Comparative Study on Triaxial Test of Undisturbed and Remolded Loess

The diﬀerence of sample preparation methods will directly lead to the diﬀerence of the structure of the sample, and the mechanical response characteristics under external load will be diﬀerent. Based on the conventional tests such as particle analysis, speciﬁc gravity, critical moisture content, and consolidation test of undisturbed loess from Xi’an, this paper focuses on the static triaxial consolidated drained shear and consolidated undrained shear tests on undisturbed and remolded loess. The results show that: (1) the structure has an obvious inﬂuence on the stress-strain relationship of the sample, the stress-strain curve of the undisturbed sample is a strain softening curve, and the stress-strain curve of the remolded sample is a weak strain hardening curve; (2) under the same loading conditions, the triaxial shear strength of the undisturbed soil is signiﬁcantly higher than that of the remolded soil; (3) the change rule of the consolidation ratio of undisturbed and remolded samples is basically the same, but the ﬁnal consolidation ratio of remolded soil is greater than that of undisturbed soil.


Introduction
e measurement methods of mechanical characteristic parameters of soil are divided into in-situ test and laboratory test [1,2]. In the in-situ test, the stress state of soil is close to its real stress state. However, its boundary conditions cannot be accurately controlled and cannot reflect the stress-strain relationship of soil in other stress-strain states. But by comparison, the boundary conditions of indoor test are relatively clear and easy to control, and the research results of soil indoor test are quite rich [3][4][5][6][7]. Remarkably, the preparation method of soil sample is an important factor affecting the determination results of soil laboratory test. Soil samples are divided into undisturbed soil samples and remolded soil samples. e original structure of cohesive soil can be well maintained during sampling and sample preparation, so undisturbed soil samples are used comprehensively. For soil samples with less cohesion, the original shape of undisturbed samples cannot be maintained during sampling and sample preparation. erefore, remolded soil samples are mostly used in laboratory tests.
Natural soil forms structural deformation due to the influence of soil deposition history and surrounding environmental changes. Mitchell et al. defined the structure of a soil mass as "tectonic" and "joint," which shows the arrangement of soil particles and interactions between soil particles [8]. Burland insisted that, like the initial void ratio and stress history, the structure had a significant impact on the mechanical properties of undisturbed soil [9]. Ovalle and Arenaldi-Perisic introduced the behavior of natural diatomite under compression, shear and cyclic loads, and the new opinion of undisturbed high plastic diatomite sludge in Mejillones Bay, Northern Chile [10]. Sousa et al. gave the column test results on large undisturbed samples [11]. Freitas et al. introduced the coefficient of pore expansion to correct the possible differences in measuring the total pore volume under dry or water-saturated conditions [12]. Xu et al. proposed a new method for measuring the preconsolidation pressure of structural loess soil [13]. Aiming at the collapse potential of Gorgan loess, Haeri et al. studied the undisturbed and remolded samples by using the consolidation instrument test [14]. Gao et al. investigated the strength and deformation characteristics of undisturbed unsaturated loess under different moisture content conditions [15]. Chu studied the engineering characteristics of coarse-grained soil, such as strength and deformation [16]. Airey conducted conventional and stress path triaxial tests on cemented carbonate soil obtained from the northwest shelf of Australia and found that the dry density and cementation degree of natural calcareous rock samples vary greatly [17]. Chen et al. systematically studied the mesostructure evolution characteristics of expansive soil and loess under different stress paths, dry wet cycle, immersion expansion, and inundation collapse conditions by using CT triaxial apparatus [18]. Nokande et al. tested different types of soil [19]. Wang et al. studied the different dry densities of undisturbed soil and remolded soil through oedometer test, hydraulic conductivity test, and field emission scanning electron microscope [20]. Jiang et al. studied the effect of structure on secondary compressibility during pressure change [21].
Based on the conventional tests such as particle analysis, specific gravity, limit moisture content, and consolidation test of undisturbed loess from Xi'an, we focus on the static triaxial consolidated drained shear and consolidated undrained shear tests of undisturbed and remolded loess. And in order to explore the influence of structure on the results of indoor triaxial test, we analyze the differences between them in the test results of stress-strain relationship, stress path, shear strength index, and consolidation ratio.  Table 1. We collect about 30 kg remolded samples and 6 undisturbed samples, including 3 tunnel faces and 3 open excavation foundation pits. All the samples are sealed with plastic film after on-site collection.

Specimen Preparation and Test Plan
According to the requirements of test contents in Table 2, the axial principal stress of triaxial compression test sample is set by the sampling depth. erefore, the total overburden stress at the sampling point is calculated based on the field sampling depth, overburden type, thickness, and density. e specific calculation results are shown in Table 2. Table 3, firstly the triaxial compression test sample shall be subject to K 0 consolidation, and the K 0 shall be determined based on the survey report. Secondly, the consolidation test results shall be sorted out with basic physical property parameters. e test scheme is shown in Table 3.

Water Content Test.
According to the standard for geotechnical test methods (GB/T 50123-2019), the moisture content of this test adopts the drying method, and the temperature is 105°C, and the weighing accuracy is 0.01 g. e test results are shown in Table 4. e parallel difference of water content of sample D14-01 is 0.26%, less than 1.0%, which meets the specification requirements. e parallel difference of moisture content of sample D14-02 is 0.38%, less than 1.0%, which meets the specification requirements.

Density Test.
According to the standard for geotechnical test methods (GB/T 50123-2019), the ring knife method is adopted in this test, and the weighing accuracy is 0.01 g. e test results are shown in Table 5. e density parallel difference of sample D14-01 is 0.01 g/cm 3 , and the density parallel difference of sample D14-02 is 0.01 g/cm 3 , less than 0.03 g/cm 3 , which all meet the specification requirements.

Limit Moisture Content Test.
According to the standard for geotechnical test methods (GB/T 50123-2019), the liquid plastic limit combined tester method is adopted in the limit moisture content of this test, and the moisture content adopts the drying method. In addition, the drying temperature is 105°C, and the weighing accuracy is 0.01 g. e test results are shown in Table 6.

Particle Analysis Tests.
e better size 2,600 laser particle size analyzer (wet method) is used for particle size analysis in this experiment. e analysis results are shown in Figure 1.
e particle compositions and contents of sample D14-01 and sample D14-02 are shown in Table 7.

Consolidation Test.
According to the standard for geotechnical test methods (GB/T 50123-2019), the standard consolidation test is adopted for the consolidation test of undisturbed samples, and the deformation measurement accuracy is 0.01 mm. e specific test scheme is shown in Table 8.
e undisturbed specimen has a diameter of 61.8 mm and a height of 20 mm. e test instrument is WG (GDG-4S) horizontal bar consolidation instrument (triple high pressure). e stability standard is based on 24 hours of consolidation under each level of pressure or the hourly change of sample deformation shall not be greater than 0.01 mm. e compression modulus E s1-2 of undisturbed sample CT-02 of sample D14-02 is 9.18 MPa and the compression coefficient is 0.18 MPa −1 . It belongs to medium compressible soil, and its preconsolidation pressure PC is 302.0 kPa determined by the CASA grant graphical method.

Pilot Protocol.
According to the test requirements, the triaxial compression tests of undisturbed and remolded samples are carried out. e specific test scheme is shown in Table 9. Testing of the basic physical properties of soil Moisture content, density, limit moisture content, particle analysis, and consolidation test 2 ree-axis compression test (TC) Stress-strain relationship, consolidation ratio, and stress path

Specimen Preparation and Test Instruments.
e diameter of undisturbed and remolded specimens is 39.1 mm and the height is 80 mm. e remolded sample shall be prepared by controlling the dry density (dry density of undisturbed sample), and the prepared sample shall be placed in the petri dish for more than 48 hours. e test instrument is a SLB-1A stress-strain controlled triaxial creep instrument. As shown in Figure 4, the test instrument consists of shear system, pressurization system, and data acquisition system from sitting to right. e set shear rate is 0.2 kPa/min, the deformation is controlled by 20 mm, and the data recording interval is 0.1 mm.

Test Results of D14-01 Triaxial Consolidated Undrained
Shear Test

Stress-Strain
Relationship. In our test, the lateral stress is controlled to decrease gradually, the axial stress is increased gradually, and the average principal stress is kept unchanged. Figures 5 and 6 are the stress-strain curves of the triaxial test of axial loading and lateral unloading of sample D14-01, undisturbed sample TC-01, and remolded sample TC-02 after K 0 consolidation (consolidation ratio is 0.43). As illustrated in Figures 6 and 7         sample TC-01 is 120.85 kPa and that of remolded sample TC-02 is 92.6 kPa. Figures 9 and 10 show the change curves of consolidation ratio during the test shear of sample D14-01, undisturbed sample TC-01, and remolded sample TC-02, respectively. It can be seen from the figure that from the initial state, the consolidation ratio K 0 is 0.43, which gradually decreases with the strain and remains in a certain value range. e final consolidation ratio K 0 of undisturbed sample TC-01 is about 0.10 and that of remolded sample TC-02 is about 0.21.

Test Results of D14-02 Triaxial Consolidated Undrained
Shear Test Figures 11 and12 show the stress-strain relationship curves of axial loading and lateral unloading triaxial test of sample D14-02, undisturbed sample TC-03, and remolded sample TC-04 after K 0 consolidation (consolidation ratio is 0.42). It can be seen from the figure that the stress-strain curves of undisturbed sample TC-03 are strain softening curves, and the stress-strain curves of remolded sample TC-04 are weak strain hardening curves. When the strain exceeds 0.5%, the strain of the undisturbed sample TC-03 begins to increase, and the effective principal stress difference quickly reaches the peak value of 314.2 kPa. When the strain exceeds 4%, the effective principal stress difference begins to decrease to 266.3 kPa (the strain is 13.0%), with an obvious steep drop process, and then increases. When the strain exceeds 0.3%, the strain of remolded sample TC-04 begins to increase, and the corresponding stress increases little, which maintains at about 225.0 kPa, without obvious steep drop process. e undisturbed sample shows the structure clearly. Figures 13 and 14 are the test effective stress path p'-q' curves of sample D14-02, undisturbed sample TC-03, and remolded sample TC-04, respectively. As is shown in the figure, the change law of the effective stress path curve of the undisturbed sample TC-03 and the remolded sample TC-04 is basically the same, and the curve remains at the line with the average effective principal stress of 193.5 kPa. After the sample is damaged, the stress path deflects, and the average effective principal stress p' and the average effective principal stress difference q' decrease. When the sample is damaged, the average effective principal stress difference of undisturbed sample TC-03 is 158.2 kPa, and the average effective principal stress difference of remolded sample TC-04 is 113.4 kPa. Figures 15 and 16 are the change curves of consolidation ratio during test shear of sample D14-02, undisturbed sample TC-03, and remolded sample TC-04, respectively. As the graphs show, the consolidation ratio K 0 is 0.42 in the initial state, which gradually decreases with the strain and remains in a certain value range. e final consolidation ratio K 0 of undisturbed sample TC-03 is about 0.10 and that of remolded sample TC-04 is about 0.25.

Pilot Protocol.
Due to experimental requirements, the effective cohesion c' and effective internal friction angle of undisturbed and remolded samples are determined according to the test requirements ϕ'. erefore, we need to make four test schemes, and the effective confining pressure of each group is 100, 200, and 300 kPa. e test scheme is presented comprehensively in Table 10.

Test
Methods. the consolidated drained shear test (CD test) has two stages, and one is consolidation stage and the other is shear stage. In the consolidation stage, constant surrounding pressure is applied to the soil sample to measure the displacement and the dissipation process of pore water pressure. After consolidation, the soil sample is sheared at a constant strain rate under the condition of drainage, and the shear failure resistance of the soil sample is measured.

Mathematical Problems in Engineering
Sample installation: at first, take out the cylindrical sample to be tested (Φ39.1 × 80 mm) and wrap the rubber membrane on the outside of the sample with a membrane bearing cylinder and a suction ball. Secondly, clamp the filter paper and water permeable stone up and down, tie the lower end of the rubber membrane with the base and the upper end with the upper cover with a rubber ring, and install the pressure chamber cover. en, open the water filling switch, fill the pressure chamber with water, close the upper exhaust valve, seal the pressure chamber, and prepare for the test;       difference, axial deformation, pore water pressure, and axial deformation, and record the data every 0.1 mm, and test termination condition happens when axial deformation is greater than 20 mm.

Stress-Strain
Relationship. Figures 17 and 18 are the stress-strain curves of triaxial consolidated drained shear test of sample D14-01, undisturbed sample, and remolded sample, respectively. As can be seen from the Figures 18 and 19, the stress-strain relationship curve of undisturbed sample is a strain softening curve, and the stress-strain relationship curve of remolded sample is a weak strain hardening curve.

Stress-Strain
Relationship. Figures 21 and 22 are the stress-strain curves of triaxial consolidated drained shear test of sample D14-02, undisturbed sample, and remolded sample, respectively. As it can be seen, the stress-strain relationship curve of undisturbed sample is a strain softening curve, and the stress-strain relationship curve of remolded sample is a weak strain hardening curve.

Conclusion
Firstly, the structure has an obvious influence on the stressstrain relationship of the sample. In the triaxial undrained and drained experiments, the stress-strain relationship curve of the undisturbed sample is a strain softening curve and the stress-strain relationship curve of the remolded sample is a weak strain hardening curve. Secondly, in the same loading conditions, the shear strength of undisturbed soil is significantly higher than that of remolded soil in triaxial undrained and drained experiments. Finally, in the triaxial test of undisturbed and remolded samples, the change law of consolidation ratio is basically the same. e consolidation ratio decreases gradually with the axial strain and finally stabilizes at a fixed value. But the final consolidation ratio of remolded soil is greater than that of undisturbed soil.
Data Availability e data supporting the conclusion of the article are shown in the relevant figures and tables in the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this article.