Generalized Deadbeat Solution for Model Predictive Control of Five-Phase PMSM Drives

Abstract

Model predictive control (MPC) has been widely emerged and applied on five-phase PMSM drives. However, the conventional MPC schemes have employed the high-complexity enumeration process and/or virtual voltage vector(s) (V <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> (s)), which are designed offline in one subspace. Thus, the traditional MPC schemes may have a large computational burden, low tracking accuracy, and/or limited DC-bus voltage utilization. This article proposes a generalized low-complexity deadbeat (DB) solution for MPC schemes of the five-phase drives, which have a double-space control problem. In the proposed solution, the double-space control problem of the five-phase PMSM drive is formulated in a linear system of equations. To have a unique solution, five candidate real vectors are defined and solved for their duties during every sample interval. The optimal voltage vector(s) are directly selected without any enumeration process according to the highest duty cycle(s), which are designed online considering the control objectives of the two-subspace components. Hence, five different MPC schemes can be obtained from this solution. Compared with traditional MPC schemes, the salient features of the proposed DB-based schemes are a low computational burden, improved steady-state performance, high DC-bus voltage utilization, and a simple generalized execution. These features are verified using a five-phase PMSM drive experimental prototype.