Design And Implementation Of Switched Reluctance Drive System With Low Torque Ripple And Low Noise

Due to the simple structure, high reliability, wide speed range and suitable for harsh environments, switched reluctance machine(SRM) has attracted more and more attention, and has been widely used in electric vehicles, general industrial, household appliances, textile machinery and many other fields. However, its doubly salient structure and highly non-linearity of magnetic field lead to serious torque ripple and noise problems, which have constrained the application of SRM deeply. In this paper, theoretical research and experimental verification were adopted to reduce the torque ripple and noise problems of SRM.Radial force is the main cause of the SRM noise, so the analysis of radial force is the base of the research of noise. In this paper, the analytical equation was deduced based on the magnetic equivalent circuit(MEC), and obtained the static radial force curve. The natural frequency was calculated with FEM based on energy theorem. With the radial force and natural frequency obtained, the transfer function between the radial force and acceleration was deduced based on electromechanical analogy method; furthermore, the sound power level was calculated based on the acoustic fluctuation equation.According to the strong nonlinear electromagnetic characteristics of the SRM, a modeling method based on the genetic algorithm optimized wavelet neural network(GA-WNN) was proposed. The simulation model was built in Matlab/Simulink with the flux-linkage and torque characteristics measured, and the dynamic simulation was accomplished. A torque hysteresis control based on the torque sharing function was proposed aiming to reduce the torque ripple, based on which an integrated method considering both the vibration and torque ripple was proposed.The drive system was built based on the DSP TMS320F28335+CPLD, and the proposed method was tested on it, which validated the verification of the proposed method.