Temperature Effects on Unconned Compressive Strength of Clay Soils: Experimental and Constitutive Study

: Unconﬁned compressive strength ( S u ) is one of the soil engineering parameters 11 used in geotechnical designs. Due to the temperature changes caused by some human activities, it 12 is important to study the changes in S u at diﬀerent temperatures. For this purpose, kaolin, illite 13 and montmorillonite clays with a liquid limit (LL) of 47, 80 and 119 respectively, were tested in a 14 temperature-controlled cell in temperature range of 20 to 60 ℃ . The results showed that the pore 15 water pressure is a function of temperature and by heating, pore water pressure in the samples 16 increased. In all three types of clay, the S u decreased linearly with increasing temperature. The 17 reduction of S u in kaolin is more than illite and in illite is more than montmorillonite. The reason 18 for this reduction, might be due the diﬀerence in the mineralogy of the clays. The results of 19 unconﬁned compressive tests at diﬀerent temperatures were simulated using hypoplastic model. 20

In order to investigate the effect of temperature on unconfined compressive strength (S u ) of 122 kaolin, illite and montmoriollonite, an apparatus was designed and manufactured. Fig. 1 shows 123 the schematic of the apparatus and its various sections. 124 To raise the temperature in the sample, a Plexiglas cell with high coefficient of thermal 125 resistance was used. The transparency of this cell allows the sample to be seen during the test   Table 1. 150 Table 2 shows the abundance of minerals in the three clay samples based on XRD analysis.

151
In kaolin, about 60% of the mineralogical composition of the sample is kaolinite. The second 152 sample contains about 51% of illite mineral. In the third sample, montmorillonite mineral with 153 40% abundance is the highest mineral constituent of the sample. Comparing Tables 1 and 2, it 154 can be seen that the difference in the mineralogical composition of the three selected samples 155 has caused differences in the liquid limit (LL), plastic limit (PL) and activity (A) of the samples.

162
The flowchart of the preparation of soil samples and steps of the test that has been conducted 163 are shown in Fig. 3. 164 To prepare standard sample (length twice the diameter) and to saturate the samples before 165 the test, a cylindrical mold was made that consisted of two cylinders. The small cylinder 166 contained the clay slurry with two porous stone and the large cylinder had a retaining role. To   Table 3. The sample was removed from the sampler using a jack and then 178 a rubber membrane was installed on the sample. In order to reach the desired temperature,  During the heating phase, the upper drainage of the cell was closed and the bottom drainage 185 of sample was connected to the barometer, so the sample was heated in undrained conditions. 186 The heating rate was 5 ℃/h.   2008. 197 The general rate formulation of hypoplastic follows: whereσ andǫ are the objective stress rate and the Euler stretching tensor, respectively. L 199 and N are the fourth-and second-order constitutive tensors, f s and f d are two scalar factors.

200
The model has five parameters to be calibrated, φ c , λ, κ, N and ν which have similar (but where T 0 is the reference temperature and in this study is considered to be 20 o C. In the sample preparation step, a vertical stress of 150 kPa was applied to the samples for 220 24 hours in order to reach the desired void ratio (see Table 3). For kaolin clay, the before 221 consolidation water content (ω) was 70% with a void ratio (e) of 1. 8. These values reduced to 222 ω=41% and e=1.1 after consolidation. For illite the water content and void ratio before and after 223 consolidation were 96% and 63%, 2.4 and 1.6. For montmorillonite initial water content was 224 143% which decreased to 91% after consolidation and the void ratios were 3.1 and 1.9 for before 225 and after consolidation.

247
In undrained heating, with increasing temperature, pore water pressure was generated in 248 all three clays. The generation of pore water pressure with increasing temperature was due to  The pore water pressure at 20℃ (room temperature) was equal to zero. Then, with every 255 5℃ of heating, the increase in pore water pressure was recorded. Fig. 5 shows that for all three 256 types of clays with increasing temperature, the pore water pressure increased from zero to 4, 5 257 and 5 kPa for kaolin, illite and montmorillonite. The increase continued until the temperature 258 of about 50°C and then the rate of increase with temperature was slow down. The slope of 259 the pore water pressure versus temperature was different before and after 50°C. This slope was 260 higher for temperatures below 50°C and lower for temperatures above 50°C.

261
With a heating rate of 5°C/h, each heating increment would take 2 hours. Fig. 6      to negative pore water pressure. On the contrary for illite and montorillonite, the void ratios 320 were higher (1.2, 1.6) therefore they were less stiff compare to kaolin clay and the pore pressure 321 increased while application of axial loading.

322
The changes in pore water pressure in the heating and shearing process could lead to the

345
Elastic shear modulus reduction for montmorillonite can be seen in Fig. 11c. The same 346 trend as other clays was observed in this test but the reduction of initial elastic shear modulus 347 at 60 o C was more significant in montmorillonite. The initial elastic modulus at 20 o C was 15.5

348
MPa and it reduced to 14 and 5.9 at 40 and 60 o C respectively.  was not replicated by the model but the stress corresponds to larger strains were reproduced 368 correctly by the model (Fig. 14b).

387
-Increasing temperature reduced the initial elastic modulus in three types of clay.

388
-The stress-strain curve of kaolin at different temperatures had no peak, while for illite and 389 montmorillonite samples a clear peak at strains around 2% was observed.

390
-In the heating phase with a rate of 5℃/h, the pore water pressure in the samples increased.

391
The increase in pore water pressure in illite and montmorillonite was greater than kaolin.

392
-During the application of axial loading, the pore water pressure in kaolin first slightly increased 393 and then decreased. In illite and montmorillonite, the increase in pore water pressure at the 394 beginning of the axial loading was greater than in kaolin.

395
-Using single set of parameters for each soil, by taking into account the impact of temperature : where ν, λ * and κ * are model parameters, p = -trT/3, and 1 and I are second-and fourth 405 order unity tensors respectively. The factor f d reads with α = 2 and the equivalent pressure where N is a parameter and p r is a reference stress equal to 1 kPa. The factor f A d reads where F m is the Matsuoka-Nakai factor calculated from 409 F m = 9I 3 + I 1 I 2 I 3 + I 1 I 2 (12) and the exponent ω reads Finally, the asymptotic strain rate direction d is calculated as with the Lode angle θ