Simultaneous optimization of voltage vector and duty cycle in model predictive torque control of PMSM drives

Model predictive torque control (MPTC) is attracting wide attention in the control of permanent magnet synchronous motor (PMSM) drives owing to its conceptual simplicity and flexibility to incorporate nonlinear constraints. However, the steady torque ripple is still relatively large because of the limited voltage vectors in two-level inverter. Recently the concept of duty cycle control was introduced in MPTC to achieve steady state performance improvement. However the low speed performance is still not satisfactory due to the separate processing of voltage vector selection and duty cycle. This paper proposes an improved MPTC by simultaneous optimization of voltage vector and duty cycle in the cost function. As a result, much better steady state performance in terms of torque ripple and current harmonics can be obtained compared to prior methods. Furthermore, the proposed method is less insensitive to the weighting factor variations. The presented simulation and experimental results confirm the effectiveness of the proposed scheme.