摘要
为了提高高速列车车轮模型修正效率和降低模型修正误差,考虑不确定性影响因素,提出了一种基于DREAM和Kriging模型的贝叶斯模型修正方法,并将其应用于高速列车车轮有限元模型修正。首先,根据设计参数建立某高速列车车轮有限元模型,初始待修正参数样本,由拉丁超立方抽样产生,分析不同样本所对应的频率函数并提取小波系数。其次,以车轮待修正参数初始样本为输入,小波分解后所得小波系数为输出构建Kriging模型。最后,使用满足精度要求的Kriging模型代替高速列车车轮有限元模型进行有限元分析,并使用DREAM算法对高速列车车轮有限元模型待修正参数进行修正。结果表明,基于DREAM和Kriging模型的贝叶斯模型修正方法,可以节约大量模型修正时间,提高模型修正效率,且修正后的高速列车车轮有限元模型误差较小。
To improve the model updating efficiency of high-speed train wheel and reduce model updating errors, considering the influence of uncertainty factors, a Bayesian model updating method based on DREAM and Kriging models is proposed and applied to the finite element model updating of high-speed train wheels.Initially, a finite element model of a specific high-speed train wheel is established based on design parameters, and the Latin Hypercube Sampling method is used to design the initial sample of parameters to be updated for the wheel.The frequency functions corresponding to different samples are analyzed, and wavelet coefficients are extracted.Subsequently, a Kriging model that meets accuracy requirements is constructed with the initial sample of parameters to be updated as inputs and the wavelet coefficients as outputs.Finally, the finite element model is replaced with the constructed Kriging model for finite element analysis, and the DREAM algorithm is used to update the parameters of the high-speed train wheel finite element model to be updated.Results show that the Bayesian model updating method based on the DREAM and Kriging models can save a significant amount of model updating time, enhance model updating efficiency, and the errors in the updated high-speed train wheel finite element model are small.
作者
曹明明
彭珍瑞
CAO Mingming;PENG Zhenrui(School of Mechanical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China;Nanjing Vocational Institute of Railway Technology,Nanjing 210031,China)
出处
《兰州交通大学学报》
CAS
2024年第4期45-52,共8页
Journal of Lanzhou Jiaotong University
基金
国家自然科学基金(51768035)。