A simplified multi-mode model predictive current control method for a novel neutral-point-connected open-end winding induction motor driven system

This paper presents a simplified multi-mode model predictive current control (MPCC) scheme for a novel NPC-OEWIM powered by dual three-phase four-leg inverters with a 2:1 DC-link voltage ratio. Conventional MPCC requires evaluating 91 effective voltage vectors (VVs), resulting in high computational complexity. To address this, a cascaded split-sequence strategy is proposed that projects the 3D VVs onto the αβ plane, selects candidate VVs for the first inverter, refines the candidate set for the second inverter, and finally chooses the optimal VVs based on zero-sequence components. This method reduces predictive iterations from 91 to 23 – a 75 % reduction – and eliminates the requirement for weight coefficients. Additionally, a multi-mode switching strategy is integrated to reduce the switching frequency over a wide speed range: under low-speed conditions, only the low-voltage side inverter is active; during medium-speed operation, the high-voltage side inverter is engaged; and at high speeds, both inverters operate concurrently. Experimental results demonstrate significant computational efficiency and robust performance, making the proposed MPCC be well suited for high-performance industrial applications.