Parameter Recognition And Sensorless Control Of Permanent Magnet Synchronous Motor

The permanent magnet synchronous motor has a strong application prospect in the all-electric aircraft drive system.In the operating condition,the motor pa-rameters are susceptible to change due to various factors such as temperature,which leads to the degradation of the control system performance.In addition,the motor system also requires accurate motor RPM and the location data,therefore,this paper designs an improved mayfly algorithm for motor parameter identifica-tion and improves the sliding mode observer to achieve better control performance of PMSM:Firstly,the design structure of PMSM was analyzed theoretically,mathemat-ical models were established in three coordinate systems,and i_d~*=0 vector con-trol was selected to build the PMSM control system in Matlab/simulink.In the controller,the PI parameters are rectified using the sparrow search algorithm,and simulation experiments are conducted using the rectified parameters.Secondly,to address the problem of PMSM parameter identification accu-racy,we design an improved parameter identification scheme for the mayfly op-timization algorithm.The disadvantages of the mayfly algorithm are analyzed,and three improvement schemes are designed by designing a new velocity update mechanism and introducing inertia weights into the position update equation of the mayfly algorithm.Simulation experiments on the iterative optimization per-formance of the three improvement schemes are carried out by means of standard test functions.After analyzing the mathematical principle of parameter identifica-tion,a full-rank fourth-order parameter identification model is established,and parameter identification simulation experiments are conducted using the im-proved mayfly algorithm and the other two algorithms.It is confirmed that the revised mayfly algorithm converges faster than the other two algorithms and can identify the motor parameters more accurately.Finally,for the dithering problem of the conventional sliding mode observer,a segmentation function is set instead of the symbolic function as well as the introduction of an integral sliding mode surface,which is combination with the superhelix convergence rate,is improved.On this foundation,a proportional in-tegration module is added to further jitter weakening of the inverse electric poten-tial signal for stability analysis.The optimization effect of the two improvement schemes is verified through virtual simulations in Matlab/simulink.