An Improved Modulated Model Predictive Torque and Flux Control for High-Speed IPMSM Drives

This paper proposes an improved modulated model predictive torque control as an alternative of linear controller based direct torque and flux control (DTFC) for high-speed IPMSM drive. The complicated controller design and tuning in PI-regulators based DTFC are eliminated. Moreover, the transient performance is greatly improved while maintaining similar steady-state performances. It also proposes analytical solutions for machine trajectories, which combine with modulated model predictive control (M2PC) to achieve high electrical efficiency in wide speed operation. M2PC is modified to select an optimal vector combination among one-active, two-active, one-active-one-zero, or two-active-one-zero voltage vectors based on their corresponding cost function. The control precision is improved further by introducing virtual vectors (VVs) associated with real VVs in the control set, especially in high-speed operation. A two-stage optimization is then employed to reduce the computational burden added due to the extended control set. Extensive experimental results are shown to validate the effectiveness of the proposed method.