基于递推最小二乘法的永磁伺服系统参数辨识

PMSM Parameters Identification Based on Recursive Least Square Method

为使永磁同步电机(PMSM)控制系统在复杂环境中具有较好的动态性能,伺服系统必须具有参数辨识和参数自整定的功能,而转动惯量与负载转矩辨识是其首要解决的问题。采用零阶保持器对电机运动方程进行离散化建模,考虑了摩擦系数对辨识结果的影响,将基于遗忘因子递推最小二乘辨识算法应用于该离散模型可以同时辨识出系统转动惯量、负载转矩和摩擦系数。同时,针对Matlab/Simulink中库模型参数不能在线动态修改的缺点,提出改进型PMSM模型,以此搭建了伺服系统的仿真控制模型,完成了定参数与变参数的动态仿真。最后,在stm32微控制器上进行了实验验证。仿真和实验表明该文提出的电机离散化模型和参数辨识方法具有一定的准确性和实时性,仿真结果验证了改进型PMSM模型在变参数仿真研究中的实用性。

The permanent magnet servo control system must have features o f parameter identification and parameter self-tuning to get better dynamic performance in a complex environment. Among them, the identification of inertia and load torque has the first priority. In this paper, the zero-order holder is introduced to discretize the motion equation of the motor. And the influence of the friction coefficient on the identification results is taken into consideration. The inertia, the load torque, and the friction coefficient can be identified simultaneously using the recursive least square method with a forgetting factor. In the Matlab/Simulink library model , the parameters cannot be modified on-line dynamically. So the improved PMSM model is proposed. Then the servo system simulation control model is built up. The constant and the variable parameters of dynamitic simulation are conducted. Finally, the experiments are carried out on the stm32 micro controller. The simulation and experimental studies show that the motor discrete model and the parameter identification method proposed in the paper have a certain accuracy and real-time property. The simulation results show the usefulness of the improved PMSM model in the variable parameter simulation study.