Research On Parameter Identification Strategy Of Permanent Magnet Synchronous Motor

Permanent magnet synchronous motors(PMSMs)have made great breakthroughs in various fields such as transportation,aerospace and military in recent years due to their competitive advantages such as high torque density and good reliability.Since the difference between the real parameter values of the motor and the actual parameter values during the system operation will lead to poor and unstable system performance,the high performance control of PMSM needs to accurately grasp the parameter changes.In order to solve the above problems,this paper investigates the online parameter identification strategy of permanent magnet synchronous motors with and without position sensor control.Firstly,this paper realizes the online identification of the resistance and d-axis inductance or q-axis inductance and magnetic chain of the motor by using the recursive least squares method under the control of a position sensor.For the problem that the given voltage of the inverter does not match the actual voltage due to the dead zone effect,this paper completely counteracts the influence of the dead zone effect on the parameter identification by feed-forward compensation of the given voltage.For the problem that the current and voltage sampled in the identification algorithm are not the same moment due to the system delay in the digital control system,this paper analyzes the relationship between the current and voltage sampling moments to compensate for the delay caused by the zero-order holding link of the inverter,so as to offset the influence of this delay on the parameter identification.For the current sampling method of deadband compensation caused by the error of the amount of deadband compensation,this paper analyzes the mechanism of the double sampling and double updating algorithm to get the size of the deadband compensation at this time,in order to feed-forward compensation of the reference voltage to avoid the impact of the current sampling method of parameter identification.Secondly,to address the problem of under-ranking in multi-parameter identification under position sensor-less control,this paper theoretically analyzes the observability of motor parameter identification under position sensor-less control according to the observability determination theory,and realizes the online identification of cross-axis inductance using the recursive least squares method.For the problem that the original parameter identification model is no longer applicable when there is an error in the estimated position,this paper re-derives the equations of parameter identification when there is an error in the position estimation,and theoretically analyzes the observability of motor parameter identification at this time.Finally,simulations and experiments are carried out in MATLAB/Simulink simulation platform and 2.2k W permanent magnet synchronous motor experimental platform to verify the compensation method of non-ideal factors and the theoretical analysis of the system observability when there is an error in estimating the position under the control of no position sensor.The simulation and experimental results show that the online parameter identification strategy under position sensor control can completely avoid the dead zone effect,the control system delay and the current sampling method’s influence on the parameter identification,and verify the correctness of the theoretical analysis of the system sizability when there is an error in estimating the position under position sensor-free control.