| 引用本文: | 易晓山,任均国,周建平.有限元结构分析神经网络计算.[J].国防科技大学学报,1996,18(4):27-30.[点击复制] |
| Yi Xiaoshan,Ren Junguo,Zhou Jianping.Structure Analysis By Neural Network[J].Journal of National University of Defense Technology,1996,18(4):27-30[点击复制] |
|
| |
|
|
| 本文已被：浏览 6517次 下载 6135次 |
| 有限元结构分析神经网络计算 |
| 易晓山, 任均国, 周建平 |
|
(国防科技大学 航天技术系 湖南 长沙 410073)
|
| 摘要: |
| 采用精确积分四边形单元刚度矩阵，构造一种多层反馈神经网络以实现平面问题有限元结构分析，其中间层神经元联接权值为求解问题的单元刚度矩阵。分析和数值仿真结果表明，该神经网络稳定收敛于在限元模型解，为有限元结构分析并行计算提供一种新的方法。 |
| 关键词: 神经网络，有限元，结构分析 |
| DOI： |
| 投稿日期：1996-06-17 |
| 基金项目:国家自然科学基金、国家教委优秀青年教师基金和校预研基金资助 |
|
| Structure Analysis By Neural Network |
| Yi Xiaoshan, Ren Junguo, Zhou Jianping |
|
(Department of Aerospace Technology, NUDT, Changsha, 410073)
|
| Abstract: |
| A feedback multi-layer neural network model is proposed for the structural analysis. The weight coefficients of middle-layer neural cells of the stiffness matrix given by finite element method using an exact integrated 4 node element. Numerical simulation shows that the method is consistent with FEM. This method presents a new way for the parallel structural analysis. |
| Keywords: neural network,finite element method,structure analysis |
|
|
