Active torque ripple reduction based on an analytical model of torque

Abstract

The vibration caused by torque ripple may degrade the performance of drivetrain, thus the reduction of torque ripple are necessary for hybrid electrical vehicle/electrical vehicle application, while an analytical description of torque will be the basis. In this study, two-dimensional Fourier series supplemented by polynomial fitting is introduced to reconstruct the numerical solution of magnetic coenergy (MCE) from finite element analysis. An analytical torque model, which can describe torque ripple precisely at all working points, is derived from the reconstructed MCE. On the basis of the torque model, the expressions of harmonic currents used to reduce torque ripple are obtained. The proposed expression also satisfies the principle of maximum torque per ampere. Then, a feedforward torque controller based on the expression of harmonic currents is introduced. Meanwhile, a feedback torque controller based on a torque observer is developed as a comparison. The results of simulation and experiments show that both two controllers can reduce the torque ripple in steady state, but the feedforward torque controller has obviously better effects when motor speed increases. Furthermore, the feedforward torque controller is also beneficial to decline torque ripple during transient process of torque response.