Research On Control Algorithm Of Permanent Magnet Synchronous Motor Based On Single Current Sensor

Permanent magnet synchronous motors are widely used in civil,aerospace and military fields due to small structure size,high power density and large moment of inertia.Real-time and accurate three-phase current feedback is the key to the control of AC servo system.Compared with the conventional driver of multiple current sensors,using single current sensor to realize phase current reconstruction can simplify the system structure,enhance reliability,and avoid the influence of different gains between multiple current sensors on sampling accuracy.However,the current reconstruction scheme has the problems of reconstruction dead zone and error,which limits its application in practical systems.In this thesis,the permanent magnet synchronous motor control system based on single dc-link current sensor is taken as the research object.The model predictive control is used to reduce the influence of two core problems in the phase current reconstruction scheme and improve the steady and dynamic performance of the system.Firstly,different installation positions of single current sensor are introduced in this thesis.By using the principle of space vector pulse width modulation,the corresponding relationship between the DC side and each branch current and phase current of the inverter under different switching states is studied.The basic principle of three-phase current reconstruction based on dc-link current sampling is analyzed.Based on the actual sampling and reconstruction situation,a detailed analysis is conducted on the dead zone of current reconstruction in each sector,as well as the time-sharing sampling error caused by current sampling at multiple different positions.At the same time,the basic principle of conventional switching state phase shift method was specifically introduced,and the existence of reconstruction dead zone and reconstruction error was verified through simulation,as well as the effectiveness of switching state phase shift method.Model predictive control is widely used due to strong robustness,fast dynamic response,and simplicity to deal with nonlinear constraint problems.In this thesis,the basic principle of model predictive current control is analyzed,and the single vector model predictive current control is studied.Aiming at the problem of poor steady-state performance,conventional dual-vector predictive current control is studied,and the improvement effect of the scheme is verified by simulation.Due to the characteristics of phase current reconstruction based on single dc-link current sensor,at least independent two-phase current information needs to be known in each switching period in model predictive control scheme.Aiming at the problem of reconstruction dead zone and current reconstruction error,a single vector model predictive control using sub-optimal replacement principle and a model predictive control using two active voltage vectors are studied.A variable vector model predictive control method is proposed.Firstly,fast vector positioning and voltage vector selection are performed by expected voltage prediction.Different vector synthesis methods are used to re-divide the linear modulation region,and the action time of the selected vector is flexibly allocated to reduce the voltage vector tracking error and complete reliable current sampling.At the same time,based on the modulation characteristics of model predictive control,on the basis of the known selected vectors and action times,the action sequence of vectors is freely allocated,and the specific sampling times are set.The two types of time-sharing sampling current errors are classified as the same error,which simplifies the current analysis and compensation process.Combined with instantaneous current compensation,the reconstruction accuracy is improved to complete the high-precision reconstruction of three-phase currents.On the existing motor experimental platform,the control performance of the designed scheme was tested and analyzed by setting different steady and dynamic operating conditions,verifying the effectiveness of the scheme described in this thesis.