Model Predictive Control of PMSM Drives with Reduced DC-Link Capacitance

The permanent magnet synchronous motor (PMSM) drive system with reduced dc-link capacitance will have the problem of instability of the drive system due to the small damping of the LC filter and the negative impedance characteristics of the constant power load. The input LC filter can play a role in suppressing interference and is essential in the normal operation of the drive, but it may not match the impedance of the post-stage inverter and cause the drive system to become unstable. In this paper, by constructing the dc-link voltage predictive model, combined with the PMSM current predictive model, the weighted cost function of the model predictive control (MPC) strategy is used to realize the motor optimization control and the dc-link voltage fluctuation suppression control. To avoid the introduction of additional sensor when constructing the dc-link voltage model, a source state observer that can estimate the dc-link voltage and input current is constructed. The experimental results show that the proposed control strategy can effectively reduce the dc-link voltage ripple and ensure the stable operation of the system.