(a) and (b) sho¥v the plots of the B-valueagainst Vp and V as well as v obtalned from the testseries A vith D*=20-300/0. Another set of data for thetest series A for the relative density of D*=35-450/0 issho¥vn in Fi**s. 7(a) and (b). In these series of tests, speci-300/0. The plots of the B-value against the overallmens ¥vere prepared by pouring the dry sand into thePoisson's ratio, v, are also sho vn in Fi**. 4(b). The simi-silicate-based solution, and lvithout furt.her permeation : .b;0'45: ・;i:;1::o: /TSUKAN,10TO ET AL.23820002000I' ITest series OAir pluviation(a)l Ifll')f0 (> - - QVp* 1 600 -・ 1 700(m/s)11Toyoura sand (Df400/0, . (T'(f98kPa)l ',IO Vp CI! Vs ¥rl -1500e4 2( i vV2=V2 -+[Iooopds 3 3(1-2v )(l-B)ll/ ll!>_Or>500,e} C_'-¥,* 9._ :-/1000;/ ///// ///_//O T1_¥._."cO_////¥'b=0 4 vlfO.35(>(l,,=I ll!'J=Test series A>*P- eIx)::f1500J1492(m/s),,(m/s) 711,/ tHSedimentation into siiicate solution 'Toyoura sand (DpZ0-30"'., c: '0=9SkPa)o Vp Cl Vs500. (a)vb O '5 vb=0.- n J[: r*[I- IVs= I S 2(m/s)/rlC:= L ' !LJdlc- -} - - - - -Vs=2 1 6(m/s)ooO O.4 0.6 0.8 lo0.2DO.4 0.6 O.8o.,_1B-valueB-value0.5Yb=0'4 ¥'b O :))0.4oc:1)0,3:lpOleee 'l/_el O'- '/ le'el Q//// / / //I le¥lIV OL1)3(-2Vb)BOe>o/ /ce:/-5:o*v) 0.4//¥vb=0 45/ 3vb+ l*2vbB3-(1-2vb)B(,)v:::c (o.2ce.e:$5)>>O0.5(,/ 3 Vb ( V=I -2Yb)BOv)Op*e l__ eell_ 'l-ll//_'/eO>,:; llO. 1O(b)(b) Toyoura sand (DF40 . , cF'0=9SkPa) IoTest series ASedimentation into silieate solutionTest series oi Air pjuviation lO O.2 O.40.6 O 8 1B-valueFig. 5. Resillts of non-destructive tests in tesseries O (D*=40 ): (a)V p & ,・', against B-vaiue ond (b) overall Poisson's rario agninstB-valueT0.1/0ura sand (Df20-300/0, c '0=9SkPa)0.3o0.20.4 0.6 O.81B-valueFig. 6. Results of non-dcstructive tests in tcst series A (D* = 20-300, ):(a) V, & V. against B*value and (b) overal Poisson's ratio ngainsB-valueor grouting of the silicate solution, the specimens ¥verehighly likely that, in the case of the silicate-sedimentedkept under an atmospheric pressure during the curin_sand ¥vith enough liquid, ¥vell-developed structures ofperiod. The results of the tests sho¥v the ¥'alue of V*=15,_ m/s and 171 m/s for the relati¥'e densities of D*='_O300/0 and 35450/0, respecti¥'ely. Similar to the case ofvater-saturated sand sho¥vn in Figs. 4(a) and 5(a), thevelocity of S-¥va¥'e propagation, V*, for the silicatesedimented sand is sho¥vn also to stay constant regardlessof the B-value, as sho¥vn in the plots of Figs. 6(a) and7(a). It is easy to under'stand that the shear ¥vave effective-ly propagates through fabrics of soil aga_regates, and its¥relocity is independent of ¥vhether the void is saturated¥vith ¥vater or occupied by gelled substance. Holvever, the¥'elocity of P-¥vave propagation, Vp, appears to be con-stant at about Vp= 1650 m/s regardless of the B-¥'alue,except for some of the test specimens lvhich follo¥v thetheoretical relation of Eq. ('-) assuming vb=0.487. It is'*elled substance prevailed ¥vithin fabrics of sand parti-cles, throu*'h which thevave may travel primarily¥vithout interaction ¥vith pore ¥vater. Therefore, the valueof Vp became constant at about 1650 m/s.Figures 8(a) and (b) sho¥v the plots of the B-¥'alueagainst Vp and V* as ¥vell as v, obtained from the testseries B ¥vith D*=250/0. The set of data prepared by thesame method is sho¥vn in Figs. 9(a) and (b), for the caseof the relative density of D,=400/0. In these series oftests, silicate-based permeation or grouting lvas performed on ¥vet-tamped nearly dry sand specimens ¥viththe relative densities of D*=250/a and 400/0, and the soilspecimens were kept under an atmospheric pressure ¥vithno surcharge during the curing period. It is seen in Figs. 8and 9 that the B-value ran_"..es bet¥veen O and 0.45, and :.PERN'IEATIO.N GROUTING ON SAi¥'D20002392000( :0 : {._( p._.1500l!Vp= 1 600 - 1 700(m/s)( a)Test series BocGrouting in wet-tam ed near-dry sand._. _ _ _8 o(>ll71Vw*1492(m/s):VP' ,'=V[+ 2(1+vJll1000Test series ASedimentation into silicate soiution500500'Toyoura sand (Dr35-45 0, c;'I 98kPa)C io0.5p erVs=20)tm/s'o0.20.5O_1 -¥ice0.4 0.6 0.8_ ._ vtfO.4g7Lii1B-valuee ;1i:, : i:vb=0 5>oJ3 3(1-2vb)(1-B)o0.40.6 0.8 1B-value0.2_ / ¥J Li ij v2=V.2[・-・4- 2(1+v( a)o Vp C! sD { h - , O{II Vs*}-1 7 -1 (m/s)-oll_ ' (_/_ # // ¥io ood)o3 3(1-2vb)(l-B)d;llll_CrO>llo Vp n,_ Vs15001000!!Toyoura sand (Df250/0, c;'0=98kPa)Yb=0 4871,,>oc* ',._p, ,,__ ,= ..:o'c:s3vb ( I -2vb)B;1)V::ov). .'':')3 - ( I -21 b)B:ocl)//) 0.4'o O 4'53 vb+(C (V s$3-(-2vb)B-2Vb)Ba)>OOTest series BGrouting in wet-tamped ne r-dr¥.' sand'rest ser{es ASedimentation into silicate solution '(b)Toyoura sand (Dr=35-450/0, (;'0=98kPa)o.3o(b) i Toyoura sand (Dr=250/0, ( '0=9SkPa)o0.4O.6 0.8 1B-valueFig. 7. Results of non-destructive tests in test series A (D*= 35-4501 ):(a) Vp & V* against B-value and (b) overall Poisson's ratio againstB*valuedoes not go up beyond 0.45, because of the dry nature ofthe samples. It may be seen also t.hat the ¥'alue of Vpapproximately follo¥vs the theoretically deri¥'ed curvesassumin_"._ the values of about vb=0.43 to 0.44. Since theB-value is the pore pressure response during undrainedisotropic cornpression, it can be said that the B-value ofthe specimens prepared as above is incapable of risingabove the value of 0.45, due to a f'airly large amount ofair bubbles entrapped ¥vithin the voids. The primary ¥vaveis therefore thought to propagate through the solidifiedstructure of gelled substance ¥vith trapped pore air bubbles, and the skeleton Poisson's ratio of such inner structures takes a value of' vb=0.43 to 0.44.Figures 10(a) and (b) sho¥v the plots of the B-valueagamst V and V as well as vb, obtained from the testseries C* ¥vith a relative density of D* = '_50/0. The same setof data for the relati¥'e density of D*=400/0 is shown in0.3o0.2O.4 0.6 0.81B-valueFig. 8. Results of non*destructive tests in test series B (D*=25): ( )Vp & ;( aga nst B Ialue and (b) o erall Porsson s ratro agalnstB*valueFigs. 1 1(a) and (b). In these series of tests, silicate-basedpermeation was performed on ¥vet-tamped nearly saturated sand specimens ¥vith the relative densities of D* = '_5o/o and 400/0, and the soil specimens ¥¥'ere kept under anatmospheric pressure ¥ 'ithout surcharge during the curin_"_period. It may be seen that the value of Vp stays constantat about 16,50 mls, and is independent of the B-value.From this observation, it can be said that there seem to bewell-developed solidified structures of gelled substanceoccupying the voids continuously from the bottom to thetop of the soil specimens, through which the primary¥vave was transmitted. On the other hand, the B-valueranges fully from O to I . This may be due to the fact thatthe structure of gelled substance vas existent to_"*,ether¥vlth partly saturated lvater, and the latter must ha¥'eacted independently as a rnedium to r'espond to porepressure gauge, producing various B-¥'alues. In fact, TSUKA*¥,iOTO ET AL.240・_ooo・_oooTest series B ! 'lVp= 1 600 - 1 700(m/s)( a),,Grouting jn wet-tamped near-dry sand / /o Vp i Vs ll15001500Vw-1492(m/s) Jf/ / !/:ellSll1000lll <tii-d ;1000d// ¥¥¥l:L iit:I>500s' [Vpl =V*( a)2( l+v+3Test series CGrouting in 1 'et-tampednear-saturated sand-J500Toyoura sand (Dr250/0, cT'o*98kPa)3( I -2vb) ( 1-B).TII^1rH: i]Ti_! h= io Vp1 Vs]:;}Vs=204(m/s)ooB-value0.5'eee/// ¥// LVb on43 & Orf44o0.43Vb+(V=o o.2dee O1etpii;:ee t::://¥/vb:;O '45/// //¥i::ol)1)' O 4/43 l ( 1'2¥b)B0.8B-value>o:''12¥b)B::CI} :}O.4 0.60.5h/')/1)o: (t: i:;L::/1'>o}CC!Q¥ Ve=2 1 O(m/s)O.8 10.4 0.6o.・_i ooo_ ) ) )c ) J)Toyoura sand (DF400/0' o'o*98kPa) / /3Vbt( 1-2¥'b Bv:::3-( I -2¥'b)B' *c:s.ceS>e)>OOTest series B(b) jo.3oo.?_Groutin_Test series CGrouting in wet-tampedin wet-tamped neaF-dry sand(b)Toyoura sand (Dr400/0, c;'o*98kPa)0.4 O.6o.3o.8B-valueFig. 9. Results of non-destruct ve tests in tcst series B (D*=40 ・ ): (a)J/' & 1 against B Iallie and (b) oleral] Po]sson s rat o aga!nstB-vaitreo o._,near*satu,ated sandToyoura sand (D 250/0, O'o*9gkPa)0.4 0.6o.81B-valueFig. lO. Results of non-destrucrive tests in test series C (D,= 250,lo): (a)Vp & V> against B*value and (b) overail Poisson's rario ngainstB-value¥vithout any interaction ¥vith the air in the pores, throughmain difference appeared in the P-1vave response asthe continuous path formed by the gelled substance, theprimary vave seems to ha¥'e propagated.Figures 1'_(a) and (b) sho v the plots of the B-valueagainst Vp and V* as ¥vell as vL,, obtained from the testclassified as belo¥v.series D ¥vith a relati¥*e density of D* = 250/0 . The same setsimilar to those observed in the test series O and B, ¥vherethere ¥vas no grouting or grouting into neari¥_" dry sand.of data for the relative density of D*=400/0 is sho¥vn inType-A Respollse.'P-¥va¥'e velocity changes ¥vith the B-value, ¥vhich isFigs. 13(a) and (b). In these series of tests, silicate-basedgrouting ¥vas conducted on ¥vet-tamped nearly saturatedsand specimens and the.v ¥1'ere subjected to the sustainedapplication of an over'burden stress of 98 kPa over thecuring period of about one month. Although each specimen ¥vas prepared under the identical procedures, there¥vere t¥vo different types of response that ¥vere obser'ved ineach of this test series. In each type, the shear ¥va¥'eresponse ¥vas the same, ¥vith a ¥'alue of V,= ?_09 m/s forD*= 250/0 and a value of V, = 2,_O m/s for D*=400/0. TheType-B Respor7se.'P-¥vave velocity stays constant at about Vp = 1650 m/s,and is independent of the B-¥'alue, ¥vhich is the responsesimilar to that observed in the test series A and C, ¥vherethere ¥vere abundance of ¥vater and silicate at the time ofthe sample preparation.The difference in the obser¥*ed response bet¥veen TypeA and T"vpe-B is illustrated in Fig. 14. It is to be mentioned that there is no definite reason accountable for a ;T,.g j_ as ;7p-p/ SS' : ' ipPER 'IEATION GROUTING ON SAND・_412000・_ooo/x ;( *).81500lj/Vp*1 600- 1 700(In/s)- ,VP1::'s- - # ;;;i< '<';i'{'==.'.{ii{{i/_> :e i' *1000=Test series C1'et-tampednear-saturated sandO_47o o._-cl'Type*Bo VpVse Vp¥oe e o3 v +( 1 1'2¥'b)BY::::3 -( I -2v )Bce(rP'_'t ];t:L) -( I I : - ->o/ce(,):/c/) O.4.//0.3o o._Fig. Il. Results of non-destructive tests inest seFies C (D*= 4001 ): (a)¥ L3 v5+( I -2vb)Bv=3 -( I -2vb)BType of responseTest series DGrouting in wet-tampedOToyoura sand (Df40o/o, (;'o*98kPa)0.40.6 0.8 1B-value/////(b)>Test series CGroutin5a in lvet-tampednear*saturated sand1/1)>E L;// /ce)(b)Vs0.8.vb O 47 & O 48cl)') 0.4*5OlB*valueoCl?L0.4 0.60.2// / /¥ vb:/ O'joL('Type-AToyoura sand(Df250/., o'0=98kPa)o0.40.6 O.8 lB*value_/¥ ii//¥Jl ,.Lc s' (a)* O:1:: }e oe>ovw: 149 _(m/s)Type of responsenear*saturated sandeured under pressure5000.50.5'Test series DVs*2 1 3 (m/s)o/yGrouting in wet-tampedo Vp C] Vs- C : I* }';' - " r vbc>Toyoura sand (Dr400/0, e:'(f98kPa){_j'・ ' ;;,;d;Grouting" in' i ; 1 crtx)o o ri$ /;S500:4s #'Type-A : 3 samples ./':ec/'/ a1 500C 1000""s i s;ifo I_Co:>J ' =1 600 'v 1 700(m/s)near-saturated sandcured under pressureTyp e -AToyoura sandO(Df250/Q, cr'0=9SkPa)0.3o0.20.4 0.6Type*B0.81B-valueFig. 12, Results of non-destrnctive tests in tcst series D (D*= 25?・b): (a),', & ,i against B-vaiue and (b) overall Poisson's ratio againsiB*valueVI' & V* against B-value and (b) overall Poisson's ratio againstB*value**iven sample to distinguish whether it ¥vill exhibit Type-Abetween Type-A and Type-B is presented in Table 1.Various types of' behaviour of sand samples preparedby different procedures are surnmed up, as sho¥vn inor Type-B response. A conceivable scenario in explainingthe t vo dift r'ent types of response ¥vould be that in someof the soil specirnens, the structures of gelled substanceft'actur'ed due to the sustained application of an overbur-Fig. 14, in t.he f'orm of a schematic diagram illustratin*"the relations bet¥veen the P- ¥'ave velocity and the B-valueobserved in all of the test series employed in this study.den stress, and the fractured structure was filled ¥vithLooking over at the diagram, one can initially noticeimperfectly saturated free ¥vater, thus sho¥ving thethat the P-¥vave veloclty of non-grouted norrnal sand isresponse of Type-A, ¥vhere the Vp-value changed ¥vith thelo¥vest o¥'er all the B-values. The value of Vp is seen tending to increase if the sand is improved by silicate inclusiondeveloped durin*' the curing periodvere destroyed andB-value. On the other hand, the other specimens lvereexempted from such fracture and sho ved the response ofType-B keeping the Vp-value unchanged, where the gelledsubstance remained intact acting as the continuouspath vay for the P-1vave propagation. It is to be noticedthat for the specimens sho¥ving the response of Type-A,the skeleton Poisson's ratio was as high as vb=0.46 to0.48. The summary of the observations distinguishin_"._in any ¥vay. When there are bubbles in the pores in thegel-formation, the Vp-value tends to increase from600 m/s up¥vards. If the gel-formation due to silicate isfractured, the P- vave velocity is seen rising further up.Eventually ¥vhen the *'el-formation is not destroyed, the¥'alue of Vp takes the highest value of about 1650m/sirrespective of the B-value. It is also of interest to note 242TSUKAi¥,10TO ET AL,,2000Vp=1600-1700(m/s) / /ll2000!Tes series A & C I ! '1 il I/ /'111 ifrell-developed gel-formation-r '- hce eee e ; 4s# s1500/e o A:/ 'ji# ? ': '/¥ vtf0.47/¥ 'TTT¥l Lii:1)>*T),:pe*Anear*saturated sand500-o VpD Vs¥lc1 OOOe Vp(SS{# "'i s' ;;'f""/:";'i:i' *; :;;''';'S0.5Oe ': :' ': .: c:.1 _a- -Q ->o)i * ;¥ j:{};L:i:i lsaturated sand// 3 vb+( I:/ot/) O.4'- - =v:O O.4 O.6 O.8 12000Fractured ¥Veil-developed /el-forrnation el*formation /# F " i#L!;; ・ /,/"" /2vb)B*v =04S : ;!;:l500'0.46(b)e*ceOType of responseToyoura sand::o*7 /O.'_ 0.4 0.6 0.8 }13. Results of mon-destructive tests in test series D (D.= 4001la): (a); & ,・( against B-value and (b) overall Poisson's ratio againstB*valuethat the limiting value of Vp near B=0 takes a value ofabout 400 m/s for non-grouted sand ¥vith vb = 0.35. ThisV -value becomes about 600 m/s ¥vith vb=0.43 for thesamples containing air bubble in the geo-formation. Forthe grouted samples with fractured gel-formation, theVp-¥'alue at B=0 is higher up in the range of 700-1000m/s ¥vith vb=0.46-0.48. For the specimens ha¥*in*'//-:43 /.3 // /// ///// //S/!/ 5:c/B-valueFig; : /¥ :iL;'¥/ 044f';::;1,, i;,;; j//ooJAir-bubble// /z_ _d,¥containin ; j "r_ _ _ _ _ ___ I> 500o;i*ige -foTTnation:beQO'))Type*B; :{i;;:f::c:t:; 1000Type*Ai (DF400/0, cf'e 98ki)a)0.3rSummar¥_' of the relations of Vp and V* against B*va ueFig. 14.3 - ( I -2vb)BTest series DGroutin_ jn wet-tampednear-saturated sandcured under pressurerLn' 'rvs=200Lm/s)B-value//Cl)/lT0.2i iiLi ///c'/C5:C -1 / '¥i/vb//;':: )// 'd :;;i'-Ho;" ":/ f '#;'iAir-j:eujb:}:;"' ' ':i;:ioiningO.2 0.40.6 0.8 lB-valueo'S//_ i// !- - - gel-forTnation o// #*# ;'1"!!!ill*. * "500Vs'T ' '/ -/';'#/O'**# 3s' '#' 1'*'="'; -s_ - ;{Type-BElr ' 'l o'44// / J/ ii'>f Grouting in wet-tampedi cured under pressureToyoura sand(Dr400/0, ("0=98kPa),i; ' ";':'f' #'=c#' '*s "'"#'# #' * '' {;*"= **' ;' ;"" #'! i;/'; !' !'# ';! ';. 7lType of responseTest series D< '' *'*#* -^Test series D ## ss;;raotured gel-formation ' ;' SS //"ss"i -! "_ //*' s"" '" i;*; * * *1 500iV1 '-1492(m/s) 1I ____ _ (a)" Type-A :.t I' samplee:# "" ;-' # "" ' */ l._r_ foe/'/--**d S; T..__//>" 1000/ / A; ILiliiii# "*";# #" s **""'-''*"#*'+^"*Type B 4 samples/In Perfectly S^i{;-'-*-,saturatedI;d j//1 OOOO15005002000Vp ofnon-grouted imperfectly saturated sand (m/s)Fig. 15. Relation between velocities of P-wave propngation throughimperfectly saturated sand and silicate-grouted sand¥'elocity of non-*'routed sand is plotted against theP-¥vave velocity of the sand grouted under varyingconditlons. Such a diagram can be provided with refer-well-de¥'eloped gel-formation, the value of Vp takes theence to the diagr'am in Fig. 14, ¥vith the result shown inFig. 15. It is assumed here that the ¥'alue of V* = 200 m/shighest value of 1650m/s ¥vith the skeleton Poisson'sand vb=0,35 are taken for non-grouted sand. It may beratio of vbgenerally kno¥vn (Ishihara et al., ,_004), that imperfectlyseen that the P-¥vave velocity of grouted sand is al¥¥*aysgreater than that of non-grouted sand, but it may changedependin*" upon the conditions under lvhich the grouting:saturated soil layers often prevail do¥vn to a depth ofis conducted in-situ producing different gel-formationsabout 5 metr'es belo¥v the _",_round¥vater le¥'el, and these¥vithin the pores in sand deposits.layers are usually a target of ground impro¥'ement b.¥"means of silicate-based grouting. Therefore, it would beof interest to provide a diagram in ¥vhich the P-¥vaveIn the technical report published by the technicalcommittee of the Japanese Geotechnical Societ,y (JGS,1993), there are many methods recommended to evaluate0.50.In the practice of in-situ soil improvements, it is ,PER 'IEATIOi¥'GROUTING O*N SAND24' Vthe field performance of chemical grouting, includin_",_standard penetration tests, d_vnamlc cone penetrationTest series ASedimentation into silicate solutiontests and laboratory triaxial tests. Other field explorationtechniques are also recommended, including electricalresistivity method, electrical logging, ¥'elocity logging andothers. Based on the chart sho¥vn in Fig. 15, correlatingthe propagation velocities of prirnary ¥vaves throu*'h non-D At)oo/*Dr20-30 /0DF35 l50/0Dr50-609/a(C)_ O.6¥ogrouted sand and grouted sand, it lvould be possible torefine the procedure of e¥'aluating field performance ofchemical **routing through velocity logging tests as fol-v)}olvs. Since it ¥vas found frorn the present study that thereois no change in the velocity of shear vaves V* induced bypermeation *'routing, the velocity of primary wa¥'es Vp isonly used. In doing so, the results of' t vo independent¥'elocity logg:ing tests are necessary. One of them needs tobe conducted at a site located ¥vithin a zone of chemicallygrouted areas. The other needs to be conducted at a sitelocated outside of' the chemically grouted areas. Since thesoil layers located just beneath a ground¥vater level arekno¥vn to be in partially saturated conditions and thoseToyoura sand (( 'o9SkPa)e 0.8tO1/:sfJ 0.4¥¥¥¥¥ov)¥¥]¥¥5>1 O.'_(JDr = 50-600/0+< i-¥'z:]' '5- 50/0¥_ l_.l^^L¥JPl:iLiL LoO. llOll OONumber ofcycles, Nrig. 16.Plots of cyclic stress rario againstseFiesnumbeF ofcycles in testAlayers are usually targets of permeation grouting, the0.8¥'elocity of primar'y¥'aves Vp measured at a given depth ata site of non-grouted areas lvould indicate values less than1500 m/s, which corresponds to a fully saturatedcondition. On the other hand, vhen the permeationgrouting is perfor'med successf'ully and the gelled soilstructures ar'e ¥vell developed ¥vithin gr'ounds, the value ofVp measured at a site of grouted areas ¥vould becomeJ2Lie O.6( ]*¥-of the Vp-value measured at the same depth at a site ofnon-grouted areas, as implied in Fig. 15. The interpreta-¥¥(¥,¥ot)¥;rc:J Li Z /S 04close to about 1650 m/s. When the gelled soil structuressuft r from any f'racturing durin*" the process of groutingand the subsequent curing periods, the value of Vp ¥vouldbe likely to stay lolver in a range betlveen 800 and 1650 m/s, ho vever, it ¥vould still be greater than the counterpart/Dr = 400/0o ,oa/);;)1; O.'Test series B>VGrouting in wet-tamped near-dry sandToyoura sand (c '( 98kPa)D ADr2jo/.Dr40"/oDr55'/ol:o/ootion of the data of velocity loggin*' tests thus describedlOOlOl¥vould serve as a refined procedure for evaluating theNumber of cycles, Ncperfor'rnance of permeation grouting ¥vith good engineer-ing judgement.Fig. 17.sericsPlots ofc¥. clicstress ratio againstnumber ofc) cles in testBCYCLIC TRIAXIAL 1'EST RESULTSAfter' non-destructive measurements of the velocitiesof' P-¥va¥'e and S-1vave pr'opagation, the undrained cyclictriaxial tests vere conducted on the identical specimens.and 17 versus the number of cycles. It is seen that thecyclic strength defined above is highly dependent on thedensity. The data for the test series C and D are sho¥vn inFig. 18. The relation is also found to be densityC yc!ic Stress Ratio against Number of C'yc!esThe data of undrained cyclic triaxial tests are summarized in the f'orm of' cyclic stress ratio ( d/(2(T6) producing40/0 double amplitude (DA) axial strain plotted againstthe number of cycles, N*. In the usual practice, the DAde pendent .From the observation of the non-destructive test resultsmeasuring the B-value, P-¥¥'ave and S-¥va¥'e velocities, it¥¥'as found that the soil specirnens of' the test series Bcontained a sufficient amount of pore air bubblesaxial strain amplitude of 50/0 is taken in such plots, but inentrapped ¥vithin the structure of gelled substance andthe case of silicate-strengthened sand, some of the cyclictests did not produce the DA axial strain of 50/0. Thus,the 40/0 DA axial straln vas considered appropriate as anindex indicative of cyclic strength of the sand strengthened by addition of silicate.The cyclic stress ratio causing the 40/0 DA amplitudeallo¥ved the air to respond so as to increase the compressibility as a vhole. On the other hand, the soil specimensof the test series A, C and D Ivere found to contain onlydispersed air attached to the structure of gelled substanceaxial strain in the test series A and B is plotted in Figs. 16and the air did not contribute to the change in the Vpvalue. Particularly in the test series D, there are t vodifi rent types of response of the *"elled substance as TSUKA ,IOTO ET AL244o.8lToyoura sand (Df35 :50/0* ( 'o98kPa)l-O.6( 1' './t). * + LQ*... 04+* . - ,* -,o_v)l))'>1)oc'o Dr25"/o (Testseries C) jnear-saturated sandToyoura sand ((s'0 )8kPa)D A.=40/aU>1O.'_n. Dr250/. (Test series D)e Dr40'/* (Test series C)Dr lOo/. (Test series D)l ,in lvet- aTnpedC ? neer r¥. & near*satu 2ted sandC ;:' S-,¥{oSed!mGntntioninto ¥si i i cate selution*;; .. .¥¥ ; * ' :i { {:]o¥'¥-¥e-+**+' +*'4-rest senes O 1> - _¥ ret.tempeti water-ss ufstednon- outed sandooo.1lOO10lo.l18. Plots ofC} Clic stressllOlOONumber of cycles, NNumber of cycles, NFig.O 4v)Test series C & DGrouting in wet-tampedGroutin_Test series A ¥*cl)oo O.'Test series BTest series C&DTesl series B. C & Dt). 0,6O1)cl)Test series O ,e Test series A*-/¥*t** 9 ' H.t+ ****:'D A =4'/* icb¥J*t)' 'e 0,8n ch Dr :s 400/0Dr = 250/0ratio against ntlmber of c . cles in testserics C and DFig. 20. Summars piots of c .'clic stress ratio against number of c .'cles(D, = 35-45?/o)mens prepared by the method of the silicate-solutionlToyoura sand (DF20-30"/.. c '0=98kPa)D Ao Test series Oo Test series A' 'e O 8t)c¥sedimentation are seen bein*' slightly greater than that of'/.:theTest series Bi Test series C&Dmens in the test series B, C and D gi¥'es the cyclic stressratios t¥vice as great as that of the ¥vet-tamped non-grouted saturated sand. In particular, the inclusion of pore airTest series C & DTest series itLt). O.6Groutiil _ in IYct・taJTlpedl. near sstursted sandSedirnentst'on intoos'licate soiu on C:iTest se es Be¥¥ Groutin_ in Tet-tamped¥¥ **ncar i,) sendl')l¥) O.4_ ¥¥o¥u)+t-_¥ -I _,-o O_eO>1(.O.2bubbles ¥vithin voids formed by grouting on nearly dry:o¥¥oTest series O*et-te;mpe water*ssturatednon+ routed sandoo1llOOlOvet-tamped non-grouted saturated sand at anynumber of cycles, N*. Grouting on the ¥vet-tamped speci-Number of cycles, Ncsand in the test series B, makes the soil specimens some¥vhat more r'esistant and leads to the cyclic stress ratios¥vhich are greater than that of the grouted saturated sandin the test series C and D. Ho¥vever, there is no discernable difference among the data of the test series C and D.Therefore, it may be mentioned that there is ne_2:ligibleinfluence of the sustained application of an o¥'erburdenstress applied during the curlng period on the cyclicresistance, no matter ¥vhether the gelled substance in thesample ¥vas of fractured or non-fractur'ed type. Therig. 19. Summar¥. plots of cyciic stress ratio agninst nu lber of c)cles(D, = 20-30C )comparison in the same vein is sho¥vn in Fig. 20 for thesamples ¥vith the relative density of D,=400/0^ For thedense sand specimens, the difference as above disappearsdescribed above, presumably induced by the sustainedand ther'e are no significant differences in the cyclic stressapplication of an overburden stress during the curin_..period, ¥¥'hich resulted in the fractured or non-fracturedstructures of gelled substance. With this fact in mind, itratios amon_g: the test series B, C and D, regardless ofwhether groutin*' or permeation ¥vas conducted on nearlydry. sand or saturated sand.lvould be of interest to examine ¥vhether these differencesin the non-destructive tests are refiected in the cyclicphase of loading which is destructive in nature.The comparisons of the data in this context for all thetest series ¥vith the relati¥'e density of D* = 250/0 are sho¥vnin Fig. 19. Since it is kno¥vn that the method of samplepreparation affects the cyclic resistance of soils dueIn looking at the test data sho¥vn in Figs. 16 to '-O, thecyclic strength is defined as, (Td,f/(,-(T ),vhich is the cyclicstress r'atio causing 40/0 DA axial strain in the course of,_O cycles of load application. The cyclic strength thusdefined is no¥v plotted against the relative density, assho¥vn in Fig. '-1 . Overall, it is found that at any value ofprimarily to different structures of soil fabric, it ¥vould bethe relative density, the specimens prepared by themethod of sedimentation (test series A) exhibited thepreferable to compare the results of grouted and nongrouted sand specimens prepared by the same method.Ho¥ve¥'er, the results of non-grouted sand specimenscyclic resistance t¥vice as large as that of non-groutedsaturated sand (test series O). In the case of the lvetdamped grouted specimens (test series B, C and D), theprepared by the method of ¥vet tamping are only availablecyclic resistance ¥vas three times lar'ger than that of non-in the present study. The c.¥'clic stress ratios for the speci-9:routed saturated sand. ';f{PERNIEATION GROUTING ON SAND245lACK_NOWLEDGEMEi¥'TSToyoura sand ((T'0=9SkPa)i D A 4'/..Nc20e Test series Oo Test series Ar o O,8bTest series BTest series C&Dc lel' 0.6Test series B. C ; DYet-ts!nped)oGrQutin_g in{,S?'- O.4',J,' Test se es AScience, M. Shimane and J. Sato for their cooperation inconducting the laboratory tests reported in the presentpaper. Thanks are also extended to Mr. A. Yoshida andMr. T. Hada of Raito Kogyo C o., for their generous support on the pr'esent study. This research ¥vas funded byMinistry of Education, Science and Technofog.v.'r t4s'#'+ Sedi:Tlentation into silicate selutionc:)5>1(Jnear*dr) & near-satura!ed sandThe authors ¥vould like to express their sincereappreciation to the past students of Tokyo Uni¥'ersit..v ofO.2#s ,i.;._*・'"'"7.;-S.,; ;Test serles O¥Vater-satureted non* ,outed sand_¥TOT'ATIONcB: B-valueoO _'O 40 80 10060Relative density, Dr ('/.)Fig 21. Plots of c}clic resistance agninst relative dcusit)CO_NCLUSIONSUsing the triaxial test apparatus, t¥vo phases of tests¥vere conducted on ¥*ariously prepared sand samples, i.e. ,non-destructive lva¥'e propagation tests and destructivecyclic loading tests. A group of' sand samples ¥vas prepared in the condition saturated ¥vith ¥vater. The secondC : initial shear modulus.: number of c}cles in undrained cyclic triaxial: ¥'elocit of propagatio of longi udinal vaveV.: ¥'elocity of propagation of shear wave_*¥r.testsV8=,: axial slraina*,: eft cri¥'e confining stresscrd: amplitude of cyclic axiai siressad ,1(2g ): cyclic resis ance, defined as ihe cyclic stress rallO atD.A. axial sil'ain and !V*=20l': overail Poisson's ratiol' : skeleton Poisson's ralio4O!/OREFERENCESl) Hatanaka, ¥+,{., Uchida, A., .¥,latsulTrura, ivl. and Imazato, Tgroup of samples ¥vas formed by sedimentation of sand(2002): EYaluarion of chemical grouted area by resistivityinto silicate solution. The third group ¥vas prepared by¥vet tamping of silicate-mixed sand. The f'ourth group¥vas prepared by the method of lvet tamping of' silicate-tomography method, Soi!s ancl Fbimc!ations, 42(4), 69-75.2) Ishihara, K_ and Tsukamoto, Y (2004): C)clic streng h of irnperc ly satura ed sands and analysis of liquef ction, P/'oc. Jpn.Ac'ac!einl', Ser B, 80(8), 372391mixed sand, but subjected to an effective overburden3) Ishihara, K., Tsukamoto,Y. and Kamada, K_ (2004): Undrainedpressure dur'ing the curing per'iod. The results of the testsho v that, ¥vhile the shear wave velocity is not affected sobehaviour of near-saturated sand in cyclic and monotonic loadinP/"oc_ Ilu. C'on.f. C _vc!ic Behaviour of Soi!s ancl Liqu fac'!iollmuch, the longitudinalva¥'e velocity is influencedsignificantl)* b), the B-value and also by the method in¥vhich the specimens ¥vere prepared.In the phase of the destructi¥'e tests, the specimensprepared by the method of sedimentation into silicatesolution sho¥ved the cyclic resistance about 1.1 times asmuch as that of non-grouted saturated sand. Grouting onnearly dry sand seems to produce specimens containing asufficient amount of air bubbles trapped ¥vithin the struc-tures of gelled substances. Such samples exhibited thecyclic strength about '_.5 times greater than that of thenon-grouted samples. G routing on saturated sand appears to produce soll specimens containing partly satur'ated lvatervithin the structure of g:elled substance andsuch specimens sho ved the cyclic resistance, vhich isabout t vice greater than that of non-grouted saturatedsand. On the other hand, the sustained application of' anoverburden stress over the curing period of one monthdid not exhibit significant influence on the cyclic strength.Pllenonlen(!, Bochum, Germany, 31 hiarch02 April 2004; C}clicBehaviol r of Soils and Liquefac ion Phenomena (ed. by Triantafyllidis, 'Th.), Taylor & Francis Group, London, 27-39.4) Japanese Geotechnical Sociely (1993): Prediction and evalualion ofperfbrmance of chemicai grouting, T chniccl! C*ommiuee Report,ll44 (in Japanese)_5) Kaga, 'I. and Yonekura, R ( 991): Estimation of strength ofsilicate-grouted sand, Soi!s anc/ Fbuncla!ions, 31(3), 43596) Kokusho, T (2000): Correlation of pore-pressure B-value withP-wave ¥'elociiy and Poisson's rario for imperfectly saturated sandor gravel, Soi!s anc! Founc!arions, 40(4), 95-102_7) 'lori, A_ and Tamura, .¥,1 (1986): Effect of dilatancy on permeability in sands stabilized by chernical grout, Soils anc! Fbuncl(uions,26(1), 96-104.8) ,Iori, A., Tamura, ,1. and Fukui, Y. (1989): Distribuiion of groutsin solidified region on chemicai grou ing, Soi!s anc! Founc!ations,29(4), 127-1349) Nakazawa, H , Ishihara, K., Tsukarnoto, Y_ and Kamata, T_(2004): Case studies on evalua ion of liquefac ion resistance ofimperfectly sa urated soil deposits, Proc'. IjltConf・ C;'c!ic Behcl *-iour o.f Soi!s clnc! Lic!uefac!ion Phenomena, Bochum, Cerrnany, 3 1N,Iarch02 April 2004; C'yclic Behaviour of Soils and LiquefactionPhenomena (ed, by Triantafyllidis, Th_), 'Taylor & Francis ( *roup,London, 295304.lO) Tsukamoto, Y_, Ishihara, K , Nakazawa, H , Kamada, KandHuang, Y. (2002): Resistance of partly saturaled sand to liquefaction with reftrence to longitudinal and shear ¥vave velocities, Soi!sa/Id Fol!nc!ationj, 42(6), 93- 104.
  • ログイン
  • タイトル
  • Load Transfer Characteristics of Socketed Piles in Mumbai Region
  • 著者
  • s. S. Basarkar・D. M. Dewaikar
  • 出版
  • soils and Foundations
  • ページ
  • 247〜257
  • 発行
  • 2006/04/15
  • 文書ID
  • 20903
  • 内容
  • = 7SOILS AND FOUNDATIOi¥lTSVol46 ,No.,-,247 257, Apr. 2006Japanese C; eotechnical Societ}LOAD TRANSFER CHARACTERISTICS OF SOCKETED PILESIN MUMBAI REGIONS. S. BASARKARi} and D. M. DE¥VAIKARii)ABSTRACTThis report uses load-transf'er approach for analyzing the load-displacement response of piles socketed in ¥veatheredrocks typically found in Mumbai region. The field information used in the load-transfer analysis is obtained f'romcon¥'entional site investig:ations and does not necessitate elaborate tests. The load transfer beha¥'iour of each stratum isexpressed as a non-linear function. Empirical relations are used to ex'press rock mass modulus and limiting values ofunit side and base resistance in terms of unconfined compressive strength of intact socket material, ¥vhich irnplicitlytake into account the site-specific conditions. More than t¥venty pile load test datavere back-analyzed beforegeneralizing the go¥'erning parameters applicable to this region. This report demonstrates a vide z'ange of applicationsof load transfer approach lvhich includes the load-displacement beha¥'iour and separation of elastic and net piledisplacements.Ke¥ ' words: base resistance, Ioad-displacement cur¥'es, Ioad transfer, pile load test, side shear resistance, socketed pile(IGC: E2/E4)QIN1'RODUCTIONBored cast-in-situ piles socketed in rocks are no¥vadaysTamongst the widely used variety of deep foundations. InMumbai region of' vestern India, ¥vhere the present studyis focused, such piles pass through a soft clayey deposit,and through a highly veathered straturn, before beingsocketed in discontinuous rocks, Iike basalts, volcanicbreccia and tuff (Fig. 1). Such sub-surface conditions''f'avour use of socketed piles as a very con¥'enient foundation alternati¥'e in this metr'opolis and its suburbs.To ascertain the field per'formance of socketed piles,in-situ pile load tests are conducted, ¥vhereupon piles maybe subjected to static maintained or cyclic loads. SuchLptests give load-displacernent information upto somepre-assigned pile loads. Ho¥vever, considering theL weathnumber of piles in¥'olved at a construction site, these t.estsfcannot be performed on e¥'ery single pile because of timeand cost constraints. In practice, the frequency of pile10ad test in this region varies from one-half to t¥vopercent of the total piles installed, ¥vhich may beTiincreased depending on the nature, type of structure andthe local sub-soil conditions. This frequency, ¥ 'hichL socobviously considers the pr'actical constraints in¥'ol¥'ed infield tests, may not give sumcient idea to judge theperformance of remaining piles founded under different- 1 D l<-sub-soil conditions, albeit around the same site. Becauseof the kentledge limitations and high capacity of socketedFig. 1.ldealized sub-surface profile and nomenclaturepiles, failure load is not ahvays reached, and conseliiResearch Scholar, Ci¥'il Engincering Depanment, Indian Institule of Technology Bomba) , Po¥vai, Nlumbai 400 076, India.Professor, ditto (dmde@civil.ii b.ac.in)_The manuscript for this paper vas received fbr revie¥v on Ju}y 6, 2004; approved on December 1, 2005Vri en discussions on this paper should be submitted before No¥'ember l, 2006 to the Japanese Geotechnical Society, 4-38-2,Bunkyo-ku. Tokyo I i2-001 1 , Japan. Upon request the closing date may bc extended one month.247Sen_"*ok, _・4sBASARKAR AND DE¥¥,AIKARquently these tests have not been able to furnish thetransfer approach is a form of subgrade-reactionbehaviour' of such piles at or near failure.method, vhich assumes that displacement at a pointdepends only on the stress at that point. This methodinvolves modeling the pile as a member supported byClosed form solutions are available (Randolph andWroth, 1978; Kraft et al., 1981; Kodikara and Johnston,1994; Motta, 1994; CJuo and Randolph, 1998) that givethe load-disp]acement response of piles. Ho¥vever thesehave not been frequently applied to the field situations,since they rely on the parameters and properties that ar'enot a part of the routine site investigations. Loaddisplacement analysis usin*" elastic theories are alsoavailable (Mattes and Poulos, 1969; Pells and Turner,1979). These approaches at best, yield solutions at lo¥vervalues of ¥vorking loads, ¥vhere the behaviour of the rocksockets are expected to be elastic. In contrast, field load-displacement relationships of socketed piles of Mumbairegion are observed to be non-linear, and hence it isunlikely that the elastic theories ¥vould yield a goodagreement, particularly. at higher r'ange of loadsLoad-deformation studies have been carried out forpile embedded in a variety of materials exhibiting nonlinear stress-strain relationships like clays (O'Neill andReese, i970), hard rock (Webb, 1977), soft rocks such asshale or clay-shale (Goeke and Hustad, i979), rocks ofvarying stren_g:ths like marl, gypsum, diabase (Car'rubba,1997) and ¥veathered rocks of granite-gneiss variety (Kimdiscrete springs, ¥vhich represent resistance of the soil inside shear and in end bearing.The response of these springs in fact, represents theload transfer curves. The numer'ical analysis star'ts ¥viththe assumption of a unit base resistance. Using the baseload transfer r'elationship, the corresponding basedisplacement, wp is computed. V Jith this displacement theanalysis proceeds up¥vards, in ¥vhich the forces and thecorresponding displacement at each point are computed¥vith the help of unit side shear versus pile movement(f-Tt*) relationships for the shaft resistance. Thecor'responding elastic compression of se*'ment is alsoaccounted by using appropriate stiffness properties. Theload and displacement at top of the pile provide one pointon the load displacement curve. Working ¥vith differentunit base resistance values, data sets of load-displacementat the pile head are obtained.FIEl.D LOAD TEST INFORMATIONDur'ing late nineties se¥'eral flyover projects ¥vere in-et al., i999). Hirayama (1990) employed Kondner-typeitiated in Mumbai invol¥'in9: extensive use of boredhyperbolic functions to model side and base ioad transfercast-in-situ piles socketed in rock. The contract conditionapplicable for bored piles in sands and clays. Thespecified by the controlling agency, Maharashtra StateRoad De¥'elopment Corporation (MSRDC) requir'ed pilefoad tests to be conducted as per the guidelines of theparameters used for defining these functionsvere basedon the results of conventional in-situ and laboratorytests. Fleming (i99'_) proposed a novel method, vhichenables the determination of interaction parameters byback-analysis of a pile load test data. Ho¥vever, it cannotbe applied to piles socketed in ¥veather'ed rocks offerin_"..high resistance, since it requires large pile displacementsto mobilize significant portion of end bearing.Empirical/semi-empirical relations exist that *・iveload-deformation beha¥'iour of piles or that specify theunit side shear and unit base resistance of rock sockets(Vi.jayvergiya, 1977; O_ 'Neill and Hassan, 1994; Zhan_"._and Einstein, 1998). Empirical relationships, that expressthe unit pile resistance to the unconfined compressionstrength of intact rock, implicitly take into account theconstruction technique and are lvell suited for socketedpiles (Carrubba, 1997).Althou*'h extensive ¥vork has been reported on boredpiles, no ¥vork has been reported so far giving the loadtransfer characteristics of bored piles in ¥veathered rockslike basalt, breccia and tuff, found in Mumbai region.Indian code (IS: ?_91 1, 1985). Each site involved lnitia!and Routille load tests Initia! Ioad tests ¥vere performedto ascertain the ultimate pile load. In such tests, the testpiles are subjected to a maximum test load equi¥'alent to2.5 times the desi :n load. ¥ fhereas Routil?e load testsperformed on ¥vorking piles, involve a maximum testload of I .5 times the estimated desi :n load.A detailed data of about t¥venty pile load tests ¥vascompiled from information collected from fiyoverprojects (MSRDC, 1999). The analytical results from the10ad-transfer formulation are compared ¥vith the fieldload-displacement beha¥'iour deri¥'ed from static pile loadtests.The site locations ar'e sho¥vn in Fi*_. ,_. Other details ofload test data, ¥vhich include pile and sub-surfaceconditions, are reported in Table I .GEOLOGICAl. AND SUB-SURFACEIn ¥*ie¥v of these observations, this study is focused onINFORMATIONapplying load-tr'ansfer concepts to simulate loaddisplacement response of piles socketed in weatheredThe city of lvlumbai originally comprlsed of sevenislands. Extensi¥'e reclamation of the submerged areasrocks of lvlumbai region.has linked these islands, resulting in a uniform heartlandl,OAD TRANSFER MF,THODas seen today (Fig:. 2). Its geological features aredescribed in detail by Sukesh¥vala and Polder¥'aart(1958); and Sethna (1999). The parent rocks of MumbaiThe load transfer method for estimation of the loaddisplacement response of an axially loaded pile ¥vasare volcanic in nature and depart from the normalsuggested by Coyle and Reese (1966). E ssentially the load(Sethna, 1999). The ¥'olcanic acti¥'ity in this region isuniform plateau basalts of Deccan peninsula in India ':',SOCKETED PILESrJ . .,c Y jiL "/ Trll I !Jt¥:;¥rihC ::I l!'ll¥!!'' h T ONAL1' / 71"'i¥Ji'lL ' 'h,!/r IB RiYL{1 PP!vl:;1-J0Hiaivi}+are follo¥ved by a residual soil derived from the ¥veather'-r' F )1" '1'ing of bedi・ock. The bedrock may be ¥veak ¥*olcanic tuff,) -'PH Rbreccia or hard basalt. Compact or amygdaloidal basaltr )c vfound in abundance in this region may be extrusive,>.¥ *'1"hypabyssal or plutonic type. Such rocks are often fractured and jointed. Tuffaceous breccia is kno¥vn to be: TH' ' E ]: r((*,?O 1lIerelatively porous and in many cases yields a core recoveryof less than 50?/o even after 6 rn drilling into such rock(Datye, 1990). If such rock contains clay, the chiselingxtft ''rDatye and Karandikar (1988) state that, due to change insea bed le¥'el in the ¥'ery recent geological past, marineclay and silty sand deposit sequence is found to abruptlychange both in lateral extent and thickness. These layers;.f Cff : _ _I-" ' :e_C249IN_ ivlU iBAI--'.:1/ : 'Ie')(sN" il -procedure adopted for sockets lea¥'es a plastic muck at)Jfj "I':;;.' !lOADthe base of the bor'ehole. Ground vater table is hig:h andLCfi lfiuctuatesc¥vith ride le¥'elsIt also contains high sulphates(1300 ppm) and chlorides (15000 ppm).The load test sites related to the present study areconcentrated along the express¥vays ¥vithin the lvlurnbaimetropolis. Load test sites D, G, J and K (Fig. 2) areLNXPo 30SS'Y!;:j.*:r< ;! ll ( } ?; ii uil lhL 1 rl.:iiL}tl[L r Cs!I]]* ,j :I ( i lTi ii iE:r c: if)r, iFig. 2.Tr L; (FCh lL: isisnlocated along the ¥vestern expr'ess high¥vay (WEH), ¥vhilesites A. B, C*, E, H and I are located along the easternexpress high¥vay (EEH). Load test site F is located at theexpress¥vay linkin*" Mumbai suburbs to its heartland. Thedetails of the sub-surface conditions at these sites arereported in Table 1.Asr.ciTa!e}ei: '; F iv,** i;L:, ,sL(e=1Fl i( ji r} ; i t l GI s 1_t;pSr (Dir rt( l;LP*ET LsiYLit; i cti,1Ftii l . '* eFe"***.erl(i Dsr 1.pad ; : f: ii !1 ; t ,' , Lfu e lenMumbai map showing load test !ocationbelieved to represent a much younger phase in theeruptive sequence of the Deccan volcanic province.Pyroclastic rocks are common and sho¥v abrupt changesin thickness indicating that, fissure eruptions ¥vereaccompanied by central type eruptions.Prominent ridges are seen along ¥vestern and easternsides in Mumbai (Sukeshwala and Poldervaart, 1958).The basalts are ¥vell exposed along most of the easternridge. Up vard it *・i¥'es ¥vay to a red-ash breccia which isfollolved b}, a fine grained, ¥vell stratified yello¥vish ashbed. Tuft and intrusive rocks are also exposed intermittently along the eastern ridge, ¥vhich are conspicuouslyabsent along the ¥vestern rid**e. The pyroclastic beds arethe lolvermost exposures and commonly exist alon*・ thewestern ridg:e and central lowlands. Along: the ¥vesternridge medium grained basalt exists, ¥vhich is follo¥ved byyellowish tuff, coarse and compact at the base but becoming fine *"rained and stratified up¥var'ds. The youngestrocks exposed in Mumbai region are hyalopilitic basalts¥ 'hich form hard capping of the western ridge.In general, almost all test sites testify the generalizedversion of the sub-surface profile (Fig. l). The top stratum at almost every site comprised fill material overlyin__'natural marine deposits and also includes ¥veathered soil(N: 5-30). In all the test sites, use of steel liners deniedany possibility of resistance being offered by this stratum.The ¥veathered stratum represents a transition iayer in the¥veathering process of rocks. The criterion used to defineveathered str'atum is based on sub-surface information.This stratum is indicated by a negligible value of corerecovery and rock quality designation (RQD); and isassumed to extend beyond the depth corresponding to theRefuSa/ of SPT. The projected N values for this stratum,estimated on the basis of the chiselin*' energy cr'iterion, Iiein the range of 60 to 100. The socket portion is indicatedby the straturn yielding sufficiently high core recovery(at least 45010). In some sites (Load test no. 5, 7, 16, 17,'_1, '_3, 37 and 48) a good socket material ¥vas availableimmediately af'ter the top stratum. Major'ity of the socketmaterial encountered at load test sites ¥vas basalt rangingfrom compact to arnygdaloidal basalts (locations A, B,E, G for' example). Other rocks like volcanic tuff andbreccia ¥ver'e encountered in locations D and H. Theser'ocks are highly jointed, and in exceptional cases yieldcore reco¥'ery and RQD values beyond 900/0. A majorThese rocks sho v various de*・rees of ¥veathering andhence occur in the states ranging from fresh to highlyfraction of the socket material encountered indicated acore recovery in the range of 40 to 600/0 and a very lo¥vveathered and disintegrated forrn, ¥vhich is converted toa residual soil matrix, colloquially known as murrum.Datye (1990) briefiy describes the sub-surface conditions of' Mumbai region. Typical sub-surface stratigraphyRQD (O to 350/0). The stratum pref'erred f'or pile socketingin Mumbai consists of a heterogeneous fill follo¥ved bysoft, compressible marine deposits in the creek areas.in load tests had the unconfined compressive strength(UCS) in the range betlveen 5.0 IVIPa for relati¥'elyveakertuffeceous breccia to as high as 40.0lvlPa for' hardvarieties of' basalt. t\⊃OloTIlble l91曲.(》adles重     Sile(k)calion aS pel’1?ig.2)Load にsl lype匿io.ll)1』1「l EVik[’〔)財u江1Cl1o蓮1(A, 3Vikl’oli lullc縫OK1(A)Rl)LTユ 4KalljU郎1al’glしmcnOl1(B}Rl)LT 5Kan餌K’ma監’gjtEIlcl10n(B》RPl 6AGl.、R・号1しmCliOn(QRPLτ、T P員cide監瞭y夏)es独npile load(MN)L(猟hes“粟馳dsul》一s“rfacdl監for匪奮1a繰(レ臓Sockct  6 σじmaにrlal(Ml)a)CR7(%)ゐ【}1’iD(m}51.331曾(}12.705.6(}4、602.50P4/03/w 2.50 Gl−eyishjointcdl3aSと虚19、4374.0(}10.0(}i.O14.206.604.503.10l)7/05/w 2.50 Mod.Wealh.Basa賛1i.7762、66 8.0()i.O14.855.905,953.(x}l)8/(}2/W 2.50 Weathered l3ξ匙salt12.4663。(}7 o.o{)I.O11.909。500.002.4(} 2.5(} V㌧1e“thered AInyg.8asa1148.(}029.96l.O13.757、803.402.559〔),(}(}37.(}0l、213.155.3063.0(}3(}.0(}1.211.303.1(}3.854.358(}.{x}2(}.o(}1.219,lo5.259.354、4537.00i.2I2.108.503.200.4084.0〔) AT/037.07 .loi瓢ed Yel1〔)wish TulrH.49 8.IVLR−Jaicoach.阻.(1)}Rl)LT Pl8/! 2.83 Greyisパi『しIlr28.50 11.IVl認R−Ja1c(}ach jn.(1))RPL.T  Al/4 2.83 Vdca田c l3recciε毛RI)1、、T 至)7E/l 3.40 Mod,Wea由.Amyg.13貸sa蓑39.4(}47.00 3.4(} Wea由,へmyg.Bas蝋35.7091.(}oP27w/l8.(X)lGO.{)(}1.25.654.i5o5(}o!.509.257.0(}l。251.o(}2.251.lo i7Cadbul−y jn、,Tha“e(氾}RP【」Tl)35W/i 3.40 Fl’esl1GrcyAmyg.BaSak27。60 !8Talo.le l、R jしmc穏OI1(F)I l)LT Test l)He 3.40 G[℃yishjoi搬edl3aSak28.0457.0(}1(}.oo1.2 l)20/E−4 1.90 しloi瓢ed B貧sa猛35.709().0037.000,91(}.357.00 TeSl i)ile 1.90 .101med BaSak 0。(X}0.910。256.75 23AAREY.iヒmClk}11(G}RI)1’T Pl/W−4 1。90 .k)lnted Basall 29KurlajしmClio盲1(H)Rl)14丁  P目1、、 3。39 Mod、Wealh Bl・eccia 3747 48 5(}(〕hiヨedanagal’.獅.(1)1)allaPad縦lullcti〔川(、1)GMLR㍉ul1面ol1(K)1)a臓しP該da,油.(」》聾)1.ぽTRl)1’TRl)1.ぽTiPl、T Test l)HeAM/Z/EI)1(512−420》 Pl「esl l)ile 3.75 .萎oi:試cd Anlyg。Basai星8,4821.835.3640.80l(}o,o(}55.00(}.o〔)3.5027.{)G0.911甲i(}8.8082.0072.001、222,659.201噸114.o(}6。(}00.008.001.0514.o(}4.oo6.0(}4.OG4.(X)0.007.206.201.752.4096.0042.oo (1【’ey1曲蓋3asak15、3070.9042、72 2.60 (h−eyishBasa員11。8045.21 0.{X}(}.6 0.85 (}1’eylslヨBasa駐!4.2483.3339.47(},6U.2010.350.0(}13.352.300.!0℃・rega・蓑1−Mし蓋kll・dl−inkR・ad,‘1Unc・1・1111edCGlま・Pl・essK)nStゴenglh,7CorcRec〔)vcry,gRock Quaiity Desi9【1alk}銭,サi)iie Lellg山,耳ol)ei)こh ofTop C()n}presslble Slral篭lm,Ill)cPしh(》fWe猟hel℃d slra吐uE11,IbcP由ol−1く(》ck Sockcl.NO!e:RefeI¶g.1E’Orge江1e【’aiiz“1Sub−SUrfacePE『of玉le。(}.〔)o60.0(} 2.60I h撤ial PHe至.oad「ギCSl,ユROllt組e l)ile Loる1d Tesl,零Alldherl一(}ilatk(〕par I」i総k R(}ad,’ .logesllw撒一Vi1《rolほ.」艮蝋Road,1.oo1.2RP韮ぜ「至『l i)i’T7、859.(}(}CadburyjB.,Thane(E}AAR狂Y junc重iOI1(G}(15(lo切>しG6.40 162!(111)58.0(} 2.83RPl、、T乙、隻,c巳14.i4 TeSl PileAAR氾Y jtmc星lon(G}(m) G1’ey[sh画n【cdBξ一sa駐鍍)1』ぎ2(}乙、、c,1竈1毒目 きり(m} 2.50.IVLR卜㌧Jaicoach.隅,(粟)、Cadb田一yjll.,Thalle(E)∂(m}TeSほ)ilc 7 15RQl)8(%)>刀ズ>カ>zoo閏ダ>囚>男 -s .xSOCKETED PlIES IN251, 'IU 'IBAI(/T )./L'ST'-iRI'LOAD TRANSFER FUNCTIONSThe field load-displacement cur¥'e of a socketed pileindicates a non-linear trend. Hence non-linear load trans-fer f'unctions are used to compute side shear and baseresistance under a given pile movement.The unlt side shear transfer function, j; for pile shaftand socket vall interaction based on the criterion ofO'Neill and Hassan (1994) is given as;J 2.5D=Iv,Tv. ( I )E, J''",=**¥vhere,Tv.= pile movement at any depth z along pileD= pile diarneterE*= effective Young's modulus of rock andf ,=**=ultimate shaft resistance at socket interfaceThe above load transfer model is meant for most typesof rock sockets ¥¥"ith regular and rigid interface asperitypattern .E* is expressed in terms of the unconfined compr'essivestrength of Intact lock core, (7* using the expressionE* = C* (( *)o 5 (_,)vhere, C* is a dimensionless parameter relating Fock massmodulus to cr* so as to closely simulate the load displace-ment curve.Ro ve and Armitage (1987) recommend C*= _15.0 intheir expression for rock mass modulus, givin,_._O, E*='_15.0 (a*)o 5 and this expression has been confirrned fromthe results of the field load tests by Radhakr'ishnan andLeung (1989).The proposed relation forJ' in rock also expressed in*,*** 'terms of (7*, isJ *** = C*r ((T*)o 5 (3)¥vhere, C*f is a parameter relating rnaximum unit sideshear in rock socket ¥vith (T*.According to Zhang and Einstein (1998), C'r=0.4 forsmooth sockets, and C*f=0.8 for r'ough sockets.The pile base perf'ormance is modeled using thefollo¥ving hyperbolic expression (Fleming, 1992).rig. 3.Flow sequences for load transfer programO_p 0.6 wpTE* DB O_ *,,**against an abscissa of.,,., a linear plot is obtained and the¥vhere, E*= rock mass modulus at base; DB = diameter ofinverse slope J/m gi¥'es an asymptotic value of Q. Thepile base; Q**,**=ultimate base resistance=q***.AB;Qp=qp.AB, q.,..=maximum base resistance mobilized;qp=unit base resistance under base mo¥'ement wp andlinear plot is based on the assumption that Q exists eitheras shear or as end bearing resistance and not as a combination.AB = cross-sectional area of pile base.This relation has been derived by equating the hyperbolic and linear elastic functions at a load value of Q ,**/4(Fleming, 1992).Equations (1) and (4) are generalized Kondner typehyperbolic form (Kondner, 1963) expressed as wlQ=n7.}v.+ c, ¥vhere }v, is pile displacement at depth z due topile load Q and c and n7 are constants. If }v./Q is plottedThe proposed relation for "***q usedin Eq. (4), is,expressed in terms of cr* as;q*.. = C,1 ((T.)o 5 (5)¥vhere, Cq is referred to as a parameter elating maximurnbase resistance for rock socket ¥vith cT*. All the variablesused in Eqs. (2), (3) and (5) are expressed in MPa.Zhang and Einstein (1998) recornmend the lo¥ver and BASARKAR AND DE¥VAIKAR252fLeTest No O )Sub-Svr e PtofiieFp l:ed LQsd ejO f t,e , ( S5I r_.FTgP=!e Ne d F NFi[lT7ate 3r*13 Omd gnn8el y}5nl5 emo'F > eYelle,** lshh,,eBihef droek10 2mPl e B:seGr ylshC *S3elntedi :i2 Orn RCi: slBBSE,tLX;S*IF 3C_omparative toad*displacement curves (Site: Vikroli junction,Fig. 4.Fig. 7. Comparative load-d;splacement curves (Site:tion, Goregaon, Pile TP)AAREYjunc-Pile Test Pile)Su Surf ee PrQ ie(Load Test No 48)PiApplied Lo d ot7 Pi:e He d MNOa o flea HlgO CfT [ eY }FF steri31s 5 021 OmWBathBredSoOsr{* ts3 55m4 QmWeatheredL $=2・; F*Basa tClc3to 5QAQO I:2510.3mereyishBaS :tP !e baBe VeS=1 1 E,*11 2 mSC; 5 2i'peFio. 8. Comparative ioad*displacemeut curves (Site:Fig. 5. Comparative load*dispiacement curves (Site:JVLR-JaicoachGMLR junction,Pile P1-512-420)junction, Pile: Test Pile)segment for mo¥'ement, tv, at depth z,f **=maximum unit side shear, and(Load Test No 110 d on Pi:l'v* = critical movement of the pile segment at ¥vhich f*** isSub*Surf oe P[o [eP ieo 2 1Appi[edhlead l. NeO O T, i ・Y OSSErimobiliz,ed and tv,Si:ly sBndv*.Based on several trials, the limiting side shearf in+ ",**''!Ilh gr9Yel5 25m5 1'5mthis equation is found to be in the range of iO0.0 to 200.0kPa corresponding to a limiting displacement, }v*= l0.0mm.Hiah:yrJ > soweatherEd:1SsDii/roekOXpile modulus Ep expressed in MPa is14 sm+*/eathered2SL1,IcsnioCA *Plie base ReBrecc2Ep = 5000(,f,k)o 5 (7)e= EO OVC: S=19sr* PsOmFig. 6, Comparative toad-displacement curves (Sitc: JVl.R- Jaicoachwhere, f*k = 28 day characteristic compressive strength ofconcrete used in pile, in MPa.Using the above parameters and load transfer functions, a theoreticai load displacement curve is obtainedby choosing appropriate values of parameters C*, Cf andjunction, Pile Al/4/E)upper bound values of Cq as 3.0 and 6.6 respectively.Weathered stratum of the typical Mumbai sub-surfacestratigraphy is modeled using Vijayvergiya's (1977) shearload transfer function expressed as;f =f (' O !r = It' /1Throughout the analysis, pile behaviour under the testloads is assumed to be elastic. According to IS:456, ・_OOO,)-tv /}t (6), .* . ','** ' . * .¥vhere, f.=unit slde sheal mobillzed along the pileCq ¥vhich gives best match ¥vith the measured loaddisplacement response of the pile. This procedure isapplied and compared vith se¥'eral field load displacement cur¥'es before eneraliz,in9: the values of theseparameters applicable for socketed piles in the Mumbaire*'ion.A Ioad transfer program, AXPII,E incorporating thefiow sequences sho¥vn in Fig. 3 is prepared and the results 、調丁三嚢ble2贋 Prope当’“es ofω1∋s重r癬註AIterberg lh11震Gl’aia slze(%)Sl.11(》.Ref.i、9ure  S吐r斜UmFi蕪m組el・lalSikGayLkluid40.023.(1lO.027、0 {).512.530.556.5i6.(}3{).(143、011尋027。()Salldiimitφ、、21)1哉stic!1mll 8.0Siky sξmd w巽hgl−ave!7{).(}48、(}38.0(kPa)Remarks 25.4i5.σ『 (o)20.(}盤 46.(15.o、27.0 13しり  0Oo湾 16削屑踊○1マi9.5Conlpletely35−50wea吐hered l・ock」.SPT,N 〔’uFi9.4Marinecl“y2、Parameters聖1甲監、5‘%}  deSCl’iPllOn(}r貧vei1. Shα毫星’   (%)Fig.6Sil【y sand Wilh27.(}24.040.() 3.O23.042.0gr註vel9.046.026.032.044.026.(}CR3(%}:0−17RQD’ξ(%):o 20℃r属しりzン⊂:Yeik}wisll clayeysik4. 2(}.06{).(}35=577寧卜至igl擁y℃(}hesio跳,閃>Fig.7 35weε鷲he:・ed朕)ck  5.ζ01〔)31F搬malel・iai l7,037、029.0Vレノeε鷲he1’ed SoΩ 7.083,08.017.090.o61.o5k)20 74.0F19.82、o12.0額29.5(}盤 35ユ〈ngie・fshαミringresis{allce,諄Bξ篭sed・11dll・eclsl}eI顧es蓬,曳Base(1()11vallesllearlesl,;1=1螂ed・崖11rklxl31c()mPress1・1・lest,3C()rerec()very,一llく()ckquε曲y⊂lesigllalKm,5NこUuralwaterc〔)監11enl.【)σ1し⊃ BASARKAR AND DE¥¥*AIKAR254Table 3. Load transfer parametcrs¥¥*ea heredaximum resisLanceRock¥socketLoad stratum in rock socketno._ q*****(mm)It(¥・IPa)( IPa)tes, r* ./ *n=**lO O O 10fni= *r_ * Cf C'! ( hi P a )07 5 O 4 *Oi 5003 10 O 0.lO 107_5 O_4 3.0 13.2204 10 O O )O * 5.0 O 4 3 O lO.29O_value of rock mass modulus as suggested by Rolve andArmitage (1987) over'estimate the ield value of E* inmany of the cases. It also indicates that, constructiontechniques, presence of discontimrities and jointing pat-tern in rock mass of Mumbai region ha¥'e pronouncedeffect in reducin : the in-situ rock mass modulus.Cases 4, 5, 7, 8, 16, 17, ,_1 and ,_3 use C*=,_15.0, ¥vhile1 . 76those like 37 and 48 use C*=14:3 3. lvlost of the casel .37studies (Cases 1, 3, 6, 1 1, 15, 18, ,O, '_9, 47 and 50) indi-2 5.0 O.4 3.0 10 59I .41cate a choice of C*= 107.5. This is in conformity ¥vith the7.98recommendations of VVyllie (1992) vho applied a factorof safety of 2 to the sua_gested ¥'alue of C*=215.0 by07 10 O O )O 21_ .O O.4 3.0 1 .36 IO.17Ro¥ve and Armitage (1987). Hence, in absence of anylO O O 10 107.5 0.4 *O08 100 O l) )1)O 04 30 2.14 16.02l I 10 O O 10 107,5 O.4 3 O I.Oi7.5915 O O O 20 107.5 0.4 3,0 2 5i 8 8316 15.0 O.12 0.9 0.7217 215.0 0.12 O 9 06318 10 O O O 107 5 O.4 3 ,O 1 Il I10 O O 10 107 ・ O 4 3_O5 38¥'alues suggested by Zhang and Einstein (1998). Excep-.89tions to these ¥'aiues ¥vere observed in cases 16, 17 and ,_1 ,2^39.8719 10 O O l) 107 5 O 4 3.0 O.9337 143 3 O 4O 2 55 196.95347 i .56 1 1 ,7310 O O 10 lO!O 4 3 O481001OIO lO/)3704 13.0 O. 3 l4srespecti¥'ely, ¥vhich is in accordance ¥vith the lo¥ver bound4,732_61ll 215 O O.1' O 9 O 35il5 O O 4 3 14.02_O10ad test information around a _,_ iven site in Mumbai,simulation of load-displacement curve can be based onC*= 107.5.Table 3 further sho¥vs that, in majority of the casesstudied, the ¥'alues of Cf and C* used lvere O.4 and 3.0O 4 3.051vhere ver'y lo¥v values of Cr and C*, ¥vere used (0.17_ andO.9 respectively). This could be due to factors likevariation in adopted and true values of rock massstrength or due to inthience of ¥veaker Tuff rock belo¥v thepile base. Ho¥vever, the simulated cur¥'es for these casescan be applied for parametric studies in¥'olving resistanceof strata and determination of displacement components.From the abo¥'e studies, it is seen that, a close matchbet¥veen field and load transfer simulated cur¥*es is obtained by choosing appropriate ¥'alues of the parametersC.. C and C*}. Such a close match can gi¥'e vital information on the load-deformation response and can assist inquantifying var'ious factors that affect the behaviour ofare compared ¥vith mor'e than t venty fielci }oad test datasocketed piles.of Mumbai region befor'e generaliz,ing the applicableparameters.JVfobi!i atioJ7 of Sic/e allc/ Base ResistancesThe load shared in shear at pile-rock interface dividedRESULTS AND DISCUSSIONSSinlu!ation of Pi!e Loac/-Disp!aceme/71 Re!ationsllipFigures 4 to 8 present some of the comparati¥'e loaddisplacement cur¥*es along lvith the relevant sub-surfaceprofiles. Detailed properties of the top strata in the loadtest sites are reported in Table '_. A close match in the10ad-displacement curve is obtaineci by choosing appropriate ¥'alues of,f**** and T ** for the ¥veathered stratum;and C*, Cr and C*} for the socket material. These compari-sons confirm that, non-1inear load transfer functionsemplo_ved in the present study successfully map the loaddisplacement beha¥*iour obser¥'ed in the field. E¥'en incases like AAREY, Goregaon (Fig. 7) ¥vhere excessivedisplacements lvere involved, a good match could beobtained. The case studies indicate that, correct choice ofthese parameters is the key to successful mapping of fieldcurves.Table 3 gives values of abo¥'e referred parameters for'various case studies. It is seen that, the parameter C*¥*aries from 107.5 to '_15.0 thereby implying that, theby its surface area gives mobilized unit side shear insocket portion. This value ex'pressed in terms of f*,,=,*(Eq. 3) gives percentage mobilization of unit side shear inrock socket uncler any applied load. O_n similar lines,from the values of pile load reaching the base (Eq. 4) andq*** (Eq. 5), percentage mobilization of base resistance iscomputed. Table 4 presents the mobihzed values atmaximum test load.From this table it is observed that, in almost e¥'ery casea very high reserve capacity of pile exists in the form ofside shear anci base resistance in socket. The load transferanalysis indicates that, in bulk of the cases side shearmobilization under maximum test load is lo¥ver than 200/0(Cases 3, 4, 7, 8, 1 1, 15, ,_3, 29, 48 and 50). While in a felvcases, ¥vhere ¥veathered stratum is absent, higher shearmobilization in the range, 62.25 to 93.30/0 is observed(Cases 5, 16, 17, ,_1 for example). High mobilization inthese cases ma}' be attributed to a lo¥v ¥'alue of maximumside shear ,f**=,* (Cr=0.12) emplo _'ed to mode] the loaddisplacement curves.The unit base resistance mobilizeci under maximum test F ".SOCKETED PILESlable 4. iMobilization of resistance under maximum tesI ' ¥IU ,1BAi" Oload)A pfied L d on Piie HB d MN*o O,-o. resistanceBaseLoad lesthlaximumloadshearest no, ( iN)mobilized mobilizedSide50F:ol6 2029, 1 316 72033 739*335.05043,736 sos 54053,3469.0645 4106.7321 26l I _80077.0513.lO5.56084 367.563 94ll4,369 Il4 73155.0914.938.25165 0962 6440.5917s.0962.2539. I O188 6434 5920.33202_9050_333 1214.8093 3068.01,-32.9016 028.83295. s8_8037: OO;2e*O -・sQ -03Fig. 9. 'T} eoretical pile load-displacemeni relationsllipcomputed as 19.67 MN and ,-0.67 lvlN respecti¥'ely.Based on the above criteria, the lo¥ver ¥'alue, that is, 19.67MN is adopted as the safe pile load.The above example indicates that, for the load tests notcarried to failure, the ultimate and hence safe loadsand corresponding displacement information can begenerated from the extrapolated curves developed fromAXPILE program.4 76Separatiol7 ofE!asric anc! Net Sett!ements9.0036.817.23473 . 9036.9621 12compared to the elastic method (Pells and Turner, 1979).4s3 . 9016.808 54502.1316.8s8 90¥¥fhile the elastic method gives only settlements resulting:from elastic compression of pile and the base, it fails toThe load transfer method has se¥'eral advantagesaccount for the net or permanent settlement. At any loadapplication, net settlement ma_¥' be defined as the totalloads is also lo v. In absence of the ¥veathered stratum, itspile head displacement minus the elastic settlementvalues are obser¥'ed to be high (68.010/0, Case '-1).measured at pile head. The load-transfer analysis ¥vithHo vever, for other cases ¥vhere ¥veathered stratum exists,mobilization of base l"esistance is lvithin 45.410/0.appropl'iate choice of base load transfer functions is ableA Io¥v mobilization of these resistances suggests anample scope for an economic pile design.cornputed as a numerical difi rence bet¥veen the pile headto o¥'ercome this limitation. The net settlement can bedisplacement obtained using Fleming's (1992) hyperbolicbase-load transfer function (Eq. 4) and Boussinesq's'rheol'etica! Pi!e Head Load-Displacement CurvesUsing the AXPILE Ioad transfer program, the theoreticai load-displacernent relationship for JVLR-Jaicoach(Poulos and Davis, 1980) elastic base-load transf'er function expressed as;junction slte (Case 7) is sho¥vn in Fig. 9. This relation mayc/p = 8E* , It*L_ (8)be considered as a generalized form of load-displacementcur¥'e f'or the socketed piles of lvlurnbai region. Theultimate pile load, Q**, is indicated by an asyrnptotic¥'alue in this cur¥'e. Such ext "apolated cur¥'es using theload transf'er approach can be eft cti¥'ely used to define¥vhere, D is the pile diameter; E, and v* are the averagedeformation pararneters of the material beneath the pilebase, vhich may be estimated experimentally or throughestablished relationships. This equation assumes that, thethe ultimate pile load and hence its safe ¥'alue.As Fig. 9 indicates, Ioads as per the asymptotic ¥'aiue,Q*>. and corresponding to a specified settlement criterion(12 mm) are 49.15 MN and 31.0 MN respecti¥'ely. Thesafe load from this cur¥'e is defined as lo¥ver of the.7rD(1 - vF)base load isvithin the elastic limit of' the socket materialand that the elastic la¥vs are applicable.Figure 10 sho¥vs load displacement cur¥'es obtained byelastic and hyperbolic base-load transfer function, andexplains ho¥v the net settlement is computed.In another exarnple pertaining to GlvILR junctionfollolving: cases:head displacement of 1'_mm (IS 29il, 1985) and (il)(Case 48), the measured pile head displacement under thetest load of 3.90 ¥.4N is 2.485 mm, of ¥vhich 2.i55 mmasymptotic load, Q**comprises an elastic rebound and 0.33mm of net(i) tlvo-thirci of' the pile load, Q12 corresponding to piledl¥'ided ¥vith a f'actor of' safety, F=2.5.The safe load values based on (i) and (ii) abo¥'e aresettlement. The load transfer analysls using Fleming'shyperbolic base load transfer function (Eq. 4) gives pile BASARKAR AND DE¥¥,AIKAR25610 20Loadon Pii8 He d MNoe525Oethey can be applied for simulation of load-displacementbehaviour of ¥vorkin_ piles under similar sub-surfaceconditions, ¥vhose diameters and socket lengths aredifferent from the test piles. Such analyses can giveinformation on the likely performance of ¥vorkin_g: pilesand ¥vhether or not they comply ¥vith the design specifica-ICetions.S 15Q(iv) Such exercise can lead to the optimization in the:design of socketed piles.2 e: 2SOCONCLIJDING REMARKS3! OThis report presents a load transfer approach for simulation of load displacement curves for piles socketed in3'oFig. lO. Separation of elastic and net settlementshead displacement of ,_.45 mm. On the other hand, usingBoussinesq's elastic base-load transter function (E.q. 8),this displacement is computed as '_.13 mm. Thus thediff rence (2.45-2. 1 3) = O. 32 mm r'epresents the netsettlement, Ivhich matches ¥'ery ¥vell lvith the obser¥'edvalue.MF,RITS AND APPl.ICATIONS¥veathered rock typically found in Mumbai region. Inaddition to load displacement simulation, it provides anin-depth information on the behaviour of each str'atum,¥vhich includes side shear and base r'esistance mobilization, and the load sharing under any applied load. Themethod is also capable of separating elastic and netsettlements under any applied load with the use of anappropriate base-load transfer relationship. The extrapolated field curves can g:ive information of ultimate andhence the safe pile loads.In absence of load test data at any particular site inMumbai region, f ** and w* for' ¥veathered strata may beThe foregoing discussions demonstrate that, a closeadopted as 100.0KPa and l0.0mm respectively. Thematch bet¥veen the field and simulated load displacementcurves can be obtained by using appropriate parametersin the load transfer program. The applicable parametersvalue of rock mass modulus parameter for rock socket,¥ 'hich influence the load-displacement simulations are C*,Cf and Cq for rock sockets, in addition to f**,*,* and s'*,¥ 'hich 9:overn load transfer in a veathered stratum.3.0 (cr*)o 5, may be used to generate design load-displace-The comparison of results based on a limiteci numberdetermined from the back-analysis of this data. Oncethese parameters are established, they may be used forof load tests on socketed piles sho¥vs that, the ran_ge ofabo¥'e referred parameters is not ¥vide. Cr and C*1 arefound to be 0.4 and 3.0 respecti¥'ely, except for threecases ¥vhere they are of the order O. 1'_ and 0.9 (Table 3).Similarly, the parameter C* is found to be in the range of107.5 to '_15.0. The choice of f*,** for ¥veathered stratumC* recommended is 107.5. The relations for maximumside shear and base resistance f*,** = O.40 ((T*)o , and q*,,** =ment curves. For other sites, ¥vhere the load test dataexists, the applicable par'ameters C*, Cf, and Cq can be*'enerating load displacement cur¥'es, ¥¥*hich can assist inthe design optimization of a socketed pile.Althou_9:h the parameter's for load transfer areranges from 100 to 200 kPa, keeping vt**= lOmm. Theapplicable to the socketed piles of Mumbai region, theseparameters can be obtained at similar sites from the backanalysis of load test results and applying them to theanaly. ses performed indicate that, these parameters can be¥vorkin*" piles.arrived at ¥vithin a fe¥v tr'ials. lvloreover, other input datalike E*, f . q *** for lock and E for plle do not reqmre. ",**, 'an.¥' sophisticated tests and are determined from simplerelations (Eqs. 2, 3, 5 and 7).NOTATIONA 13The load transfer method can be used as both apredictive and an analytical tool for the socketed piles. In_"._eneral, this method is ¥'ital in the follo¥ving field situarions:(i) For optimal sizing of the test piles: The load transferanalysis allo¥vs the selection of appropriate diameter, andsocket length for Initia! pile load tests.(ii) For establishing the load transfer parameters, backanalysis of lllitia/ Ioad test results may be carried out bychoosing a smaller pile diameter. This ¥vould imply substantial pile displacements at lo¥ver testing costs, allo¥vinggood determination of all the relevant parameters.(iii) Once the appropriate parameters are established,CeCCC_RDDBbase area of pileC.S area of pileparameter for rock mass modulusparame er for maximum uni side shearparameter for maximum uni base resistanceCore Recoverypile diameterbase diame erEpYoung's modulus of elas icity of pile materialE.a¥-erage Youn_ 's modulus of elastic.ity ofsocket maLerialcharacteristic compressi¥'e s rength of con-f*cretef T1*Ff;maximum unit side shearunit side shear at depth clength of pile in comact lvithheop stralum 驚257SOCKETED PILES蓋N−N韮UN叢B、森1 ゐ  ロpile lengtわ11)Kim,S、,Jeong,S.,Cilo,S、and Park,1、(i999):Shear Ioad汀ansfer 乙、o¢depth of ro¢k socket  charac【erisこics of driiled shaf芝s in wea樋1ered rocks, /、 (3θo∼θ‘ゾ∼、五∼、c二瓦【hde妻》th of weathered sこratum in con【ac【with  Gθoθノ∼vヂ、だノ∼9ノ召、,125(11),999−1010、pile曲af【12)Kodikara,」,K.andJol}ns芝oΩ,1,W、(1994)IAnalysisofcompressi− !v SPT blow count  ble axiaHy loaded piles in rock, ∫1∼’、 ノ、 八『1’〃∼, 刃1∼‘71、 A/(∼’h.  ηnumber of pile segmen【s  Gθ01刀θぐ17,18,427−437、 9aPPlied load on piie head13)KoΩdner,R.L、(1963):Hyperbo巨csτress−sζrainresponse:cohesive(2m息xgP‘7max σPglユ          9巳s》uR1mate base resis【a勲ceload a{pile base  SQils,ノ、5A/αノ∼ゴF五五)’v・,ASCE・89(S銭{1),115−143.14)Kraft,L、M,,Ray,R,P.and Kagawa,丁,(1981):Theoretical[一zmaxlmum u鷺i【base res1Stance  c艮rves・/、Gθo’θc1∼・」E1∼gヂg・五)ハ・.,ASCE,107(GT11),孟543−156Lunitbaseresistanceoftilepilel5) 汽董a硫es, NS、and Poulos, }{,G.(玉969}=Seuleme田 of singleload correspo【1ding to pile賄ead displacemen乳  compresslble pile,ノ、∫Mα1∼4F五〇’y、,ASCE,95(i).189−207,of12mm16) λ10寝a,E、(1994):ApProximaτe eias芝ic−Plastic solu{ion for axiallyasymr)£oζic value oピ1P玉1e head ioad          9b。L load at bottom of出e plle segmen乞loadauopoftllepilesegme飢  ioadedpiles,/.(}80’θch。五129∼9,,ASCB,120(9),1616−1624、17)MSRDC(1999):Reports on geo1eclmical investigations and pile          RgoRock Quali[y Designa韮ion  loadteStsfor租yoversites,ρ旅θoゾ/∼.∫∀σ照9”∼9加創oヂ,  Ms良DC,Mumbai,hldia.           事t’ccr王tical nlove珪1ent in p員e segnlent18) 0’Neili,}vl,∼V、and Hassan,K、氏{、(1994):Dri温ed shafts:EH台c[s of           、∼’zver亘ca王pile movemer亘  conslruction o厳performance and design cr耽eria,P’昌oc,∫∼∼’、Co’∼ゾ、           WP1)llebasedisplacement          9こ。p(」}OE,(∠h∼’)2,  (jL∼’)n  01r Oθ∫’g1∼01∼ゴCo11∫’ノ・、oゾ0θθρ∫=o燃ゴ、.Federal Higi生way Adminis一depth along pile  £rat1QI1,0rlando,Fla,.1(1)137−187.elaS{iCdefOrmatlOnatmidpOinこOfSegmenτS19) O’NeiH, N蓋、 、V、 and Reese, L、 C、 (1970): Be}1aviour of axia1量y1,2,、_,η  loadeddrilledshafl{inBeatmo環ciay,Rθ∫.Rθρ、89−8,C’ヂ、.戸oノー」σP incrernen【in base resis{ance  κ、∼ツ.Rε5、,U錘versi£y of Texas,Aus{in,τX、 VrPolsson,s ra{io of socket ma【erial20)Pells,P、1、N、andTumer,R,M.(1979):aas【icsolu【ionsfortlle σ菅u馳con丘ned compressive strengtむ of ir}こact  design and analysis of rock−socke[ed piles,Cαη,Gθo’θ(ソ∼,ノ、,16,rockspecimen  481−487、21)Pouios,H,G、and Davis,E、H、(王980):ρ’1θFα’ノ1ゴα”o’∼擁’∼醍ソ5Z5REFERENCES  αnごノ0θ5’91∼,Jo!1n V♂iley and Sons,New York.22)Rad!1akrisllnan,Ra鷺dLeung,C,F、(1989)ILoad【ransfer  behaviQur of rock−sockeこed p11es,/.Gθo’θcノ∼.五1’∼91習、,ASCE,1151)Carrubba,P、(1997):Skinfrictiolloflarge−diameこerPilessocke【ed  1n葛o rock,Cσ刀,Gεo’θc1∼・ノ∼、,34,230−240.2)Coyle,H、NLa自dReese,L、C,(1966):Load芝㈱sferforaxla!1y  至oaded p貢es in clay,ノ、5∫、4α’∼4FE OA㌧,ASCE,92(S∼12),1−26.3)Datye,K.R、(1990):Bored piling in Bombay region,P1』oc、∫’∼ゴ’α’1  Gθo’θcノ∼.Co1払‘/GC”990,Bombay,571づ88.4) Datye, K、 R、 and Karandikar, D. V、 (1988): Bored piiillg in  Bombay reg1on, P’『oご, 1n’. Gθo’θc1∼、 Sθ〃11nα1闇on ∠)θθp Foμnゴ、  80rθゴo∼κ1,由’gε1’P’!θ5,A ,A、Balkema,Rotterdam,315−323.5)FiemiRg,W.G.K.(豆992):Anewme由odforpilesettlemem  predictionandaΩalys1s,(3θo’θchη’(1με,42(3),41玉一425.6)Goeke,P.M、aadgustad,P。A.(1979):hlstrumenteddrilledsbafこs  in claydlale,P1』oc、⑤γ’ηρ、ρθeρ∫7α’η4、,ASCE,Adanta,Ga、,  182−214、7)Guo,∼V、D、and Randolp員,M,F、(1998):Rationality of load  transferar)proacilforp11eanalysis,Co〃1ρ∼”.0θo’θc1∼.,23,85−112.8)Hirayama,卜{.(1990):Loadsettlementanalysisforboredp11esusing  llyperbolic鷺ansf−er fPunc“ons,50〃5α’∼ゴFoμ1∼ゴα’io’∼∫,30(王),55−649)至S:2911,Part4(1985)l Indians[aodard codeofpractice fordesign  and construc【ion ofl pile fou鷺da【io1》:Load tesτon p㍊es、10)玉Sl456(2000):lnd1&nsこandardcodeofpracticeforplalnand  re玉nf−orced coΩcrete.  (6),755−76823)Randoゆh,N・L F、aad Wro由,C、P、(1978):Analysis ofdeforma〔ion  of verεica韮y loaded 罫)i!es,ノ、0θo’θc11 ,五1∼918、五)ハ『,.ASCE,104  (Gτ12),1465−1480、24) Rowe,R  K,aPd Armitage,}{、H、(1987)=A desig自 nletわod for  drilled piers in soft rock,Cαn、Gθo’θご1∼、/.,24,126−142.25)Se【ilna,SF,(王999):GeoiogyofNlun玉baiandsurro縁ndingareas  anditsposiエionimheDeccaavolcan1cs芝rat1graphy,ノ、Gθ0109、  Soぐ./1∼4’θ,53.359−365、26) SukesllwaIa,R、N、and Poldervaart,A、(}958):Deccan basalts of  こhe Bombay area, Ind玉a, βμ〃、 0θ0109, 50c、 .4〃∼θ1”‘α, 69,  1475−1494、27)Vijayvergiya,VN、(1977):Loadmoveme【1しcbaraαeristicsofpiles,  P”oc.4f/1⑤v’ηρ.1レh’θハt・の’,Poノ響’,Coσ5∫α1姻ゴOcθα110’V.,ASC厩,  Long Beacll・C&11foraia,2,269−284,28)Webb,D、L.(1977):Pillnglnweakrock、Sess1on2,P〃ε5’n晩αん  Ro欲.TilelnstitutionofICivi1E:1gl鷺eers,209−21029) ∼Vyllie, D、C、(1992):Fo召n4σ”o’∼5 0∼∼Rocん,E and 類N Spon,  LOIldon,30)Zhang,L、and Einstein,H.H (1998):End bearing capacity of  dr1lledsllaftsiarock,ノ、Gθo∫εc1∼、Gεoθ11w−、五ノ∼919.,ASCE,124(7),  574−584、
  • ログイン
  • タイトル
  • Generalized Coulomb's Criterion for 3-Dimensional Stress Conditions
  • 著者
  • Mitsutoshi Yoshimine
  • 出版
  • soils and Foundations
  • ページ
  • 259〜266
  • 発行
  • 2006/04/15
  • 文書ID
  • 20904
  • 内容
  • SOILS AND FOUNDATIONS46,¥*olNo,-,259-266. Apr 2006Japanese Geotechnical Soclet}GENERAL.IZED COULOMB'S CRITERION FOR 3-DIMENSIONALSTRESS CONDITIONSMITSUTOSHI YOSHlivllN 'Ei)ABSTRAC1'The Coulomb's Failure Criterion for materials is comrnonly and implicitly applied to the stress components onfailure plane. In this paper, it ¥vas attempted to apply the Coulomb's Criterion not only to the failure plane but also togeneral planes ¥vhich have ¥'arious directions relati¥'e to the principal str'ess courponents at failure conditions, and it¥vas sho¥vn that the previously proposed ¥vell-kno¥¥'n criteria for materials could be expressed by an unified simpleequation. This fact indicates that the direction of the plane for the Coulomb's Criterion is a material property as lvellas the friction ratio and cohesion. The r'e¥'ised Coulomb's Criter'ion ¥¥'as applied to simulate test results, and it ¥vassho¥vn that the model could reproduce the test data fittingly.Ke¥_' words: failure, failure criterion, shear strength, yield (IGC:D6)speakin*', if R and C in the Coulomb's failure criterionare pure constants vhich represent the strength propertyINTRODUCTIONCoulomb (1773) postulated the liner relationship be-of the material, Coulomb's Criterion should not bet¥veen the magnitude of the normal and shear stresscomponents on a plane at failure condition of materials.Failure criterion based on this hypothesis is kno ¥'n asapplied to the failure surface, but it should be applied onanother special surface in ¥vhich direction is related to thestrength characteristics of the material. This special sur-Coulomb's Criterion, and given by the simple equationr=Ra + C, ¥vhere T is the maximum shear stress compo-in this study.nent on a plane, cr is the normal stress component on thesame plane vhich is positive in compression, and R and Care the friction ratio and the cohesion of the rnaterial,THE CRITERlor lT BASED ON THE "RMP"respectively.If the strength characteristic of the rnaterial is isotrop-The revlsed failure (or yielding) criterion for isotropicmaterials is expressed by;face is named as "Referentlal Mobillzed Plane (RMP)"ic, and if t.he failure plane should satisfy the failurerR¥. Il' = RoR IP + C (1 )cr'iterion earlier than any other planes during the loadingprocess, the failure plane should have a special directionrelative to the directions of the principal stress cornpo-where(7R lP = s (T + s (T + s (7 (2)nents, i,e, it should be parallel to the intermediateprincipal stress component (T2, and the angle to theis the normal stress component on the RMP,maximum principal stress component should be equalrR 1P =to 45' - (1 /2) tanlR. Application of this failure planeto the Coulomb's C*riterion yields Mohr-Coulomb's,. + s (rsj( ' 1 + s (7sls.((71 - 02)2 +- cr' R¥IPs;((- (T3)2 + s s (g al) (3)Criterion expressed by (TI - a3 = ((71 + ( 3) sin c + 2C cos c,where cTl, (T3 are the maximum and minimum principalstress components ¥ 'hich are positive in compression,respectively, and c = tan lR is the internal friction angleis the maxirnum shear stress component on the RMP, andsl, s'2, s3 are the directional cosine components of theRMP to the directions of crj, (r2 and (r3, respectively asof the rnaterial.illustrated in Fig. 1, in ¥vhich s"i +s +s ;= I is satisfied.Although this Mohr-Coulomb's Criterion is ¥videlyused in geotechnical engineering and soil mechanics,many researchers have reported that the value of c isparameters sl , sl and s3 that indicate the direction of RMPThe feat.ure of this criterion is that it contains thein geometric space. When si becomes smaller, the inclination of RMP to¥vards the (7i direction becornes steeper.It can be said that these directional parameters sl, s2 andaffected by the magnitude of (T2 relati¥'e to (TI and a3. Thismeans that the strength parameter c or R is not a materialconstant, but a function of stress conditions. Inverselys3 indicate the efi ct of' (TI, a2 and (T3, respectively, on the=) Associate Proressor, Department of Civil Engineering, Tokyo ,Ie ropoliian UniversilJapan () oshimine-mi su oshie^c.metro-u.ac.jp).The manuscript fbr this paper ¥vas received tor review on April 19, 2005; apl)roved ou Januar)' 13, 2006.¥¥'ritten discussions on this paper should bc submi ied before No¥'ember l, 2006 to lhe Japanese Geolechnical Socie )', 4-38-2, Sengoku,Bunk)'o-ku, Tokyo 1 1-'-ool l, Japan. Upon request the closing date may be extended one montll.259 YOSHllvIINE9_60with R=0, the criterion is reduced to rOct=C ¥vhich iskno¥vn as the von Mises' Criterion (von Mises, 1913). If' l" aldlsllf ¥the material has non-cohesive nature ¥vith C=0, the' l¥criterion becomes T *t =R(TO*1 ¥vhich is so-called as Extend-/f '¥ a'L¥・/s x+s T+s /c/ ! C'・¥f'_/;/(( RMP¥ '¥ ' ¥.i/ )/ed (or Modified) von Mises' C_riterion., ¥ J/1:R1ld:P Para!!e! to (TI Tvit/7 I/7c!inatiol7 of 4.5' Re!ative to (Tland (T3.¥'( ia 3Substituting the directional parameters for the RMP ofRMP ¥sl =s3 = (1 /2) /2 and s2= O into the Eqs. (1) to (3) yields acriterion denoted by (Ti-(T3=R((7]+(T3)+'-C. If they / d/s3material is non-frictional ¥vith R=0, the criterion isreduced to Tresca's Criterion (Tresca, 1864) ¥vhich isd/S)expressed by (Ti - a3 = '_C.Fig. 1. The Referential Mobilized Plane (RMP) in ti]e spatialcoordinate sl. stem in the case of (T== (Tl, a* = a2 and a*=3RMP Iclentica! to Failu/'e P!ane (the P!ane Tvl7ic/7 Satisfiesthe Cl'itel'ion First!y)In this case, the direction of RMP is uniquely relatedwith the fr'iction ratio of the material R=tan c, ¥vithcriterion. Especially, s2= O provides the RMP parallel toparameter's sl = cos (45 ' + c l,-) = { (1 - sin c) /2 } I !2 =the direction of (T2, and no effect of cr2 on the criterion. It[(1 /2){ I -R/(1 +R2)1/2}] !2 s2= O and s3 = cos (45' - c12)= {(1 + sin c)/2}1/2 = [(1 /2){ I +R/(1 + R2)1/2}]l,"2. Sub-can be also said that the directional parameters representthe linearity of the criterion. If one of the directionalcosine components is zero, then the criterion (1) becomesa linear' function respect to the principal stress components, thus the envelope of the criterion is formed by acombination of fiat planes in the stress space, or astituting these parameters into Eqs. (1) to (3) resulted in(71 - a3 = ((TIcr3) sin c + 2C cos c ¥vhich is identical to theMohr-Coulomb's Criterion. If the material is non-fric-tional ¥vith R=0 and c=0, the Mohr-Coulomb'scombination of straight lines on any section in the stressspace. From the physical point of vie¥v, the ma*"nitudesof sl, s, and s are related to the deformation and failureCr'iterlon is reduced to Tresca's Criterion. On the otherhand, if the material has non-cohesi¥'e nature lvith C= O,the criterlon becomes sin c=((TI cr3)/((7! +(T3) and thusthe directional cosine of RMP ( = failure plane) is relatedmodes of the materials. Especially, s3= I and sl =s2=0pro¥'ides the RMP perpendicular to (T3 and the criterionto the magnitude of the principal stress components inthe manner as sl= {(1 -sin c)/2}1,'2={a3/((Ti+cT3)}1"2(73 =- C/R. This exceptional materia] fails or yields onlys2 = O and s3 = {(1 + sin c)/2} l/2 = {(71 /((71 + (T )} ll2by tensile clackin_g: but never fails in shear mode. Theseconsiderations sug :est that the direction of RMPrepresents the strength char'acteristics of the material,and the parameters sl, s2 and s3 are the material propertiesas ¥vell as its cohesion and friction ratio.In the next section, the criteria for' several directions ofRMP Iclelltica! to the SJ d:PMatsuoka and Nakai (1974) and Matsuoka (1974)had proposed a criterion regardin*' to the stress components on a plane ¥vhich had the directional cosine of s} ={ al(Ts)1((TiCT2 + (T2CT+ (73(Tl)} ! ,'2, s2 = { ((T3(T1)1(cTlal + (T2(T3 +RMP ¥vill be examined to verify the capability anda3(7i)}]/2, s3= {(al(T2)/((rl(J2+(T2(73+ a3(71)} !2 and namedconcrete concept of the revised Coulomb's Criterion. Itthe plane as "Spatially Mobilized Plane (SlvlP)". If thisSMP is adopted as RMP and the cosine ¥'alues ar'e applied¥vill be sho¥vn that the ne¥v representation of theCoulomb's Criterion embraces ¥vell-knolvn criteriaestablished by former scholars.to Eqs. (2) and (3), then the stress components on theRMP (=SMP) are expressed by (TR lP=crs lP=3crlcr2(T3/(al(T2 + cr2cr3 + (T crl) and rR¥.(3cTs 1p)CRITF.RIA FOR MATF.RIALS ¥VITH SPECIAL RMPP = rs¥. Ip = as lp{((TI + (T2+ (:r3)/1}1!2, and the failure criterion (3) yields rsp=RPM IcJentica! to tlze Octa/7ec!l'a/ Pla/7eRcTs¥ p+ C. If the material has non-cohesive natur'e lvithC= O, the criterion coincides ¥vith the Matsuoka-Nakai'sFailure C_r'iterion (lvlatsuoka and Nakai, 1974; Matsuoka,O_ ctahedral plane is a special plane ¥vhich has the sameinclination to all of the directions of the principal stressrs¥. Ip =Rcrs¥. fP or ((Tl(72 + (T2(T3 + (T3al)(cTl * CTI + (r3) = 9 (R2DIRECTIONScomponents. If the octahedral plane is adopted as RlvlP,the directional parameters of the RlvlP are s{ =s2=s3 =(1 /3)1/2. In this case, the stress components on the RMPderi¥'ed from Eqs. ('-) and (3) are (TR IP= (T..* = (1 /3)((71cr2 + cfi) and rRP = T*,* = (1 /3) { ((71(T )1 + ((72 - (T3)2 +1974) for granular materials ¥vhich is expressed byl)(CTla2a3).CRITERIA FOR MATF.RIALS W'ITH GF.NF,RAl,DIRF.CTIONS OF RMP((T3-cTl)2}1!2, and the criterion (1) yields r.**=R(7.**+ CIntroc!uction of the l/7cJepelldei7t Dil'ectional Pai'ai77eter¥vhich is the Drucker-Prager's Criterion (Drucker and(x and pPrager, 1952). If the material has non-frictional natureAlthough the direction of the RMP is expressed by the GENERALIZED COULONIB'S CRITERIONDirectional parameters of RMP for ,,ell-known failure cFiteria'Iab!e l,ss+/l(von hlises')/1 { Rl l! '_=1 1/'1 R-i/¥j(general)/ (T//¥! (rl al¥, ,iorh-Coulomb(for C= O)Nakai' (Tlr(1lcrlft/3fio! 2 "fi =/Ifl+ Rr・・a-(;...'c'(/r (T_ -( u/¥/ (Tl(T'(T1(T/¥lcf (r'(T (Tlolls' +s' =1. In order to decrease the number of theparameters for the direction of RMP, ne¥v parameters,namely,(aO(X :. /¥/'CL = 1.0 /Cf :: O.8 /0.6 /CL ::0.4 //r'fia.(TCL :::: O. I // ¥vill be introduced. Apparently the directional cosines ofRMP are related to c and fi by equation;o( r(ri( l(X_ S:and fio(r;(T Cr.dir'ectional cosine components sl' s2 and s3, only t¥vo ofthem are independent because of the r'estriction of si= )' Sai - RlfrFl R//1 Rol'TII, cr_' (r ;__iatsuoka-l S1fi : -'lr .1! '¥,Iorh-Coulombslo1/2Dracker-Pra ar(x::: ! sl !slTresca261¥///,,・u)' ' ¥s'(x+fi+1'/¥/' fce+ +1' /¥,/ (x+fi+1(sl's')=and the stress components on RMP are expressed by;(xcrl + fia2 + 03(TR IP a +fi + I (6)and(Fig. 2. Failure criteria plottcd on lz planef ordiffereut a = (s /s3)2values (effect of the angle of RMP to (Tl)(7)seen from FiO_. 7_ that the c -value reproduces the "triangularity" of the en¥'elope on lz plane. Larger (x-valuemakes the shape of the envelope more tri-angular. On thein terms of the ne¥v parameters ce and fi. These parametersalong the axis of triaxial extension, in ¥vhich the inter-oe and fi represent the magnitude of sl and s , r'espectively,mediate principal str'ess coefficient b= ((T2- (;3)1((TI - a'3)r'elative to s3. Smaller cx and fi ¥'alues make the inclinations of RMP Iarger to the direction of a'l and o'2, respec-triaxial compression (b=0). It should be noticed thattively. For example, fi=0 provides the RMP parallel tonon-con¥'exity of the envelope appears if ce> I ./cefi((T - (72)2R¥IP =fi(cr2 - (73)2c ((T3 - cfl)2c +fr lother hand, ¥vhen ce-value becomes smaller, the radius= I , expands compared ¥vith the radius in the dir'ection ofon the criterion. The directionalFigure 3 depicts the effect of fi= (sl/s3)2 ¥'alue (effect ofparameters of RMP for vell-kno vn criteria ¥verethe inclination of RMP in the direction of (T2) on thefailure condition of the materials. In this display theCT2, and no effect of (Tsummarized in Table i.R-value is fixed to O (non-fr'ictional material). C-value isCritel'ia f'ol' RMPS }vith Col7stant DirectionsPfotting envelopes of the criteria on the 7r plane (thesection of 3p = ol + 02 a3 = const, in the principal stressspace) is a simple but effective manner to investigate theconstant and a-value is fixed to I (RMP of 45' to al andcr3). Apparently the P-value reproduces the "angularlt} "or "roundness" of the envelope on 7r plane. The settingof a= I and fi= O provides a right-hexagonal criterion onstrength and deformation characteristics of mater'ials in3-dimensional stress conditions. Fig:ure ?- describes theeffect of (x=(s'l/s'3)2 ¥'alue (effect of the inclination ofthe criterion shifts to a more round shape and fi=1RMP in the direction of (Tl) on the failure (or yielding)condition of the rnaterials. In this display, the C-value isfixed to O (non-cohesi¥'e material), R-value is constantand -value is fixed to O (RMP parallel to cT2). It can be7T plane (TFesca's Criterion). V,rhen fi-value is increasedresulted in a true-circle (von Mises' Criterion). Furtherincrease of fi-value creates non-convex en¥'elopes.The problem concerning the convexity condition forthe revised Coulomb's Criterion is some¥vhat complicated and remained unsolved in this study. It seems that the 262YOSHlivlli¥TE= 1'O a_= 0.7 rl _2iJ :2 0.5 / /:= 1 5!'..-:::-- ¥¥/ '5ai1¥ = O.2_¥' /¥ ¥¥ ii ¥r-- ¥hlhFOk_iP..o 51'1!rl',:¥¥¥A O_i/ ¥¥)¥,¥¥'¥¥ :,{¥o/O 7'+/ /tR*OC=0,lh RO llt ¥¥¥¥¥I ///ll'fO::lIlk=i 'CTa1';(Fig. 4. Failure enveiopes on It plane ¥vith various /*-values (for loosesands vith R=0.5 and C=0)Fig. 3. Failure criteria plottcd on lr plauc for tiifferent P=(,s._ l,s3)lvalues (effect of the angle of RMP io (12)repr'esent a str'ength characteristics of materials, as ¥vell asthe friction ratio R and the cohesion C. Setting: of k=0con¥'exity condition requires some limited range of thedir'ection of RlvlP. The condition should be expressed byprovides the Drucker-Prager's Criterion ¥vith righta function of the model par'ameters ai, fi (or sl , s2 and s3),Generaliz,ed (or Extended) Matsuoka-Nakai's CriterionC and R-values in future.(Matsuoka et al., 1990). If the material has non-fr'ictionalR1ld:Ps Satisfying tlle Col7c!it!oll .for Synllnetry a/7dSn700tll nessIt is reasonable to assume that mechanical beha¥'ior ofcircular envelopes on Jz plane, besides k= I provides thenature ¥vith R=0, E.q. (9) constantly pro¥*ide (x= I andP= I ir'respective to the k-value, RlvlP identical to theoctahedral plane, and the von lvlises' Criterion.On the other hand, if the material has non-cohesi¥'ematerials including yielding and failure phenomenanatureshould be spatially symmetr'ic if stress conditions aresymmetric and if the material is isotropic. This entailsthat ¥vhen t¥vo principal stress components approach eachfied and expressed by;other as (7iH'(7j, the RlvlP direction should be symmetricin terms of i-axis and j-axis and therefore si- sJ. Thiscondition requires that the directional parameters ofRMP should be functions of principal stress componentsin such manner as;sl=sl (and fi=ai) ¥vhen (Tl= (TI ands2=s3 (and P= 1) ¥vhen cr2= (73. (8)If this symmetric condition is satisfied, the shapes of theenvelopes of the criteria in stress space or fz plane becomesmooth ¥vithout any singularity.There are infinite number of stress functions for ai and¥vhich satisfy the smoothness conditions expressed byEq. (8). Among these, Iet us select the simplest functionsR(7 + C RcT3 )CR(71 + C R(T' Cvith C= O, the c,lirectional functions (9) are simpli-=(x(71(ii))andk (=fi(7(T'(T3 (10)This setting of parameters for non-cohesive materialcan be also ¥vr'itten as sl = [((T2( 3)k/{((Ti(T2)k+((T2(73)k+((T3(TI)k } Il "2, s2 = [((T3CTl)k / { (( lcr2)k + (cr2(73)k(ai(71)k } I I "2and s3= [((Tl(72)k/{((JI(T2)k + ((7 a3)k + (cr3(7i)k}]1,jn termsof the directional cosine components of the RMP, andthe criterion (1) for non-cohesive material ¥vith C=0 isexpressed by{ ((7i CT2)k + (CT2(T3)k + (cr3(71)k } { (T}t2 k} + (7](2= (R2 + I )((71(T2(T3)k{(Tl(1k) + a.(1k+ (f2- k)}* }k) + cf {i k) 2 (1 1 )in ¥vhich the setting of the parameter k= I provides theorig:inal Matsuoka-Nakai's Criterion for non-cohesivematerials ¥vith C=0.If lve adopt the functions (9) or (10), the degree of free-dom of the model is reduced, because the number ofdirection of RlvlP should be fixed parallel to the (71 axisvith constant parameters of a= O and sl = O. If (73(7 <,parameters ¥vhich define the direction of RlvlP is reducedfrom t¥vo to only one, hence the (x-¥'alue and the -valuecannot be controlled independently. Nevertheless, ¥ve canobtain a variety of envelopes for the criter'ion on fz planeby selecting the combinations of the parameters R, C andk', as sho¥vn in Fig. 4. This figure depicts the en¥'elopesin the follo¥vin_a* analysis. It should be noted that thefunctions (9) are applicable for the stress conditions ¥vitha3 Iarger or equal to - C/R. If cr3< - C/R then the- C/R occurs then the direction of RMP is al¥vaysonly for non-cohesive materials ¥vith C=0, ¥vhereas theperpendicular to the (73 axis ¥vith constant parameters ofenvelopes for non-frictional materials (R=0) al¥vaysci= O, fi = I , Sj =s2 = O and s3 = I , ¥vhich resulted in tensi]ebecome right circles as pre¥*iousl.¥* mentioned.clacking mode as explained before.The ne¥v parameter k is a material constant that GENERALiZED COULO ,IB'S C RITERIONQ,i6*:sc :CL:i4Dense Monterey No O Sand(/)( 3' = 58 8kPaData trom Lade and Duncan(i 973)1.2k=0 75k=0 5Based on the test data froma i4=o(L)o::12E:oSDp.1,1R =i_132a6=(Qk=0c:o008ag aQ)Rav006:;O O O 2 O.4 O.6 O.8Ine mediate principai stress coefficient, b=(a2-a3)/((10-(;3)004002oiOIS;*k=0 250430*:Lade and Duncan(1973)o oiO>O 8ot;?o_ Oi2c:olOoc' O 16Dense Monterey No O Sand,ok=1 OO(:Go4 O(=>a'O 181 .8C26 *.nQ CTSvv >c:k=0 629SD , =0 03000,7OO 02 04 06 08 10Directional parameter forCQe,eferencial mobilized piane, kFig. 6. Detection of the fittest direction of R 'lP (k-1=alue) and frictionratio (R-value) for dense ¥^ {ontere, Sandrig. 5. Friction ratio of dense_ ionterey Sand for various directions ofRMP (k・values) at failLlreaj lpAPPLICATION OF THE REVISED COULOMB'SCRITERION TO TEST RESULTSMor}terey No O SandtJlaO Dense s nd (e=a 5T)CI Loose s nd (e=0 78)L e i nd Dunean (1 g73)k o e2gThis section describes the attempt of simulations of thefailure conditions from previous test results on cleansands by means of' the generalized Coulomb's Criterion¥¥05/defined by the combination of the Eqs. (1), (2) and (3), orthe combination of Eqs. (1), (6) and (7). The modelparameters (material constants) are onl}' R and k,aMonterey SandLade and Duncan (1973) conducted a series of truetriaxial tests on cubic specimens of dense and looselvlonterey Sand. The ¥'ertical principal stress component(71 ¥vas increased and a horizontal principal stresscomponent ¥vas kept constant (a3=58.8kPa), ¥vhileanother horizontal principal stress al ¥vas adjusted so t.hatthe intermediate principal stress coefficient b= ((7 -(73)/R=a 724c=0because the clean sands are considered as non-cohesivematerials vith C-¥'alues equal to zero, and the criterion(1 1) Ivas practically employed. It is also assumed that thestrength characteristics of the tested sands were isotropicthough this assumption is not necessarily true in reality./k=0 4e7(; !pC,lprig. 7. Failure envelopes of dense and toose Montere) Sands for thebcst*fittcd R _ (P and fricrion ratiodirection of' RlvlP. It ¥vas found that k=0.629 pro¥'idedthe best-fitted direction of RlvlP ¥vith the minimum stand-ard deviation of SDR=0.0300 and the a¥'eraged frictionratio of R= 1.132. The test results of dense Montere¥'Sand ¥vere plotted on ;T plane in Fig. 7. It should be noted(al -cr3) ¥vas kept constant. Consequently the principalthat the (T] ¥vas fixed to vertical ((rl =(7,) in the tests thusstress conditions ((T1, (T , (T3) at failure ¥vere observed un-the real test results exist only in the range of O = O' to 60',der difi rent b-¥'alues. From these test results, the frictionratios on RMP (R = rl . Ip/al l ') at failure ¥vere calculated¥vhere e is the angle from the cT, axis on the fz plane. Theby means of Eq. (1 1) for various directions of RMP vithdifferent k-values. Figure 5 is the plot of the calculatedR-vaiues from the test result of dense Monterey Sandversus the b-values. This figure suggests that the Extend-ed von Mises' Criterion (RMP for k =0) and MatsuokaNakai's Criterion (k = 1) are not applicable to the testresults since R-value is a material constant that should beunchanged by the b-value during testing, but one may seefroul the figure that the RMP of dense lvlonterey Sandshould have a direction of a k-value bet¥veen O.5 andO.75. The standard de¥'iation of the calculated R-valuesfrom the test results, namely SDR, ¥vas plotted versusk-values in Fig. 6 for the purpose of detecting the fittestplots in the range of 60' < e< 360'are the mirror imagesassuming the isotropy of the material. In the same figure,the failure envelope generated by the re¥'ised Coulomb'sCriterion vith the fittest model parameters R = I . i3 _ andk = 0.629 is also indicated. It may be seen that the revisedC*oulomb's Criterion fairly simulated the test results. Thesame procedure ¥¥'as applied to the test results of looselvlonterey Sand, then the fittest parameters of R =0.724and k =0.467 Ivith the minimum standard de¥*iationSDR=0.0383 vere detected. The test data of looseMonterey Sand and its model simulation vere alsoplotted in Fig. 7. YOSHllvIINE264(;,(kPa)(5_(kPa)Undra nedestsFuji River Sandp' = 98 1 kPa,on Toyoura Sand,Res dua stateData fromat p' = 90 kP ,DafrornYoshimine ( 996)Yarnad¥¥¥(1 979)¥5402a//¥¥60/R=0 58 1R=0 79 1C=0c=0k=0 287k=0 5aoCcf(;l(;Fig. 8, Failure envelope of Fuji River Sand for the best-fitted RMPand friction ratiorig. 9. Failure envelope of Toyoura Sand afitted RMP antl friction r ltioresidual state for the best-The same pTocedure for detecting the modelFuji River SandYamada (1979) and Yamada and Ishihara (1979) con-parameters, explained in Figs. 5 and 6, ¥vas applied to theundrained residual state from the test results of Toyouraducted a ser'ies of true tr'iaxial tests on cubic specimens ofSand, then the fittest parameters of R=0.'_87 and k=loose Fuji River Sand. The three principal stress components ¥vere adjusted to maintain the b-value constant, as0.581 ¥vith SDR=0.0219 ¥vere obtained. The model¥vell as the p= (T.** = 98. I kPa in these tests. The test con-simulation of the stress envelope for the undrainedresidual state of Toyoura Sand usin*' the detecteddition covered the range of the direction of str'ess pathparameters is also plotted on the 7z plane in Fig. 9 ¥vith thefrom e=0' to 180' on 7r plane. The same proceciure fordetecting the model parameters described in the previoustest data.subsection ¥¥'as applied to the failure state from the testresults of Fuji River Sand, and the fittest parameters ofR = O.791 and k= 0.500 ¥vith SDR = 0.0,_35 were obtained.The test data of Fuji Ri¥'er Sand and its model simulationby using the above fittest parameters ¥vere plotted inFig. 8 on ft plane. This plot indicates that the concept ofRMP reproduced the 3-dimensional failure conditionsfittingly.Residual State of Toyoura Sanc! in Undrail7ec! Sllea/'Yoshimine (1996) and Yoshimine et al. (1998) performed undrained shear tests on hollo¥v cylindricalspecimens of loose Toyoura sand ¥vith relative density ofD*=400/0, in lvhich the direction of (TI ¥vas kept constantinclination of 45' to the vertical of the bedding plane ofthe specimens, and the intermediate principal stresscoefficient h ¥vas fixed to different values during loadin_",_schemes. When undrained shearing started, the samplestended to contract and pore pressure increased to¥vardsthe points of phase transformation ¥vher'e the effectiveconfining stress, namely p' =((;i' + (72' + a3')/3, becamethe minimum. After' that the dilatancy of the samplesturned to be expansi¥*e and the effective confining stressincreased, ¥vhereas the effecti¥'e principai stress ratios¥vere more or less constant ¥ *hen p' r'ecovered 90 kPa andRELATION BETWF.F.N THF. DIRECTION OF RMPAND THF. STRF.SS RATIO ON + RMPIn the prevlous section, it ¥vas attempted to simulatethe test results by means of the proposed concept of RMPbased on the settings of the directional parameters forRMP expressed by Eq. (9). All of the simulated materials¥vere clean sand, thus C= O ¥vas assumed, hence the directional parameters expressed by Eq. (10) and the criterion(11) were utiliz,ed, where the direction of RMP was notexplicirly related to the friction ratio on RMP. Ne¥'erthe-less, the relationship bet¥veen the fittest parameters,namely, the k-values that represent the direction of RMPand the R-values for the four clean sands vhich ¥vereevaluated in the previous section, was examined andplotted in Fig. 10. Accordingly, a trend was observedsuch that the k-values ¥vere larger for the materials ¥vithlarger friction ratio, andk = O. 5 82R ( 1 2)¥vas the best linear approximation of the relationship.The number of model parameter's no¥v becomes only t¥vo,i.e. R and C, if the relation bet¥veen the direction of RMPand friction ratio such as Eq. (12) is assumed.the subsequent hardening in larger deformation. Accord-Finally. , the test results presented in the previoussection ¥vere simulated again based on the assumption ofingl_v the stress states at p' = 90 kPa ¥vere adopted as thethe formulation of directional parameters for RlvlPresidual state in terms of stress ratio, and these stressstates ¥vere plotted on Jr plane and sho¥vn in Fig. 9. Itshould be noted that the real test data ex'ist only in theran9:e of e=0' to 60'.denoted by;t(71) and=t 2) GENERALIZ D COULO¥. (B'S CRITER ONO:o(08als, and it ¥vas attempted to apply the criterion on various07planes (RlvlP) ¥vhich have various directions relative tothe principal stress components at failure conditions. Thefundamental concept of the presented failure criterion isnot a ne¥viy proposed one, but merely an extension of thesimple and classical C*oulomb's Criterion to 3-D stresscondirions. Nevertheless, it ¥vas sho¥vn that the pre¥'ious-06-o 055E 04(o_foly proposed famous failure criteria including Tresca'sCriterion, ¥'on Mises' Criterion, Extended von Mises'Criterion, Drucker-Prager's Criterion, Mohr-Coulomb's03r:'c:26502oZS O1Crite 'ion, Matsuoka-Nakai's Criterion and Extendedlvlatsuoka-Nakai's Criterion could be expressed by theoooo02 0406 08 10 12 14Friction ratioon RMP RFig. 10. Relationshipetween the directiona! parameter for RMP(k-value) and tl]e friction rario on R 'lP (R-value)Based on the ¥¥ Dense Monte eyOassumption of / ¥ ¥¥ Sand. Rsl 151k: O 582RllFuji River S nd, // / ¥ ¥:¥// /R: sO 778¥¥///Loose¥¥¥ MontereySand. R=a 71 1Toyoura Sand, / O 5 - ¥¥¥Residuais{ate,R:sO 59i¥/ ¥¥//rh //f <¥///' ¥¥ ¥)¥/ 7 / _+¥ Y¥///'!!1,1i/vi!r¥r¥ 1 ¥¥!/ r¥?'1 1///I i /'1! I1 1 l, !l//' '/I04''/'+/''I/lI.I ¥¥¥R-03///!//!05/I * _ -T --_ / f ¥¥¥// fL __-'_LOl /n__ Tij"j(CT.r/pr R'I=20-l I Oll// /The shape of the failur'e en¥'elope on 7z plane forvarious directions of RMP ¥ 'as examined to clarify theg:eneral effects of the direction of RMP on strengthcharacteristics in 3-dimensional stress conditions. Consequently it ¥vas sho¥vn that the direction of' RMP is relatedto the linearity and the effects of the magnitude of(5_ IpClean Sands C=0 ///unified simple forrnulation based on the concept of' RMP.¥ ¥1¥t ¥ ¥ a2 r1/R=1 OFig. Il. Failure envelopes of the clean sands based on the assumptionof k = 0.582Rintermediate principai stress on the criterion, i.e. theangularity and r'oundness of the envelope on 7T plane. It¥vas also sho¥vn that the syrnrnetry of the forrnulation offailure c 'iteria refiects the srnoothness of its envelope instress space.The revised Coulomb's Criterion ¥vas applied to thef'ailure and residual stress states of clean sands observedin the pre¥'ious experirnental studies. It ¥vas sho¥vn thatthe re¥'ised Coulomb's Criterion could simulate the testresults fittingly, although the direction of' RMP (k-valuein the model) is just fitted to the test results in the simula-tions vithout any physical background. In addition, it¥vas suggested that the direction of RMP could have somer'elationship ¥vith the friction ratio of the materials.ACKNOWLEDGMENTThis study lvas carried out as a part of the jointresearch pr'ogram organized by Xi'an Jiaotong Universitylvhich ¥vere obtained by the corrrbination of forrnulae (10)and Tokyo Metropolitan Uni¥'erslty. Discussions ¥vithand (12) for non-cohesi¥'e materials, in which the modelparameter is only R-¥'alue. The best-fitted R-values underthe assumption of (13) ¥vere re-evaluated and they veref'ound to be i.15i 0.711 0.778 and 0.591 for the denseMonterey Sand, the loose lvlonterey Sand, the Fuji RiverSand and the Toyoura Sand, r'espectively. The re-evaluated stress envelopes on ,T plane are plotted in Fig. 11Professor Yu Mao-Hong of Xi'an Jjaotong Universityto*'ether with the test results. It should be noted that theduring the joint research program. The author lvould alsorelation such as Eqs. (i2) and (13) is only tentative andcould not necessarily be applicable for general materials,because only four rnaterials were examined in this study.It may be possible that the relationship bet veen k-valueand R-value could be different depending on some factorslike to thank Professor Hajime Matsuoka of NagoyaInstitute of' Technology and Prof'essor Yuji Kishino ofTohoku University for their discusslons and encouragements.such as mineralo :_y, shape of the soil particles, grain sizedistribution of sands, and the confining stress levels.REFERENCESCONCLUSIONSThis paper stated that the C oulomb's Criterion shouldnot be necessarily applied on the failure plane of materi-¥vere most helpful for developing the concept of failurecriteria of materials descr'ibed in this paper. The author isgrateful f'or the assistances by Dr. Che Ailan of ShanghaiJjaotong University, Professor L.iu Feng-Yin of Xi'anUniversity of Technology, Professor Takahiro lwatateand Dr. Yoshiya Oda of' Tokyo Metropolitan Universityl) C'oulomb. C A (1773): Essai sur une applicalion des r gles desmaximis et minimisquelques probl mes de slatique rdlatif al'architecture, *1l nl Acac! Roy. Pres_ Divers Sal'ants, 7, 343382.2) Drucker. D. C. and Prager, ¥V_ (1952): Soil mechanics and plasticanalysis or lim}t design, O. A!)p!. ,Vlech., 10(2), 157l75,3) Lade. P_ V, and Duncan. J. lvi. (1973): C*ubical triaxial tests on 266YOSHINIINE  collesionless soil,So顧Mechanics and Foundations Div1slo自,ノ. 。45Cだ,99(SM10),793−812.  GesellschaftderWisseascllaftzuG6疑ingen,582−592。4) 真・1atsロoka,H.(玉974)Sしress−stra1員relationships of sands based on  underε11ree−d1mensiopal stress cond玉tions,P1∼1)7=hθ5Z5,Un圭vers叢ty  tぬemob1lizedp!ane,So1なθη4Fα!11伽10n5,14(2),47−61.5) レlatsuoka, H. and Naka叢, 丁. (1974)= Stress−deforma盛on and  s甘engl恥 c1璽aracterist玉cs of so員 under l}1ree di齪ereRt prlnc呈pal  stresses,P”o‘、/ρ11,50‘、αv’!E1∼gi11θθ∼下,(232),59−70.6)Ma菰suoka,H、,}{oshlkawa,T。a且dUeno,K、(1990)IAge職eral  failurecriterlonands【ress−strainrelationforgranularmaterialsto  metals,50〃∫σ1∼ゴFα’11ゴσ”o’15,30(2〉,119一玉27.9)Yamada,Y.(1979)l Deformation c鼓aracτerlstlcs of loose sa貸(i  ofTokyo、正0)Yamada,Y.and Ishillara,K、(1979):An三sotropic deformatlo艮  cllaracτer玉stics of saIld under t蝕ree d玉mens玉onal s{ress cond疑iolls,  So’なαη(ノ170∼〃1ゴθ”01∼∫,19(2),79−94・11)Yos1}imine,M.(1996):UI1(玉ra1ned負ow deformation of sa田rated  sa疏dundermono芝onicload量簸gcondltlons,Ooα01”oゾどη9’πθθ111119  r/rθ∫な,Univers1tyofTokyo,7)Tresca,H.(1864):}vl6mo玉r sur L’ecou至eme猷des corps so11des12)Yos願mine,M.,王shlhara,K.andVargas,W.(1998):跡ectsof  soumisadeforIespress1on,Co1ημε5Rθ顧’5擁ωゴ.5ci.Pθ∼P’5,59,  Principal stress direct叢on and l【}termediaIe pr1ncipal stress oa  754−758曙  undra玉aed shear be雇av1or of sa鷺d,50〃5απゴFo’”7ゴσ’10η5,38(3),8)VQnMlses,R、(玉9B):MechanlkderfeStenK6rper1mpiaStlsc員一  deformablen Zusξaad,nachric卜ten von der k6nig鮭cilen  177_玉86.
  • ログイン
  • タイトル
  • Two Case Studies of Consolidation Settlement Analysis Using Constant Rate of Strain Consolidation Test
  • 著者
  • J.-C. Chai・Norihiko Miura・Shigenori Hayashi
  • 出版
  • soils and Foundations
  • ページ
  • 267〜268
  • 発行
  • 2006/04/15
  • 文書ID
  • 20905
  • 内容
  • ログイン
  • タイトル
  • Two Case Studies of Consolidation Settlement Analysis Using Constant Rate of Strain Consolidation Test(closure)
  • 著者
  • Koji Suzuki・Kazuya Yasuhara
  • 出版
  • soils and Foundations
  • ページ
  • 268〜269
  • 発行
  • 2006/04/15
  • 文書ID
  • 20906
  • 内容
  • ログイン
  • タイトル
  • JGS NEWS
  • 著者
  • 出版
  • soils and Foundations
  • ページ
  • I〜II
  • 発行
  • 2006/04/15
  • 文書ID
  • 20907
  • 内容
  • ログイン