On the basis of the one-dimensional fractal model, a two-dimensional fractal model is developed to investigate the particle size distribution (PSD) of heterogeneous geomaterials. Based on the simulation results for PSD curves of geomaterials, it is proved that the two-dimensional fractal model is superior to the one-dimensional fractal model. The PSD analyses indicate that two major mechanisms may govern the formation of the geomaterials and these characteristics can be captured by the two-dimensional fractal model. In this respect, the two-dimensional fractal model is better than the three-dimensional model. The two-dimensional fractal model is further adopted to investigate the relationship between the mechanical properties and PSD of geomaterials.

Solid waste formed during the excavation of soft red stratum rock is often encountered in engineering practice. However, its reuse has been limited because it often shows a gradual degradation mechanism during water–rock interactions. Similarity simulation experiments of geotechnical materials have been developed to be environmentally friendly; however, their application in soft rock mechanics is still limited. Based on these limitations, this study aims to prepare red-bedded soft rock-like materials by referring to the diagenetic process of sedimentary rocks using low-melting-point glass powder (STGP) and high-temperature and vertical stress to accurately simulate the progressive disintegration properties of red-bedded soft rock. For this purpose, a series of laboratory tests were conducted to verify the function of STGP in the embedment of skeleton particles of soft rock as a cement material for resisting the dry–wet cycle. Micro-scanning electron microscopy, disintegration experiments with dry–wet cycles, and basic physical and mechanical property tests were conducted for the synthetic red soft rock-like material. Finally, the synthetic and natural materials were compared based on their density, microstructure, disintegration breakage, and uniaxial compression mechanical properties. The results showed that adding STGP promoted embedded solidification between aggregate particles. The simulated material exhibited the same characteristics of gradual disintegration breakage as natural red-bedded soft rock. Meanwhile, the basic physical and mechanical properties were in substantial agreement when the STGP content was 0.5~2%.

The capillary rise of shallow mineralized groundwater can contribute to the salinization of the soil layers. The excessive salt amounts adversely affect soil physical and mechanical properties, as well as the heat transfer performance, all of which are key factors with regard to the design of geothermal-related earth structures such as geothermal energy piles (GEP), ground source heat pumps (GSHP), and earth-air tunnel heat exchangers (EATHE). Therefore, in this study, the thermal-mechanical properties of saline soils are systematically investigated. A series of thermal and mechanical response tests were carried out under different salinity conditions, and the shear wave velocity-stress behavior of saline soil was measured using a modified oedometric cell coupled with an anchored bender element pair. Experimental results showed that saline soils generally have higher dry density and lower optimum moisture content at higher salt contents. The shear strength of saline soil increased about 5% while the salt concentrations of bulk solution increased from 0 mol/kg to 6 mol/kg, and the shear wave velocity increased by 50% to 83% when the normal load increased from 12.5 kPa to 250 kPa for sodium chloride- (NaCl-) treated soil and 39% to 52% for calcium chloride- (CaCl₂-) treated soil. In addition, the thermal conductivity decreased by 0.121 W m^-1 K^-1 for NaCl-treated soil and 0.129 W m^-1 K^-1 for CaCl₂-treated soil on average when the salt concentration increased from 0 mol/kg to 6 mol/kg. Finally, an elastic shear modulus (G0) model and a thermal conductivity (K) model were formulated for saline soil for the first time, and the effectiveness and feasibility of the proposed models were validated by comparisons of the model predicted values and experimental data.