RBF神经网络在岩爆预测上的运用

Prediction Forecast of Rockburst Based on RBF Neural Network

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。

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.