面向PMSM参数在线辨识的引入惯性因子的改进模型参考自适应方法

在线辨识永磁同步电机的参数可以实现PI控制器的实时整定，有利于电机的解耦控制以及弱磁的计算。针对模型参考自适应法在辨识永磁同步电机参数时出现的波动问题，本文引入了电机参数变化趋势的惯性因子，使参数辨识值在迭代更新时，更符合参数的变化趋势。针对永磁同步电机电气参数辨识的方程欠秩问题，本文重新设计了分步辨识的步骤，减少了定子电阻误差对永磁体磁链辨识的影响。仿真实验显示，改进后的模型参考自适应算法无噪音条件下电感、磁链、电阻辨识的绝对误差分别为改进前的30.30%、19.77%、14.12%；噪音条件下，电感辨识精度接近，磁链、电阻绝对误差分别为改进前的24.17%、37.42%，表明改进后算法辨识精度得到提升。

Improved Model Reference Adaptive Systemby Introducing Inertia Factor for PMSM Parameter Online Identification

Online identification of permanent magnet synchronous motor (PMSM) parameters can not only realize the real-time tuning of PI controller, but also be conducive to the decoupling control of the motor and the calculation of flux weakening. To solve a fluctuation problem in using model reference adaptive method to identify parameters of permanent magnet synchronous motor (PMSM), an inertia factor describing changing trend of motor parameter is introduced in this paper. The inertiafactor makes the parameter identification value according well with changing trend when the parameter is updated iteratively. To solve the degenerate-rankproblem of equation in PMSM parameter identification, this paper redesigns procedure of step-by-step identification to reduce the influence of stator resistance error on permanent magnet flux identification. Compared with the results before the improvement, simulation results show that: under noiseless condition, absolute identification errors of inductance, flux and resistance are reduced to 30.30%, 19.77% and 14.12%; under the noise condition, accuracy of the inductance identification is closed, and absolute identification errors of the flux and resistance are reduced to 24.17% and 37.42%. It implies that t he identification accuracy of thealgorithm is improved.