论文部分内容阅读
提出并探讨了基于神经网络本构模型的智能有限元的算法及其实现问题 ,主要有直接法和间接法两种嵌入方式以及神经网络材料模型表达和训练的方法 ,自向前算法和基于修正拉格朗日列式法的算法迭代格式 ;并就智能有限元实现时采取的方法及编程语言等进行了讨论。
An intelligent finite element algorithm based on neural network constitutive model and its implementation are proposed and discussed. There are two kinds of direct method and indirect method, as well as neural network material model expression and training methods. Since the forward algorithm and based on the amendment Lagrange method algorithm algorithm iterative format; and the realization of intelligent finite element method and programming languages are discussed.