Research On Phase Current Reconstruction Technology Of Permanent Magnet Synchronous Motor Based On Model Predictive Control

In the field of motor control,reducing control system cost,decreasing system volume,and improving system reliability are currently hot research topics.For the phase current reconstruction technology of permanent magnet synchronous motor,using a single current sensor to reconstruct three-phase currents can not only reduce the cost of the control system,but more importantly,compared with three-phase current sensors,a single current sensor can significantly reduce the failure rate and improve system reliability.Therefore,using a single current sensor for three-phase current reconstruction is of great significance.However,most of the research on phase current reconstruction technology is based on space vector pulse width modulation.With the widespread application of model predictive control,it is necessary to use model predictive control strategies for three-phase current reconstruction.However,there is a significant difference between the effects of space vector pulse width modulation strategy and model predictive control strategy on the voltage vector of the inverter,so further research is needed for phase current reconstruction technology based on model predictive control.This thesis focuses on the phase current reconstruction technology research using the optimal duty cycle model predictive control as the object of study.Within one pulse width modulation cycle,the optimal duty cycle model predictive control can output an effective voltage vector and a zero voltage vector.However,using the direct current bus current sampling method will cause information loss,so only the coupled current sampling method can be used.However,there are various forms of coupled current combinations,and by using the coupled sampling branch selection algorithm to search for the optimal combinationG_1-8=（1,0,0,0,0,1,-1,0）,the three-phase current reconstruction can be completed without further algorithm processing in the reconfigurable area.Moreover,the circuit change for this combination is convenient and meets practical application requirements.Analyzing the phase current reconstruction scheme based on the optimal coupled current sampling reveals that the scheme has a dead zone and reconstruction error problem.A detailed analysis of the reconstruction dead zone shows that it can be divided into two parts:low modulation dead zone and high modulation dead zone.The low modulation dead zone is caused by insufficient effective voltage vector action time,while the high modulation dead zone is due to insufficient zero voltage vector action time.Through a detailed analysis of the reconstruction error,it was found that the reconstruction error mainly consists of three parts:sampling point position error,delay error between two sampling points,and phase current decoupling calculation error.Aiming at the current reconstruction dead zone of the proposed scheme,different solutions are proposed based on the characteristics of low modulation dead zone and high modulation dead zone.In the low modulation dead zone,the compensation voltage vector insertion method is adopted and its effectiveness is verified through experiments and simulations.However,this method has the problem of high switching frequency.To address the issue of high switching frequency,an improved compensation voltage vector insertion method is proposed,which not only reduces the switching frequency but also maintains a fixed switching frequency when the reference voltage vector completely enters the low modulation dead zone.In the high modulation dead zone,the phase shifting approximation method and the zero voltage vector insertion approximation method are adopted.The phase shifting approximation method can ensure that the switching frequency remains basically unchanged before and after adjustment,but it will increase the current harmonic content.The zero voltage vector insertion approximation method has a better effect on suppressing the current harmonic content,and using a fixed sampling time can simplify the algorithm complexity,but it will increase the switching frequency.Both methods have their own advantages and disadvantages.The phase shifting approximation method can be applied to high-power scenarios that are sensitive to switching losses,while the zero voltage vector insertion approximation method can be applied to high-precision scenarios that are sensitive to current harmonic content.Aiming at the current reconstruction error of the proposed scheme,an incremental current prediction model is proposed to compensate for the current error by predicting the current value at the end of the cycle.According to experimental and simulation analysis,this approach can reduce the current reconstruction error to a low level,confirming its feasibility and effectiveness.