A Model Predictive Control Method for Three-level Permanent Magnet Synchronous Motor Considering Vector Partition

Abstract

The invention discloses a three-level permanent magnet synchronous motor model prediction control method which takes into account the vector partition. Based on the permanent magnet synchronous motor powered by the NPC three-level inverter, the value of constructing the dq-axis stator flux linkage as the control object function, the weight coefficient is eliminated, and a vector partition method is used to reduce the optimal candidate vectors of the value function to 1-3. While controlling the effect, the calculation amount of the second vector selection is reduced, and the balance of the mid-point potential is taken into account.

Description

Technical field [0001] The present invention relates to a predictive control method of permanent magnet synchronous motor, which belongs to the motor drive and control. Background technique [0002] Model Predictive Control, MPC is a computer control algorithm produced in late 1970s. In the motor drive system, it can be divided into model predictive current control (MPCC) according to the control target. Current Control (MPCC) And model predictorque control (MPTC). MPTC uses value functions online optimistic idea to obtain the optimal voltage vector output of the inverter, not only improve the dynamic response of the system, but also reduce the torque pulsation to a certain extent, improve the steady state properties of the system. However, the traditional MPTC needs to design the weight coefficient, and the design of the weight coefficient lacks a unified guiding policy. Therefore, a model predictive flux control (MPFC) is proposed, and the stator magnetic chain and electromagne...

AI Technical Summary

Problems solved by technology

However, since there are 27 basic voltage vectors in the NPC three-level inverter, 27 judgments need to be made when selecting the value function, especially in multi-step prediction and multi-vector control, the calculation amount of the system increases exponentially , a large amount of calculation will cause the processing delay of the system, resulting in inaccurate prediction

In addition, in order to reduce the torque ripple and improve the steady-state performance of the control system, the dual-vector control strategy is a good method. However, for the traditional control strategy in which the second vector is fixed as the zero vector, in some cases , the role of the zero vector may be invalid, and the second vector is not fixed as the zero vector, although it can improve the control accuracy, but there is a problem of the second vector selection, which will increase the calculation burden of the system

Images