Investigation of Shale Mesoscale Mechanical Properties Based on a Coupled Statistical-Clustering Analysis Method

The mechanical behavior at the microscale of shale is crucial for determining the macroscopic deformation and failure characteristics. In this study, we proposed a method for predicting the macroscopic modulus of shale through the nanoindentation test. The study was carried out using grid nanoindentation, scanning electron microscopy (SEM)-energy-dispersive spectroscopy (EDS), and scanning electron microscopy mineral quantitative analysis (QEMSCAN). The mineral composition of the indentation points was described in terms of clay mineral, organic matter, silicate, and carbonate content using SEM-EDS and QEMSCAN, and the shale mineral phases was determined. The variation in Young`s modulus was analyzed based on the recorded indentation depth curves and modulus distributions. And the mechanical phases of the shale were determined using a clustering algorithm. The relationship between the mineral phases and mechanical phases of the shale was optimized using a coupling statistical-clustering analysis method to upgrade the microscale mechanical scale model of the shale. The results show that the shale grid nanoindentation points are multi-mineral inclusions, resulting in large discrete results of indentation. The optimal equivalent mechanical model of the shale to simplify the discrete results of indentation is a three-phase model. Further, the upgraded model determined by nanoindentation results using Mori-Tanaka, Dilute, and optimized Z-Mori methods is satisfactory when compared to the uniaxial compression results, and the optimized Z-Mori model has a better fitting performance for the macroscopic elastic modulus calculated by uniaxial compression.