节点文献
基于径向基神经网络的有限元模型修正研究
Model updating based on radial basis function neural network
【摘要】 模型修正属于反问题的一种,针对其非线性、计算量大等不足之处,提出一种基于径向基神经网络的有限元模型修正方法,并把反问题归结为正问题进行研究。该方法将特征量作为自变量输入、设计参数作为因变量输出,用试验设计构造样本,以径向基神经网络逼近两者之间的非线性映射关系,利用神经网络的泛化特性直接输出设计参数的修正值。某空间钢结构模型的计算结果验证了该方法的有效性。
【Abstract】 Model updating is part of inverse problem which calls for large amount of calculation.This paper presents a new method which treats the model updating as a positive problem.Features and design parameters are regarded as independent variables and samples are designed by test design.The trained radial basis functional neural network is utilized as a map function.The target value of the design parameter can be directly estimated due to the generalization character of the neural network.A steel model is employed to verify the effectiveness of the proposed method.
【Key words】 model updating; neural network; radial basis function; finite element model;
- 【文献出处】 武汉科技大学学报 ,Journal of Wuhan University of Science and Technology , 编辑部邮箱 ,2011年02期
- 【分类号】TP183
- 【被引频次】15
- 【下载频次】190