Study On Multi-step Model Predictive Current Control Of Permanent Magnet Synchronous Motor For Vehicle

Permanent Magnet Synchronous Motor(PMSM)has become an important part of new energy vehicles with its excellent performance.The motor control quality directly influences the performance of vehicles.With its Advanced features,in solving complex control problems with nonlinear,multiple input and multiple output(MIMO),and constraints,model predictive control(MPC)is a high-performance control strategy.For reducing current harmonic content and switching losses,multistep Finite Control Set Model Predictive Current Control(FCSMPCC)has significant advantages,but it also has the drawback of exponential growth in computational complexity.This paper focuses on the improvement of control performance and reduction of computational complexity in the PMSM control system of FCS-MPCC.Establish the FSC-MPCC system for built-in PMSM in the d-q rotational coordinate system.The harmonic content and the switching frequency are used as performance indicators to measure the comprehensive control performance.Establishing a simulation model to verify the control effect of model predictive control.A multi-step predictive model is established for FSC-MPCC of PMSM.According to the current state quantity at each control cycle,current reference values at multiple future times are calculated,and a multi-step cost function is established to solve the optimal voltage vector.According to different mathematical forms,discrete recursive expression prediction models and discrete state space prediction models are established,with current error and switching times as cost functions,and the two are weighed by weight coefficients.An exhaustive search algorithm is used to solve the optimal voltage vector for multi-step FSC-MPCC.Compare the control performance of FSC-MPCC through simulation: in the high switching frequency region,the control effect of vector control has absolute advantages over model predictive control,and the effect of multi-step MPCC is equivalent to that of single-step MPCC;In the middle switching frequency region,the control effect of vector control and multi-step MPCC is similar,and the control effect of multi-step MPCC is better than that of single-step MPCC;In the low switching frequency region,vector control is out of control,mode MPCC operates normally,and the effect of multi-step MPCC is better than that of single step MPCC.Therefore,for traditional MPCC,multi-step prediction is more suitable for medium and low switching frequency regions.In order to improve the control effect of multi-step MPCC,an LC filter is added between the PMSM and the inverter to upgrade the MPCC control system from a first order system to a third order system.The LC filter inductance parameters and capacitance parameters are determined based on the PMSM drive system.Based on the PMSM mathematical model and the LC filter mathematical model,a three-level MPCC control system is established to predict the inverter current,capacitor voltage,and motor current.The inverter current error,capacitor voltage error,motor current error,and switching times are used as cost functions,and the optimal voltage vector is solved with exhaustive search algorithm for the three-level MPCC.Compare the control performance of the third-order FSC-MPCC through simulation.At a lower switching frequency,the third-order MPCC results in model prediction out of control and performance inferior to the first-order MPCC;At higher switching frequencies,the performance of third-order MPCC improves as the number of prediction steps increases.At higher switching frequencies,the current THD of the three to five step three step MPCC are smaller than those of vector control,indicating that the three step multi step MPCC has better control effects than vector control.Therefore,three-stage multi-step MPCC has better control effect.It is necessary to adopt a more efficient optimal solution algorithm.The multi-step MPCC problem is transformed into an integer least square problem,and an improved spherical coding algorithm is used to solve the optimal voltage vector.spherical encoding solves the inverter switch state,continuously reducing the spherical radius during the solving process,based on the event triggering mechanism,the spherical encoding algorithm reduces the amount of multistep MPCC operations while ensuring the optimal solution.Compare the control effects of spherical encoding algorithm and exhaustive search algorithm.As an optimal solution algorithm,the control effect of spherical encoding algorithm is completely consistent with that of traditional exhaustive search algorithm.Based on the STM32H7 computing platform,the calculation time of the algorithm was tested,and the algorithm was run in the form of C code in the MCU to verify the simplification effect of the spherical coding algorithm on the calculation amount.FPU and DSP acceleration had a significant improvement in the calculation performance;Based on Model Based Design(MBD),using the STM32H7 computing platform to perform Model In The Loop(MIL)simulation,Software In The Loop(SIL)simulation,and Processor In The Loop(PIL)simulation for MPCC.The effectiveness of the control algorithm is verified during the entire operation process.