Rockburst failure time prediction based on a fuzzy comprehensive evaluation method using the acoustic emission

Rockbursts have become one of the most serious disasters in underground engineering around the world, which seriously threaten the construction safety of underground engineering. The effective prediction of rockbursts is of great significance for the safe production management of deep engineering. In this study, the uniaxial compression tests were carried out on sandstone and granite specimens with different shapes and sizes. A multi-index fuzzy comprehensive evaluation model was established based on the acoustic emission (AE) characteristic parameters to quantitatively evaluate the possibility of rock failure. In the fuzzy comprehensive evaluation model, the exponential distribution function in reliability theory was introduced, and the membership function was constructed by Gaussian distribution. The analytic hierarchy process (AHP) and entropy weight method (EWM) were utilized to determine the subjective and objective weights of each index respectively, and the distance function was employed to obtain the synthesized weight. Thereafter, the comprehensive prediction results were obtained by variable fuzzy pattern recognition (VFPR). The results show that for both sandstone and granite specimens with different shapes and sizes, the time advance (Δ t ) of rock failure forecasting is in the range of 145-491 s, and the forecasting point is 0.761-0.889 of the total loading time of rock failure. The prediction results are mainly affected by lithology, while the impact of rock shape and size is relatively insignificant. The sensitivity of fuzzy comprehensive evaluation index is: granite>sandstone. This research can provide a useful reference for the prediction of rockburst. • A multi-index fuzzy comprehensive evaluation model was established using the characteristic parameters obtained from acoustic emission monitoring to quantitatively assess the possibility of rock failure. • The uniaxial compression tests were carried out on granite and sandstone samples of different shapes and sizes in the laboratory to verify the accuracy and universality of the model. • A fuzzy comprehensive evaluation index is constructed, and the sensitivity of lithology and sample size to the fuzzy evaluation index is analyzed.