Logo do repositório
 
Publicação

Differentiable neural search architecture with zero-cost metrics for insulator fault prediction

2026-03Artigo científico Acesso aberto
Ver registo simples
dc.contributor.authorSeman, Laio Oriel
dc.contributor.authorBuratto, William Gouvêa
dc.contributor.authorGonzalez, Gabriel Villarrubia
dc.contributor.authorLeithardt, Valderi Reis Quietinho
dc.contributor.authorNied, Ademir
dc.contributor.authorStefenon, Stefano Frizzo
dc.contributor.institutionCTS - Centro de Tecnologia e Sistemas
dc.contributor.institutionUNINOVA-Instituto de Desenvolvimento de Novas Tecnologias
dc.contributor.pblElsevier BV
dc.date.accessioned2026-07-15T14:28:02Z
dc.date.available2026-07-15T14:28:02Z
dc.date.issued2026-03
dc.descriptionThe APC was supported by the project Self-adaptive platform based on intelligent agents for the optimization and management of operational processes in logistic warehouses (PLAUTON), PID2023-151701OB-C21, funded by MCIN/AEI/10.13039/501100011033/FEDER, EU. Buratto would like to thank UDESC for the financial support for his interuniversity exchange doctorate at the University of Salamanca. Also, the authors would like to thank the Coordination for the Improvement of Higher Education Personnel (CAPES - Brazil) for the scholarship to Buratto. This study was financed (i) in part by CAPES under the doctoral scholarship number 88887.808258/2023-00, and (ii) by Council for Scientific and Technological Development (CNPq) under grant numbers 305910/2024-8 and 307858/2025-1. A realização desta investigação foi parcialmente financiada por fundos nacionais através da FCT - Fundação para a Ciência e Tecnologia, I.P. no âmbito dos projetos UIDB/04466/2025 e UIDP/04466/2025. Publisher Copyright: © 2026 The Author(s).
dc.description.abstractReliable monitoring of high-voltage insulators is critical for maintaining the stability of electrical power systems, particularly under environmental contamination that can lead to flashover. Traditional inspection techniques struggle to anticipate degradation dynamics, while data-driven models often rely on fixed neural architectures that inadequately capture the complex temporal patterns in leakage current signals. This work proposes a Differentiable Neural Architecture Search (DARTS) framework, based on zero-cost metrics, tailored for time series forecasting in insulator monitoring. The method based on DARTS integrates a mixed encoder-decoder design with learnable selection over long short-term memory, gated recurrent units, and transformer components, coupled with a cross-attention bridge featuring temporal bias and gating mechanisms. To ensure efficient architecture exploration, the search leverages metrics such as SynFlow and Jacobian covariance for early candidate screening, followed by a bilevel optimization stage with entropy and diversity regularization. Experiments on real-world leakage current data demonstrate that the discovered architectures outperform manually designed baselines, offering improved forecasting performance.en
dc.description.versionpublishersversion
dc.description.versionpublished
dc.format.extent7710066
dc.identifier.doi10.1016/j.rineng.2026.109716
dc.identifier.issn2590-1230
dc.identifier.otherPURE: 168368277
dc.identifier.otherPURE UUID: 37ada2cf-0bab-4b31-a20e-1d554db965cd
dc.identifier.otherScopus: 105032047509
dc.identifier.urihttp://hdl.handle.net/10362/204545
dc.identifier.urlhttps://www.scopus.com/pages/publications/105032047509
dc.language.isoeng
dc.peerreviewedyes
dc.relationinfo:eu-repo/grantAgreement/FCT/Concurso de Projetos de IC&DT em Todos os Domínios Científicos 2023/2023.16583.ICDT/PT
dc.subjectDifferentiable neural architecture
dc.subjectForecasting
dc.subjectNeural network architectures
dc.subjectPredictive maintenance
dc.subjectGeneral Engineering
dc.titleDifferentiable neural search architecture with zero-cost metrics for insulator fault predictionen
dc.typejournal article
degois.publication.titleResults in Engineering
degois.publication.volume29
dspace.entity.typePublication
rcaap.rightsopenAccess

Ficheiros

Principais
A mostrar 1 - 1 de 1
Miniatura
Nome:
Seman_et_al._2026_.pdf
Tamanho:
7.35 MB
Formato:
Adobe Portable Document Format
Ver/Abrir

Coleções

Home collection (FCT)