Dual-Vector Located Model Predictive Control With Single DC-Link Current Sensor for Permanent-Magnet Linear Motor Drives

In this article, a dual-vector located model predictive control (DL-MPC) method based on a single dc-link current sensor is proposed for the permanent-magnet linear motor drive system. In DL-MPC, the phase currents are reconstructed by sampled dc-link current and selected active voltage vector (VV). Meanwhile, the sunk number of phase current is defined to evaluate the accuracy of current reconstruction. The first-optimal and second-optimal VVs are determined by dual-vector location. The second-optimal VV is selected for the voltage source inverter when the sunk number exceeds the threshold. Compared with traditional MPC schemes, DL-MPC has smaller current tracking error and better dynamic performance. Meanwhile, DL-MPC has more accurate current reconstruction and lower computation burden than that of the traditional MPC. The effectiveness of DL-MPC is verified by both theoretical analysis and experimental results.