BP神经网络在煤矿瓦斯安全事故预测中的应用

Application of BP Neural Network in Coal Mine Gas Safety Accident Prediction

煤炭安全开采对我国的能源安全战略实现具有非常重要的意义,对煤矿瓦斯安全事故进行预测可以有效提升煤矿开采的安全性。对煤矿瓦斯安全事故中的关键因素和影响机理进行了分析,阐述了BP神经网络的原理及特点,建立了基于BP神经网络的煤矿瓦斯安全事故预测模型,给出了BP神经网络训练流程,确定了输入层、隐含层以及输出层参数,研究了隐含层不同神经元数量时的训练结果,提出了一种自适应学习速率的方法来缩短训练时间,使用25个样本数据对模型进行验证,建立的预测模型可以有效预测煤矿瓦斯安全事故,且误差较小。

Safe coal mining is of great importance to the realization of China′s energy security strategy,and the prediction of coal mine gas safety accidents can effectively improve the safety of coal mining.The key factors and influencing mechanisms in coal mine gas safety accidents are analyzed,the principle and characteristics of BP neural network are described,the prediction model of coal mine gas safety accidents based on BP neural network is established,the training process of BP neural network is given,the parameters of input layer,hidden layer and output layer are determined,the training results of different number of neurons in hidden layer are studied,an adaptive learning a method of adaptive learning rate is proposed to shorten the training time,and the model is validated by using 25 samples of data.