基于遗传算法—BP神经网络的突出强度预测

煤与瓦斯突出强度的预测对研究煤与瓦斯突出,保证矿井安全正常生产有着重要意义。本文提出采用遗传算法结合BP神经网络的模型来预测突出强度,采用遗传算法对BP神经网络的权重和阈值进行优化,将优化好的权重与阈值作用于网络进行训练,直至性能函数符合要求。实际计算表明,该模型有较好的预测精度,且克服了普通BP神经网络训练时间长、收敛速度慢的缺点,在已知瓦斯膨胀能和煤层厚度的前提下,可以用该模型对突出强度进行预测。

Based on Genetic Algorithm-BP Neural Networks Outburst Intensity Prediction

Coal and gas outburst intensity forecasting has important significance to study coal and gas outburst and ensure safe and normal production in the mine.This article proposes using genetic algorithms and BP neural network models to predict the outburst intensity,the use of genetic algorithm on BP neural network’s weights and thresholds optimizing,optimizing weight and threshold will effect on the network for training,until performance functions to meet the requirements.The actual calculation means that this model has the better forecast accuracy,and overcomes the shortcomings of the general BP neural network long training time and slow convergence speed,under the premise of known gas expansion energy and seam thickness can use this model to make predictions on the outburst.