Discrete-Time Modeling of PMSM for Parametric Estimation and Model Predictive Control Tasks

Abstract

This paper presents novel implicit and explicit discrete-time permanent magnet synchronous motor models. Both derived models solve the problem of numerical instability and poor precision of motor currents' discrete-time prototypes formed using the forward Euler method and preserve the tolerable complexity of resulting descriptions. Discrete-time models of currents are derived based on the linear time-varying systems approach, considering the electrical angular velocity time-varying parameter. Angular velocity and angle are discretized by using the linear multistep methods. The implicit variant of the model is dedicated to parametric estimation tasks, and the explicit variant is to model predictive control. The derived descriptions are validated within the simulation by comparing the original continuous-time model and Euler approximation with the explicit model. Furthermore, the prediction capabilities of the explicit model and Euler approximation are compared as well.This paper presents novel implicit and explicit discrete-time permanent magnet synchronous motor models. Both derived models solve the problem of numerical instability and poor precision of motor currents' discrete-time prototypes formed using the forward Euler method and preserve the tolerable complexity of resulting descriptions. Discrete-time models of currents are derived based on the linear time-varying systems approach, considering the electrical angular velocity time-varying parameter. Angular velocity and angle are discretized by using the linear multistep methods. The implicit variant of the model is dedicated to parametric estimation tasks, and the explicit variant is to model predictive control. The derived descriptions are validated within the simulation by comparing the original continuous-time model and Euler approximation with the explicit model. Furthermore, the prediction capabilities of the explicit model and Euler approximation are compared as well.