Research On Low Speed Performance Of PMSM Based On Improved DTC

Permanent magnet synchronous motor has been widely used in electric vehicles,rail transit,aerospace and other fields because of its high work efficiency and excellent running performance.With the development of permanent magnet synchronous motor,the new high-efficiency control strategy for permanent magnet motor has become the current research hotspot.Especially in the conventional control methods,the permanent magnet synchronous motor has problems such as unstable operation at low speed and large pulsation.It is necessary to analyze the mechanism of the problem and propose a reasonable solution.According to the problems existing in the low-speed area of permanent magnet synchronous motor based on DTC system,this paper deduces the mechanism of the large torque ripple in detail,discusses the effect of the flux control strategy on the electromagnetic torque,and analyzes the effect of changes in motor parameters on the performance of the motor.In order to solve the problems existing in the traditional DTC system at low speed,this paper uses a variable flux direct torque control strategy based on id = 0 to avoid demagnetization and saturation of the magnetic circuit,which greatly improves the output efficiency of the hidden-pole synchronous motor.The control strategy adopts a modified second-order generalized integrator to improve the flux observer.The space vector pulse with modulation is used to synthesize the desired voltage.Through MATLAB/Simulink simulation,it is verified that the improved control strategy not only keeps the traditional DTC system responding quickly but also restrains the system's pulsation problem when the motor runs at low speed.In order to ensure that the position-free sensor based on the pulse high-frequency signal injection method still maintains high position extraction accuracy when the motor runs at a low speed,a signal extraction method based on wavelet packet transform is adopted in the paper.The method uses multi-scale decomposition of the current signal containing position information by wavelet packet,selects the maximum and minimum criterion thresholds to process the wavelet packet decomposition coefficients,and then reconstructs high-frequence band signals to extract the position information.The improved pulsed high-frequency voltage signal injection method achieves that the position-free observer can still maintain efficient and accurate signal extraction under low signal-to-noise ratio,which greatly reduces the torque ripple caused by high-frequency signal injection.Because the improved DTC strategy based on id =0 control needs real-time accurate motor parameters to participate in the control calculation,this paper proposes an improved identification algorithm that combines the cloud model theorv and the particle swarm optimization algorithm.This improved algorithm uses the advantages of the cloud model theory in terms of converting between fuzzy and accurate.This method effectively solves the problems of early maturity and loss of diversity among particle in the late period of traditional PSO algorithm.In order to improve the real-time and accuracy of the identification model,the identification algorithm also uses a new motor identification mathematical model to build a fourth-order online real-time identification model.The simulation shows that the improved motor parameter identification algorithm can effectively jump out of premature convergence in different working conditions and can track the actual parameter values quickly and accurately.