Approach for classifying continuous control set-predictive controllers applied in AC motor drives

Abstract

In this article, basic concepts, operating principles and important characteristics of continuous control set-predictive controllers (CCS-PCs) applied to AC motors are explained in detail. Based on the existence or absence of cost function as well as the method used to find the optimal control action, CCS-PCs can be categorised into the following three categories: predictive controllers without cost function, predictive controllers with a cost function (or model predictive controllers) and deadbeat controllers. To identify the advantages and disadvantages of each category, one of the recent algorithms of each category is selected and implemented experimentally. Then, various experimental tests are conducted where performances of all the algorithms are assessed in a comprehensive manner by evaluating stator-flux ripple, torque ripple, stator current harmonics, robustness against parameter changes, computational complexity, memory requirement and torque dynamic response. To achieve general and meaningful conclusions, the analysis is performed on a traditional three-phase two-level voltage source inverter used to control an interior permanent-magnet synchronous motor. Based on the theoretical discussions and experimental results, it is indicated that which category of CCS-PCs can be adopted in high performance AC motor drives.