RBF神经网络在岩爆预测上的运用

岩爆是煤矿生产中常见的动力灾害,岩爆的预测一直是矿井灾害预测中的难点。由于岩爆的发生突然、机理复杂、干扰和影响因素多,因此很难对岩爆进行有效的预测。人工神经网络具有并行分布处理与存储、高度容错、自组织、自适应和自学习功能,能分析较为复杂的非线性系统。文中以MATLAB7.0为开发环境,利用其神经网络工具箱的RBF网络,通过一套权值和阀值来仿真模拟岩爆对岩爆进行预估。通过实际模拟表明利用RBF神经网络对岩爆进行预估是一种成功的的方法,其预测结果与实际情况相一致,说明利用该网络模型对岩爆的预测是可靠的。

Prediction Forecast of Rockburst Based on RBF Neural Network

Rockburst happen frequently in the production of coal mine disasters momentum.Rockburst forecast is the prediction of mine disasters difficult.Because a rockburst occurred suddenly,the mechanism of complex,interference and influence factors and more features,it is difficult to carry out effective rockburst forecast.Artificial neural networks are parallel distributed processing and storage and a high degree of fault-tolerant,self-organizing,adaptive and self-learning function to analyze the more complicated nonlinear systems.t By using its neural network toolbox RBF network and through a set of weights and the threshold to simulation rockburst forecast of a rock burst in MATLAB7.0 development environment.Through simulation shows that the actual use of RBF neural network to forecast the rock burst is a successful method,the results are consistent with the actual situation,the use of the network model of the rock burst forecast is reliable.