Improvement of Current Control Performance and Stabilization of Neutral-Point Voltage in Parallel-Connected Back-to-Back Three-Level Inverter with a Small Sharing Inductor

Abstract

The parallel connection of three-level inverters is a good way to increase power rating and system reliability. High efficiency and low total harmonic distortion (THD) can be achieved through the introduction of a three-level inverter. When adopting a distributed control method where each modular inverter regulates its own output current, current control performance can be improved. In addition, flexibility in increasing system capacity can be secured. However, the output current may be distorted by an uncontrollable differential-mode circulating current (DMCC) with a frequency higher than the designed control bandwidth. Additionally, it should be noted that the stability of the control system cannot be guaranteed. Furthermore, there is a potential issue in this system configuration not only zero-sequence circulating current but also fluctuation in the neutral-point voltage. Each control objective can be managed through zero-sequence voltage injection, respectively. However, the controllable degree of freedom is limited to only one, which may lead to a potential conflict between the two controllers. In this paper, a novel control strategy for back-to-back parallel three-level inverter-fed permanent-magnet synchronous machine (PMSM) drives with a small sharing inductor is proposed. This control strategy consists of two parts: the first part addresses DMCC reduction, and the second part focuses on neutral-point voltage balancing of three-level inverter. In the first part, the stability and cause of DMCC are analyzed. Based on analysis, a novel current control strategy for effectively reducing the DMCC and improving transient response is proposed. Experimental results with a back-to-back 2-parallel 2-level inverter with a small sharing inductor are presented to verify the analysis and proposed control method. In the second part, a novel method for neutral-point voltage balancing without zero-sequence voltage injection is proposed. The voltage vector reference is split into positive and negative vectors, and each split vector is optimally relocated to regulate neutral-point voltage. Simulation and experimental results are provided to verify the effectiveness of the proposed control method. Finally, to validate the overall control strategy, which combines the two proposed methods, hardware-in-the-loop simulation (HILS) testing with a real-time simulator is conducted. The real-time simulator includes a back-to-back parallel voltage source inverter implemented using Simulink. The overall control strategy is executed through a control board. The HILS results are presented to verify the effectiveness and feasibility of the overall proposed method.