{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T15:10:22Z","timestamp":1778339422167,"version":"3.51.4"},"reference-count":50,"publisher":"Oxford University Press (OUP)","issue":"4","license":[{"start":{"date-parts":[[2026,3,11]],"date-time":"2026-03-11T00:00:00Z","timestamp":1773187200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100012632","name":"Nazarbayev University","doi-asserted-by":"publisher","award":["040225FD4734"],"award-info":[{"award-number":["040225FD4734"]}],"id":[{"id":"10.13039\/501100012632","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,4,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>This work investigates the use of physics-informed geometric operators in tandem with latent variable models to support surrogate learning, dimensionality reduction, and generative design of airfoils. A baseline dataset was constructed via a NURBS-based airfoil parametric model with physically interpretable design variables, and discretized using two schemes: a uniform parametric and a uniform arc-length sampling. A hybrid Variational AutoEncoder (VAE) with the addition of convolutional layers was developed and iteratively refined to evade shape invalidities. A systematic comparison of reconstruction accuracy, robustness, and diversity showed that a loss function based on the mean sum of squared distances had the best performance with sufficient stability during model optimization. However, this can be only established when the reconstruction and Kullback\u2013Leibler terms in the $\\beta$-VAE objective function are weighted via an appropriately selected $\\beta$ value. Additionally, augmenting geometry with physics-correlated high-level descriptors, such as geometric moments, further improves latent-space quality. Among the tested operators, third-order geometric moments yielded the most consistent robustness gains. Discretization and achieved diversity proved to be linked, with uniform arc-length spacing achieving the best reconstruction accuracy, but with many near-identical designs that degraded the resulting diversity. In contrast, uniform parametric spacing exhibited higher diversity without the need for any special treatment of design distributions and diversity quantification measures. This study consolidates practical guidelines on architecture, loss-function scaling, physics-informed features, and quantification protocols for reliable, data-efficient airfoil generative design with VAEs.<\/jats:p>","DOI":"10.1093\/jcde\/qwag025","type":"journal-article","created":{"date-parts":[[2026,3,11]],"date-time":"2026-03-11T12:57:08Z","timestamp":1773233828000},"page":"51-66","source":"Crossref","is-referenced-by-count":0,"title":["Physics-informed geometric operators in generative airfoil design with hybrid VAEs"],"prefix":"10.1093","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-0484-3582","authenticated-orcid":false,"given":"Zahid","family":"Masood","sequence":"first","affiliation":[{"name":"Nazarbayev University Department of Mechanical and Aerospace Engineering, , 53 Kabanbay Batyr Ave. 010000, Astana ,","place":["Kazakhstan"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0298-9089","authenticated-orcid":false,"given":"Shahroz","family":"Khan","sequence":"additional","affiliation":[{"name":"Compute Maritime , 566 Chiswick High Road London W4 5YG ,","place":["UK"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1052-3329","authenticated-orcid":false,"given":"Konstantinos","family":"Kostas","sequence":"additional","affiliation":[{"name":"Nazarbayev University Department of Mechanical and Aerospace Engineering, , 53 Kabanbay Batyr Ave. 010000, Astana ,","place":["Kazakhstan"]},{"name":"University of West Attica Department of Naval Architecture, , Ag. Spyridonos Str., Egaleo, 12243, Athens ,","place":["Greece"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1843-8815","authenticated-orcid":false,"given":"Panagiotis","family":"Kaklis","sequence":"additional","affiliation":[{"name":"Ocean and Marine Engineering, University Strathclyde Department of Naval Architecture, , Glasgow ,","place":["UK"]},{"name":"Institute Applied & Computational Mathematics, FORTH, Crete (GR) ARCHIMEDES Unit, ATHENA Research Center , 15125, Athens ,","place":["Greece"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2026,3,11]]},"reference":[{"key":"2026050910314799900_bib1","doi-asserted-by":"publisher","first-page":"2897","DOI":"10.1109\/TPAMI.2017.2784440","article-title":"Information dropout: Learning optimal representations through noisy computation","volume":"40","author":"Achille","year":"2018","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"2026050910314799900_bib2","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1804.03599","article-title":"Understanding disentangling in\u00a0-vae","author":"Burgess","year":"2018"},{"key":"2026050910314799900_bib3","doi-asserted-by":"publisher","first-page":"107909","DOI":"10.1016\/j.asoc.2021.107909","article-title":"MO-PaDGAN: Reparameterizing Engineering Designs for augmented multi-objective optimization","volume":"113","author":"Chen","year":"2021","journal-title":"Applied Soft Computing"},{"key":"2026050910314799900_bib5","doi-asserted-by":"publisher","first-page":"031703","DOI":"10.1115\/1.4048626","article-title":"PaDGAN: Learning to Generate High-Quality Novel Designs","volume":"143","author":"Chen","year":"2020","journal-title":"Journal of Mechanical Design"},{"key":"2026050910314799900_bib6","first-page":"167","article-title":"Metro: measuring error on simplified surfaces Computer graphics forum","author":"Cignoni","year":"1998"},{"key":"2026050910314799900_bib7","doi-asserted-by":"crossref","DOI":"10.1137\/1.9781611973860","article-title":"Active Subspaces: Emerging Ideas for Dimension Reduction in Parameter Studies","volume-title":"SIAM spotlights","author":"Constantine","year":"2015"},{"key":"2026050910314799900_bib8","doi-asserted-by":"publisher","first-page":"106852","DOI":"10.1016\/j.oceaneng.2019.106852","article-title":"Design-space assessment and dimensionality reduction: An off-line method for shape reparameterization in simulation-based optimization","volume":"197","author":"D\u2019Agostino","year":"2020","journal-title":"Ocean Engineering"},{"key":"2026050910314799900_bib9","doi-asserted-by":"publisher","first-page":"1525","DOI":"10.1016\/j.cma.2014.10.042","article-title":"Design-space dimensionality reduction in shape optimization by Karhunen\u2013Lo\u00e8ve expansion","volume":"283","author":"Diez","year":"2015","journal-title":"Computer Methods in Applied Mechanics and Engineering"},{"key":"2026050910314799900_bib10","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-642-84010-4_1","article-title":"XFOIL: An Analysis and Design System for Low Reynolds Number Airfoils","author":"Drela","year":"1989"},{"key":"2026050910314799900_bib11","doi-asserted-by":"publisher","first-page":"1347","DOI":"10.2514\/3.9789","article-title":"Viscous-inviscid analysis of transonic and low reynolds number airfoils","volume":"25","author":"Drela","year":"1987","journal-title":"AIAA"},{"key":"2026050910314799900_bib12","doi-asserted-by":"publisher","first-page":"139","DOI":"10.1145\/3422622","article-title":"Generative adversarial networks","volume-title":"Commun. ACM","author":"Goodfellow","year":"2020"},{"key":"2026050910314799900_bib13","article-title":"beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework","volume-title":"International Conference on Learning Representations","author":"Higgins","year":"2017"},{"key":"2026050910314799900_bib14","doi-asserted-by":"crossref","first-page":"504","DOI":"10.1126\/science.1127647","article-title":"Reducing the dimensionality of data with neural networks","volume":"313","author":"Hinton","year":"2006","journal-title":"science"},{"key":"2026050910314799900_bib15","doi-asserted-by":"publisher","first-page":"77","DOI":"10.1109\/MSP.2010.939747","article-title":"Preimage problem in kernel-based machine learning","volume":"28","author":"Honeine","year":"2011","journal-title":"IEEE Signal Processing Magazine"},{"key":"2026050910314799900_bib16","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4757-1904-8_7","article-title":"Principal Component Analysis","author":"Jolliffe","year":"1986"},{"key":"2026050910314799900_bib17","doi-asserted-by":"publisher","first-page":"076103","DOI":"10.1063\/5.0097740","article-title":"Physics-aware reduced-order modeling of transonic flow via $\\beta$-variational autoencoder","volume":"34","author":"Kang","year":"2022","journal-title":"Physics of Fluids"},{"key":"2026050910314799900_bib18","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2311.10921","article-title":"Compact and intuitive airfoil parameterization method through physics-aware variational autoencoder","author":"Kang","year":"2023"},{"key":"2026050910314799900_bib19","doi-asserted-by":"publisher","first-page":"27","DOI":"10.1093\/jcde\/qwaf002","article-title":"Intuitive and feasible geometric representation of airfoil using variational autoencoder","volume":"12","author":"Kang","year":"2025","journal-title":"Journal of Computational Design and Engineering"},{"key":"2026050910314799900_bib20","doi-asserted-by":"publisher","first-page":"103327","DOI":"10.1016\/j.cad.2022.103327","article-title":"Shape-supervised dimension reduction: Extracting geometry and physics associated features with geometric moments","volume":"150","author":"Khan","year":"2022","journal-title":"Computer-Aided Design"},{"key":"2026050910314799900_bib21","doi-asserted-by":"publisher","first-page":"116051","DOI":"10.1016\/j.cma.2023.116051","article-title":"ShipHullGAN: A generic parametric modeller for ship hull design using deep convolutional generative model","volume":"411","author":"Khan","year":"2023","journal-title":"Computer Methods in Applied Mechanics and Engineering"},{"key":"2026050910314799900_bib22","doi-asserted-by":"publisher","first-page":"102937","DOI":"10.1016\/j.aei.2024.102937","article-title":"Physics-informed geometric operators to support surrogate, dimension reduction and generative models for engineering design","volume":"63","author":"Khan","year":"2025","journal-title":"Advanced Engineering Informatics"},{"key":"2026050910314799900_bib23","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2017"},{"key":"2026050910314799900_bib24","article-title":"Auto-encoding variational bayes","author":"Kingma","year":"2022"},{"key":"2026050910314799900_bib25","doi-asserted-by":"publisher","first-page":"79","DOI":"10.1016\/j.cad.2016.07.002","article-title":"Shape-optimization of 2D hydrofoils using an Isogeometric BEM solver","volume":"82","author":"Kostas","year":"2017","journal-title":"Computer-Aided Design"},{"key":"2026050910314799900_bib26","doi-asserted-by":"publisher","first-page":"107020","DOI":"10.1016\/j.oceaneng.2020.107020","article-title":"Parametric model for the reconstruction and representation of hydrofoils and airfoils","volume":"199","author":"Kostas","year":"2020","journal-title":"Ocean Engineering"},{"key":"2026050910314799900_bib27","doi-asserted-by":"publisher","DOI":"10.3390\/jmse11071470","article-title":"Machine-learning-enabled foil design assistant","volume":"11","author":"Kostas","year":"2023","journal-title":"Journal of Marine Science and Engineering"},{"key":"2026050910314799900_bib28","article-title":"Fundamental parametric geometry representations for aircraft component shapes","author":"Kulfan","year":"2006"},{"key":"2026050910314799900_bib29","doi-asserted-by":"publisher","first-page":"142","DOI":"10.2514\/1.29958","article-title":"Universal parametric geometry representation method","volume":"45","author":"Kulfan","year":"2008","journal-title":"Journal of aircraft"},{"key":"2026050910314799900_bib30","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.paerosci.2017.05.003","article-title":"Review of design optimization methods for turbomachinery aerodynamics","volume":"93","author":"Li","year":"2017","journal-title":"Progress in Aerospace Sciences"},{"key":"2026050910314799900_bib31","doi-asserted-by":"publisher","first-page":"105391","DOI":"10.1016\/j.compfluid.2022.105391","article-title":"Aerodynamic design optimization: Challenges and perspectives","volume":"239","author":"Martins","year":"2022","journal-title":"Computers and Fluids"},{"key":"2026050910314799900_bib32","doi-asserted-by":"publisher","first-page":"1851","DOI":"10.3390\/jmse11101851","article-title":"Shape-informed dimensional reduction in airfoil\/hydrofoil modeling","volume":"11","author":"Masood","year":"2023","journal-title":"Journal of Marine Science and Engineering"},{"key":"2026050910314799900_bib33","doi-asserted-by":"publisher","DOI":"10.3390\/jmse12040566","article-title":"Generative vs. non-generative models in engineering shape optimization","volume":"12","author":"Masood","year":"2024","journal-title":"Journal of Marine Science and Engineering"},{"key":"2026050910314799900_bib34","article-title":"An introduction to convolutional neural networks","author":"O\u2019Shea","year":"2015"},{"key":"2026050910314799900_bib35","first-page":"8024","article-title":"Pytorch: An imperative style, high-performance deep learning library","volume-title":"Advances in Neural Information Processing Systems 32","author":"Paszke","year":"2019"},{"key":"2026050910314799900_bib36","doi-asserted-by":"publisher","first-page":"2925","DOI":"10.1016\/j.cma.2007.02.008","article-title":"Reduced basis approximation and a posteriori error estimation for a boltzmann model","volume":"196","author":"Patera","year":"2007","journal-title":"Computer Methods in Applied Mechanics and Engineering"},{"key":"2026050910314799900_bib37","doi-asserted-by":"publisher","first-page":"686","DOI":"10.1016\/j.jcp.2018.10.045","article-title":"Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations","volume":"378","author":"Raissi","year":"2019","journal-title":"Journal of Computational Physics"},{"key":"2026050910314799900_bib38","doi-asserted-by":"publisher","first-page":"2323","DOI":"10.1126\/science.290.5500.2323","article-title":"Nonlinear dimensionality reduction by locally linear embedding","volume":"290","author":"Roweis","year":"2000","journal-title":"Science"},{"key":"2026050910314799900_bib39","doi-asserted-by":"publisher","first-page":"112951","DOI":"10.1016\/j.jcp.2024.112951","article-title":"Uncertainty quantification in autoencoders predictions: Applications in aerodynamics","volume":"506","author":"Saetta","year":"2024","journal-title":"Journal of Computational Physics"},{"key":"2026050910314799900_bib40","doi-asserted-by":"publisher","first-page":"1299","DOI":"10.1162\/089976698300017467","article-title":"Nonlinear component analysis as a kernel eigenvalue problem","volume":"10","author":"Sch\u00f6lkopf","year":"1998","journal-title":"Neural Computation"},{"key":"2026050910314799900_bib41","doi-asserted-by":"publisher","first-page":"115776","DOI":"10.1016\/j.cma.2022.115776","article-title":"Parametric model embedding","volume":"404","author":"Serani","year":"2023","journal-title":"Computer Methods in Applied Mechanics and Engineering"},{"key":"2026050910314799900_bib42","doi-asserted-by":"publisher","DOI":"10.1007\/s11831-025-10349-x","article-title":"A survey on design-space dimensionality reduction methods for shape optimization","volume":"33","author":"Serani","year":"2025","journal-title":"Arch. Comput. Methods Eng"},{"key":"2026050910314799900_bib43","article-title":"Probability, Random Processes, and Estimation Theory for Engineers","author":"Stark","year":"1986"},{"key":"2026050910314799900_bib44","doi-asserted-by":"publisher","DOI":"10.2514\/6.2024-0686","article-title":"Towards universal parameterization: Using variational autoencoders to parameterize airfoils","volume-title":"AIAA SCITECH 2024 Forum","author":"Swannet","year":"2024"},{"key":"2026050910314799900_bib45","doi-asserted-by":"publisher","DOI":"10.3390\/math12101417","article-title":"A review of intelligent airfoil aerodynamic optimization methods based on data-driven advanced models","volume":"12","author":"Wang","year":"2024","journal-title":"Mathematics"},{"key":"2026050910314799900_bib46","doi-asserted-by":"publisher","first-page":"1350","DOI":"10.1093\/jcde\/qwad046","article-title":"Airfoil GAN: encoding and synthesizing airfoils for aerodynamic shape optimization","volume":"10","author":"Wang","year":"2023","journal-title":"Journal of Computational Design and Engineering"},{"key":"2026050910314799900_bib47","doi-asserted-by":"publisher","first-page":"108409","DOI":"10.1016\/j.ast.2023.108409","article-title":"Three-dimensional aerodynamic shape inverse design based on isomap","volume":"139","author":"Xiao","year":"2023","journal-title":"Aerospace Science and Technology"},{"key":"2026050910314799900_bib48","doi-asserted-by":"publisher","first-page":"240","DOI":"10.1016\/j.patcog.2007.05.001","article-title":"Geometric moment invariants","volume":"41","author":"Xu","year":"2008","journal-title":"Pattern Recognition"},{"key":"2026050910314799900_bib49","doi-asserted-by":"publisher","first-page":"2239","DOI":"10.1007\/s00366-022-01617-6","article-title":"Inverse design optimization framework via a two-step deep learning approach: application to a wind turbine airfoil","volume":"39","author":"Yang","year":"2022","journal-title":"Eng. with Comput."},{"key":"2026050910314799900_bib50","doi-asserted-by":"publisher","first-page":"105883","DOI":"10.1016\/j.compfluid.2023.105883","article-title":"Airfoil-based convolutional autoencoder and long short-term memory neural network for predicting coherent structures evolution around an airfoil","volume":"258","author":"Zhang","year":"2023","journal-title":"Computers and Fluids"},{"key":"2026050910314799900_bib51","doi-asserted-by":"publisher","DOI":"10.3390\/aerospace12050443","article-title":"Aerodynamic optimization of airfoil in wide range of operating conditions based on reinforcement learning","volume":"12","author":"Zhang","year":"2025","journal-title":"Aerospace"}],"container-title":["Journal of Computational Design and Engineering"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jcde\/advance-article-pdf\/doi\/10.1093\/jcde\/qwag025\/67300962\/qwag025.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jcde\/article-pdf\/13\/4\/51\/67300962\/qwag025.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jcde\/article-pdf\/13\/4\/51\/67300962\/qwag025.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,9]],"date-time":"2026-05-09T14:31:55Z","timestamp":1778337115000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jcde\/article\/13\/4\/51\/8514302"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,11]]},"references-count":50,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,4,1]]}},"URL":"https:\/\/doi.org\/10.1093\/jcde\/qwag025","relation":{},"ISSN":["2288-5048"],"issn-type":[{"value":"2288-5048","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,4]]},"published":{"date-parts":[[2026,3,11]]}}}