基于PSO算法优化的自组织竞争神经网络在煤与瓦斯突出预测中的应用研究

Application of optimized self-organizing competitive neural network by PSO algorithm for coal and gas outburst prediction

为了提高煤与瓦斯突出预测的准确率和方法的合理性,在分析突出机理的基础上,提出了运用经PSO算法优化的自组织竞争神经网络来预测煤与瓦斯突出的方法。由于影响突出的因素是多方面的,根据可以反映其相互关系最大钻屑量(S)、煤屑解析指标(K1)、钻孔瓦斯涌出初速度(q)和煤的坚固性系数(f)来建立预测模型,并且运用PSO算法对网络权值进行优化。结果表明,该方法结果稳定,准确率高,因此运用该方法来预测煤与瓦斯突出是可行的。

To improve the accuracy rate and the rationality of the prediction of coal and gas outburst, the method for predicting the coal and gas outburst with the optimized self-organizing competitive neural network by PSO algorithm was put forward on the basis of the analysis of the outburst mechanism. The prediction model was established with the maximum drilling cuttings weight （S）, the drilling cuttings analytical index （K1）, the initial velocity of gas outburst through the boreholes （q） and the coefficient of coal rigidity （f）. Besides, the network weights were optimized by PSO algorithm. The results showed that this method has good performance of accuracy and stability and it is feasible to predict the coal and gas outburst.