Thrust Ripple Suppression Based on Negative Current Control for Short-Primary Low-Speed Large LIM Under Transient Operation

Abstract

End effect of linear induction motor (LIM) leads to magnetic asymmetry and thrust ripple, which seriously restricts the performance and application of LIM. Existing methods for thrust ripple modeling and compensating have mostly focused on the steady state, which suffer from poor transient performance. Moreover, the influence of asymmetric currents on output thrust is not considered. In this paper, the thrust ripple characteristics and impact factors of a low-speed large LIM with negligible dynamic longitudinal end effect are analyzed in details, firstly. Then, a data-driven extracting and modeling method for transient thrust ripple is proposed. The non-stationary time-varying thrust ripple signal is extracted through wavelet decomposition and reconstruction, the polynomial parameters of thrust ripple model are trained off-line by particle swarm optimization algorithm. Finally, an improved suppressing scheme for transient thrust ripple is proposed, considering the decoupling and feedback control of positive and negative currents, based on the dynamic model of six-phase LIM. Full-scale experimental results have fully demonstrated the accuracy of proposed data-driven transient thrust ripple model is more than 90%. During the high-thrust operating stage, the axial vibration acceleration level of LIM within 100Hz is reduced by more than 12dB, with the adoption of improved control scheme.