基于新型滑模参数观测器的异步电动机MPC策略

Model Predictive Control Strategy Based on Novel Sliding Mode Parameter Observers for Induction Motors

异步电动机转子自感及定转子互感是不断变化的,传统的离线测试方法不能反映实时参数变化,导致基于离线参数的模型预测控制性能下降。为解决该问题,提出了一种基于滑模转子电感与定转子互感观测器的模型预测控制,抑制由参数失配带来的控制性能下降。创新性体现在利用M轴电流模型及机械模型构建滑模参数观测器,并将实时参数用于模型预测控制计算中。考虑到滑模观测器的抖振问题,采用一种新型双曲正切函数作为滑模观测器的开关函数,避免使用低通滤波器所带来的延时效应。实验结果表明,提出的滑模观测器能够估算实时转子电感和定转子互感值,基于观测参数的模型预测控制方法具有更好的稳态性能,与传统控制方法相比,相电流谐波畸变率降低16%。

The rotor-related parameters of induction motors(IMs),such as rotor self-inductance and mutual inductance,change frequently in practice. However,the traditional offline measurement methods for these parameters cannot reflect the real-time values,leading to the fact that the performance of the model predictive control(MPC)methods based on the offline parameters is not high. In order to solve the problem,an MPC strategy based on sliding mode(SM)self-inductance and mutual inductance observers was proposed to suppress the performance degradation caused by parameter mismatch. The novelty is that the SM observers based on M-axis current model and mechanical model are established,and the observed parameters are used for predictive calculation. Considering the chattering phenomenon,a new hyperbolic tangent function was used as the switching function,and the delay caused by using low-pass filters was avoided. The experimental results show that the proposed SM observers are able to estimate the real-time rotor self-inductance and mutual inductance. Moreover,the proposed MPC controller based on parameter observers has better steady-state performance,and the total hormonic distortion(THD)of phase current is reduced by 16% compared to the traditional control strategy.