Rockburst prediction and classification based on the ideal-point method of information theory

A rockburst is a sudden dynamic process under high geostress conditions where rocks spontaneously explode. This is an important geological problem for underground construction processes. A rockburst could lead to equipment damage, casualties, and construction delays. Therefore, rockburst prediction and classification are extremely significant. A prediction and classification model is established by introducing the basic theory of the ideal-point method, considering the rockburst mechanism. Three parameters are selected as evaluation indexes, including the rock stress coefficient (sigma(theta)/sigma(c)), rock brittleness coefficient (sigma(c)/sigma(t)), and elastic energy index (M-et). To eliminate any correlation between the parameters, a principal component analysis based on mutual information (MIPCA) for the rockburst feature selection is used to calculate a new group of parameters. Then, using the information-entropy theory, the weight coefficients of these new evaluation indexes are confirmed. Finally, using statistics-related projects, engineering-case analyses show the feasibility and applicability of the proposed model. A computer evaluation program with a rockburst-classification interface was developed, based on the proposed model. This model and computer software can be used for other similar engineering practices in the future.