基于BP神经网络预测岩石导热系数

Prediction of rock thermal conductivity based on BP neural network

采用光学扫描仪测试了南昌地区出露的主要岩石样品的导热系数,并分析导热系数与孔隙率、含水率、密度、温度及压力之间的关系。在此基础上,提出基于BP神经网络的岩石热导率预测模型。结果表明,BP神经网络模型具有较高的预测精度,相关系数可达0.993 82,满足工程精度要求。

The thermal conductivity of exposed rocks in Nanchang area was measured by thermal conductivity scanning,and the relationship between thermal conductivity and porosity,moisture content,density,temperature and pressure were analyzed.Furthermore,the prediction model of rock thermal conductivity based on BP neural network is proposed.The results show that the BP neural network model has high prediction accuracy,and the correlation coefficient can reach 0.99382,which meets the engineering accuracy requirements.