遗传算法与神经网络的结合在煤与瓦斯突出危险性预测中的应用研究

On the combined application of genetic algorithm and neural network in the prediction of coal/gas outbursts

将反映煤与瓦斯突出的重要特征指标:最大钻屑量(S)、钻屑解析指标(K1)、钻孔瓦斯涌出初速度(q0)和煤的坚固性系数(f)通过神经网络与遗传算法有效结合,建立煤与瓦斯突出危险性预测模型,并通过现场实测数据进行突出危险性预测。结果表明:两种算法的结合对煤与瓦斯突出危险性预测是有效的,它与传统的预测方法相比收效速度更快,容错能力更强,预测精度更高。

Indicators reflecting the important characters of coal/gas outbursts such as maximum drill chips （S）,drill chip analytical index （K1）,initial velocity of gas outflow through drill hole （q0） and the coefficient of coal rigidity （f） are effectively combined with genetic algorithm via neural network to form a coal/gas outburst prediction model.With the help of the data and test results obtained on site,outburst danger is predicted.Results indicate that the combination of the two algorithms produces an effective prediction of the danger of coal/gas outbursts.Compared with traditional prediction methods,this method produces quicker results with higher fault tolerance and better accuracy.