Fig. 2 Granules used in this study through crushing and dry-wet cycles

2.2 Test setup for limited shear displacement

The progress and development of ring shear apparatuses have provided an effective approach for exploring the dynamic shear process of quartz granules under large shear displacement and high normal stress conditions (Fukuoka et al., 2007; Kimura et al., 2019; Mair et al., 2002; Tanikawa et al., 2012). Due to the difference in lithology between mudstone and quartz, the shear behavior of soft interlayers may rely more on the water content and weathering intensity rather than the shear displacement and normal stress. To select appropriate shear box dimensions, ASTM D 3080 (2001) suggests a minimum specimen width of at least 10 times the maximum particle size diameter, a minimum width-to-thickness ratio of 2, and a minimum initial specimen thickness of six times the maximum particle diameter. With a maximum particle size of 5 mm, a self-developed digital direct shear test apparatus (shear box dimensions of 200*200*200 mm3) was used in this study; the bottom of the shear box was filled with quartz grains 2~5 mm in size to ensure that the sheared samples were 40 mm thick. Two digital force sensors were equipped in the vertical and horizontal directions, and the shear and normal forces were loaded with a hydraulic system. With this system, the shearing rate can be controlled in the range of 0.05~10 mm/s.
Considering the strength reduction of mudstone under the action of water, normal stresses ranging from 200 kPa to 1400 kPa were adopted in the direct shear tests with a constant shear speed of 0.1 mm/s until a shear displacement of 20 mm was reached. The highest normal stress of each group decreased with increasing dry-wet cycle number. Two conditions, dry and wet (water content equal to 15% of the dry specimen mass), were introduced to discuss the attenuation behavior of strength due to water.

2.3 Testing procedure

Each specimen was filled in the shear box in three equal layers, each subjected to 25 blows with a light compactor (2.5 kg hand rammer falling from 305 mm) to preset the target dry density of 1.337 g/cm3. The top of each layer was loosened before the next filling and compaction to keep the sample homogeneous. The prespecified normal stress was applied to consolidate the specimen until the vertical displacement was constant. After reaching the preset shear displacement, the cutting ring (with a 61.8 mm diameter and 40 mm height) was gently pushed into the sheared specimen from a vertical direction to measure the permeability with the TST-55 falling-head permeameter.
To analyze the variation in particle shape with dry-wet cycles, PCAS software (Liu et al., 2011) was adopted to obtain the shape coefficients by statistically analyzing 480 particles with different sizes. Five coefficients were taken into analysis, including the uniformity coefficient (UC), sorting coefficient (SC), curvature coefficient (CC), average form factor (AFF), and fractal dimension (FD). Coarse granules were broken to obtain the cross-section for observing their microstructure in Thermo Scientific Prisma E environmental scanning electron microscope (ESEM).