Torque ripple minimization in PM synchronous motor using iterative learning control

Abstract

Permanent magnet synchronous motor (PMSM) drives are increasingly used in highperformance motion control applications where smooth torque is an essential requirement. However, presence of the air gap flux harmonics gives rise to undesirable torque pulsations. These torque pulsations result in performance deterioration in high-performance drive applications. In this paper, we present a systematic way to design and implement a new instantaneous torque controller using discrete iterative learning control (ILC) scheme with the objective of achieving torque ripple minimization. In the proposed scheme, the ILCbased dynamic torque controller compares the desired and the instantaneous motor torques and generates the reference q-alris current (iqs-,#) that is required to produce ripple-free torque. The instantaneous motor torque is estimated by using an estimator based on MRAS technique. The proposed ILC torque controller is simple to implement from computational point of view. The effectiveness of the proposed torque controller is demonstrated through computer simulation studies and results obtained verify the effectiveness of the proposed scheme.

A discrete-time current controller for permanent magnet synchronous motor drives

“…A repetitive controller is designed in discrete-time for torque ripple minimization of PMSMs with harmonics on back emf in [5]. Considering parasitic torque pulsations caused by nonsinusoidal flux density distribution around the air-gap, iterative learning control structures are also applied [6], [7]. A review paper of iterative learning controller schemes is presented in [8].…”

Review of torque ripple minimization in PM synchronous motor drives

“…al. [18] and [19] have proposed the use of ILC to minimize the torque ripples in PMSM drive and have used the time domain ILC in their implementation. The time-domain ILC has the problem of getting unstable unless the learning mechanism is turned-off after initial learning.…”

Torque Ripple Minimization in Interior PM Machines using FEM and Multiple Reference Frames

“…On the other hand, in this torque control schemes, the parameter estimation and the state observation require heavy data processing and cannot be easily implemented. An alternative technique is to use a torque transducer with high bandwidth output to measure the instantaneous torque signal [7], which can then be applied as the feedback information. This, however, increases the cost of drive system.…”