用反向传播神经网络求解J积分

A METHOD FOR CALCULATING J-INTEGRAL THROUGH NEURAL NETWORKS

提出了一种基于反向传播神经网络 (Back Propagation Neural Network)建立专家系统求解 J积分的方法。该方法以构件或材料的性能和形状参数、裂纹的几何参数为神经网络专家系统的系统输入 ,利用神经网络所具有的自学习、自组织能力 ,来实现系统输入与输出参数之间的非线性映射。通过应用实例表明 ,该方法可以用来有效求解

A method for establishing the specialist system to calculate J-integral through back propagation neural networks is proposed in the paper. The property and shape parameters of a component or material as well as the shape parameter of a crack are taken as inputs of the neural network specialist system. The non-linear mapping between the inputs and outputs can be realized by the self-learning and self-organizing ability. Examples are given to show the effectiveness of the proposed method.