摘要
根据某煤矿多年动力现象的分析研究资料,利用广义回归神经网络(GRNN)模型,并借助Matlab软件编程,对该矿冲击地压发生情况进行了模糊综合评判。结果表明,GRNN神经网络能智能的学习地质构造、煤层倾角变化、煤厚变化、顶板管理情况和采前卸压情况与冲击地压的映射关系,并能很好的预测冲击地压。
Based on the GRNN neural network, real data and Matlab, rock bursts of this coal mine are predicted. Results shows, the GRNN model can build mapping relation among rock burst and structure, change of coal seam dip angle and coal thickness, roof management and pressure relief, and it can predict rock burst.
引文
[1]蒋金泉,李洪.基于混沌时序预测方法的冲击地压预测研究[J]岩石力学与工程学报,2006,25(5):889-895.
[2]赵洪波,茹忠亮.冲击地压预测的PSO-SVM模型[J].岩石力学与工程学,2007,26(增1):3479-3483.
[3]陈刚,潘一山.基于异步迭代算法的冲击地压预测[J].岩土力学2004,25(3):446-450.
[4]尹光志,代高飞,闫河,等.冲击地压预测的遗传神经网络方法[J].岩土力学,2003,24(6):1016-1020.
[5]尹光志,谭钦文,魏作安.基于混沌优化神经网络的冲击地压预测模型[J].煤炭学报,2008,33(8):871-875.