As the single-vector model predictive control (MPC) for permanent magnet synchronous motor (PMSM) has large current ripples, an improved MPC strategy for PMSM with a spatial geometry-based cost function is proposed to reduce the current ripples. First, a new spatial geometry-based objective function is presented, which is aiming at reducing the geometrical distance between the given current vector and the calculated actual one. Next, 12 adjustable virtual voltage vectors are constructed based on the 8 basic voltage vectors with the assumption that the duration time of each basic voltage vector is inversely proportional to its corresponding cost function value. By using these adjustable virtual voltage vectors, the current ripples are reduced accordingly. Additionally, a new voltage vector preselection method is also proposed based on the defined virtual voltage vectors aiming at reducing the calculation amount. Moreover, a detailed theoretical analysis is carried out based on the spatial geometry of the voltage vectors as well as the presented cost function, and the effectiveness of this strategy is validated theoretically. Contrastive simulation analyses verified the effectiveness of this strategy.

Model predictive control (MPC),permanent magnet synchronous motor (PMSM),theoretical analysis,effectiveness validation