Research On Optimal Model Predictive Control Strategy Without Weighting Factor For Permanent Magnet Traction Synchronous Motor

Permanent Magnet Synchronous Machine(PMSM)has been widely applied in high-speed railroad transportation,new electric vehicles,aerospace industry,and other fields due to its advantages,such as simple structure,small structure volume,improved efficiency and high torque-current ratio.Model predictive control(MPC),as a classic control strategy in motor control,can be generally divided into model predictive current control(MPCC)and model predictive torque control(MPTC)according to different objectives.It has been widely studied because of its fast dynamic response,the possibility of multi-objective control and reliable performance.However,there are still problems such as large calculation amount,complex weighting coefficient adjustment,and large torque and flux ripples.This paper focuses on the model-based torque control strategy and optimization scheme of a permanent magnet synchronous motor.Specifically,simulation modeling and performance optimization are carried out for weight coefficient tuning,high computational complexity,and magnetic flux pulsation issues.Based on the "Model Predictive Torque Control" system modeling to "Model Predictive Flux Control",the weight coefficient is eliminated.The research roadmap includes "1: Optimizing the three-level structure combined with the elimination of weight coefficient in the model predictive flux control strategy,reducing computational complexity,improving the reaction speed of target quantity,and optimizing the steady-state effect;2: Optimizing the three-vector structure combined with the elimination of weight coefficient in the model predictive flux control strategy to optimize the steady-state effect".The main work focuses on the optimization of the model predictive control without weight coefficient,which includes the following steps:1)Firstly,a complete mathematical model of d-q-axis permanent magnet traction motor is established,and the mathematical analysis is carried out from three perspectives of permanent magnet synchronous motor,two-level inverter and coordinate transformation respectively,so as to lay a good foundation for the next step of model prediction torque and magnetic chain control model.2)Considering the weight coefficient adjustment issue in MPTC,this paper proposes a MPFC strategy.By constructing the load angle and transforming torque and flux control into a single control of stator flux based on the relationship among given torque,load angle,and reference flux target,the objective function of the reference flux target value and the predicted flux value on the d-axis and q-axis is built to eliminate weight coefficients and solve the problem of complex weight coefficient adjustment.However,the issue of large fluctuations in torque and flux is still unresolved.3)The NPC three-level MPFC strategy is an improved solution proposed to reduce computational complexity and improve steady-state performance.This paper focuses on dividing the sectors by virtual slope and optimizing vector selection to reduce computational complexity,while studying the strategy of controlling midpoint potential through small vector positive and negative differences for the problem of maintaining midpoint potential balance in NPC three-level inverter and complex model predictive calculations.4)The improvement of multi-vector mainly focuses on adjusting the optimal vector amplitude and direction to achieve arbitrary range control of vector adjustment and optimize steady-state performance.The duty-cycle strategy adds a zero vector as the second vector and uses duty cycle control to adjust the amplitude,selecting the best effective vector and zero vector combination within one cycle to construct a virtual vector and synthesize a PWM signal with a constant switching frequency.Based on duty-cycle control,this paper proposes an Optimized Three-Vector MPFC method.In each sampling period,the first optimal vector is selected through the value function,and then the effective vectors except for the optimal vector are traversed to combine with the zero vector to obtain the second vector.The duty cycle of the three vectors is calculated using the principle of no deadbeat,and the virtual vector is constructed again through the value function selection,thereby establishing a composite vector with three vector directions and adjustable amplitude within one cycle.To demonstrate the feasibility and effectiveness of the three-vector model’s predicted magnetic flux control strategy,this paper conducted simulation and experimental research on both Matlab simulation software and semi-real Starsim equipment.Simulation and experimental results show that the model-predicted magnetic flux control strategy can avoid the adjustment of weight coefficients.Both the optimal duty cycle model-predicted magnetic flux control strategy and the three-vector model-predicted magnetic flux control strategy can reduce torque and magnetic flux pulsation while maintaining good dynamic response performance of the model-predicted magnetic flux control strategy.However,the three-vector model-predicted magnetic flux control strategy is more effective in improving steady-state performance.