A novel algorithm for model predictive control of AC electric drives

Abstract

The paper introduces a numerical solution of the cost function in model predictive control of a voltage source inverter, considering a continuous control set. In each switching instance, it calculates a feasible voltage vector that minimizes a predefined cost function. The developed algorithm is tested on a case study represented by the flux and torque control of an induction motor drive. The investigated technique allows including all nonlinearities, cross coupling as well as constraints, existing in the drive system. The effectiveness of the proposed approach is studied for 5 kHz sampling frequency. Practical implementation of the developed technique is also a key point being discussed in this paper.