利用BP网络预测结合面基础特性参数

The Prediction of the Basic Characteristic Parameters of Joint Surfaces by BP Neural Network

分析了影响结合面基础特性参数的众多因素，首次提出利用ＢＰ网络建立结合面基础特性参数与其诸多影响因素之间的非线性关系，并对其进行正确预测的有效方法，从理论上证明了结合面基础特性参数与其诸连续变化影响因素之间的映射关系可由一个三层神经网模型来精确实现。给出了实例分析。

This paper has analysed many factors influencing the joint surfaces′ basic characteristic parameters, and presented a correct method to develop the nonlinear relationship between the joint surfaces′basic characteristic parameters and their influencing factors, and then predicted their values under different contact conditions by BP neural network .It has theoritically proved that the relationship can be precisely reialized by a neural network′s model having three layers.The example has proved it′s validity.