Iterative Learning Based Torque Ripple Suppression of Flux-Modulation Double-Stator Machine

Due to the double layer airgap configuration, the even-order harmonic components exist in the phase back electromotive force (EMF) of flux-modulation electrically-excitation double-stator synchronous machine (FMEE-DSM), thus resulting in relatively high torque ripples. To promptly and effectively suppress these torque ripples, this paper proposes and implements a second-order iterative learning control (ILC) strategy to minimize the main harmonic torque components based on the periodically repetitive characteristic of torque ripples in the FMEE-DSM. Via harmonic analysis on the torque ripples caused by the cogging torque and the non-sinusoidal induced EMFs, the dominative harmonic component, namely, the 7th order component, can be precisely identified and then involved in the proposed ILC for torque ripple suppression. To verify the feasibility of proposed control strategy, a test bench of the FMEE-DSM control system was built based on dSPACE 1103. Accordingly, both the static and dynamic performances of the FMEE-DSM were investigated under the conditions of no-load and rated load. The theoretical analysis, computer simulation and hardware experimentation are given to verify the proposed second-order ILC.