基于改进型平方根UKF算法的永磁同步电机状态估计  

State estimation of permanent magnet synchronous motor using modified square-root UKF algorithm

作　　者：曲智勇[1] 姚郁[2] 韩俊伟[1] 

机构地区：[1]哈尔滨工业大学机电工程学院,黑龙江哈尔滨150001 [2]哈尔滨工业大学控制与仿真中心,黑龙江哈尔滨150001

出　　处：《电机与控制学报》2009年第3期452-457,共6页Electric Machines and Control

基　　金：国家科技支撑计划(2006BAF01B12-03)

摘　　要：针对永磁同步电机驱动系统的状态估计问题,提出一种改进型平方根UKF(SRUKF)的状态估计算法。为避免增加sigma点带来的计算量大问题,依据UT变换理论,采用超球体单形采样方法,使得sigma点的数量减少,从而在与SRUKF算法估计精度相当的情况下,计算量大大减少。考虑系统的非线性,采用SRUKF估计方法研究系统的状态估计问题,避免了扩展卡尔曼滤波(EKF)产生的线性化误差。同时在滤波过程中采用Cholesky和QR分解,以协方差平方根阵代替协方差阵参加迭代运算,有效地避免了滤波器的发散,提高了滤波算法的收敛速度和稳定性。仿真表明,与EKF、SRUKF估计方法相比,该方法能减少估计过程中的计算量,提高估计精度。Concerning the problem of permanent magnet synchronous motor state estimation, an estimation method based on modified square root UKF （SRUKF） is derived. To avoid the problem of significant calculation caused by increasing the amount of sigma points, based on the UT transformation, the spherical simplex sampling method was put forward. So the amount of calculation was lessened greatly with greater performance of UKF. With regard to the non-linearity of system, the SRUKF estimation method was adopted to solve the state estimation and avoid the linearization error of extended Kalman filtering（EKF）. What＇ more, to avoid the divergence of filter and raise the velocity of convergence and stability of filter algorithm, the Cholesky, QR decomposition and the covariance square root matrix instead of covariance matrix were used in the process of estimation. Simulation results show that the method can reduce the amount of calculation and raise the estimation precision in contrast to extended Kalman filtering and SRUKF.

关 键 词：永磁同步电机 SRUKF滤波 超球体单形采样 非线性估计 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]