Summary The crack initiation stress, which serves as the boundary between linear elastic deformation and nonlinear deformation in rock materials, plays a crucial role in rock mechanics and engineering. Numerous scholars have developed various methods for identifying crack initiation stress based on the deformation and failure mechanisms of rocks. In this study, building upon previous research, we propose a novel method for determining crack initiation stress that adheres to principles of simplicity, clear physical interpretation, and absence of artificial influence. This method leverages the energy dissipation characteristics of rock materials during compressive deformation, defining the crack initiation point as the location with the least dissipation energy ratio (LDER). Theoretical analysis demonstrates that this approach is inherently valid for hard rocks with minimal primary fractures, and the crack initiation stress obtained using this method is independent of the Young’s modulus and Poisson’s ratio. To validate the proposed method, uniaxial compression stress-strain curves of granite, marble, and dense sandstone were analyzed in comparison with acoustic emission (AE), lateral strain response (ΔLSR) method, crack volume strain method (CVSM), volumetric strain response method (VSRM), and LDER. The results indicate that the crack initiation stress determined by the LDER method aligns closely with that of the value of the AE method and exhibits superior reliability in distinguishing the boundary of linear elastic and nonlinear deformation. Additionally, the influence of confining pressure on the LDER method was investigated through a series of stress-strain curves of Longmaxi shale under varying confining pressures. The findings reveal that the LDER method accurately reflects the influence mechanism of confining pressure. Further experimental data analysis confirms that the LDER method is independent of the elastic modulus and, within a certain range, independent of Poisson’s ratio. These findings enhance the robustness and applicability of the LDER method, providing sufficient tolerance for practical applications.
He et al. (2026) studied this question.