基于改进中心差分卡尔曼算法的滞回模型参数识别OA
Identification of hysteresis model parameters based on improved central difference Kalman filter algorithm
在动荷载作用下土木工程结构会表现出显著的非线性滞回特性,中心差分卡尔曼滤波(CDKF)算法能够较好地进行系统状态和模型参数的识别,传统的CDKF算法通过乔里斯基分解(Cholupdate)来求解协方差矩阵的平方根,必须确保协方差矩阵为正定,可能导致递推过程中断.为此本文提出了一种改进中心差分卡尔曼滤波算法,该方法使用QR分解替代Cholupdate,克服了协方差矩阵必须正定严格要求,使递归计算过程无条件数学稳定.对不同噪声条件下的Bouc-Wen模型系统进行数值计算,结果表明,改进的中心差分卡尔曼滤波算法能准确识别滞回系统的状态和模型参数,并有较强的稳定性和抗噪性.
Under the action of dynamic loads,civil engineering structures exhibit significant nonlinear hysteresis characteristics.The Cenral Difference Kalman Filter(CDKF)algorithm can effectively identify system states and model parameters.Traditional CDKF algorithms use Cholesky decomposition(Cholupdate)to solve the square root of the covariance matrix,which must ensure that the covariance matrix is positive definite and may cause interruption in the recursive process.This paper proposes an improved center difference Kalman algorithm,which uses QR decomposition instead of Cholupdate to overcome the strict requirement that the covariance matrix must be positive definite,making the recursive calculation process unconditionally mathematically stable.Numerical calculations were performed on the Bouc Wen model system under different noise conditions,and the results showed that the improved central difference Kalman filter algorithm can accurately identify the state and model parameters of the hysteresis system,and has strong stability and noise resistance.
王向平
中国铁建昆仑投资集团有限公司,成都 610000
建筑与水利
Bouc-Wen滞回模型中心差分卡尔曼算法QR分解参数识别
Bouc-Wen hysteresis modelcentral differential Kalman algorithmQR decompositionparameter identification
《计算力学学报》 2026 (1)
145-150,6
重庆市建设科技计划(城科字2024第5-4号)资助项目.
评论