See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/345470486

Cyclic and creep combination effects on the long-term undrained behavior of

overconsolidated clay

Article in Acta Geotechnica · April 2021

DOI: 10.1007/s11440-020-01078-5

CITATIONS READS

21 698

4 authors, including:

Zhen-Yu Yin Christophe Dano

The Hong Kong Polytechnic University Laboratoire 3SR Grenoble
429 PUBLICATIONS 11,867 CITATIONS 123 PUBLICATIONS 1,911 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Christophe Dano on 11 June 2021.

The user has requested enhancement of the downloaded file.

 1 Cyclic and creep combination effects on the long-term undrained behavior

2 of over-consolidated clay

3 Jian HAN1, Zhen-Yu YIN2, Christophe DANO3 and Pierre-Yves HICHER4

1
4 School of Materials Science and Mechanical Engineering, Beijing Technology and Business University,
5 Beijing, 100048, China
2
6 Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hung Hom,
7 Kowloon, Hong Kong
3
8 3SR Laboratory, Grenoble Alpes University, UMR CNRS 5521, Grenoble, 38000, France
4
9 Research Institute of Civil Engineering and Mechanics (GeM), Ecole Centrale de Nantes, UMR CNRS 6183,
10 Nantes, 44300, France

11 Correspondence to Dr. Zhen-Yu YIN (Phone: +852 34008470, Email: [email protected];

12 [email protected])

13

14 Abstract: Soft soil subjected to cyclic loading typically exhibits an increase of excess pore pressure under

15 undrained condition which brings the soil to an over-consolidated state. Then, under a subsequent large number

16 of cycles (e.g. more than one million) which also takes time, the creep at over-consolidated state influences the

17 cyclic effect and thus results in a complicated long-term undrained behavior. This paper aims to clarify this

18 long-term undrained behavior of over-consolidated clay. The reconstituted samples are prepared from natural

19 specimens retrieved in the north of France. First, the shear strength characteristics along monotonic triaxial

20 stress paths are identified. Then load control cyclic tests on over-consolidated samples are conducted in fully

21 saturated and undrained conditions. Small cyclic deviatoric stresses are applied in order to investigate more

22 particularly the behavior under a very large number of cycles, during which an unusual pore pressure evolution

23 is observed. To explain this, undrained triaxial creep tests are performed on reconstituted samples with different

24 values of OCRs under some specified stress states. The evolutions of axial strain, excess pore pressure, stress

25 ratio, stress path, plastic strain rates and stress dilatancy during undrained creep are discussed. The additional

26 undrained creep tests also show that two processes are simultaneously acting in a competitive manner: increase

27 of the pore pressure due to the cyclic loading, decrease of the pore pressure because of creep.

28 Keywords: cyclic loading, creep; triaxial test; soft clay; constitutive relation; stress dilatancy

1
29 1. Introduction

30 Various design rules are proposed to geotechnical engineers to calculate both shaft capacity and base resistance

31 of piles under axial or lateral loads. Most of these rules are calibrated from in situ tests (e.g., cone penetration

32 test, pressuremeter test) that allow estimating the static capacity. However, there are more and more situations

33 where geotechnical engineers may have to consider the effects of repeated loads with variable amplitude with

34 time. Wind, waves, tides are examples of physical phenomena that induce cyclic loadings on structure

35 foundations with very large number of cycles over the time. For soft clay, the cyclic behavior has been widely

36 investigated during last decades (Andersen et al.[1]; Vucetic and Dobry [2]; Biarez and Hicher [3]; Hyodo et al.

37 [4]; Li and Meissner [5]; Boulanger and Idriss [6]; Cai et al. [7]; Yin et al. [8]; Mortezaie and Vucetic [9]; Qian

38 et al. [10,11]). However, long-term undrained cyclic behavior (e.g. more than one million) is still a critical issue

39 which was not studied so much in the literature.

40 As well known, soft soil subjected to cyclic loading typically exhibits an increase of excess pore pressure under

41 undrained condition which brings the soil to an over-consolidated state. Then, under a subsequent large number

42 of cycles which also takes time, the creep at over-consolidated state generating negative excess pore pressure

43 will definitely influence the cyclic behavior in long-term. Thus, the cyclic and creep coupling effect at

44 over-consolidated state may result in a complicated long-term undrained behavior, which needs to be clarified.

45 Thus, to investigate this coupling behavior, the undrained creep behavior of over-consolidated clay may also be

46 separately clarified. Up to now, the undrained creep behavior has been widely investigated for normally

47 consolidated clay (e.g., Walker [12]; Arulanandan et al. [13]; Holzer et al. [14]; Vaid and Campanella [15];

48 Adachi and Oka [16]; Hinchberger [17], Yao et al. [18, 19]; Yin et al. [20]; Yin and Hicher [21]; Yin et al. [22,

49 23]; Wang and Yin [24]; Zhao et al. [25]). Overconsolidated clay was considered as relatively stiff foundation

2
50 soils and few studies on undrained creep (e.g., lightly overconsolidated clay by Tavenas et al. [26]) are available.

51 However, besides of the above mentioned cyclic and creep coupling effect, the progressive failure of

52 overconsolidated clay structures or foundations is also mainly due to the undrained creep under the undrained

53 shear condition. Thus, the research in this respect cannot be ignored.

54 Therefore, this paper aims to clarify the long-term undrained behavior of over-consolidated clay to investigate

55 both the response of soil specimens subjected to a very large number of cycles (about one million), and the

56 undrained creep behavior of overconsolidated clay. Reconstituted samples are first prepared from natural

57 specimens. First, the shear strength characteristics along monotonic triaxial stress paths are identified. Then load

58 control cyclic tests on over-consolidated samples are conducted in fully saturated and undrained conditions.

59 Small cyclic deviatoric stresses are applied in order to investigate more particularly the behavior under a very

60 large number of cycles. Beside of this, undrained triaxial creep tests are performed on reconstituted samples

61 with different values of OCRs varying from 2 to 14 under the same stress path having the initial stress state for

62 creeping lying almost on the critical state line in p'-q plane. Finally, the cyclic and creep processes

63 simultaneously acting in a competitive manner are discussed: increase of the pore pressure due to the cyclic

64 loading, and decrease of the pore pressure because of creep.

65 2. Tested clay

66 2.1 Description of clay and clay samples

67 The tested natural clay was sampled in the northern France (at Merville), at a depth between 5 and 11 meters. Its

68 origin is similar to the London clay, more studied in the literature. This highly plastic, stiff clay (wL = 97 %, IP =

69 59 %, OCR ≈ 30 in-situ) with dominant illite and smectite minerals exhibits a uniform appearance in the form

70 of a grey brown material. However, because of their specific geological history, samples also have cementation
3
71 and structuration in the form of pre-existing fissures, which prevent any comparative subsequent analysis.

72 To obtain more identical and homogeneous samples without influence of inter-particle bonds and occurrence of

73 fissures, dry powder from natural samples was mixed with water at an initial water content equal to 1.5 times of

74 the liquid limit (wL) according to Burland [27] and then consolidated under one-dimensional condition for one

75 month which is more or less one week longer than the duration of the primary consolidation based on Taylor’s

76 square root of time method. A vertical stress of σ'v0 = 100 kPa was applied for the consolidation.

77 2.2 Physical and mechanical properties

78 Some typical physical properties of the tested clay were measured and summarized in Table 1. Eight undrained

79 triaxial tests in compression and two undrained triaxial tests in extension with isotropically consolidation and

80 unloading stages were performed on the reconstituted samples of Merville clay with different values of OCRs

81 (here OCR = p’max/p’0, where p’max is the maximum effective mean pressure during isotropic consolidation and

82 p’0 is the mean effective stress at the onset of shearing) varying from 1 to 14. The applied axial strain rate was

83 2.57 %/h for all undrained tests, with all results presented in Fig. 1. The slope of the critical state line (CSL) was

84 obtained from stress paths in p’-q plane, based on which the slopes of CSL in compression (Mc = 1.0) and in

85 extension (Me = 0.96) were measured. The compression and swelling indexes were obtained from e-log(p’)

86 curves. The assumed critical state line in the plane of e-log(p’) is obtained with the same slope as the isotropic

87 compression line in Fig. 2. The results of overconsolidated samples did not reach the critical state line, since

88 axial strains at the end of tests are not large enough for reaching the critical state. All these mechanical

89 properties are also summarized in Table 1.

90

4
 91 3. Undrained cyclic tests

92 3.1 Test program and test procedure

93 All specimens for undrained triaxial cyclic tests are 35 mm in diameter and 70 mm in height. A device with the

94 computer-controlled GDS dynamic triaxial testing system (10 HZ/1 kN) was used for all cyclic tests. After the

95 saturation stage verified by B-check (B value no less than 0.96 in this paper), all specimens were isotropically

96 loaded up to a constant mean effective stress of 400 kPa with a back-pressure of 100 kPa within 3 days, and then

97 consolidated at an isotropic effective stress of 400 kPa for another 7 days. After that, all specimens were

98 unloaded isotropically to a mean effective stress of 100 kPa for obtaining an OCR of 4 within 2 days, and then

99 stayed for another 2 days before undrained cyclic loading.

100 The undrained cyclic shearing was load-controlled. The mean deviatoric stress was null for all the tests.

101 Therefore, the sine shape cycles are defined by the frequency set to 1 Hz and the cyclic stress half-amplitude qcyc.

102 Table 2 presents the initial characteristics of the four tested samples. Rc is the ratio qcyc/qmax, where qmax = 194.5

103 kPa is the undrained peak strength from a monotonic CIUC test on a sample with OCR = 4. Ncyc is the total

104 number of cycles attained for each test. Note that the frequency of 1 Hz was selected to ensure the test duration

105 is reasonably not too short and not too long, e.g. for 1 million cycles requiring 12 days if 1 Hz is adopted.

106 3.2 General experimental results

107 All experimental results of four tests are presented in Fig. 3. At low stress ratio (i.e. Rc = 0.22), the cyclic

108 effective stress paths did not reach the critical state lines, neither in compression nor in extension (Fig. 3(a)). For

109 Rc = 0.44, the stress path exceeded the critical state line in extension only (Fig. 3(d)). For Rc = 0.48 and 0.52, the

110 stress paths reached the two critical state lines in both compression and extension (Figs. 3(g), 3(j)). A precise

5
111 analysis of the experimental data for the tests CYC3 and CYC4 shows that the specimens failed before the end

112 of the cyclic sequences: it was obvious for CYC4 in which the test was stopped after 6990 cycles. As observed

113 during field tests on piles, it seems that the over-consolidated clay, even if the OCR values are different, exhibits

114 a threshold value Rc of approximately 0.45 below which cycles do not lessen significantly the soil resistance.

115 3.3 Evolution of axial strain

116 The axial strains versus number of cycles during cyclic loadings are presented in Fig. 4 for cyclic stress ratios

117 from 0.22 to 0.52, in which the evolution of axial strain expands more significantly for higher cyclic stress ratio

118 in general. Furthermore, permanent (mean) axial strain (εaper = (εamax+εamin)/2) and cyclic axial strain (εacyc =

119 (εamax-εamin)/2) are plotted with the number of cycles, shown in Fig. 5.

120 For comparison shown in Fig. 5(b), the test results on reconstituted Merville clay (main mineral: illite, PI ≈

121 58.6 %, clay fraction CF = 26%, OCR = 4) were compared with data obtained by Andersen et al. [1] on

122 Drammen clay (hydromica and feldspar, PI = 27 %, CF ≈ 50%, OCR = 4) and by Hicher [28] on both Black

123 clay (kaolinite and illite, PI = 30 %, CF = 54 %, OCR = 4) and Bentonite clay (smectite, PI = 54 %, CF = 69 %,

124 OCR = 4). As shown in Fig. 5(b), the sensibility of clay specimens to cyclic loading, through the evolution of

125 the cyclic axial strains, depends on the mineralogy related to the plasticity indexes PI. Clays (Drammen, Black

126 clays) with smaller PI are thus more sensitive than clays with higher PI (Bentonite, Merville clay) because they

127 present a weakest resistance to the evolution of their microstructure.

128 3.4 Evolution of excess pore pressure

129 The excess pore pressures versus number of cycles during cyclic loadings are presented in Fig. 6 for cyclic

130 stress ratios from 0.22 to 0.52, in which the evolution of excess pore pressure expands with more significant

6
131 amplitude for higher cyclic stress ratio in general. Furthermore, the evolutions of permanent (mean) pore

132 pressure (∆uper = (∆umax+∆umin)/2) and cyclic pore pressure (∆uper = (∆umax-∆umin)/2) were plotted against the

133 number of cycles in Fig. 7. Except for Test CYC4 with Rc = 0.52, the excess pore pressure started to increase as

134 usually observed in cyclic tests but decreased until the end of the loading at large number of cycles. For this

135 unusual phenomenon, the test procedure and equipment was carefully re-checked, e.g. consolidation was fully

136 achieved before shearing and that no unexpected water leakage occurred.

137 Since the duration of the cyclic tests, with more than one million cycles, is about 12 days, the question of creep

138 might be addressed. For this, the investigation of undrained triaxial creep behavior of over-consolidated clay

139 was followed.

140 4. Undrained triaxial creep tests

141 4.1 Test program and test procedure

142 Reconstituted samples for undrained creep tests series followed the same preparation procedure than previously

143 described in section 3.1. All specimens were isotropically loaded up to a constant mean effective stress of 700

144 kPa with a back-pressure of 100 kPa within 3 days, and then consolidated for another 7 days. After that, all

145 specimens were unloaded isotropically to different mean effective stresses (=50 kPa, 100 kPa, 200 kPa and 350

146 kPa, respectively) for obtaining different values of OCR (=14, 7, 3.5 and 2, respectively) within 2 or 4 days, and

147 then stayed for 2 days before undrained creep.

148 Four undrained creep tests were first performed under four deviatoric stress levels (Tests referred to be

149 CRE1/OCR=2 to CRE4 / OCR=14 in Table 3), so that the initial stress state for creep converged quickly

150 towards the critical state line in pꞌ-q plane. The loading rate for the deviatoric stress for all tests was 50 kPa/min.

7
151 For all creep tests, the vertical loadings were applied up to their target deviatoric stresses keeping the confining

152 stress constant, which is a conventional way adopted by many researchers. To make sure whether the evolution

153 of excess pore pressure is positive or negative for the test on the sample with OCR=2, an additional test (named

154 CRE5/OCR=2*) was performed by vertical loading applied simultaneously with confining stress unloading and

155 keeping the constant mean effective stress. The creep test program is listed in Table 3. The slight difference

156 between target values of stresses and real measured values (# symbol) is due to the stress control system of the

157 triaxial equipment.

158 The criterion of creep test duration in this study is the axial strain rate decreasing at about 10-7 /min, which is

159 relatively small, so that the axial strain can be considered as stable, as for the tests with OCR=3.5, 7 and 14. For

160 the case of axial strain rate increasing (tests at OCR=2), the test was stopped when the sample failed.

161 4.2 Evolutions of axial strain and excess pore pressure

162 As mentioned above, the loading rate of deviatoric stress for all tests was 50 kPa/min. Thus, the vertical

163 loadings of all creep tests reaching their target deviatoric stresses needed several minutes, i.e. the CRE1/OCR=2

164 needed 7 minutes to attain the target deviatoric stress. Most of the test results presented in the following figures

165 include both loading stage and creep stage. However, in some cases, only the creep stage is presented, and the

166 time on the x-axis is therefore marked as “elapsed time of creep” in the figure, i.e. Figs. 9 and 12.

167 Regarding to the first four conventional creep tests as shown in Fig. 8(a), the axial compressive strain increases

168 generally with time for each test, which is similar to the same type of test on normally consolidated and slightly

169 overconsolidated clays. Beside of this, higher OCR involves smaller strain levels which is consistent with the

170 phenomena observed during conventional shear tests (Fig. 1(d)), i.e. higher OCR results in higher stiffness of

171 “q/p’-εa”. The test “CRE1/OCR=2” goes to failure in a few hours due to a high deviatoric stress level on a
8
172 slightly overconsolidated clay sample. The test “CRE5/OCR=2*” goes faster to failure than the test

173 “CRE1/OCR=2” due to its higher initial stress ratio for creeping. The filled symbols presented in Fig. 8

174 represent the end of loading stage, in other word, the creep stage starts.

175 For the time evolution of excess pore pressure as shown in Fig. 8(b) during creep stage, the more heavily

176 overconsolidated clays (OCR=3.5, 7 and 14) exhibit first a quick increase with an amount almost equal to

177 one-third of the applied deviatoric stress at which point the stress ratios (q/pꞌ) are on the critical state line in p’-q

178 plane, and then a long duration of decrease up to negative values (namely dilation). Higher OCR results in more

179 negative excess pore pressure or dilation. For the lightly overconsolidated clay (OCR=2), the excess pore

180 pressure of the test “CRE1/OCR=2” develops quickly resulting in the stress state on the critical state line. The

181 excess pore pressure of the test “CRE5/OCR=2*” changes slightly because of stress control system during the

182 loading and remains constant with the stress state on the critical state line. To distinguish the initial minutes on

183 the curves of the results presented in Figs. 8(a) and (b), these curves are enlarged in Figs. 8(c) and (d),

184 respectively.

185 4.3 Evolutions of stress ratio and stress path

186 The relationship between the stress ratio q/pꞌ and the elapsed time of creep is presented in Fig. 9. At the

187 beginning of the creep stage for all the overconsolidated clays, the stress ratios reach the critical state line

188 (q/pꞌ=1.0) immediately and even are slightly beyond the critical state line. Then there is a long duration of

189 decrease below the critical state line for heavily overconsolidated clays (OCR=3.5, 7 and 14). Except for the test

190 "CRE5/OCR=2*", the elapsed time of creep for reaching the critical state line will be longer for the lower OCR.

191 As mentioned above, the overconsolidated clay with higher OCR has higher stiffness. Thus, the stress ratio of

192 the overconsolidated clay with the higher stiffness (due to higher mean effective stress) reaches the critical state

9
193 line faster during creep.

194 To understand the evolution of stress ratio clearly, we divided creep tests into two processes. One is from the

195 beginning of loading to the maximum value of stress ratio (Fig. 10(a)), which may include the creep stage for

196 some tests, the other from the maximum value of stress ratio to the end of creep test (Fig 10(b)). In Fig. 10, each

197 value of time shown above the critical state line represents the elapsed time of creep when the stress ratio

198 reaches the maximum value for each creep test. For example, the value of time "300 min" represents the elapsed

199 time of creep when the stress ratio reaches the maximum value for the test "CRE1/OCR=2". And values of time

200 shown below the critical state line, three values (100 min, 10 min and 1min) in Fig. 10(a) and four (100 min,

201 1000 min, 10000 min and end of creep) in Fig. 10(b), represent the elapsed time of creep for the thick dot-dash

202 lines (for tests CRE1/OCR=2 to CRE4/OCR=14). The solid point for each test result shown in Fig. 10(a)

203 represents the end of vertical loading, that means the creep beginning at constant vertical pressure.

204 At the end of vertical loading, the stress state for test “CRE5/OCR=2*” nearly reaches the critical state, and in

205 few minutes, the sample is broken as shown in Fig. 10(a). The three time lines (1 min, 10 min and 100 min) are

206 more and more close to the critical state line as the creep time increases. These lines all intersect with the critical

207 state line. As the value of OCR increases, the tangent point moves down along the critical state line. That means

208 creep test with higher OCR can reach the same target stress ratio (q/pꞌ=1.0) faster. Whereas, in Fig. 10(b), the

209 time lines (100 min, 1000 min and 1000 min) are far away from the critical state line as the creep time increases.

210 The points representing the end of creep tests, corresponding to the axial strain rate decreasing at 10-7 /min, are

211 also shown in Fig. 10(b). A unique curve is obtained based on these points.

212 Likewise, we divided creep tests into two processes in e-log(p') plane, except for the test “CRE5/OCR=2*”. One

213 is from the beginning of loading to the maximum value of stress ratio (Figs. 11(a), 11(b) and 11(c)), the other

10
214 from the maximum value of stress ratio to the end of creep test (Fig. 11(d)). In Fig. 11(a), the stress state of the

215 test "CRE1/OCR=2" nearly reaches the critical state immediately after the creep stage beginning, conversely,

216 the points representing the results on highly overconsolidated clays (OCR=3.5, 7 and 14) have some distance to

217 the critical state line. As the creep time increases, all the thick dot-dash lines (representing the position after a

218 same test duration) are far away from the critical state line (Figs. 11(a), 11(b) and 11(c)). That indicates that

219 values of effective mean stresses decrease with the creep time increase, until the value of stress ratio (q/pꞌ)

220 becomes maximum. As mentioned above, when the maximum stress ratio is reached, the stress paths reach the

221 critical state line in p'-q plane, except the result of the test “CRE5/OCR=2*”. That implies that though the stress

222 ratio of overconsolidated clay reaches the critical state line, the sample does not reach the critical state.

223 Inversely, in Fig. 11(d), the time lines (100th min, 1000th min and 1000th min) are more and more close to the

224 critical state line as the creep time increases. The points representing the end of creep tests, corresponding to the

225 axial strain rate decreasing at about 10-7 /min, are also shown in Fig. 11(d). A unique curve is obtained based on

226 these points. This curve has some distance to the critical state line, indicating that highly overconsolidated clays

227 (OCR=3.5, 7 and 14) do not reach the critical state. This phenomenon is consistent with the phenomena of the

228 non monotonic evolution of excess pore pressure as well as the stable evolution of axial strain (Fig. 8(b)).

229 4.4 Evolutions of plastic strain rates

230 How to obtain the incremental plastic deviatoric strain and incremental plastic volumetric strain is presented in

231 the Appendix of this paper. Thus, the evolution of plastic deviatoric strain rate with creep time is shown in Fig.

232 12(a) and the evolution of plastic volumetric strain rate versus creep time in Fig. 12(b).

233 At the beginning of the creep stage, the plastic deviatoric strain rate is very large for the test “OCR=2*”, and the

234 sample fails in few minutes. For highly overconsolidated clays (OCR=3.5, 7 and 14), the plastic deviatoric strain
11
235 rate decreases linearly in a log-log plot with the increase in time at the beginning of the creep stage. The slope of

236 this relationship is independent of the creep stress. As the creep time increases (beyond 104 min), all three lines

237 are shifted vertically downward, as indicated in Fig. 12(a). The onset of failure for the lightly overconsolidated

238 clay (OCR=2) is signaled by the reversal in slope, as shown by the two topmost curves. As mentioned above, all

239 the creep tests have the same target stress ratio and can reach the critical state line, but just samples with lightly

240 overconsolidation ratio (OCR=2) are broken during the creep. This phenomenon indicates that, at the same

241 stress ratio, the applied vertical constant stress is the most important factor influencing the failure of samples.

242 The more important OCR is, the more important the plastic volumetric strain rate at the beginning of the creep

243 (Fig. 12(b)) is. Then, the strain rate decreases in a log-log plot with the increase in time, except for the constant

244 strain rate of the test “CRE5/OCR=2*”. The incremental plastic volumetric strain rate at the end of the test

245 “OCR=2” is 3.2E-4 %/min, which indicates that the sample is in the dilative state all the time during creep,

246 although the volumetric strain during undrained creep test is equal to zero. This phenomenon of the

247 reconstituted Merville clay in this paper has also been observed from the undrained creep tests on a soft clay –

248 Wenzhou clay (Wang and Yin [24]). The plastic volumetric strain rates are -2.6E-5 %/min, -7.4E-5 %/min and

249 -8.6E-5 %/min at the end of creep stages for the samples with OCR= 3.5, 7 and 14, respectively presented in Fig.

250 12(c). The filled symbols represent the end of creep stages. For highly overconsolidated clays, the samples are

251 firstly in the shear shrinkage state and then in the shear dilatancy state during creep.

252 4.5 Evolutions of stress dilatancy

253 Based on the plastic strain rates in the previous section, the evolution of dilatancy defined as
d = dε vp dε dp can

254 be obtained. Similarly to the stress-dilatancy relationship, the

d = dε vp dε dp is plotted against the stress ratio

255 starting from creep points in Fig. 13 for all five creep tests.

12
256 During the short duration after rapid loadings, all samples start with positive values of d representing the

257 contractive behavior and corresponding to the quick increase of excess pore pressure in Fig. 8(b). Then, d

258 becomes zero whereas the stress ratio q/p' = 1 corresponding to the stress states lying on the critical state line in

259 Fig. 10(a). The dilatancy d continuously decreases up to a minimum value representing a maximum dilation, and

260 then increases towards zero with the stress ratio decreasing. The points representing the beginning of dilatancy

261 increasing is marked by filled symbols with the creep time (Fig. 13). It is shown that higher OCR has a potential

262 of smaller value of stress ratio for d = 0. The higher OCR specimens require more time for the dilatancy d

263 reaching the minimum value shown in Fig.13.

264 4.6 Discussion of combined effects of cyclic loading and creep

265 Fig. 14 shows comparisons of excess pore pressures versus number of cycles or time between undrained triaxial

266 cyclic tests under different cyclic stress ratios and undrained triaxial creep tests under different

267 over-consolidation ratios. It can be seen after certain moment (e.g. approximately 300 minutes from the figure) a

268 continuous decrease of the pore pressure due to creep, which can also explain the specific evolution of the

269 excess pore pressure during cyclic tests where cycles and creep can be considered as two opposed mechanisms

270 over the test period. Since these additional creep tests showed that two processes were simultaneously acting in

271 a competitive manner: increase of the pore pressure due to the cyclic loading, and decrease of the pore pressure

272 because of creep, the two mechanisms “increase of the excess pore pressure due to cycles and simultaneous

273 decrease due to creep” should be considered for the interpretation of results.

274 5. Conclusions

275 In order to clarify this long-term undrained behavior of over-consolidated clay with both cyclic effect and creep

276 effect, the reconstituted samples are prepared from natural specimens retrieved in the north of France. Before the
13
277 investigation of long-term behavior, the shear strength characteristics along monotonic triaxial stress paths are

278 identified on samples with different ratios of over-consolidation and in both compression and extension

279 conditions.

280 Then, load control cyclic tests on over-consolidated samples are conducted in fully saturated and undrained

281 conditions. Small cyclic deviatoric stresses are applied in order to investigate more particularly the behavior

282 under a very large number of cycles. It was shown that even after one million cycles, failure did not necessarily

283 occur in most of the samples unless the cyclic stress level was greater than a critical value. By comparison with

284 previous data published in the literature, at similar over-consolidation ratio of 4, we show that mineralogy and

285 plasticity indexes play a fundamental role on the resistance to cyclic loads. This also confirms the possible

286 existence of a threshold stress ratio below which cycles have only a small effect on the soil resistance.

287 Moreover, during cyclic undrained triaxial tests, with an unusual number of cycles larger than one million, at a

288 frequency of 1 Hz, an unusual pore pressure reduction is observed. This reveals that two mechanisms have to be

289 considered: increase of the excess pore pressure due to the cycles and simultaneous decrease due to creep under

290 highly over-consolidated state. To explain this, undrained triaxial creep tests are performed on reconstituted

291 samples with different values of OCRs under some specified stress states. The evolutions of axial strain, excess

292 pore pressure, stress ratio, stress path, plastic strain rates and stress dilatancy during undrained creep are

293 discussed: (1) Higher OCR has smaller creep strain level corresponding to higher stiffness. (2) The stress ratios

294 reach and can be beyond the critical state line at the beginning of the creep stage, and the overconsolidated clay

295 with higher stiffness reaches the critical state line faster during creep. (3) For highly overconsolidated clays

296 (OCR=3.5, 7 and 14), the strain rate decreases linearly with increase of time in the log-log plane. The slope of

297 this relationship is independent of the creep stress. The relationships are concave upward as shown with the

14
298 failure for the lightly overconsolidated clays (OCR=2). This phenomenon indicates that, at the same stress ratio,

299 the applied vertical constant stress is the most important factor influencing whether the samples fail or not. (4)

300 Though the incremental axial strain rate reaches the criteria of creep test duration, regarded as stable at the end

301 of creep, the evolution of excess pore pressure non monotonically changes with the time increasing. That makes

302 highly overconsolidated clays (OCR=3.5, 7 and 14) not reach the critical state. (5) Higher OCR has a potential

303 of smaller value of stress ratio for d = 0. To reach the criteria of creep test duration takes much more time for

304 higher OCR.

305 The additional undrained creep tests also show that two processes are simultaneously acting in a competitive

306 manner: increase of the pore pressure due to the cyclic loading and decrease of the pore pressure because of

307 creep, which should be considered when interpreting results.

308 Further works will be carried out on modelling the combined effects of cyclic and creep for engineering

309 practice.

310

311 Acknowledgements

312 The financial supports provided by the RIF project (Grant No. PolyU R5037-18F) from Research Grants

313 Council (RGC) of Hong Kong are gratefully acknowledged.

314

15
315 Reference

316 1. Andersen, K.H., Pool, J.H., Brown S.F., and Rosenbrand W.F. (1980) Cyclic and static laboratory tests on

317 Drammen clay. Journal of the Geotechnical Engineering, 106 (GT5): 499–529

318 2. Vucetic, M., and Dobry, R. (1988) Degradation of marine clays under cyclic loading. Journal of Geotechnical

319 Engineering, 114(2): 133–149

320 3. Biarez, J., and Hicher, P.Y. (1994) Elementary mechanics of soil behaviour: saturated remoulded soils.

321 A.A.Balkema Publishers, Rotterdam

322 4. Hyodo, M., Yamamoto, Y., and Sugiyama, M. (1994) Undrained cyclic shear behaviour of normally

323 consolidated clay subjected to initial static shear stress. Soils and Foundations, 34(4): 1–11

324 5. Li, T., and Meissner, H. (2002) Two-surface plasticity model for cyclic undrained behavior of clays. Journal

325 of Geotechnical and Geoenvironmental Engineering, 128(7): 613–626

326 6. Boulanger, R.W., and Idriss, I.M. (2006) Liquefaction susceptibility criteria for silts and clays. Journal of

327 Geotechnical and Geoenvironmental Engineering, 132(11): 1413–1426

328 7. Cai, Y., Gu. C., Wang. J., Juang, C., Xu, C.J., and Hu, X.Q. (2013) One-Way Cyclic Triaxial Behavior of

329 Saturated Clay: Comparison between Constant and Variable Confining Pressure. Journal of Geotechnical

330 and Geoenvironmental Engineering, 139(5): 797–809

331 8. Yin, Z.Y., Xu, Q., and Chang, C.S. (2013) Modeling cyclic behavior of clay by micromechanical approach.

332 Journal of Engineering Mechanics, 139(9): 1305–1309

333 9. Mortezaie, A., and Vucetic, M. (2016) Threshold Shear Strains for Cyclic Degradation and Cyclic Pore Water

334 Pressure Generation in Two Clays. Journal of Geotechnical and Geoenvironmental Engineering, 142(5):

335 04016007. doi:10.1061/(ASCE)GT.1943-5606.0001461.

336 10. Qian, J.G., Wang, Y.G., Yin, Z.Y., and Huang, M.S. (2016) Experimental identification of plastic shakedown

337 behavior of saturated clay subjected to traffic loading with principal stress rotation. Engineering Geology,

338 214: 29–42

339 11. Qian, J.G., Du, Z.B., and Yin, Z.Y. (2018) Cyclic degradation and non-coaxiality of soft clay subjected to

340 pure rotation of principal stress directions. Acta Geotechnica, 13(4): 943–959

16
341 12. Walker, L.K. (1969) Undrained creep in a sensitive clay. Géotechnique, 19(4): 515–529

342 13. Arulanandan, K., Shen, C.K., and Young, R.B. (1971) Undrained Creep Behavior of a Coastal Organic Silty

343 Clay. Géotechnique, 21(4): 359–375

344 14. Holzer, T.L., Hoeg, K., and Arulanandan, K. (1973) Excess pore pressure during undrained clay creep.

345 Canadian Geotechnical Journal, 10(12): 12–24

346 15. Vaid, Y.P., and Campanella, R.G. (1977) Time-dependent behaviour of undisturbed clay. Journal of

347 Geotechnical Engineering, 103(7): 693–709.

348 16. Adachi, T., and Oka, F. (1982) Constitutive equations for normally consolidated clay based on

349 elasto-viscousplasticity. Soils and Foundations, 22(4): 57–70

350 17. Hinchberger, S.D. (1996) The behaviour of reinforced and unreinforced embankments on rate sensitive

351 clayey foundations. Thesis at University of Western Ontario, Canada

352 18. Yao, Y.P., Hou, W., and Zhou, A.N. (2009) UH model: three-dimensional unified hardening model for

353 overconsolidated clays. Géotechnique, 59(5): 451–469

354 19. Yao, Y.P., Kong, L.M., Zhou, A.N., and Yin, J.H. (2015) Time-dependent unified hardening model:

355 three-dimensional elastoviscoplastic constitutive model for clays. Journal of Engineering Mechanics,

356 141(6):1–18

357 20. Yin, J.H., Zhu, J.G., and Graham, J. (2002) A new elastic viscoplastic model for time-dependent behaviour

358 of normally and overconsolidated clays: Theory and verification. Canadian Geotechnical Journal, 39(1):

359 157–173

360 21. Yin, Z.Y., and Hicher, P.Y. (2008) Identifying parameters controlling soil delayed behaviour from laboratory

361 and in situ pressuremeter testing. International Journal of Numerical and Analytical Methods in

362 Geomechanics, 32(12): 1515–1535

363 22. Yin, Z.Y., Chang, C.S., Karstunen, M., and Hicher, P.Y. (2010) An anisotropic elastic viscoplastic model for

364 soft soils. International Journal of Solids and Structures, 47(5): 665–677

365 23. Yin, Z.Y., Yin, J.H., and Huang, H.W. (2014) Rate-dependent and long-term yield stress and strength of soft

366 Wenzhou marine clay: experiments and modeling. Marine Georesources & Geotechnology, doi:10.1080/

367 1064119X.2013.797060.
17
368 24. Wang, L.Z., and Yin, Z.Y. (2015) Stress-dilatancy of natural soft clay under undrained creep condition.

369 International Journal of Geomechanical, 15(5):1–5.

370 25. Zhao, D., Hattab, M., Yin, Z.Y., and Hicher, P.Y. (2018) Dilative Behavior of Kaolinite under Drained Creep

371 Condition. Acta Geotechnica, doi:10.1007/s11440-018-0686-x

372 26. Tavenas, F., Leroueil, S., La-Rochelle, P., and Roy, M. (1978) Creep behavior of an undisturbed lightly

373 overconsolidated clay. Canadian Geotechnical Journal, 15(3): 402–423

374 27. Burland, J.B. (1990) On the compressibility and shear strength of natural soils. Géotechnique, 40(3): 329–

375 378

376 28. Hicher, P.Y. (1979) Contribution à l’étude de la fatigue des argiles. PhD thesis, Ecole Centrale de Paris,

377 France (in French)

378

18
379 Appendix. Determination of plastic strain rates

380 The undrained triaxial condition implies null volumetric strain ( dε a + 2dε r = 0 ). Thus, during undrained

381 creep the incremental deviatoric strain can be obtained:

2
382 dε d = (dε a − dε r ) = dε a (A1)
3

383 Because of the null change of the deviatoric stress during undrained creep test, the elastic deviatoric strain

384 increment is null. Therefore, the Eq.(A1) can be used for the incremental plastic deviatoric strain.

385 The null volumetric strain condition also implies the relationship "dε vp = −dε ve " between plastic and elastic

386 volumetric strains. Thus, the increment of plastic volumetric strain can be obtained by the change of mean

387 effective stress due to the generation of excess pore pressure (dpꞌ=-du), as follows:

du
388 dε vp = (A2)
K

389 with the bulk modulus K = (1 + e0 ) p '/ κ and p ' = p0 + ∆q / 3 − ∆u .

390

19
391 Tables

392 Table 1. Physical and mechanical properties of Merville clay

Liquid limit (wL) 97.2±1 %

Plastic limit (wP) 38.6±1 %
Plasticity index (PI) 58.6±2 %
Specific gravity of solids (Gs) 2.65±0.2
Percent finer than 2 µm (CF) 26 %
Initial water content* (w0) 50.1 %
Initial void ratio* (e0) 1.28
Compression index* (Cc) 0.6
Swelling index* (Cs) 0.091
Slope of CSL in compression* (Mc) 1.0
Slope of CSL in extension* (Me) 0.96
393 * for reconstituted samples in this study.
394
395
396
397
398
399 Table 2. List of undrained triaxial cyclic tests on reconstituted samples of Merville clay

Test w (%) before cyclic qcyc (kPa) Rc OCR Ncyc

CYC1 37.4 41.7 0.22 4 1,060,200

CYC2 37.0 84.7 0.44 4 1,051,480

CYC3 37.7 92.8 0.48 4 1,050,520

CYC4 38.5 101.0 0.52 4 6,990

400
401
402
403
404
405
406
407
408
409
410
411
412
413
20
414
415 Table 3. List of undrained triaxial creep tests on reconstituted samples of Merville clay

Consolidation stress (kPa)

Test Applied stress for
w(%) before creep OCR
Number First stage Second stage creep ∆q (kPa)
pꞌmax pꞌ0

CRE1 /
36.2 700 350 (350#) 350 (344#) 2.0#
OCR=2

CRE2 /
37.0 700 200 (203#) 200 (198#) 3.4#
OCR=3.5

CRE3 /
40.9 700 100 (102#) 100 (100#) 6.9#
OCR=7

CRE4 /
41.0 700 50 (52#) 50 (52#) 13.5#
OCR=14

CRE5 /
36.3 700 350 (350#) 350 (349#) 2.0#
OCR=2*
416 # for real measured value during tests.

417

418

21
419 Figure captions

420 Fig. 1 Results of undrained triaxial tests on samples with OCR=1 in compression, with OCRs from 1 to 14 in

421 compression, and with OCRs of 1 and 4 in extension respectively: (a,c,e) deviatoric stress and

422 normalized stress (q/pꞌmax) versus axial strain for tests of (b,d,f) stress path in the plane of p’-q and

423 p'/pꞌmax -q/pꞌmax

424 Fig. 2 Results of undrained triaxial tests on samples with different OCRs in compression and in extension with

425 isotropic compression stages plotted in e-logp’

426 Fig. 3 Results of undrained triaxial cyclic tests on samples with OCR=4 under different cyclic stress ratio Rc

427 from 0.22 to 0.52: (a, d, g, j) deviatoric stress versus mean effective stress, (b, e, h, k) deviatoric stress

428 versus axial strain, and (c, f, i, l) excess pore pressure versus axial strain

429 Fig. 4 Axial strain versus the number of cycles under different cyclic stress ratio for (a) Rc = 0.22, (b) Rc = 0.44,

430 (c) Rc = 0.48, (d) Rc = 0.52

431 Fig. 5 Results of undrained triaxial cyclic tests on samples with OCR=4 under different cyclic stress ratio Rc

432 from 0.22 to 0.52: (a) permanent axial strains versus the number of cycles, and (b) cyclic axial strains

433 versus the number of cycles

434 Fig. 6 Excess pore pressure versus the number of cycles under different cyclic stress ratio for (a) Rc = 0.22, (b)

435 Rc = 0.44, (c) Rc = 0.48, (d) Rc = 0.52

436 Fig. 7 Results of undrained triaxial cyclic tests on samples with OCR=4 under different cyclic stress ratio Rc

437 from 0.22 to 0.52: (a) permanent excess pore pressure versus the number of cycles, and (b) cyclic

438 excess pore pressure versus the number of cycles

439 Fig. 8 Results of undrained triaxial creep tests: (a) axial strain versus time, (b) excess pore pressure versus time,

440 (c) enlarged initial early parts of the curves of axial strain versus time shown in Fig. 8(a), and (d)

441 enlarged initial early parts of the curves of excess pore pressure versus time shown in Fig. 8(b)

442 Fig. 9 Relationship between stress ratio (q/pꞌ) and elapsed time of creep

443 Fig. 10 Stress path in p'-q plane for undrained triaxial creep tests: (a) from the beginning of loading to maximum

444 stress ratio (q/pꞌ) and (b) from maximum stress ratio (q/pꞌ) to the end of creep test

22
445 Fig. 11 Evolution of undrained creep test results in e-log(p') plane: (a) test time from the beginning of loading to

446 1 minute, (b) test time at 10th minute, (c) test time at 100th minute, and (d) from maximum stress ratio

447 (q/pꞌ) to the end of creep test

448 Fig. 12 Results of undrained triaxial creep tests: (a) plastic deviatoric strain rate versus elapsed time of creep, (b)

449 plastic volumetric strain rate versus elapsed time of creep, and (c) enlarged the final parts of the

450 relationship between plastic volumetric strain rate and elapsed time of creep shown in Fig. 12(a)

451 Fig. 13 Evolutions of dilatancy d versus stress ratio during undrained creep stages

452 Fig. 14 Comparisons of excess pore pressures versus number of cycles or time between undrained triaxial cyclic

453 tests under different cyclic stress ratios and undrained triaxial creep tests under different

454 over-consolidation ratios

455

456

457

23
Figure 1

1000 1000
p'0=200 kPa In compression p'0=200 kPa
p'0=400 kPa OCR=1 p'0=400 kPa
800 800 CSL
p'0=700 kPa p'0=700 kPa

600
In compression
600
COR=1
q (kPa)

q (kPa)
400 400

200 200

0 0
0 4 8 12 16 0 200 400 600 800 1000
(a) εa (%) (b) p' (kPa)

1.2 1.2
OCR=1 OCR=2 OCR=1 OCR=2
OCR=3.5 OCR=4 OCR=3.5 OCR=4
1 1 CSL
OCR=7 OCR=14 OCR=7 OCR=14
In compression
0.8 0.8
q/p'max

q/p'max

0.6 0.6

0.4 0.4

0.2 0.2
In compression
0 0
0 4 8 12 16 0 0.2 0.4 0.6 0.8 1 1.2
(c) εa (%) (d) p'/p'max

0 0

CSL
-0.2 -0.2

-0.4 -0.4
q/p'max
q/p'max

-0.6 -0.6
In extension
In extension
OCR=1
-0.8 OCR=1 -0.8
OCR=4
OCR=4
-1 -1
-16 -12 -8 -4 0 0 0.2 0.4 0.6 0.8 1 1.2
(e) εa (%) (f) p'/p'max

24
Figure 2

1.4

1.2 OCR=1
ICL
OCR=2 CSL
OCR=3.5
1
e

OCR=7 Cc
OCR=14
0.8 OCR=1 (Ext.) Cs
OCR=4 (Ext.)
OCR=4 (Com.)
0.6
10 100 1000
p' (kPa)

25
Figure 3

60 60 10
1st, 2nd, 5th, 10th, 1st, 2nd, 5th, 10th,
100th, 1000th 100th, 1000th ∆umax at
40 40 5
10000th 100000th
∆umax at
10000th 0
20 CSL 20 -0.4 -0.2 0 0.2 0.4
100000th
-5 1st, 2nd, 5th, 10th,
q (kPa)

q (kPa)

∆u (kPa)
0 0 100th, 1000th
0 50 100 150 -0.4 -0.2 0 0.2 0.4 -10
-20 CSL ∆umax at -20
10000th -15
1000000th
1000000th
-40 -40 -20 1000000th
Rc = 0.22 100000th Rc = 0.22 Rc = 0.22
-60 -60 -25
(a) p' (kPa) (b) εa (%) (c) εa (%)

120 120 40
1st, 2nd, 5th, 10th, 100th, 1st, 2nd, 5th, 10th, 100th,
1000th,10000th, 100000th 1000th, 10000th, 100000th 30
80 80 ∆umax at
∆umax at
345000th 20
345000th
40 40
10
CSL
1000000th
q (kPa)

q (kPa)

∆u (kPa)
0 0 0
0 50 100 150 -2 -1 0 1 -2 -1 0 1
CSL -10
-40 ∆umax at -40 1000000th
1st, 2nd, 5th, 10th,
345000th
100th, 1000th, -20
-80 1000000th -80 10000th, 100000th
-30
Rc = 0.44 Rc = 0.44 Rc = 0.44
-120 -120 -40
(d) p' (kPa) (e) εa (%) (f) εa (%)

150 150 80
1st, 2nd, 5th, 10th, 100th, 1st, 2nd, 5th, 10th, 100th,
∆umax at 1000th,10000th 1000th, 10000th 60 100000th
100 100
56504th ∆umax at
56504th 40
50 50 100000th 1st, 2nd, 5th, 10th,
CSL 100th, 1000th,
20
q (kPa)

q (kPa)

∆u (kPa)

0 0 10000th
0 CSL 50 100 150 -4 -2 0 2 4 0
-50 ∆umax at -50 -4 -2 0 2 4
56504th 1000000th -20
1000000th 1000000th
-100 -100 -40
Rc = 0.48 100000th Rc = 0.48 Rc = 0.48
-150 -150 -60
(g) p' (kPa) (h) εa (%) (i) εa (%)

150 100
Rupture at 6990th
1st, 2nd, 5th, 10th, 100th, 1000th
80
100 Rupture at 6990th
60
8000th 10000th
50 40
20
q (kPa)

∆u (kPa)

0
-3 0 3 6 9 12 0
-3 0 3 6 9 12
-50 -20 8000th
10000th
-40 1st, 2nd, 5th, 10th, 100th, 1000th
-100
-60
Rc = 0.52 Rc = 0.52
-150 -80
(k) εa (%) (l) εa (%)

26
Figure 4

0.4 1

0.3 ∆umax at 345000th

0.5
∆umax at 10000th
0.2

0.1 0
εa (%)

εa (%)
1E-1 1E+1 1E+3 1E+5
0 -0.5
1E-1 1E+1 1E+3 1E+5
-0.1
-1
-0.2
Rc = 0.22 Rc = 0.44
-0.3 -1.5
(a) Ncyc (b) Ncyc

2.5 9
2
∆umax at 56504th
1.5
6
1
0.5
εa (%)

εa (%)

0 3 Rupture at 6990th
-0.51E-1 1E+1 1E+3 1E+5

-1
0
-1.5 1E-1 1E+1 1E+3 1E+5
-2
Rc = 0.48 Rc = 0.52
-2.5 -3
(c) Ncyc (d) Ncyc

27
Figure 5

3 4
Rc=0.48 Rc=0.52
Rc=0.52 0.52 Rc=0.48
Rc=0.48 0.75 0.56 0.52 Rc=0.44
2 3 Rc=0.22
0.52 0.44
Rc=0.44 Drammen clay

εacyc (%)
Rc=0.22 Black clay
εaper (%)

2 Bentonite clay
1 0.35

0.42
1 0.34
0
1E+0 1E+2 1E+4 1E+6
0
1E+0 1E+2 1E+4 1E+6
-1
(a) Ncyc (b) Ncyc

28
Figure 6

10 40
∆umax at 10000th
30 ∆umax at 345000th
5
20
0
1E-1 1E+1 1E+3 1E+5 10
-5
∆u (kPa)

∆u (kPa)
0
-10 1E-1 1E+1 1E+3 1E+5
-10
-15
-20
-20 -30
Rc = 0.22 Rc = 0.44
-25 -40
(a) Ncyc (b) Ncyc

80 100
∆umax at 56504th Rupture at 6990th
80
60
60
40
40
20 20
∆u (kPa)

∆u (kPa)

0 0
1E-1 1E+1 1E+3 1E+5 1E-1 1E+1 1E+3 1E+5
-20
-20
-40
-40
-60
Rc = 0.48 Rc = 0.52
-60 -80
(c) Ncyc (d) Ncyc

29
Figure 7

50 60
Rc=0.52
40
Rc=0.48 50
Rc=0.52
30 Rc=0.44
40 Rc=0.48
20 Rc=0.22

∆ucyc (%)
Rc=0.44
∆uper (%)

30
10 Rc=0.22

0 20
1E+0 1E+2 1E+4 1E+6
-10
10
-20
0
-30 1E+0 1E+2 1E+4 1E+6
(a) Ncyc (b) Ncyc

30
Figure 8

20 100
OCR=14

Excess pore pressure (kPa)

OCR=7 50
15
OCR=3.5
Axial strain (%)

OCR=2 0
10 OCR=2*
OCR=14
-50 OCR=7
OCR=3.5
5 OCR=2
-100
OCR=2*

0 -150
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5 1E+0 1E+1 1E+2 1E+3 1E+4 1E+5
(a) Time (min) (b) Time (min)

10 50

OCR=14
Excess pore pressure (kPa)

8 OCR=14 40
OCR=7
OCR=7
OCR=3.5
Axial strain (%)

6 OCR=3.5 30
OCR=2
OCR=2
OCR=2*
4 OCR=2* 20

2 10

0 0
1E+0 1E+1 1E+0 1E+1
(c) Time (min) (d) Time (min)

31
Figure 9

1.2
q/p'=1.0

Stress ratio, q/p'

0.8

OCR=14
OCR=7
0.4
OCR=3.5
OCR=2
OCR=2*
0
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5
Elapsed time of creep (min)

32
Figure 10

33
Figure 11

1 1

ICL ICL
CSL CSL
0.9 0.9
1 min 10th min
e

e
OCR=14 OCR=14
0.8 OCR=7 0.8 OCR=7
OCR=3.5 OCR=3.5
OCR=2 OCR=2

0.7 0.7
10 100 800 10 100 800
(a) p' (kPa) (b) p' (kPa)

ICL
CSL
0.9
e

100th min
0.8 OCR=3.5

OCR=2

0.7
10 100 800
(c) p' (kPa)

34
Figure 12

1E+1 0.4
OCR=14

Plastic volumetric strain rate

OCR=14
Plastic deviatoric strain rate (%/min)

1E+0 OCR=7 OCR=7

Failure
OCR=3.5 0.3
OCR=3.5
1E-1
OCR=2 OCR=2

(%/min)
OCR=2* 0.2 OCR=2*
1E-2

1E-3
0.1
1E-4
0
1E-5
Vertically downward
1E-6 -0.1
1E+0 1E+2 1E+4
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5
(a) Elapsed time of creep (min) (b) Elapsed time of creep (min)

0.0004
OCR=14
Plastic volumetric strain rate

OCR=7
OCR=3.5
0.0002
OCR=2
(%/min)

-0.0002
1E+2 1E+3 1E+4 1E+5
(c) Elapsed time of creep (min)

35
Figure 13

5
OCR=14
OCR=7
OCR=3.5
2.5
OCR=2
OCR=2*
d=∆ε vp/∆εdp
0
3028 min
9836 min
-2.5 11068 min

-5
0 0.3 0.6 0.9 1.2
Stress ratio, q/p'

36
 Figure 14

100

50

∆uper or ∆ucreep (%)

0
1E-1 1E+1 1E+3 1E+5

-50
Rc=0.52 Rc=0.48
Rc=0.44 Rc=0.22
-100
OCR=2 OCR=3.5
OCR=7 OCR=14
-150
Ncyc or Time (s)

37

View publication stats