MODELOS DE INTELIGENCIA ARTIFICIAL GENERATIVA APLICADOS A LAS ESTRUCTURAS DE ACERO: UNA REVISIÓN SISTEMÁTICA DE LA LITERATURA
Resumen
El creciente interés en la automatización del proceso de diseño estructural ha impulsado la aplicación de modelos de inteligencia artificial generativa a tareas como la generación de layouts y la documentación automatizada de proyectos. Sin embargo, a pesar de los recientes avances científicos, las estructuras de acero siguen siendo un campo poco explorado, sin una revisión sistemática consolidada que mapee las arquitecturas empleadas, las tareas abordadas y las lagunas persistentes en este dominio específico. Este trabajo tiene como objetivo realizar una revisión sistemática de la literatura sobre la aplicación de modelos generativos de inteligencia artificial al campo de las estructuras de acero, llevada a cabo según el protocolo de Kitchenham (2004) y Kitchenham y Charters (2007). Las bases consultadas fueron IEEE Xplore, ScienceDirect, Scopus, SpringerLink y el Portal de Periódicos CAPES, con un recorte temporal de 2019 a 30 abril de 2026. Un total de 895 artículos fueron identificados inicialmente, resultando en 12 estudios incluidos tras la aplicación de los criterios de elegibilidad y calidad. Los resultados revelan una producción científica aún incipiente, concentrada en la generación de layouts de sistemas estructurales de acero, con una marcada escasez de estudios que aborden el campo de las estructuras de acero en mayor amplitud. Se concluye que el campo requiere una exploración más profunda de las tipologías estructurales, las morfologías de conexiones metálicas y la incorporación de especificaciones de proyecto en los productos sintéticos.
Biografía del autor/a
Licenciado en Ingeniería Civil por la Universidad Federal de Paraná (UFPR). Estudiante de Maestría en el Programa de Posgrado en Ingeniería Civil de la UFPR.
Doctor en Métodos Numéricos en Ingeniería por la Universidad Federal de Paraná (2011). Profesor Asociado del Departamento de Construcción Civil de la UFPR. Docente permanente del Programa de Posgrado en Ingeniería Civil de la UFPR y miembro del CESEC/UFPR.
Referencias
AZIZ, MD. T. et al. State-of-the-art artificial intelligence techniques in structural engineering: a review of applications and prospects. Results in Engineering, v. 28, p. 107882, out. 2025. Disponível em: https://doi.org/10.1016/j.rineng.2025.107882 Acesso em: 04 mar. 2026.
BROCK, A.; DONAHUE, J.; SIMONYAN, K. Large scale GAN training for high fidelity natural image synthesis. arXiv, 2018. Disponível em: https://arxiv.org/abs/1809.11096 Acesso em: 04 mar. 2026.
DU, W. et al. 3D solid model generation method based on a generative adversarial network. Applied Intelligence, v. 53, n. 13, p. 17035–17060, 2023. Disponível em: https://www.scopus.com/pages/publications/85144549278?origin=resultslist Acesso em: 30 abr. 2026.
FU, B.; GAO, Y.; WANG, W. Dual generative adversarial networks for automated component layout design of steel frame-brace structures. Automation in Construction, v. 146, 2023. Disponível em: https://www.scopus.com/pages/publications/85142721960?origin=resultslist Acesso em: 30 abr. 2026.
FU, B.; WANG, W.; GAO, Y. Physical rule-guided generative adversarial network for automated structural layout design of steel frame-brace structures. Journal of Building Engineering, v. 86, 2024. Disponível em: https://www.scopus.com/pages/publications/85186519557?origin=resultslist Acesso em: 30 abr. 2026.
GAO, S. et al. Axial compression behavior of steel tubular members with simulated splashing corrosion. Structures, v. 80, 2025. Disponível em: https://www.scopus.com/pages/publications/105012959472?origin=resultslist Acesso em: 30 abr. 2026.
GOODFELLOW, I. et al. Generative adversarial nets. In: Advances in Neural Information Processing Systems. [S.l.: s.n.], 2014. v. 27, p. 2672–2680. Disponível em: https://dl.acm.org/doi/10.5555/2969033.2969125 Acesso em: 27 abr. 2026.
HAN, Y. et al. Generation of innovative structural connection components using generative adversarial networks. Advanced Engineering Informatics, v. 68, 2025. Disponível em: https://www.scopus.com/pages/publications/105009655457?origin=resultslist Acesso em: 30 abr. 2026.
HO, J.; JAIN, A.; ABBEEL, P. Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems. [S.l.: s.n.], 2020. v. 33, p. 6840–6851. Disponível em: https://dl.acm.org/doi/abs/10.5555/3495724.3496298 Acesso em: 27 abr. 2026.
KARRAS, T.; LAINE, S.; AILA, T. A style-based generator architecture for generative adversarial networks. arXiv, 2018. Disponível em: https://arxiv.org/abs/1812.04948 Acesso em: 04 mar. 2026.
KITCHENHAM, B. Procedures for performing systematic reviews. Keele: Keele University, 2004. Relatório Técnico TR/SE-0401. Disponível em: https://www.inf.ufsc.br/~aldo.vw/kitchenham.pdf Acesso em: 27 abr. 2026.
KITCHENHAM, B.; CHARTERS, S. Guidelines for performing systematic literature reviews in software engineering. [S.l.]: Keele University; University of Durham, 2007. Relatório Técnico EBSE-2007-01. Acesso em: 27 abr. 2026.
LI, J. et al. FrameDiffusion: a latent diffusion model for intelligent layout design of steel frame-braced structures. Engineering Structures, v. 343, 2025. Disponível em: https://www.scopus.com/pages/publications/105014216066?origin=resultslist Acesso em: 30 abr. 2026.
LIAO, W. et al. Generative AI design for building structures. Automation in Construction, v. 157, n. 1, p. 105187, jan. 2024. Disponível em: https://doi.org/10.1016/j.autcon.2023.105187 Acesso em: 04 mar. 2026.
LU, N. et al. A small-sample fatigue life prediction framework for welded joints in orthotropic steel decks using data augmentation and CNN-based feature learning. Ocean Engineering, v. 341, 2025. Disponível em: https://www.scopus.com/pages/publications/105015038642?origin=resultslist Acesso em: 30 abr. 2026.
MOU, B.; CHEN, H.; FU, Y. An effective design method of high-strength steel columns with limited datasets using physics-guided conditional tabular GAN. Engineering Structures, v. 340, 2025. Disponível em: https://www.scopus.com/pages/publications/105007413497?origin=resultslist Acesso em: 30 abr. 2026.
SIRRIYEH, R. et al. Reviewing studies with diverse designs: the development and evaluation of a new tool. Journal of Evaluation in Clinical Practice, v. 18, n. 4, p. 746–752, 2012. Acesso em: 27 abr. 2026.
WU, Y. et al. Bond capacity and recover stress of Fe-SMA–steel joints: numerical, theoretical, and machine learning study. Construction and Building Materials, v. 494, 2025. Disponível em: https://www.scopus.com/pages/publications/105015145523?origin=resultslist Acesso em: 30 abr. 2026.
WU, Z. et al. A comprehensive study on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems, v. 32, n. 1, p. 4–24, 2020. Disponível em: https://ieeexplore.ieee.org/document/9046288 Acesso em: 28 abr. 2026.
YU, M.; LI, S.; XIE, X. Machine learning for ULCF life prediction of structural steels with synthetic data. Journal of Constructional Steel Research, v. 224, 2025. Disponível em: https://www.scopus.com/pages/publications/85208916446?origin=resultslist Acesso em: 30 abr. 2026.
ZHAO, P. et al. Intelligent beam layout design for frame structure based on graph neural networks. Journal of Building Engineering, v. 63, p. 105499, 2023. Disponível em: https://www.sciencedirect.com/science/article/pii/S2352710222015054 Acesso em: 30 abr. 2026.
ZUO, W. et al. Physics-informed data augmentation and explainable machine learning for parametric design of stainless steel columns. Advanced Engineering Informatics, v. 73, p. 104581, 2026. Disponível em: https://www.sciencedirect.com/science/article/pii/S1474034626002739 Acesso em: 30 abr. 2026.
Licencia
Derechos de autor 2026 RECIMA21 - Revista Científica Multidisciplinar - ISSN 2675-6218
Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.
Os direitos autorais dos artigos/resenhas/TCCs publicados pertecem à revista RECIMA21, e seguem o padrão Creative Commons (CC BY 4.0), permitindo a cópia ou reprodução, desde que cite a fonte e respeite os direitos dos autores e contenham menção aos mesmos nos créditos. Toda e qualquer obra publicada na revista, seu conteúdo é de responsabilidade dos autores, cabendo a RECIMA21 apenas ser o veículo de divulgação, seguindo os padrões nacionais e internacionais de publicação.