人工神经网络在矿井突水预报中的应用  

The application of artificial neural network theory to mine water inrush prediction

作　　者：冯利军[1] 郭晓山[1] 

机构地区：[1]煤炭科学研究总院西安分院,陕西西安710054

出　　处：《西安科技学院学报》2003年第4期369-371,393,共4页

摘　　要：突水预报是一项重要的矿井水文地质工作。借助于人工神经网络在处理非线性问题或非结构问题方面的优势,采用BP算法,基于大量矿井突水样本实例建立了突水预报神经网络模型,并将该模型用于实际预报,并取得了较好的效果。结果表明,模型具有较强的实用性。为了提高模型的预测精度,在训练样本的选择上还应具有一定的代表性。Water inrush prediction is an important mine hydrogeological work. With the advantage of artificial neural network in solving nonlinear or unstructured problems and by using BP algorithm, a neural network model for predicting mine water inrush has been built based on a large number of water inrush examples. To apply this model to practical prediction produces a better effectiveness, which shows that this model can be practically used. In order to improve the accuracy of model prediction, the training samples should be representatively prepared.

关 键 词：人工神经网络 BP算法 突水预报 矿井安全 

分 类 号：TD713[矿业工程—矿井通风与安全] TP183[自动化与计算机技术—控制理论与控制工程]