Projection pursuit model based on particle swarm optimization for rock burst prediction

Rock burst is one of the main engineering geological hazards of underground rock engineering in deep high ground stress zone. The prediction of rock burst intensity is a complex systematic problem of uncertainty. Based on the projection pursuit (PP), the particle swarm optimization (PSO) and the logistic curve function (LCF), a new model for rock burst prediction is developed, which is referred to as projection pursuit based on particle swarm optimization (PSO-PP). The ratio of the maximum tangential stress of the cavern wall to the uniaxial compressive strength, the brittleness coefficient and the elastic energy index of rock are regarded to as the discrimination indices of PSO-PP. The model, on the one hand, uses the PSO to optimize the projection index function and the parameters of LCF so as to ensure the accuracy of the parameters used in the model. On the other hand, the nonlinear relationship between projection values and empirical grades is established according to LCF. The test results of the model show a very good precision. In this study, the prediction results obtained by applying the developed model to Qinling Tunnel and Dongguashan Mine are well consistent with the practical situation. It indicates that the model is feasible and effective for rock burst prediction.