Model predictive torque control for permanent magnet synchronous motor with hybrid mode duty cycle optimization

Due to the application of only one voltage vector in a control period, conventional model predictive torque control (MPTC) suffers from relatively large steady-state ripples. The-state-of-the-art double-vector-based MPTC (DVMPTC) was proposed to improve steady-state performance. However, its duty cycle optimization method significantly affects the control performances. Thus, MPTC with hybrid mode duty cycle optimization is proposed, which realizes better performance than the universal DVMPTC, and obtains the appropriate average switching frequency, especially at high-speed domain. The proposed method utilises the universal DVMPTC method, while it equips duty cycle optimization with discrete space vector modulation (DSVM) scheme as the additional constraint for further reducing the reference voltage tracking error that exists in DVMPTC. What's more, rotation module and pre-selection strategies are involved in simplifying the design of vector selection methods. Comparative simulations and experiments verify the effectiveness of the proposed method.

Automated summary

Conventional model predictive torque control (MPTC) suffers from large steady-state ripples due to the use of only one voltage vector in a control period. The state-of-the-art double-vector-based MPTC (DVMPTC) aims to improve steady-state performance, but its duty cycle optimization method significantly affects control performances. Therefore, a MPTC method with hybrid mode duty cycle optimization is proposed, which achieves better performance than the universal DVMPTC, especially in the high-speed domain. The proposed method combines the universal DVMPTC method with duty cycle optimization using discrete space vector modulation (DSVM) scheme to further reduce the reference voltage tracking error in DVMPTC. Furthermore, rotation module and pre-selection strategies are introduced to simplify the design of vector selection methods. Comparative simulations and experiments validate the effectiveness of the proposed method.