Failure evolution and precursor prediction of fissured sandstone based on power-law behavior of electric potential

The engineering rock with defects is prone to induce serious geological disasters such as rockbursts under stress disturbance. Exploring the power-law failure prediction method of electric potential (EP) signals is effective and has great potential. In this study, the EP monitoring tests of the fissured red sandstone samples with variable crack angles under uniaxial compression were carried out. Spatiotemporal response, The Hurst statistical feature and Benioff-type power law of EPs were adopted to analyze the scale invariance of EPs and their relationship with damage evolution during rock failure. The failure prediction method based on Benioff potential and early warming assessment system were further proposed to predict the failure time range of potential rock instability in real time. The results show that with increasing crack angle, the mechanical strength and macroscopic crack number of samples reduce, and the power rate growth trend of EPs is delayed. The EP intensity increases closer to the crack propagation path, which is associated with the local damage and moving charged dislocations. Benioff potential of all samples shows power-law properties and is closely related to the internal damage evolution. With decreasing the remaining failure time, the variable power exponent m_t in the local temporal window gradually increases and approaches the upper limit of 1, indicating an increased possibility of potential catastrophic failure in rocks. The failure time range is predicted in real time through Benioff power behaviors with the upper and lower limits of m_t, and the prediction accuracy improves as the rock failure approaches.