用人工神经网络方法辨识岩体力学参数

Identifying Mechanics Parameters of Rock Mass by Using Artificial Neuron Network

研究了用人工神经网络辨识岩石工程岩体力学参数的基本原理、网络结构、网络训练样本的获得及网络训练方法·在此基础上 ,提出了用人工神经网络辨识岩石工程岩体力学参数的方法和步骤 ,并给出了一个参数辨识实例·结果表明 ,依据巷道变形观测 ,用神经网络辨识岩石工程岩体力学参数是可行的·该方法为研究岩体物理力学性质参数和岩石工程初始地应力条件提供了一个有效途径 ,尤其适用于软岩情况下岩体力学参数的辨识·

The basic principle,structure of artificial neuron network, obtaining training samples and the method of training network in identifying mechanics parameters by neuron network were studied. The method and procedure to identify mechanics parameters by neuron network were put forward. An example of parameter identification was offered. It is workable to identify the mechanics parameters of rock mass by neuron network according to displacement surveying of wall rock. This method provides an effective approach to determine mechanics parameters of rock mass and original underground stress condition of rock engineering. It is especially suitable to identify mechanics parameters of rock mass under soft rock conditions.