基于GWO改进的PCA-BP神经网络煤层底板破坏深度预测模型

Prediction Model of Failure Depth of Coal Seam Floor Based on PCA-BP Neural Network Improved by GWO

针对矿井水害防治工作中,煤层底板破坏深度难以进行准确预测的问题,将主成分分析(PCA)与灰狼算法(GWO)改进的BP神经网络相结合,建立以采深、煤层倾角、采厚、工作面斜长、煤层底板抗破坏能力、工作面内是否有切穿型断层或破碎带为主要影响因素的底板破坏深度预测模型。根据实测资料分析各主要影响因素和底板破坏深度之间的相关性,利用PCA法将影响底板破坏深度的主要参数进行降维,根据降维后的主成分对底板破坏深度的贡献率,确定底板破坏深度的主控因素。利用灰狼算法优化BP神经网络参数,建立PCA-GWO-BP神经网络模型预测煤层底板破坏深度,并与其他预测方法进行对比,结果证明该模型误差小于0.5%、准确度高,可以对煤层底板破坏深度进行较为准确的预测。

It is difficult to accurately predict the failure depth of coal seam floor in the prevention and control of mine water disaster.In view of this problem,the principal component analysis(PCA)and the BP neural network improved by Grey Wolf algorithm(GWO)were combined to establish the prediction model of failure depth of coal seam floor.In this model,main influencing factors included the mining depth,dip angle of coal seam,mining thickness,inclined length of working face,antidamage ability of coal seam floor and whether there were cut faults or fracture zones in the working face.According to the measured data,the correlation between the main influencing factors and the failure depth of the floor was analyzed.Dimension reduction was carried out on the main factors affecting the failure depth of the floor by PCA method.According to the contribution rate of the main components to the failure depth of the floor after dimension reduction,the main controlling factors were determined.By using GWO to optimize the parameters of BP neural network,a PCA-GWO-BP neural network model was established to predict the failure depth of coal seam floor.Compared with other prediction methods,the model had a smaller error and a higher accuracy,and can predict the failure depth of coal seam floor more accurately.