Coal burst is one of the most frequent and destructive dynamic disasters encountered during underground mining engineering. However, the understanding of quantitative precursor characteristics of coal burst is still in its infancy, rendering it difficult to provide effective early warning of disaster. In this study, to quantitatively study precursor characteristics and warning signs of coal burst, the coal burst experiments were carried out on coal-rock combination with a crack. The acoustic emission (AE) technique was employed to quantitatively analyse the precursor information during coal burst process. Testing results indicated that coal burst process is classified into three stages based on evolution in AE energy, i.e., early incubation stage, late incubation stage and occurrence stage. The first significant increase in AE energy could be identified as the beginning of the late incubation stage of coal burst, accompanying by the phenomenon of macro-failure initiation. AE signals during the whole process could be classified as five types according to their dominant frequency and amplitude characteristics, i.e., HF-HA, LF-HA, EHF-LA, HF-LA and LF-LA respectively. The dramatic increase in number proportion of HF-HA and LF-HA signals is highly correlated with occurrence of coal burst. In addition, a comprehensive classification criterion for the coal burst prediction was proposed under a quantitative analysis of three AE parameters, i.e., first energy index (FEI), coal burst risk indicator based on AE energy (CRIE) and frequency spectrum (CRIF−A). The findings in this study could facilitate accurate coal burst prediction.
Read full abstract