基于神经网络的声发射定位技术及在岩石破裂实验中的应用

用声发射（AE）定位技术确定岩石内部微破裂的发生位置，研究岩石的破坏过程，对预测岩石最终的破坏位置具有重要的实用价值。建立了误差反向传播学习算法的BP神经网络模型，在传统的定位计算方法的基础上，对BP神经网络进行训练，对原有的声发射定位数据进行模式识别，实现了用BP神经网络方法对声发射源的定位。实践表明，用神经网络算法进行定位计算具有更高的精度和效果。

Aelocation method based on neural network and the application in tests on rock failure

It is very valuable that localize the micro-crack location in rock by AE location method to investigate the process of rock failure and predict the final failure position.The BP neural net work of learning method with reverse transmission of error was founded.Based on conventional location algorithm,BP neural network was trained,and let's it identified the former AE location data.The AE source location and rock failure position were successfully and accurately localized by BP neural network.