基于神经网络的仿真输出方法在地应力场分析中的应用

The application of the simulation output methods based on artificial neural networks in rock stress field analysis

本文将基于计算机仿真输出分析中的人工神经网络ＢＰ算法，应用于地下工程原岩应力场模型的辨识研究中，并将网络输出结果与某金矿实测原岩应力值进行比较，显示计算值与实测值较吻合。由此训练网络便可确定出矿区内任一给定位置处的原岩应力值，从而可建立该矿区的原岩应力场分布规律模型。

The Back Propagation algorithm (BP algorithm)of Artificial Neural Networks (ANN) for simulation output analysis of computer is applied in the model recognition of in situ rock stress field of underground project.And the results of calculation of ANN coincide with the measured data of in situ rock stress of the gold mine.Thus,the stress value of in situ rock in any position given can be determined and the model of distribution regularity of in situ rock stress field of the mining area can be established.