固体力学有限元的神经计算原理

NEURAL COMPUTING THEORY ON FINITE ELEMENTANALYSIS OF SOLID MECHANICS

针对现代结构分析的特点，利用力学中的最小势能原理，把力学有限元的计算问题化为带约束的非线性优化问题。提出了基于神经网络的弹性力学有限元分析原理。

By using the mechanical minimized potential theory,the problem of FE computa-tion of mechanics is transformed to nonlinear programming with linear equality contraints. Anovel elastic mechanical FE analysis principle, based on neural networks,is proposed.