3. Experimental verification
To validate the proposed model, we compare with the results from high velocity-dependent ring shear tests of a loess landslide at different saturation (Pei et al. , 2017), as well as high velocity rotary shear frictional tests of familiar fault geomaterials concerning in quartz sandstone (Dieterich, 1978), granite (Dieterich, 1978; Di Toroet al. , 2004) , novaculite (Di Toro et al. , 2004; Di Toroet al. , 2011).
Figure 2 compares the predicted velocity effect results with the experimental results of loess at different saturation and fault geomaterials at different lithologies in a wide velocity range. The model well captures the velocity weakening effect at close saturation and saturation of loess materials. The experiment shows that for wet loess with saturation higher than 0.8 (0.83, 0.941 and 0.995), its velocity effect is obvious, which is well revealed by the proposed theoretical model (figure 2a). The dry loess, i.e., its saturation is zero, there is no observed velocity-dependent friction effects, and the proposed model can only predict its almost friction-constant behaviors at slide velocity lower 10-2 m/s (Figure 2a). The proposed model can also well predict the friction behavior of all compared fault geomaterials involving granite, quartz sandstone, and dense quartzite (Figure 2b). Generally, granite is denser with less porous than quartz and novaculite, which brings about different velocity effects for other fault geomaterials.
We also compare the results from Aharonov and Schol’s model (Aharonov and Scholz, 2018), which employs the averaging stress at the contact surface. This means that the porosity of the geomaterial is zero, which does not exist in nature. However, the new model considers the influence of temperature and velocity for geomaterials with different porosity (Figure 2c). It also precisely emerges the three modes and its zones, i.e., no thermal effects, thermal effects, and melting, of contact temperature with increasing slide velocity (Figure 2d). These have entirely consistent with Aharonov and Schol’s model (Aharonov and Scholz, 2018).
Therefore, the above results show the validity and correctness of the proposed model. It also makes us understand that the contact temperature gradually increases until it accumulates to a very high value during the slow sliding process. The high temperature further causes the phase transition of the geomaterials, in turn which results in a sharp decrease in the friction coefficient (Figure 2d). The coefficient of friction decreases with increasing saturation in loess, as the water in the pores is subject to pore pressure, which results in a lower friction due to the reduction of the normal force between the contacting asperities. In addition, the liquid also has a lubricating effect. Fault geomaterials with smaller pores have greater internal friction, which means that the actual contact area of the contact surface is bigger thus increasing the tangential force of the contact surface. Therefore, the coefficient of friction decreases with increasing porosity (Figure 2c).