基于强度折减与ANN-GA模型的采场结构参数优化

Stope structural parameters optimization based on strength reduction and ANN-GA model

为了得到合理的采场结构参数,对采场矿柱进行强度折减,采用矿柱塑性破坏区贯通作为相邻采场整体失稳的判据,参照类似矿山,确定采场结构参数选取范围,并结合FLAC3D数值分析软件得出选取范围内的数组参数组合下的安全系数,将所得安全系数作为BP神经网络的训练样本拟合各组合参数与安全系数之间的非线性关系。根据矿山工程实际情况,确定遗传算法适应度函数,从而搜索出选取范围内适应度函数的最优解以及最优解所对应的参数组合。结果表明:以某一铜矿为例,运用该方法求得最优解为0.2,对应的参数组合为:控顶高度7 m,矿房跨度12 m,矿柱宽6 m。大大减少了数值模拟工作量。

In order to obtain reasonable structural parameters of the stope,the transfixion of the plastic damage zone of pillars was adopted as the criterion,the similar mine was referred to determine the selection of structural parameters of the stope.Combined with FLAC3D numerical analysis software,the safety factor of some parameter combinations in the selection was obtained based on pillar strength reduction.The obtained safety factor was used as the training samples of BP neural network to fit the nonlinear relationship between the various combinations of parameters and the safety factor.On the basis of mining engineering,the genetic algorithm fitness function was determined,and the optimal solution and the corresponding parameters of the stope were obtained.Taking a copper mine as an example,the optimal solution of 0.2 was obtained using this method,corresponding to the combination of parameters as follows： 7 m for controlled top height,12 m for room width,6 m for pillar width.The workload of numerical simulation is greatly reduced.