基于优化BP神经网络的矿压规律预测方法研究

Research on Prediction Method of Ground Pressure Law Based on Optimized BP Neural Network

矿压显现及其控制方法对煤炭资源安全高效生产影响巨大。针对综采工作面矿压显现致灾问题,综合运用工程调研、理论分析、模型构建等方法,开展了综采工作面矿压规律预测模型的系统性研究。采用免疫粒子群混合算法优化BP神经网络(IA-PSO-BP),建立了基于IA-PSO-BP的矿压规律预测模型。选取平均绝对误差、均方误差和相关系数作为评价指标,对比分析了BP模型、PSO-BP模型和IA-PSO-BP模型的预测性能。结果表明:相较于BP和PSO-BP模型,IA-PSOBP模型的收敛速度分别提高7.44倍和1.79倍;在测试集上的预测精度显著提高,预测结果符合矿压数据的周期性变化规律。基于优化BP神经网络的矿压规律预测方法,可为预防煤矿井下矿压灾害的发生提供一定的指导借鉴。

The ground pressure appearance and its control method have a great influence on the safe and efficient production of coal resources.Aiming at the disaster-causing problem of ground pressure behavior in fully mechanized mining face,a systematic study on the prediction model of ground pressure law in fully mechanized mining face was carried out by means of engineering investigation,theoretical analysis and model construction.The BP neural network(IA-PSO-BP)is optimized by immune particle swarm optimization hybrid algorithm,and the prediction model of ground pressure law based on IA-PSO-BP is established.The mean absolute error,mean square error and correlation coefficient are selected as evaluation indexes,and the prediction performance of BP model,PSO-BP model and IA-PSO-BP model is compared and analyzed.The results show that compared with BP and PSO-BP models,the convergence speed of IA-PSO-BP model is increased by 7.44 times and 1.79 times respectively.The prediction accuracy on the test set is significantly improved,and the prediction results are in line with the periodic variation of ground pressure data.The prediction method of ground pressure law based on optimized BP neural network provides some guidance for preventing the occurrence of ground pressure disaster in coal mine.