Optimization of Stator Slot Parameters for Electromagnetic Vibration Reduction of Permanent Magnet Synchronous Motors

• Published in: IEEE transactions on transportation electrification Vol. 8; no. 4; pp. 4337 - 4347
• Main Authors: Xing, Zezhi, Wang, Xiuhe, Zhao, Wenliang
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.12.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Air gaps; Algorithms; Artificial neural networks; Coils (windings); Deep neural network (DNN); electromagnetic force waves; Electromagnetic forces; electromagnetic vibration; Electromagnetics; immune clone algorithm (ICA); Mathematical models; Model accuracy; natural frequency; Optimization; Orthogonality; Parameters; Permanent magnet motors; Permanent magnets; Prototype tests; Resonant frequencies; Stator windings; Stators; Synchronous motors; Vibration analysis; Vibration control; Vibrations
• ISSN: 2332-7782; 2577-4212; 2332-7782
• DOI: 10.1109/TTE.2022.3183074

Slight electromagnetic vibration is one of the crucial indicators that must be considered in the motor design stage. Considering the influence of complex stator slot structure and inevitable saturation at slot opening, a permeance calculation model suitable for the motor with any pole-slot combination is introduced, and the electromagnetic force waves of the motor are calculated quickly and accurately. Considering the orthogonality of stator material and the influence of enclosure and windings, the natural frequencies of the entire stator are analyzed, and the vibration accelerations are calculated by the multiphysics model including control, electromagnetic, structural, and vibration models, the accuracy of the calculation results is verified by the prototype test. Furthermore, the components of electromagnetic force waves that have great impact on electromagnetic vibration are summarized, a weakening method of electromagnetic vibration based on the optimization of the stator slot parameters is proposed, and the optimal stator slot parameters are obtained by the simplified model and the accurate model combining deep neural network (DNN) and immune clone algorithm (ICA), respectively. For the 6-pole 36-slot motor, the vibration accelerations near the six-order natural frequency are greatly reduced by optimizing the stator slot parameters based on the accurate model combining DNN and ICA, and the maximum vibration accelerations are reduced by 34.1% at no load and 36.8% at rated load.