基于遗传算法和BP网络的岩爆预测

Prediction of Rockbursts Based on Genetic Algorithms and BP Neural Network

选取影响岩爆的一些关键指标,例如脆性系数、弹性应变能指数、应力强度比等作为输入参数,采用经遗传算法优化过的BP网络,对岩爆的发生及其烈度进行了预测。针对国内外一些地下工程的实例进行了分析计算,预测结果与实际情况是符合的,说明这种方法有一定的实际意义。

With the optimization of genetic algorithms, a kind of BP neural network is established to predict rockbursts and their intensity, in which some main control factors of rock burst are chosen as input parameters such as brittleness index, elastic energy index and stress intensity ratio. According to analyses of some actual examples of the underground engineering at home and abroad, the results of prediction agree well with the practical records, which shows that the method is effective in prediction of rockbursts.