The Control Strategy Research Of Permanent Magnet Synchronous Motor Used In Artillery Charge System

Low speed high torque permanent magnet synchronous motor has many advantagessuch as the high control precision, large output torque, which has been used in aerospaceand power plant. In order to improve the performance of this motor control system, weneed in-depth study of the control system and promote its application in the field of warindustrial, which has far-reaching significance.Based on the introduction of permanent magnet synchronous motor control systemresearch and combined with shells propellant characteristics, this paper selects appropriatecontrol strategy, write DSP control system program and propose two senseless controlstrategies using neural network.Firstly permanent magnet synchronous motor mathematical model is established, thispaper introduce the basic principle of coordinate transformation, space vector pulse widthmodulation technology, determine the current control program which is id0, anddetermine PI regulation as the regulator of current and speed.Secondly, space vector pulse width modulation module and coordinate transformationmodule are established in the matlab/simulink, erect permanent magnet synchronous motorsimulation module which is id0. Using this module, load step change and speed stepchange experiments are did. The feasibility and veracity of the dual-loop control system istest by analysis the simulation wave.Thirdly, by the digital signal processor TMS320F2812as the core, main program,PWM interrupts subroutine; SVPWM subroutine, rotor position and speed collectsubroutine are programmed by C language.Finally this chapter presents two senseless control strategy based on neural networkfor permanent magnet synchronous motor. The first method is using network to makemagnet synchronous motor speed observer and designed neural network current observerto solve the online training problem of speed observer. This observer uses the currentforecast error to complete the network online training problem. Because the first method is difficult to stability analysis, this paper design a network online adaptive learning rulesusing lyapunov theory and the state observer based on this method is stable. Simulationstudy on the basis of theoretical analysis of the two methods, the results demonstrates thatthe proposed two observers have high accuracy on the estimation speed and position forpermanent magnet synchronous motor. After the initial transient process, the speed canquickly converge to the actual value. This method has a strong robustness to parametervariations and load torque disturbance.