A multi-parameter identification model for classification of rock burst risk and its application

Rock burst is one of common mine dynamic disasters in mine production activities.So, the high accurate evaluation of rock burst hazard grade should be used to ensure a safe mining production.In order to realize a rapid, efficient and accurate forecast of rock burst hazard classification, multiple factors should be considered which exist more or less correlations that lead to the overlapping of parameter information.An improved principal component analysis (PCA) method was proposed to reduce the dimensionality of the rockburst risk index data.And three new indexes were extracted to conduct an overall risk classification evaluating on the rock burst.On the basis of PCA and the distance discriminant analysis (DDA), the PCA model for rock burst hazard classification evaluation was established and applied to Yanshitai coal mine in Chongqing City.The prediction results showed that with six different training and testing samples, the PCA-DDA model still had a favorable predictive effect, and the misjudgment rates were respectively 5.71%, 5.71%, 5.71%, 5.71%, 5.71% and 8.57%.Meanwhile the prediction performance of PCA model had been proved.The model can provide a suitable approach for the evaluation of rock burst hazard grade in mining, and can be popularized in practical engineering.