神经网络理论在延时煤与瓦斯突出预警中的应用

Application of neural network in prediction of delay coal and gas outbursts

为提高煤与瓦斯突出预警的准确性,利用人工神经网络特有的非线性适应性信息处理能力,选择瓦斯涌出峰值、上升梯度、下降梯度、超限时间四个延时突出预警指标,实现煤与瓦斯延时突出的预警。瓦斯样本学习和突出预警结果与实际情况对比表明,前馈神经网络预警模型准确率很高,可以克服传统预警方法存在的瓦斯突出漏报和误报的缺陷。

Traditional methods for prediction of delay coal and gas outbursts tend to distort, or fail to report an accident due to the complexity of the process. However, this problem can be solved by using artificial neural networks with characteristics of highly non-linear mapping. In this paper, four indexes were chosen when using the BP network model to predict the outbursts. The comparison of the sample prediction and the practice results demonstrated that the model can predict delay coal and gas outbursts with high precision.