Heat dissipation analysis and multi-objective optimization of a permanent magnet synchronous motor using surrogate assisted method

Permanent magnet synchronous motors (PMSM) have been substantially used in electric vehicles (EVs) due to their advantages such as low loss, large torque, and high power density. With the continuous improvement of the PMSM performance requirements, its heat dissipation has also attracted increasing...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Case studies in thermal engineering Jg. 27; S. 101203
Hauptverfasser: Li, Yongsheng, Li, Congbo, Garg, Akhil, Gao, Liang, Li, Wei
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 01.10.2021
Elsevier
Schlagworte:
Electric vehicles
Multi-objective optimization algorithms
Optimization framework
Permanent magnet synchronous motors
Surrogate model
Utopian point method
Water oil mixed cooling system
Permanent magnet synchronous motors
Surrogate model
Optimization framework
Multi-objective optimization algorithms
Electric vehicles
Water oil mixed cooling system
Utopian point method
ISSN:2214-157X, 2214-157X
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Permanent magnet synchronous motors (PMSM) have been substantially used in electric vehicles (EVs) due to their advantages such as low loss, large torque, and high power density. With the continuous improvement of the PMSM performance requirements, its heat dissipation has also attracted increasing attention. This paper proposes a cooling system to realize the heat dissipation of the motor through internal oil circulation and external water circulation. Meanwhile, to obtain the best cooling system parameters, an optimization framework is developed for heat dissipation optimization of the motor. First, the most suitable Latin hypercube sampling (LHS) method is selected for sampling through the Coordinates exchange algorithm. Second, we separately study the modeling accuracy of thirteen surrogate models and finally select the back propagation (BP) neural network model. Then, we use six multi-objective optimization algorithms (MOOAs) to optimize the model, and select the optimal solution via the utopian point method. Finally, the motor heat dissipation situation is effectively improved, and the effectiveness and reliability of the optimization framework are proved, which provides an alternative mean for the heat dissipation design optimization of the motor and has prominent practical significance. [Display omitted]
ISSN:2214-157X
2214-157X
DOI:10.1016/j.csite.2021.101203