不攪乱砂質土の液状化特性に関する実験的研究

Experimenial Study

1)到

Un

正

l

i

s

t

u

r

b

e

d

Pherwmeno:

n

Sands

255

Tetsuo

OKUMUR

ム

and

Yoshio

OHNE

不撹乱砂質土の液状化特性に

関する実験的研究

奥村哲夫

e

大根義男

This inv巴stigationdeals with the liquefaction char旦cteristicsof disturbed and unc1is turbed sands unc1er cyclic loading conditions using by the dynamic triaxial compression d己V1Cε. The following conclusions were obtain巴d (1) Cyclic stress ratio of undisturbed samples have high value呂ffectedby cementation and over consolidation (2) Cyclic stress ratio of unc1is:lurbed samples is not proportional to rela:live density (3) The liquefaction char呂cteristicsof specim己n8remolded by di在日rentcompactlOn

procedur日sis di百erent

lNTRODUCTION

Ithas been recently recognized that if a saturated sand is subj巴ctedto cyclic shear stresses， such as that induced by earthquakes， liquefaction ph巴nomenon occurs and major damag巴ssuch settl巳mentof sand layer or a slope failure resulting from the loss of its stlength will occur. For typical example， the Niigata eartquake of1964， the Tokachi-oki earthquake of 1968， the Chilean earthqu旦keof1960and正，-laskan earthquak巴of1964 caused extensive damages to build ings and earth embankments (11， 12， 13， 14)_ The cause of sand liquefaction by a experimental way has been studied by many investigators for many y号ars_ Seed and Lee (11) conducted dynamic exper iments， as liquefied by appling cyclic shear str巴sson

the saturatεd soil specimen， by using cyclic triaxial compression test device and presented the avail呂ble test data on liquefaction characteristics of saturated sand. Bas巴don the analysis of test data， Seed and

Lee have pointed out that the occurance of liquefac tion is resulted from the decrement of巴ffectivestress due to the increment of r巴sidualpore引'aterpressure resulting from thεS色quenceof cyclic loadings induced by the earthquak巴underundrained state， and that major factors affecting the development of liquefaction are the void ratio of sand， the con自mngpressure actmg on the sand， the magnitude of the cyclic stress or strain and the number of stress cycles Since the first investigation by Seed呂ndLee， sev 巴ralstudies ha v巴be巴nundertaken so far in res巴rch laboratories to make clear the liquefaction charac teristics for saturated sand. And th巴nit is considered that the basic studies on liquefaction characteristics have almost been accomplished at present Thes色 expenmentsw巴reconducted on the dis -turbed alluvial sand， such as that diposited at circumfer -ence of river or coas.t However， it note that undis -turbed alluvial sand， diluvium sand and tertiary era sand at in-situ have a 1旦tentstrength exhibited by soil skeleton b巴twe巴ninternal grain contact， and these kinds of sand are consider巴dto have high resist乱nceto occur旦nceof liquefaction. Investig日tionon the char -acteristics of liquefaction for undisturd巴dsampl巴have not b日E孔undertakenso far Using the cyclic triaxial compression device， the authors have conducted liqu巴factiontests on fourt邑en kinds of undistmbed diluvium or tertiary era sands and sixteen kinds of disturbed sands. The investigations d邑scribedherein are on the charact巴risticsof liquefac -tion for undisturb巴dsands

SOIL USED IN INVESTIGATION

Soils used in this investigation ar巴sandy soil

ranging from sand to sandy loam in the Triangular Classification System. The physical and mechanical properties and grain size distribution curves of the sixteen kinds of sand are shown in Table 1 and Fig 1. The maximum and minimum void ratios shown in Table 1 w巴reobtained by following methods (1) maximum void ratio， emax; pouring dried sample

男

know these influences， the relationship between the白le

contents (percent finer by weight passing N o. 200

standard seive) and the maximum and minimum void

ratios for sixteen kind of samples are determined and

are shown in Fig.2， Fig.3 and Fig.4compaired with

the test results from Watanabe， et a，.lwithin the rela

tionships between emax -em;n v巴rsusmean grain size，

D5o， and emax -emin versus emax， emin (5).

義

キ 艮

夫 ・ 大 哲

into the proctor mold of known volume without

shock

(2) minimum void ratio， emln; pouring dried sample

into the proctor mold in 2.5cm thick in each layer

and hitting the mold ten times horizontally and

repeat this step five times

Itis reasonable to consider that the maximum and

minimum void ratios would bεvaried by the e任巴cts

of di百erence of grain size distribution， etc. To 村 奥 256 Properties of soil samples Samp1 e No. Specific Sand C. Silt C C1 ay C Tri angul ar Uniformity Void Ratio E...~ 1) Py2も 50 Gravity (>0.074 (0.005'" ( <0.

∞

5mm C1assification of Coef e _{emi n} mm) 0.074mm) max ( A ) 2.636 72.0% 16.0% 12.0% Sandy Loam 40.0 1.635 0.780 44.0 0.80 ( B ) 2.681 89.8 4.7 5.5 Sand 2.8 1.657 0.836 72.3 3.20(1.9，4.5) ( C ) 2.663 97.0 3.0

。

Sand 2.3 1.110 0.610 44.7 1.28(.0.93，1.62) ( 0 ) 2.657 98.0 2.0

。

Sand 2.8 1.025 0.625 88.8 3.25(3.0，3.5) ( F ) 2.672 95.0 5.0

。

Sand 3.0 1.127 0.619 103.3 士10 ( G ) 2.790 71.0 20.0 9.0 Sandy Loam 46.7 1.317 0.600 20.7 3.4 ( H ) 2.814 87.0 9.5 3.5 Sand 8.8 1.544 0.860 55.0 2.97(2.04，2‘37， 4.5) ( 1 ) 2.688 93.0 4.0 3.0 Sand 1.7 1.454 0.802 37.6 3.23(2.1，4.35) ( J ) 2.750 90.0 日日 2.0 Sand 2.1 1.602 0.830 310.0 3.93(2.60，5.25) ( K ) 2.640 99.0 1.0

。

Sand 1.2 0.970 0.630 ( L )

'

3

2.650 100.0

。

。

Sand 2.0 0.868 0.555 (円) 2.711 92.6 7.4

。

Sand 2.2 1.490 0.800 28.8 2.09(1.78，2.40) ( N ) 2.700 95.8 4.2

。

Sand 2.2 1 .430 0.700 39.3 3.20(3.00，3.40) ( 0 ) 2.675 92.2 7.8

。

Sand 4.0 1.500 0.805 44.6 2.25(1.90，2.60) ( p ) 2.665 84.0 7.0 9.0 Sand 24.3 1 .200 0.580 199.0 .10 ( Q ) 2.657 70.0 22.0 8.0 Sandy Loam 39.0 1 279 0 782フ04 1 ま10 Table 1. 1) Coefficient of Defomation obtained by Unconfining Compression Tests 2) Conso1idated Yie1d Stress 3) TOYOURA SAND 。e一 刷 司 巴引p川、刈i川白 ロ ロ 。

。

0 0 0 0 0 0 8 6 4 2 ( U F ) μ Z O 戸 ω 一 ﹂ h D L ω に F L パ V C ω U L ω 牛 。 G U 凶 z -o : ThlS Sluoy 工竺J.laton.be，etol.(1975) Fig. 2. Fine contents VS e m•x. emlx 。 。 ωE 0.50 k g ω ().25 o m 5 1 M D お Q U

1

v

m 判 ∞ m h o emax -emin Fig.3 。 。

不撹乱砂質土の液状化特性による実験的研究 257 E E 0.6

.

霊 園 0.' void ratio， e Fig. 4. ernax -ernin VS ernax， ernin EQUIPMENT AND TEST PROCEDURES

CYCLIC TRIAXIAL COMPRESSION TEST EQUIPMENT During this.investigation all tests were conducted in the Cyclic Triaxial Compression Apparatus， as shown in Fig.5. This apparatus essentially consists of a triaxial cell， two loading systems for applying axial and lateral cyclic stresses on the columnar sample， and electronic recording system for measuring the dynamic axial stress， LIσ1， lateral stress， LIσ3， axial strain， el， and the pore-water pressure， L1u.For ap -plying constant cyclic stress on the sample， two hydraulic cylinders were controlled by the electrical hydraulic servo system. The sample is 50mm in dia -meter and 125mm in height. TEST PROCEDURE Liquefaction t巴stson disturbed and undisturbed samples wer巴 conducted. The preparation of each sample was that; Oil Pressure Supp1y Triaxia1 Cell Air Pressure Oi1 Pressure Supp1y

1) For disturbed samples; each sample was test -ed under loose and dense stat巴s. For loose

state， de-aired saturated sand (voiled from two to three houres) was carefully poured by using a spoon into the water filled specimen mold fixed in triaxial cell.Dense state was obtained by shotting the loose sample using a small hummer. 2) For undisturbed samples， to avoid disturbance of soil-skeleton， which a妊ectsthe test re -sults， all undisturbed soil were carefully insert -ed into the cylinder with 70mm in inside and 300mm in height at the field， and these were trimmed by using a trimmer in according to the dimmension and placed into the specimen mold such a standard permeameter test de司

vice， which was able to take this mold to two pieces in the triaxial c巴11 at the test and

the sample was saturated by for twelve hours with moderate head (ranging from 50 to 100cm)， and saturated undisturbed sample was set on th巴 base in triaxial cell.In this way， almost of all samples had obtained value of ranging from 90 to 95 percents in degree of saturation. For both disturbed and undisturbed samples， the experiments were conducted under the same consoli -dation condition. (Ko

=

1.0condition). The cyclic axial and lateral stresses， as which if the axial stress， L1σ'3 ， ln~ creased， the lateral stress， LIσ3， d巴creasedin equal

amount simultaneouslly， were introduced by the afore -mentioned actuator on a sample at 2.0HZ. For un -disturbed sample， especially， to taking a complete saturated condition， the back pressure of1.0kg per Sq cm applied in the s出nple.Fig.6 shows the typical record of strain， excess pore-water pressure and lood during cyclic loading test. ① Specimen ② Transducer (Axi a 1 Stress) ⑦ 日 (Latera1Stress) @ " (Pore-water Pressure) ⑤ 日 (Axia1strain) ⑤ V山 meMeasuri ng Devi ce ⑦ Gui de Oevi ce ③ Regu1ating Va1ve ② Hydrau1ic Cy1inder ⑪ Servo Va 1 ve

o

Base ⑫ Transducer (Axia1 Strain，Gap Censer) Fig. 5. Schematic drawing of the dynamic triaxial aparatus

男

the same as compaired with the results of disturbed sam-ple shown in Fig.7. The reason of the difference of test result on liquefaction characteristics will refer to the later section Almost all experiments for undisturb日ds呂mples were conducted under onεkind of confining pressure (mainly，

σ

o

equal1.0kg per Sq cm). Consequently， in this paper only the experimental data points under 吋 =l.Okg p色rSq cm condition wer邑shown. 義 市 長 夫 ー 大 哲

ドー~/vv~~vWJ\M~~M\fW~

L

どこ三一

j

村 奥 258 ，VVf'l/I.!¥ 25 30 l'i

{附TE](llcompletel!qJ噌 fðct lOn ， Ru•100 (2) 'n'tial11~uefactlon ， R (3) 10lt 1 ，..，1 j quefðc~ 100， R ¥ ト 一 一_l，ec¥O 1 寸

r

_

_

_

_

_

_

_

_

。_{_o} γ~. ---、組 ~ー~下 0-一一旦一 -図二一一---匂 一一一一- 0 」一一一ご{ヶーコ一一一 @ 侃"1.Ok9/c町ドふ1.295 ーーーー 企 配=2.0 " ';:=1.271 =3.0 e-l.131 e MeanVoldR， 。ゎ 一 一

￡!n'tlal Effect下veConflnlnqPres，ure 10 30 100 JOO NumlierOfCyclesCaU51時Initial Liquefact'帆 " Typical relationship between str巴ssratlO and number of cycl巴srequired to caus巴 initial liquefaction， sample number (M) 寸一一一一丁一一一寸 寸

。

Fig.8 0.; ト ~ 04 6 P ~ 0.2ト 古 川 ト Typical records of strain， excess pore-water pressure and load during triaxial Iiquefac -tlOn t巴st TEST RE8ULTS Fig.6. ¥ イ @ぬ可町一一一 U 加p閃附e巾_h山…川叫，_山r / -._.--r>且L_屯 J、ム IJ " 町 」

一一唯一

j

(l e-0.639 a.'"O.5 ~;..，' 。 ト0.639~=1.0" () e-0.639 a::=1.5" 図 e-O.717~O.5" 回 .-0.717~l. O" [i: e-O.717 ~1.5 5 4 3 2 1 0 0 0 0 0 間 F E J -は り ¥ β } 色 刷 J 0 2 3 凶 g b M 10 30 100 300 l

IumberOf:tycle，CaUSl町lnltiolL!quefaction， 且

Typical relationship between stress ratio and number of cycles required to caus巴 initial liquired， sample number (F)

。

Fig.9. The liquefaction test is very difficulty as com pared with a standard soil test， and the value of experimental data will be a妊ect巴dby the difference of the special character of apparatus， test procedure， etc.. In order to check up the special charact芭ron the test results obtained from this cyclic triaxial compres -sion device， the test results of disturbed samples are shown in Figs. 10 and 11， together with the other resercher's results Fig.lO shows the relationship between the cyclic stress ratio， Rn 1 ~ 10， causing liquefaction at ten cycles and the relative d巴nsity，D" for the five kinds of sands (Criterionε"

The experiments were conducted under various effective confining pressure and void ratio. Fig. 7 shows th巴 typical relationship between the cyclic stress ratio， Rnh (Rn'=

r

d

/

σL，td maximum cyclic shear strεss，品:initial effective con fining pressure) and the number of cycles to cause liquefaction， N 1， as a function of void ratio for dis -turbed saturated sand (TOYOURA-SAND). Similarly， Figs. 8 and 9 show the typical relationship for undisturbed saturated sand. From Fig. 7， it can be S巴enthat the cyclic stress ratio is significantly a百ect ed by the value of the void ratio， and in case of repre -sentingthe liquefactionresistance by stressratio， there is not a very signi五cantInfluenc巴onthe initial e在ective confining pressu日 Thisfinding is in agreement with

th巴conclusionpresented by Seed， Lee and many other

investigators. On the other hand， from the test results shown in Figs.8 and 9， it can be seen that there is a high scatter in data points and the manner or the degree of influence of void ratio and initial e妊ect1ve confining pressure on liquefaction occurance is not Dist~r凶 S 岬" A 品 @ 目 白 0.; 町 石 0.4 官、 、 σ - 0.3 叫 E U f 竺"0.2 .μ 」 z L 白 b 0.1 凶 11edn'Iold RatlO

('i(，" !nitialE伴 氏tive(onf1川ngPre，5ur 10 30 100 ]OG 川berofCyclesCauslnglnltial Liquefacti帆'" Fig. 7. Typical relationship between stress ratio and number of cycles required to cause initial liquefaction， sample number (K) Toyoura -sand 0 1

不援乱砂質土の液状化特性に関する実験的研究 259

contained fine cont巴ntsof白vepercents. The full linε

is th巴 experimental equation (τ1=4.6xlO-3 _xDrx

品， Dr : %) proposed by Tanimoto. Further， in Ref.5 and 7， Watanabe and Ishihara presented available data on the liquefaction characteristics of wide range of sand type and they proposed from th巴analysisof these data that the relationship between cyclic stress ratio and density， soil typ巴 and soil gradation would be able to be shown by the factor， e-emin， in which e-emin is a valu巴ofthe difference between

void ratio of sample and minimum void ratio. Based on their proposition， the relationship betw巴en the

cyclic stress ratio causing liquefaction at twenty cycl邑s，Rnl=20， and the factor， e-邑ml町 forthe all dis

-turbed samples tested during this investigatinn is shown in Fig.l1.From the data in Figs.l0 and 11， it can be seen that in Fig.l0， the mean value of cyclic stress ratio obtained during this investigation is smal! ranging from 0.03 to 0.07 for Dr>40 percents， and 0.6 0.5 N， =10 cycles (Criterion宮ε1) Oisturbed Sample 0.1 0.4 一..-.:l~ ニ 4.6xl0'0rσ '(TANI~10TO，L(1971) Tt =3.8xl03 Sampl e F .Cく5.0%

4

_L y 勺 ~ 0.3 O H E主ロユ 凶 ω ω d 品川戸山 20 40 60 80 100 Rerative Oensity， Or (%) Fig. 10. Comparison of triaxial compression test results (N Iニ 10cycl巴s) M加山 r ﹁ C s 巴 E y c 0 2 官 ι 也 1 D¥ S turb~d S.mp 1 e ¥ lotilrabe，t oel，19iS c 0.3 N J 。 ;; 0.2 4巴 巴

，

Cえ 0.1 e-eml~ -(1-0rH巴 悶 ズ 宮 町lln) Fig. 11. Stress ratio VS e -emln， (NI=20 cycles) in Fig.ll， the cyclic str邑ssratio is also small about 0.09 INFLUENCE OF DIFFERENCE ON UQUEFACTION CRITERION As pervious mention， liquefaction ph巴nomenon may be consid巴redthat the cyclic shear stresses induced by earthquake act on saturated sand under undrained condition， and by this action， the pore-water pressure and the strain are built up to the point of sudden increase whichdenot巴sthe onset of liqu巴factionwith increasing of number of cycles， and at last， the巴ffectivestress become to zero， liquefaction state app巳ars For practical purposes， it note that damages of saturat巴dsand layers and earth structurs will occure not only under the complete zero e百ective stress condition (difined by “complete liquefaction" in this paper) but under th巴conditionof sudden increase of pore宇water pressure or strain (difined by "initial l即lefaction" in this p旦per). Then the cyclic stress ratio required to liquefyfor all disturbed and undisturbed samples were obtained by three kinds of failure criterion as following; (1) complete liquefaction when pore-water pres -sure equals to initial effεctive confining pres sure or becomes constant during cyclic loading. (Cyclic stress ratio decided with this crit巴non indicated by R"u~loo) (2) initial liquefaction when por巴-waterpressure suddenly increase during cyclic loading (Cyclic stress ratio decided with this criterion indicat巴dby RJu) (3) initialliquefaction when axial strain suddenly increase during cyclic loading. (Cyclic stress ratio decid巴dwith this criterion indicated by Rrl) The comparisons of the magnitude of cyclic str巴ss ratio obtained from three crit巴rionfor liquefaction occurance are shown in Figs. 12 and 13. Fig.12 shows the relationship between the ratio of RεI to R.1U叶 00 and void ratio， for undisturbed and loose and d巴nsedisturbed samples. From this自gure， it can be seen that almost all values of ratio， Rε1/RJu=100， are plotted below 1.0; for disturbed samples， it rangs from about 0.07 to1.0， and for undisturbed samples， it rangsfrom about 0.07 to1.15. Similarly， Fig.13 shows the r巴lationshipbetween the ratio of RoI to R"u and void ratio. It can be seen from this figure that all of the val田 ofratio， RεI !RJu， are near to1.0. From these 白1dings，it becomes clear that the cyclic stress ratio causing initialliquefaction is small about thirty perc巴nts compan巴dwith that of completely liquefaction， for both disturbed and undisturbed samples， and on the criterion of initial liqu巴factiol1， th巴 stress ratio is almost unaffected by either criterions， pore匂water pressure or axial strain

260 奥 村 哲 夫 個 大 根 義 男

INFLUENCE OF SAMPLE PREPARATION PROCEDURE ON THE LIQUEFACTION

OCCURANCE It may be considered that even though the void ratio of sample， the magnitude of cyclic stress， the effective confining pressure acting on the sarnple and th巴nurnber of cycles且rethe same， the resistance to liquefy will diff巴rin consequence of the soil skel -eton， th巴 interrocking between soil particles and aration procedure; Fig.14 shows the relationship betwe己ncyclic stress ratio to cause initial liquefaction呂ttwenty cycl芭sand relative density obtained by two sort of sarnple pr巴p -aration procedurε， (1) pouring de-aired saturated sand into the wat邑r-filledspecimen rnold using by a spoon. (2) tarnpping rnoist sand using by a tamper. It can be s色enfrom this figure that the cyclic stress ratio by tampping have a higher resistance to liquefy than that by pouring

PORE司WATERPRESSURE DEVELOPMENT

DURING CYCLIC LOADING

It is a essential particular for liquefaction analy sis to estimate the magnitude of pore-water pressure which develops in sands during earthquakes. To estirnate the magnitude of pore-water pressure development in disturbed sand which is subjected to cyclic shear stress applications， the results of the 1.1 1 1 [ C D 1 b d PI • :Undisturbed Sample 1.0 f-••••••••••• ....0

。

o 0 む場夢 @

。目

。

。

も

@ @ _@ トー0-1 @

。

c、 日、z、0.7 0.6 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 Void Ratio (e)

R" S_{strainsuddenlyincreases} tress ratio defined as the cycle when the axial 5tress ratiodifinedasth川ecycle when the excess porewaterpressurefirstequals thein下tial confin下"9 pressureorconstants ROlF-100 Fig.12 Cornparison of stress ratio obtained by di百erentcriterion for liquefaction occurrence 1.1 1.， 1.0 0.9 0.8 ~ ¥ に 戸0.7 0.6 (3u0 G、 @ 8 0 V Q G⑫ o 0 '6..0.. . ...tY...f; . ..c噛 骨 》 も

。

@

。

o】sturbedS何 伊 " ~ Undlsturbed Sa叩1， 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 ViotRatio (e) : $tress ratiode子inedil5 the cyc1e when the eχces5porewaterpre5suresuddenly increases Fig. 13. Comparison of stress ratio obtained by di妊erentcriterion for liquefaction occurrence 0.6 ~ 0.5 口 N J.0.4 日3 。 戸 判 同 区 0.2 ωL 一ザ 0.1 Samp 1 e No. ( A ) N‘=20 cycles (Criterion -I ) Init. Efec. conf. Pressυre

Rodding Specimens 20 40 60 80 100 Relative Density， Dr Fig. 14. Cornparison of the liquefaction resistance of medium to dense specimens prepared by rodding and depositting nurnb巴ぽr0ぱfcycles until c

∞

omplet民巴 l日Iqu巴factio叩n stat白E versus r巴Sl凶dua剖1pore

and r巴ela抗叩tJ悶凶lIpbetween ムu/σL/ 弘~ and N/N! for the samples having ten percents of fine contents (where， L/u: residual pore-water pr巴ssure at N cycles，σ

レ

initial e百巴ctiveconffi.ng pressure， N: nurnber of cycles under consideration， and N

，

:

number of cycles to cause complete liquefaction is plotted in Figs.15.1 to 15.4at void ratio ranging from 1.04 to 1.15， frorn 0.89 to 0.94， from 0.78 to 0.84， and from 0.65 to 0.74， respectively， and upper and lower boundery of data points ar巴alsoshown

不撹乱砂質土の液状化特性に関する実験的研究

Th巴seresults relatively have a broad scattering band

Fig.16 shows the mean curv巴shownin Fig.15 for

each mean void ratio It may be seen from these figuars that for a given value of N/Nl， valu巴ofthe ratioL1u/σo at high void ratio is higher than the value at low void ratio， and that valu色ofthe ratioL1u/σo increase with increasing value of the ratio N/N" and that the shap巴orthe

magnitude of average pore-water pressur巴dev巴lopment curves are almost the same for void ratios of 0.92 and 1.10. 1.0 0.8 ノQ-.L " ・fム ノ・匂・2 〆 -L ' / ._.1''''_， / .'7 ;.' 〆 A・， 〆 。 ノ / Upper boundaryミノ".， ';/ .)1 〆 y ・ / ト_{lean curve}ー_ーベへ_、 少〆_，.. y τ .Y ' ・ ，.A -¥ て 〆 〆 ・ ・ y '〉 ぐ " '-~ _" ー〆_"""'__'/ '.〆 ./，- ，y_{〆 ・ 〆}， 'γ 〆 / ~//く~L刷erbou油cy 0.6 ，

.

-_

.

7'い<10兎 。0.' ¥ コ q 0.2 0.01 0.03 0.1 0.3 1.0 N / tH. Fig. 15.1 i1u/ (]~ VS N /N 1 (N : spontaneous number of cycle befor liquefaction occurance) 1.0 0.8 0.6 ，g 0.4 0.2 0.01 0.03 0.1 0.3 1.0 N I r-u. Fig. 15.2 i1u/的 VS N/N1 1.0 Oisturbed Sample 0.8 0.6 e=O.78..0.84 (長o0，81 ) 74ィ，10 . e0.4 0.2 0.01 0.03 0.1 1.0 N / Nt Fig. 15.3 i1 u/σ~ VS N/N1 261 Fig.17 shows representative results for undis turbed samples. From this figuar， it may 旧 民 間that

the results have a broad band compaired with the results for disturbed sample

Fig.18 shows the results of disturbed and undis -turbed sample for sample number (1).It is evident that in spite of almost th巴samevoid ratio， the value of the ratio i1u/dofor disturbed sample is small芭r m all range of N

/n

1.0 0.8 0.6 e=O，6S_0.74 ( ; • 0，70 ) 7'μ<10 % -- 0.4 2 0.2 0.01 0.03 0.1 0.3 1.0 H I H! Fig. 15.4 i1 u/σ~ VS N/N1 1.0 0.8 0.6 <l0.4 0.2 0.01 0.03 0.1 0.3 1.0 N I N.R FIG.16目

Relationship between i1u/σ~ and N/N

，

for di任巴rentvoid ratio， disturbed sample 1.0 Undisturbed Sample 圃 ・ .園-，

・

.

・1 .

.

・ 山 " .・・. •• "it::

"

・

'

.

'

0・.. ・" e .'

.

.

.

.

_

，

.

.

_

.

a・ 一 .巴ー

・ . .

. ・ ._・.

.

0.8ト e=0.792_0.842 ( 74μ<: 1 0 ~， ) 0，6

←

σ e o = 1 G 9 K / m c 2 二0，4

←

'. -..; . ・

_.

•

•

•

•

0，2ト 4 . •

•

a

•

••

i . . 固 . e . ， r

••

•

••

•

•

•

圃 1 0，01 0.03 0，1 0.3 1.0 N / N9. Fig. 17. i1u/的 VS N /N" undisturb巴dsample

262 1.0 0.8 0.6 ' 0 o ¥ 三0.4 0.2 奥 村 SA，υleNo (J) ， :Disturbed Sample， e " 0.932_1.027 (e=O.980) 。UndisturbedSample， e " 0.925..0.971 (e=O.94S) σ; " 1.0kg/cm2 E 句 0，01 0.03 0.1 N I肌 ロ B g e @ 号 0.3 哲 。 コ ' 町 。 @ 1.0

Fig.18. Conparison of L1U/σ; and N/N1 fOT

disturbed and undisturbed samples

CYCUC STRESS RATIO司AXIALSTRAIN

RELATIONSHIP AT LIQUEFACTION OCCURANCE Th巴 plots of cyclic stress ratio， Rnl~20 ， versus axial strain，εlh at ons巴tof initial liqu巴faction(defined as axial strain suddenly increase during cyclic loading) are shown in Fig.19. This figure shows the test results of disturbed and undisturbed sample for sample num ber (C)， representatively， and shows in each void ratio， for disturb巴dsample

From this figure it may be seen that for disturbed sample， the relationship between Rnl~2o and ε1 have a constantform; the value ofaxial strain enlarge with

ll1cr巴asing value of cyclic stress ratio for a given

void ratio and with increasing a void ratio for a given cyclic stress ratio. On th巴otherhand， for undisturbed

sample completely have a di任erenttendency as com pared with the results of disturb色dsample; the val問

。

faxialstrain incr巴asewith decreasing the valu巴of cyC]ic stress ratio untill axial strain of about 0.6 per cents LIQUEFACTWN CHARACTERS OF UNDISTURBED SANDS 1n order to evaluate the lique五ngstrength of un苧 disturbed sands， the relationship b巴tweenstress ratlO and relative density for all undisturbed samples are investigated in the same manner as that of disturbed samples. This results ar巴shownin Fig.20， together with the r巴sultsof the disturbed samples shown in Fig.lO. From this figuar， it can be seen that for the experimental results of disturbed samples， the magni -tude of stress ratio increases in accordance with in -crement of relative density and the stress ratio is in proportion to the relative density. This finding is in agreement with the traditional conclusions. On the other hand， it can be seen from the experimental

夫 ー 大 根 さ 南 我 Fig.19. 男 results of undisturbed samples shown in the same figure that the value of stress ratio is far high as compar巴d with the results of disturbed samples， and there is not 旦linearrelationship b巴tweenstress ratio and relative density Thus， liquefaction charact巴rsof undisturbed sands are different compared with disturbed sands. The cause of this difference may be assum巴dthat， in th己 disturbed sample， specimens were prepared by pouring the de-aired saturated sand as described in“TEST1NG PROCEDURE"， and in this case the soil particles form the single-grain己dstructure Consequently， the liquefing strength of disturbed sam-ples will differ only the magnitude of relativ巴density

While， in the undisturbed sample， samples may hav日

had a latent strength sllch a cementation and may have been subjected to stress history for long term， for example， cyclic str己ssdue to an巴arthquakesand

change of static stress due to the change oI topog -raphic f巴呂tures. Consequently， such undisturbed sam-pl巴shave a highly resistance to liquefy and furthεr， liquefing str巴ngthis highly complex Itwill be able to consider， from the assumption described previously， that the soil parameters，五ne contents， consolidated yi巴Idstr巴ssand co巴 伍cientof deformation， are related to the liquefing strength of the undisturbed samples. Thus， th日 relationships

between such soil parameters and stress ratio were plotted. These results are shown as following Fig.21.shows the relationship between the ratio R，.d.， (the ratio of cyclic stress ratio for undisturbed sample of diluvium or tertiary ear sands， Rundie.. to cyclic stress ratio for disturbed sample at the same void ratio in undisturbed sample， Rd i，.，which obt呂ined from cyclic stress ratio-void ratio relationship) and 自n巴contents，F.C.， of the sample. In this五gure，the results for diluvial sands presented by other investi -gators ar己alsoplotted (2，8，9，). Itmay be seen from this五gurethat the value of RU.d. for diluvial sands is almost1.0 below about 10 percents in F.C.and is somewhat high above 10 percents in F.C.， How ever， for tertiary era sands it have higher value below

(M).The cyclic stress ratio at con命ringpressures of 1.0， 2.0 and 3.0kg per Sq cm for undisturbed condition are plotted with(・)，(企)and (・))the value of cyclic stress ratio for undisturbed sample tested under the initial con金 山gpress町eof 2.0 and 3.0kg per Sq cm are apparently small as compaired with the result at confining pressure of 1.0kg per Sq cm and undisturbed sample at confining pressure of 3.0kg per Sq cm and disturbed sample have almost thelsame value in cyclic stress ratio. This長ndingshows that the value of cyclic stress ratio tested under the initial con命1Ingpressure of 1.0kg per Sq cm， Fig.20， 21 and

22， are the results under over consolidation conditions， because the consolidated yield stress of the all undis. turbed samples (sample number W exp巴cted)have

higher values more than 1.0kg per Sq cm. Conse. quently， it can be concluded that the undisturb巴d samples have a highly resistance to liquefy influenced by the over consolidation ratio To investigate the influence of over consolidation ratio on cyclic stress ratio of undisturbed samples， 社le relationship between the cyclic stress ratio，

(Rnl~10)und; ，.， and Py/σb is shown in Fig.24. The

length of straight line of data points in tlris figure in -dicates the magnitude of scatter on test datas. From this負 担re，it is apparently to see that the undisturbed samples having high over consolidation ratio exhibit high resistance to liquefy. Fig.25 shows the relationship between cyclic stress ratio causing initial liquefaction at 10 cycles under the initial e妊ectiveconfining pressure of 1.0kg per Sq cm for undisturbed samples， (Rnl~10)und;" and modulus of deformation obtained by unconfined com. pression test， E50.，It may be seen from this figure that the test data protted in this figure fall witlrin som巴whatnaロowscatter band compared with the

263 不撹乱砂質土の液状化特性に関する実験的研究 10 percents in F.C.， especially， and further， the fine contents app巴arto have an insignificant influenc巴on ratio， RU.d Fig.22 shows the relationship between cyclic stress ratio， (Rnl~10)und;" and the consolidated yield stress obtained from consolidation test， Py， for undis turbed samples tested under the same con危llngpressure， ぬ =1.0kg per Sq cm

From this figuar， it can be seen that this figuar have somewhat sccatter of data points， however， cyclic stress ratio increases with increasing value of consolidated yield stress.

However， as shown in Fig.23 (Fig.23 shows the relationship of stress ratio vecsus void ratio for dis. turbed and undisturbed conditions of sample number

•

. + 1 4

・

竺 少 グ

7

ー-.-c'デ勺 午ム占J 、、Boudsofthetestresults l I I 40 60 80 100 Relative Density， Dr Stress ratio required to caus巴initial Iiquefaction in 20 cycles for undisturbed samples VS relative desity 140 120 Und;sturbedSa，rJple N~= 20 cycles (criterionε，) d.=1.0kg/印 I

•

2 0 Fig目20. 1.4 1.2 0.2

。。

巴 1 .0 E 2 {

.

t、 " ~ 0.8 巴 。 : ; 0.6 z 凶 例 ω 』 ，;;: 0.4

i

:

:

t

r

ヰ

ニ

一

一

一

N.t= 10 cycles e " 0.64 _ 1.19 σ:=1.0kg/cm~ ( Criterion : E-t )

1

1

十

l

1.2ト

•

•

ー

1.4 目 止 よ (kg/cm2) 1 I I I I 2.0 4.0 6.0 8.0 10.0 白n臼lidatedYield日 間 日 .'y

。

。

60 0， 1.0 1.0

"

'

₁₀ 1 5 _ 35 10 20 回 目

•

平 ~1危 1

20 30 40 Fine Contents， 日 ( 包 ) Geology Specimen Tertiary. S Sedimentation Oiluyial. 5 Sedimentation Diluvial. S Moist Tamping Oiluyial. S

。

回 l'e ト

一

3.0 主2.5 也 、 腕 g2.0 1.5 1.0 0 Source of Data Author Watanabe，et a1 附ulilis，etal Yasuda， etal -A O 回 consolidated yield VS Stress ratio S仕ess Fig.22 Ratio of stress ratio of undisturbed to disturbed sample VS fine contents Fig.21

ε， ) 25 N，Q'"10 cycles ( Crite門on 男 義 1.4 1.2 根 夫 ・ 大 哲 村 奥 σo"1.Okg/cm2 n;10 Oisturbed0 UndisturbedSIoo"l.Okg/cm1 a ..0;=2.0 国σ)i=3.0" D __________..

=

=

=

=

=

=

三

264 0.5 0.4 0 H μ ~O.3 ~O.2 山 』 ~O.l 0 1.1 1.0 20 15 曜 静 ぱ O m / 抗 p y r s Q U ρ し r i t Q U @ @ 曜 静 ~ 0.8 ℃ c コ H

J

0 6 0 五日4 E ω 凶 机 ω L 出 0.2 1.4 1.4

。

¥ ¥ h - - J J

十

-

-

一

一

一

一

-

下

←

i

1.3 1.3

¥寸〈:土

1.2 1.2 。0.4 1戸 μ dO.3 例 悶0.2

.

」 μ 的0.1f -0 1.1 0.4 0 戸 ~ 0.3 g : ;: 0.2 ω L :;:0サ1 0 1.1 0.5 0.5

P

y

/

σ

;

パイ 〆1. イ ペ イ T Jイ ，ィ~' パ イ / レ_)-.1' . / レ

「

a d

ぺ

_

.

I

).JJ).ル

y

s

0.4

L

舟

dfDr

「

用ジ~.v..V い，，~，.，あ..V : ; : ; _1....⑧J.唱 VS N【 :1 ucycl es (Criterion'c，) 1.0 kg/cm' Fig.24. 1.1 ' " 1.0 c 。 三 0.8 1.2 1.3 VOid Ratio e Stress ratio in disturbed and undisturbed states for sample N o. (M) VS void ratlO 1.4 Fig.23

。

。

00 ) 3 m c / q k o 5 2 0 5 E 0 0 2 n o t a m r 0 0 5 5 T 1 e o f o s u o -O U 1 d U M門 modulus of Fig. 25. Stress ratio VS d巴formation 50 3.0 _ 2.5 γ、 e u ¥ 町 三4 1.0 5 o o l -2 σ F Z H w m E U L u w w z 。 志 的 ω ﹂ a E o u -u o E F U W C O U P E

results shown in Fig.24 and the value of(Rnl~IO)u 吋"

increases proportionally with increasing value of E50' It is n巴cessary to note that the test data

shown in Fig.25 are the result under the initial effective con自ningpressure of 1.0kg per Sq cm and if the initial confining pressure varies， the relationship shown in Fig.25 will change due to th巴influ巴nceof over consoli.

dation ratio.

The relationship shown in Fig.25 may be used together with the r巴sults shown in Fig.26 and 27，

( Fig. 26 shows the relationship of unconfined compression strength， qu， versus modulus of deforma tion， E50， for undisturbed sampl邑sused in this inves. . tigation， and Fig. 27 shows the relationship of E50 versus N.Value for many kinds of soils suggested by many investigators)， to芭stimatethe magnitude of cyclic str巴ssratio for undisturb巴dsamples subjected to cyclic shear stress applications 355 100 150 200 250 30C nodulus of Oeformation， ESO ( K/c酬 ) Fig. 26. Relationship between uncon負nmg compression strength and modulus of deformation， unsaturated samples 5

0 SUMMARY AND CONCLUSIONS

The investigation described herein are on the liquefaction charact巴ristics of disturbed and undis -turbed saturated sands under cyclic loading triaxial compression conditions. Based on the aforementioned experimental血ldings，the following conclusion are drawn町 1. Th巳cyclicstress ratio obtained by using this

apparatus have somewhat small value as compared with the value obtained by many other investigators

265

at onset of liquefaction for undisturbed sample di妊er by the inf!uences described in 4

1) 日町B.Seedand K.L.L邑e: Liquefaction of Saturated

Sands during Cyclic Loading， J our. of the Soil Mechanics and Foundation Division. ASCE. Vol 92， N 0.SM6， 105-134， 1966，

2) H.B. Seed， Ignacio Apango and Clag巴nceChan

Ev且luationof Soil Liquefaction Pot巴ntialduring

Earthquakes， EERC， Report NO.EERC 75-28，1975，

3) JSSME:針。c.16th Symposium on Soil M巴chanics

Engineering， 1971，

I

.

P.lV1ulilis， H.B.Seed， C.K.Chan， ].K.Mitchell and Kondiah Arulanandan: E任ectsof Sample Prep -aration on Sand LiqueIaction， J our. of the Geo宇

technical Engineering Division， ASCE， V 01.1， N o. GT2， 91-108， 1977，

5) Watanabe， Sod色kawa，Tanaka呂ndHioki : Tuchi

no ekijyoka ni oyobosu ryudo oyobi sairyubunga -nyuritsu no eikyo， Tsuchi to Kiso， JSSIVIFE， Vol 23， No.6， 37-24， 1975， (in Japanese)

6) Ishihara， K 呂ndYasuda， S. Sand Liquefaction

du巴toIrregular Excitation， Soil and Foundations，

JSSlV1FE， Vo1.l2， N 0.4，65-77， 1972，

7) IshiharaK. : Doshitsu dorikigaku no kiso， Kashima syuppan kai， 1977， (in Japanese)

8) Ishihara and Tanaka : Sairyubun 0 fukumu fuka -kuransa no ekigyoka， Proc. 9th Annual R巴serch

Meeting oI JSSFIVIE， 379-382， 1974， (in Japanese)

9) Sakai and Yasuda : Fukakuransashitudo no eki -jyok旦 tokus巴i，Proc. 12th Annual Res日rch of

JSSFME， 389-392， 1977， (in Japanese) 10) JSSFME: Doshitsu chyosaho， 1972，

11) Seed， H.B. and Lee K.L. : Liquefaction of Satu rated Sand during Cyclic Loading， Proc. ASCE， SM6， 105-134， 1966，

Y orihiko， Ohsaki Effects of Sand Compaction during the TOKACHIOKI Earthquake， Soil and Foundation， JSSIVIFE， Vol.lO， N 0.2， 1972，

13) Seed， H.B. and Idriss， I.1VI. Analysis of Soil Liquefaction ; NIIGA T A Earthquake， Proc ASCE， Vo1.93， No.SM3， 83-108，1967，

14) Seed， H.B. Landslides during Earthquake due to Soil Liquefaction， Proc. ASCE， Vo1.94， N 0 SM5， 1053-1122， 1968， REFERENCES 不撹乱砂質土の液状化特性に関する実験的研究 4) 12) 80 ① 日 目3N， Flne Sofld (S.t)I，d=O.2-0.06， C戸，5/5配ιr=O，7日 ① [~71+4 ，9 N、FineS削(叩)，d=O.2-0.06，C戸，5/1凹;，r=O.9叩 ① E=39+4.5N， Sand¥叫)，d=1.5-0.1， Cu=2， Sr<50¥;，戸O，95d ① [:43+11.8N，5and¥'iithG加Y旧a附 (凹S即p一G刷p円)，d制=叶21ト-0.1刊5，九C怠叫4 5/50.， r=O.886 ① ド 加10.5N， Gravel(昨)，d=63-0.06， Cu=印，5/日 r=O.783 ① [=2判 3N， Silty Sand(町)，d=2.0-Q凹，CU=8，¥(85%，戸O.76Q ⑦ [=12+5.811，SilしFineSil ty Sa吋(CL)，d=O.1-O.002， C戸。 S戸田レド0.904 ① E-"4+1l.5N，Si1t，Cla戸ySilt(CL)， d"'O.l....(O凹1，Sr<s5i， IEH51rz0924 (NOTE) Cu: Coefficient of Unifonnity Sr: DegreeofSaturation CorrelationCoefficient d : GrainSiZE( 胴 ) 1VI0dulus of deformation VS N-value

This di紅巳r巴nceof test data lies in ranging from 0.03

to 0.09.

2. On the criterion of initial liquefaction， the procedures using pore-wat巴r pressure development

and axial strain development have the same value of cyclic stress ratio. This finding observe in either samples， disturbed or undisturbed sample. The str芭ss

ratio decided at initial liquefaction state has a valu巴

of about 30 percents lower in maximum di妊erenceas compar巴dwith the r巴sultsdecided at complete lique

-faction state.

3. Th巴di妊erenceof sample preparation proce -dure inf!u巴nceson soil skeleton， interrocking betw巴巴n

soil particl巴andantec巴dentstress

4. Cyclic stress ratio on undisturbed samples such as tertiary era sands and diluvium sands have high values a狂ectedby cementation between soil par -ticle and over consolidation. Further more， the rela tionship of cyclic stress ratio against r巴lativedensity

does not have any tendency for disturbed samples 5， In case of showing the liquefaction resistance for undisturbed sample as a p且rameterof Eso， there

is an almost linear r巴lationshipbetween cyclic stress

ratio and E50' For practical purpose， this experi -mental r色sultindicates the possibility of estimating

in-situ cyclic stress ratio from N-Value

6. In comparison with the liquefaction charactor of disturbed samples， the mechanism of pore-water pressur巴development乱ndmagnitude ofaxial strain

)0 60 50 30 40 N '!alue Fig.27 10 凹 佃 叩 叩 叩 S 4 3 2 U ( N E ¥ J ) 凶 一 C O Z E L O E ロ Z22uz