Current Ripple Minimisation in Deadbeat Parameter-Free Predictive Control of Synchronous Motor Drives

Abstract

This paper proposes a motor parameter-free predictive current control of synchronous motor drives. In the context of motor parameter-free controls, the distinctive feature proposed in this paper is the use of the discrete space vector modulation technique. The advantages of the proposed architecture are manifold. As a distinctive feature of the proposed control algorithm compared to model-based solutions, is the real-time self-adaptation capability to the installed motor. Two recursive least-square estimators update runtime an accurate non parametric model. Compared to previous parameter-free solutions, the proposed algorithm can achieve a significantly lower current harmonic distortion at the same control frequency, while keeping the switching frequency value at bay. The result is an efficient and smooth torque delivery. The paper gives the necessary design hints and analyses the steady state performance in terms of both current harmonic distortion and switching frequency, taking finite-set model-based and parameter-free schemes as benchmark. This analysis was previously missing in literature for parameter-free schemes. Experimental validation is completed by including dynamic test results.

Keywords

• Control of permanent magnet synchronous motor drives (PMSM)
• model predictive control (MPC)
• model-free control
• data-driven control