基于STUKF的非线性结构系统时变参数识别
发布时间：2018-07-17 02:14
【摘要】：针对非线性结构系统时变参数识别问题,传统无迹卡尔曼滤波(Unscented Kalman Filter,UKF)难以有效跟踪结构参数的变化。将强跟踪滤波原理引入无迹卡尔曼滤波,提出一种强跟踪无迹卡尔曼滤波(Strong Tracking Unscented Kalman Filter,STUKF)算法,以识别结构参数的变化。在UKF量测更新后,依据输出残差计算渐消因子矩阵;引入两个渐消因子矩阵实时调整状态预测协方差矩阵,使残差序列强行正交,快速修正结构参数估计值,使STUKF具有对结构参数变化的跟踪能力;此外,为节省计算时间,调整状态预测协方差矩阵后不再进行sigma点采样,保证了算法的高效性。数值分析结果表明,该算法能有效识别非线性结构系统的参数及其变化,并具有较强的抗噪性。
[Abstract]:The traditional unscented Kalman filter (unscented Kalman filter UKF) is difficult to track the variation of structural parameters effectively for the identification of time-varying parameters of nonlinear structural systems. A strong tracking unscented Kalman filter (STUKF) algorithm is proposed to identify structural parameters by introducing the principle of strong tracking filtering into unscented Kalman filter. After the UKF measurement is updated, the fading factor matrix is calculated according to the output residuals, and two fading factor matrices are introduced to adjust the state prediction covariance matrix in real time. In addition, in order to save calculation time, the state prediction covariance matrix can not be sampled by sigma, which ensures the high efficiency of the algorithm. Numerical results show that the proposed algorithm can effectively identify the parameters and their variations of nonlinear structural systems, and has strong noise resistance.
【作者单位】： 兰州理工大学防震减灾研究所;兰州理工大学西部土木工程防灾减灾教育部工程研究中心;
【基金】：国家自然科学基金(51578274;51568041) 教育部长江学者创新团队项目(IRT13068) 甘肃省青年科技基金计划(2014GS03277)
【分类号】：TN713
，
本文编号：2128571
[Abstract]:The traditional unscented Kalman filter (unscented Kalman filter UKF) is difficult to track the variation of structural parameters effectively for the identification of time-varying parameters of nonlinear structural systems. A strong tracking unscented Kalman filter (STUKF) algorithm is proposed to identify structural parameters by introducing the principle of strong tracking filtering into unscented Kalman filter. After the UKF measurement is updated, the fading factor matrix is calculated according to the output residuals, and two fading factor matrices are introduced to adjust the state prediction covariance matrix in real time. In addition, in order to save calculation time, the state prediction covariance matrix can not be sampled by sigma, which ensures the high efficiency of the algorithm. Numerical results show that the proposed algorithm can effectively identify the parameters and their variations of nonlinear structural systems, and has strong noise resistance.
【作者单位】： 兰州理工大学防震减灾研究所;兰州理工大学西部土木工程防灾减灾教育部工程研究中心;
【基金】：国家自然科学基金(51578274;51568041) 教育部长江学者创新团队项目(IRT13068) 甘肃省青年科技基金计划(2014GS03277)
【分类号】：TN713
，
本文编号：2128571
本文链接：https://www.wllwen.com/kejilunwen/dianzigongchenglunwen/2128571.html
