摘要
针对使用频响函数进行有限元模型修正存在的时间成本和精度问题,结合模态参与变异系数法和模态动能法分别优化激励点和测点,使获得的模态信息更完整;然后,引入径向基函数(RBF)模型减少原有限元模型计算次数,并根据均方根误差准则对所构建代理模型的参数(spread)进行优选,提高模型预测精度;最后,选用一个36自由度的二维桁架模型进行可行性验证,对比有限元法、Kriging模型及二阶响应面模型的修正精度和迭代时间,结果表明,本文方法具有较好的优势。
In order to solve the problem of time cost and accuracy of finite element model updating by using frequency response function,the incentive point and measurement point are optimized by coefficient of variance of mode participation criterion and modal kinetic energy method,respectively,so that the modal information is more complete.Then introducing radial basis function(RBF)model to reduce the calculation times of the original finite element model,and the parameter(spread)of the surrogate model can be optimized according to the root mean square error criterion to improve the prediction accuracy of the model.Finally,a 36-degree-of-freedom two-dimensional truss model is selected for feasibility verification.By comparing with the updating accuracy and iteration time of the finite element method,Kriging model and the second-order response surface model,the results show that the proposed method has advantageous.
作者
殷红
崔攀
彭珍瑞
YIN Hong;CUI Pan;PENG Zhen-rui(School of Mechanical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《计算力学学报》
EI
CAS
CSCD
北大核心
2020年第4期417-423,共7页
Chinese Journal of Computational Mechanics
基金
国家自然科学基金(51768035)
甘肃省高校协同创新团队项目(2018C-12)
兰州市人才创新创业项目(2017-RC-66)资助项目。
关键词
模型修正
频响函数
径向基函数(RBF)
模态动能法
模态参与变异系数法
model updating
frequency response function
radial basis function(RBF)
modal kinetic energy
coefficient of variance of mode participation criterion
