|
Abstract Rock burst is a complex，nonlinear dynamic mechanics phenomenon and its mechanism is very complex. Analysis and prediction of rock burst using monitored data is an important research method. But it is difficult to present the complex，nonlinear relationship between rock burst and its influence factors using conventional mathematics and mechanics methods. Forecast of rock burst based on time series analysis is a key research direction. In time series analysis of rock burst，the rock burst is seen as a time series process，and the nonlinear relationship between time series is built using support vector machine(SVM). Because of the influence of parameters of support vector machine，they are selected by particle swarm optimization(PSO). Thus the PSO-SVM method is proposed. It enhances the efficiency and capability of forecasting. The proposed method is applied to the forecast of rock burst and the results show that it is scientific，feasible and precise.
|
|
Received: 17 April 2007
|
|
|
|
