基于改进高斯径向基函数响应面方法的蜂窝板模型修正

Improved Gaussian RBF response surface method-based model updating for the honeycomb sandwich panel

用等效板理论计算蜂窝板等效结构参数建立基准有限元模型,对基准模型中结构参数施加摄动量建立待修正有限元模型。根据均匀设计方法将等效参数分为不同水平的参数组,分别计算各参数组对应的模态频率响应,然后建立基于一次式-高斯组合径向基函数(Linear-Gaussian RBF)的响应面模型。引入变异算子的改进粒子群算法修正响应面模型,将搜索到的参数摄动量(优化解)代入待修正模型,得到修正后模型。通过比较基准模型与修正前后模型的结构参数和模态频率响应的接近程度,证实了修正方法的有效性。基于响应面的模型修正避免了迭代中重复调用有限元计算,可提高分析效率。

The equivalent panel theory was employed to calculate the equivalent parameters of the honeycomb sandwich panel to constitute the benchmark FEM,and the perturbations were added to the equivalent parameters to constitute the non-updated FEM.The equivalent parameters are divided into different groups to get the corresponding modal frequencies,and the Linear-Gaussian RBF response surface model is set up,which is updated by the particle swarm optimization algorithm with mutation operator,the updated parameters is introduced into the non-updated model and the updated model is attained.By comparing the similarity of the structure parameters and the modal frequencies of the benchmark model and the non-updated as well as the updated models,the validity of the updated model is proved.