煤与瓦斯突出预测的自适应神经-模糊推理系统研究

Studies of an ANFIS-based system on the prediction of the coal and gas burst

在煤与瓦斯突出危险性预测方法中,神经网络方法存在着收敛速度慢、拟合能力差、预测精度低、训练结果不惟一等缺陷.针对这些缺陷,应用自适应神经-模糊推理系统的原理,建立了煤与瓦斯突出危险性预测的自适应神经-模糊推理方法,并应用该方法对部分实例进行了反演预测.预测结果表明:该方法具有收敛速度快、拟合能力强、预测精度高、训练结果惟一等优点,是一种优异的反演预测方法.作为一种探讨,还对煤与瓦斯突出的危险性进行了模糊划分.

ANN - based system on the prediction of coal and gas burst has shortcomings such as slow convergence speed, poor fitting capability, low accuracy of prediction and indefiniteness of the training results. In order to overcome these, Neuro -Fuzzy Inference System is used to establish an ANFIS -based system on the prediction of coal and gas burst. Furthermore, the approach is used for the inverse predictions of several examples. The inverse prediction results show that the approach has the merits of high convergence speed, good fitting capability, high accuracy of prediction and definiteness of the training results, and it is an excellent approach for the inverse prediction of coal and gas burst. On the other hand the paper made an attempt to give an fuzzy grade of the coal and gas burst.