This is the authors' abstract. We don't add key points for this paper.
Smooth speed control is crucial for high-performance electric drives in electric aircraft applications. However, inevitable torque ripple sources, such as cogging torque and current sampling errors, can induce periodic speed oscillations. To solve this problem, the online-trained neural network (NN)-based method with arbitrary relationship approximation capability is proposed in this paper to reject various harmonic disturbances of the speed loop. As the simplest structured NN, the adaptive linear neuron (ADALINE) can rapidly approximate and compensate for periodic disturbances with known harmonic orders, which however is not always the case in real applications. Therefore, roughly regarded as inserting an additional hidden layer into the ADALINE structure, the radial basis function neural network (RBFNN) directly using rotor position as the network's input is further proposed to effectively minimize speed ripple without the requirement of knowing disturbance frequencies, which greatly enhances the system's applicability and robustness. Various comparative experiments conducted on the dSPACE MicroLabBox-based permanent magnet synchronous motor (PMSM) platform validate the effectiveness of the proposed method.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Zuo et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: