矿井瓦斯爆炸超压值分布的神经网络预测方法

Overpressure Distribution Prediction of Mine Gas Explosion Based on Artificial Neural Networks

瓦斯灾害是影响煤矿安全的重要问题之一,而爆炸后的主要危害之一是冲击波的伤害。而且,煤矿防隔爆措施是否能起到有效作用也依赖于冲击波超压值的测量和预测。在前人实验分析的基础上,应用神经网络理论,分别用BP神经网络和RBF神经网络对瓦斯爆炸后的冲击波超压值和测点之间的关系进行了预测。结果表明,BP神经网络的预测误差最小,应用神经网络进行预测可以明显的减小预测的误差,适合煤矿企业实际应用。

Gas disaster is one of important problems concerning the safety of mine. The shock wave after explosion is the main damage for human. Moreover, whether the anti - explosion - proof measures play an effective role is depended on the measurements and fore-casts of blast overpressure. Based on previous experimental analysis, the paper uses BP and RBF ANN to predict the relationship of gas explosion overpressure between point measurement and overpressure value. The results show that application of BP ANN prediction can significantly reduce the forecast error, it is suitable for mining enterprises to operate.