Investigation of crack recognition and spatio-temporal evolution pattern in coal samples damage.

Understanding the evolution mechanism of cracks helps to evaluate the behavior and performance of rock masses and provides a theoretical basis for the mechanism of crack propagation and instability. For this purpose, a rock mechanics testing system and an acoustic emission monitoring system were used to conduct acoustic emission positioning experiments on coal samples under uniaxial compression. According to clustering theory, the distribution pattern of microcracks and the dynamic evolution process of multiple cracks were studied. Subsequently, the reasons for the change in the spatio-temporal entropy (H) and fractal dimension (D) of a single crack were revealed. The research results show that microcracks present a statistical equilibrium distribution, the Gaussian distribution model is applicable to cluster crack distribution patterns, and a machine learning method can effectively identify cracks. The fractal dimension reflects the spatial characteristics of three-dimensional elliptical cracks, and low-dimensional cluster cracks are more likely to develop into macroscopic cracks. The change of H is related to the formation process of cracks, and an abnormal H (sudden increase and sudden decrease) could provide precursor information for the instability of coal samples. This research provides a new method to study crack distributions and formations and shows the competitiveness of the method in evaluating the damage state of coal.