利用神经元网络预测岩石或岩石工程的力学性态

本文将人工智能中的神经元网络引入岩石力学领域,用以预测岩石或岩石工程的力学性态。其用法类似经验公式,但其优点是影响岩体力学性态的各种描述性地质因素,均可做为变量输入,故可求得离散性较小的结果。文中附有两个实例,说明此方法的实用性。

PREDICTING MECHANICAL BEHAVIORS OF ROCK OR ROCK ENGINEERING BY NEURAL NETWORK

In this paper the neural network, a method from artificial intelligence, was applied to rock mechanics and rock engineering to predict their mechanical behaviors. This method is similar to the method of empirical formula, but the distinguishing features of it are: the amount of effecting factors input to the system is unlimited and those describtive ones such as geological factors etc. can be included as input variables in analysis. Two examples, one from rock specimens in lab and another from coal mine, were calculated for demonstration.