Performance evaluation of model predictive direct torque control

Abstract

This paper presents the implementation of a previously proposed model predictive direct torque control (MPDTC) algorithm for the control of three-phase induction motor drives comprising a three-level DC-link inverter. The MPDTC scheme is designed to keep the motor torque and stator flux and the inverter neutral point potential within given hysteresis bounds while reducing the switching frequency of the inverter, in comparison with the standard direct torque control (DTC) method. Here, the MPDTC algorithm is augmented with a suitable infeasibility exit strategy which is activated when the voltage vector selection problem does not admit a solution, and provides a stable and reliable operation of the drive throughout its operating range. The algorithm is embedded in the control software environment of ABB's ACS6000 medium voltage drive, and experimental results are provided which verify the advantageous features of MPDTC in terms of average inverter switching frequency reduction. More specifically, compared to classic DTC, the proposed MPDTC scheme achieves an average (over the whole operating range) reduction of the inverter switching frequency of 16.5%, but for specific operating conditions the reduction is as much as 37.4%.