Integrated rockburst early warning model based on fuzzy comprehensive evaluation method

The development of reliable early rockburst warning models for underground mines is a challenging task considering the complex nature of rockburst phenomena. In this paper, a real-time, integrated multi-system, early rockburst warning model is developed to achieve the quantitative prediction of the rockburst probabilities at a specific time. The model input parameters are data derived from microseismic (MS) and acoustic emission (AE) monitoring systems. The early warning model is based on fuzzy comprehensive evaluation method (FCEM), confusion matrix and maximum membership degree principle. These tools integrate the advantages of both multi-system and multi-parameter into one unified model. During the early warning procedure, six early warning indices representing temporal, spatial and energy parameters are rationalized. Next, the weight of each index is determined from confusion matrix leading to a score value. Real-time early warning results can then be determined from the maximum membership degree principle. The results show that the proposed integrated MS-AE early warning model has the advantage of combining multiple systems and indices, and it has higher prediction accuracy than individual systems. The early warning model is successfully demonstrated with a case study from the Wudong Coal Mine in China. The proposed model improves the rockburst early warning efficiency and provides a solid foundation for ensuring the safety of personnel and equipment in underground mining.