Multiobjective Design Optimization Using Dual-Level Response Surface Methodology and Booth's Algorithm for Permanent Magnet Synchronous Generators

This paper studies a dual-level response surface methodology (DRSM) coupled with Booth's algorithm using a simulated annealing (BA-SA) method as a multiobjective technique for parametric modeling and machine design optimization for the first time. The aim of the research is for power maximizati...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on energy conversion 2018-06, Vol.33 (2), p.652-659
Hauptverfasser: Asef, Pedram, Perpina, Ramon Bargallo, Barzegaran, M. R., Lapthorn, Andrew, Mewes, Daniela
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper studies a dual-level response surface methodology (DRSM) coupled with Booth's algorithm using a simulated annealing (BA-SA) method as a multiobjective technique for parametric modeling and machine design optimization for the first time. The aim of the research is for power maximization and cost of manufacture minimization resulting in a highly optimized wind generator to improve small power generation performance. The DRSM is employed to determine the best set of design parameters for power maximization in a surface-mounted permanent magnet synchronous generator with an exterior-rotor topology. Additionally, the BA-SA method is investigated to minimize material cost while keeping the volume constant. DRSM by different design functions including mixed resolution robust design, full factorial design, central composite design, and box-behnken design are applied to optimize the power performance resulting in very small errors. An analysis of the variance via multilevel RSM plots is used to check the adequacy of fit in the design region and determines the parameter settings to manufacture a high-quality wind generator. The analytical and numerical calculations have been experimentally verified and have successfully validated the theoretical and multiobjective optimization design methods presented.
ISSN:0885-8969
1558-0059
1558-0059
DOI:10.1109/TEC.2017.2777397