Framework of nacelle inverse design method based on improved generative adversarial networks

Wang, C; Wang, S; Wang, L; Cao, C; Sun, G; Li, C; Yang, Y

Framework of nacelle inverse design method based on improved generative adversarial networks

Wang, C Wang, S Wang, L Cao, C Sun, G Li, C Yang, Y

Permalink

Publisher:: ELSEVIER FRANCE-EDITIONS SCIENTIFIQUES MEDICALES ELSEVIER
Publication Type:: Journal Article
Citation:: Aerospace Science and Technology, 2022, 121
Issue Date:: 2022-02-01

Closed Access

	Filename	Description	Size
	Framework of nacelle inverse design method based on improved generative adversarial networks.pdf	Published version	4.21 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Wang, C
dc.contributor.author	Wang, S
dc.contributor.author	Wang, L
dc.contributor.author	Cao, C
dc.contributor.author	Sun, G
dc.contributor.author	Li, C
dc.contributor.author	Yang, Y https://orcid.org/0000-0002-0512-880X
dc.date.accessioned	2023-04-10T23:43:11Z
dc.date.available	2023-04-10T23:43:11Z
dc.date.issued	2022-02-01
dc.identifier.citation	Aerospace Science and Technology, 2022, 121
dc.identifier.issn	1270-9638
dc.identifier.issn	1626-3219
dc.identifier.uri	http://hdl.handle.net/10453/169439
dc.description.abstract	Nowadays, inverse design has been widely used in aerodynamic design. However, the traditional inverse design method is limited to ensure the optimality, accuracy, and reality of the target distribution, which restricts the further development of the inverse design. To address these issues, the framework of nacelle inverse design based on improved Generative Adversarial Networks (GAN) is firstly proposed in this paper. The pressure distribution database obtained by CFD is used for training the GAN model, and the well-trained generator producing a large number of samples is employed for searching the optimal pressure distribution. Especially, in order to enhance the optimization ability of the traditional GAN model, an improved GAN model is proposed by modifying the loss function of the generator and discriminator. Pressure distribution-aerodynamic parameters ANN model and pressure distribution-geometry ANN model are established to evaluate the aerodynamic performance of the target distribution and obtain the nacelle geometry corresponding to the target distribution. The inverse design of the nacelle profile is carried out using the traditional GAN model and the improved GAN model and applied for the spin-forming and non-spin-forming nacelles, respectively. The results show that the inverse design framework with the improved GAN model achieves better optimal results.
dc.language	English
dc.publisher	ELSEVIER FRANCE-EDITIONS SCIENTIFIQUES MEDICALES ELSEVIER
dc.relation.ispartof	Aerospace Science and Technology
dc.relation.isbasedon	10.1016/j.ast.2022.107365
dc.rights	info:eu-repo/semantics/closedAccess
dc.subject	0901 Aerospace Engineering, 0913 Mechanical Engineering
dc.subject.classification	Aerospace & Aeronautics
dc.title	Framework of nacelle inverse design method based on improved generative adversarial networks
dc.type	Journal Article
utslib.citation.volume	121
utslib.for	0901 Aerospace Engineering
utslib.for	0913 Mechanical Engineering
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
utslib.copyright.status	closed_access	*
dc.date.updated	2023-04-10T23:43:09Z
pubs.publication-status	Published
pubs.volume	121

Abstract:

Nowadays, inverse design has been widely used in aerodynamic design. However, the traditional inverse design method is limited to ensure the optimality, accuracy, and reality of the target distribution, which restricts the further development of the inverse design. To address these issues, the framework of nacelle inverse design based on improved Generative Adversarial Networks (GAN) is firstly proposed in this paper. The pressure distribution database obtained by CFD is used for training the GAN model, and the well-trained generator producing a large number of samples is employed for searching the optimal pressure distribution. Especially, in order to enhance the optimization ability of the traditional GAN model, an improved GAN model is proposed by modifying the loss function of the generator and discriminator. Pressure distribution-aerodynamic parameters ANN model and pressure distribution-geometry ANN model are established to evaluate the aerodynamic performance of the target distribution and obtain the nacelle geometry corresponding to the target distribution. The inverse design of the nacelle profile is carried out using the traditional GAN model and the improved GAN model and applied for the spin-forming and non-spin-forming nacelles, respectively. The results show that the inverse design framework with the improved GAN model achieves better optimal results.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/169439