| 基于神经网络的复杂结构映射建模和优化方法 |
| |
| 引用本文： | 徐格宁,高有山,刘芳明.基于神经网络的复杂结构映射建模和优化方法[J].起重运输机械,2005(5):10-13. |
| |
| 作者姓名： | 徐格宁 高有山 刘芳明 |
| |
| 作者单位： | 1. 太原科技大学
2. 招商局重工(深圳)有限公司 |
| |
| 基金项目： | 山西省自然科学基金，山西省教育厅重点学科资项目 |
| |
| 摘 要： | 运用结构力学方法建立易于解析的大型结构平面模型；采用商业有限元软件建立能精确反映结构真实受力的空间模型。求解得到2种模型各单元内力，将平面模型单元内力作为输入，空间模型的内力作为输出而构造样本集。利用神经网络的非线性映射能力，通过样本训练提取其特征和本质联系，以获得具有内插和泛化能力的映射模型，为后续优化设计奠定了基础。
|
| 关 键 词： | 神经网络 优化方法 结构映射 非线性映射能力 建模 平面模型 空间模型 有限元软件 大型结构 力学方法 样本训练 映射模型 泛化能力 优化设计 内力 样本集 特征和 单元 |
| 修稿时间： | 2005年1月5日 |
Mapping modeling and optimal design of complex structure based artificial neural network |
| |
| XU Ge-ning.Mapping modeling and optimal design of complex structure based artificial neural network[J].Hoisting and Conveying Machinery,2005(5):10-13. |
| |
| Authors: | XU Ge-ning |
| |
| Abstract: | Plane and space model of complex structure are built respectively by structure mechanics and commerce FEA software.The former is simple,but could not well reflect the stress in the structure;the latter is complicated but could perfectly reflect the stress in the structure.The two models are solved and internal force of each element is obtained separately.Taking internal force of plane model as input and that of space model as output,specimen collection is established.Essential character of the two models and relationship between them are extracted through training specimen with nonlinear mapping capability of Artificial Neural Network,and thus mapping model with interpolating and extending power is obtained,providing a basis for the following design. |
| |
| Keywords: | plane model space model mapping model artificial neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
