摘要
为提高转向架构架模型的修正效率和实时性,提出了一种基于Kriging模型和无迹卡尔曼滤波的模型修正方法。首先,对构架进行模态分析,引入信息熵确定模态阶数来优选频响函数频率区间。其次,构造Kriging模型,将频响函数经过小波变换并提取第4层低频系数作为Kriging模型输出,并通过改进的灰狼算法(grey wolf optimizer,GWO)确定Kriging模型相关参数值。最后,以待修正参数作为状态向量,以Kriging模型预测的小波系数和真实响应的小波系数之差的平方和作为观测函数,通过无迹卡尔曼滤波算法求解待修正参数。结果表明,所提方法对构架模型参数修正有良好的精度、效率和鲁棒性,且在0.03 s内收敛到真实值。
In order to improve the efficiency and real-time performance of bogie frame model updating,a model updating method based on the Kriging model and the unscented Kalman filter was proposed.Firstly,modal analysis on the frame was carried out.information entropy was introduced to determine the modal order to optimize the frequency range of the frequency response function.Secondly,the Kriging model was constructed,and the frequency response function was extracted through the wavelet transform.The low-frequency coefficients of the fourth layer are the outputs of the Kriging model.The relevant parameter values of the Kriging model are were by the improved gray wolf optimizer.Finally,the parameters to be updated were used as the state vector,and the square sum of the difference between the wavelet coefficients predicted by the Kriging model and the wavelet coefficients of the real response were taken as the observation function,and the parameters to be updated were solved by the unscented Kalman filter.The results show that the proposed method has good accuracy,efficiency,and robustness,and that the true value is converged within 0.03 s.
作者
赵敏龙
彭珍瑞
张亚峰
ZHAO Minlong;PENG Zhenrui;ZHANG Yafeng(School of Mechanical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2022年第4期270-277,共8页
Journal of Vibration and Shock
基金
国家自然科学基金项目(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)。
