延时煤与瓦斯突出的实时预警理论与应用研究  

Research on Real-Time Alarm Theory of Delayed Coal and Gas Outburst

作　　者：罗新荣[1] 杨飞[1] 康与涛[1] 张爱然[1] 

机构地区：[1]中国矿业大学煤炭资源与安全开采国家重点实验室安全工程学院,江苏徐州221008

出　　处：《中国矿业大学学报》2008年第2期163-166,共4页Journal of China University of Mining & Technology

基　　金：国家自然科学基金重点项目(50534050);教育部科学技术研究重点项目(105025)

摘　　要：结合煤矿井下瓦斯涌出实时监测图,利用神经网络技术判断瓦斯异常情况.选取4个参数作为瓦斯延时突出预测的特征指标:井下瓦斯涌出峰值、瓦斯上升梯度、瓦斯超限时间和瓦斯下降梯度.瓦斯异常涌出超限3%,并持续时间超过10 s为瓦斯延时突出敏感指标的临界值.结合VB+ADO的编程及数据库访问技术,建立了人工智能神经网络的瓦斯预警理论模型、瓦斯预警模型的自学习训练方法和瓦斯预警技术.模型预测结果与实际情况完全相同.Real-time monitoring of gas discharges in a coal mine was done using neural network （NN） technology. Four parameters were found to predict delayed coal and gas outburst. They were the peak concentration of the gas emission, the rate of increase of the gas concentration, the overtime, of gas emission and the descent gradient of gas concentration. The results show that gas concentrations exceeding 3 % for duration of more than 10 s are conditions that predict a delayed coal and gas outburst. The artificial intelligence NN theory, the training by selflearning, and the technology warning gas unconventionality were established by programming with VB and ADO. The predicted results of the model are consistent with the facts.

关 键 词：瓦斯 延时突出 预警 神经网络 

分 类 号：TD713[矿业工程—矿井通风与安全]