Neural parameter calibration for reliable hysteresis prediction in bolted joint assemblies

Registro completo de metadados
MetadadosDescriçãoIdioma
Autor(es): dc.contributorUniversidade Estadual Paulista (UNESP)-
Autor(es): dc.creatorAlmeida, Estevão Fuzaro-
Autor(es): dc.creatorSilva, Samuel-
Data de aceite: dc.date.accessioned2025-08-21T22:21:43Z-
Data de disponibilização: dc.date.available2025-08-21T22:21:43Z-
Data de envio: dc.date.issued2025-03-12-
Data de envio: dc.date.issued2025-03-12-
Data de envio: dc.date.issued2025-03-09-
Fonte completa do material: dc.identifierhttps://hdl.handle.net/11449/295405-
Fonte: dc.identifier.urihttp://educapes.capes.gov.br/handle/11449/295405-
Descrição: dc.descriptionBolted joints are a common way for connecting multiple structures, and ensuring their safe operation is crucial. Changes in operational conditions, such as variations in tightening torque, can introduce hysteresis mechanisms and complex nonlinearities, making it challenging to analyze and diagnose any issues. Vibration measurements can be used to calibrate a reduced-order model that captures the effects of energy dissipation in bolted joints, such as a Bouc-Wen oscillator. However, the nonlinearities inherent in these systems make calibration problematic, typically requiring ad-hoc knowledge and considerations. In this work, we propose to use a new neural parameter calibration paradigm for this computational model, utilizing a physics-informed neural network to estimate the Bouc-Wen model parameters from time series data. The approach involves using a neural differential equation to represent the hysteresis effect and extracting coefficients from the vibration time series to inform the model. The method generates accurate predictions for the hysteresis loop in a matter of minutes, demonstrating the potential of this approach for real-time monitoring and diagnosis of bolted joint assemblies.-
Descrição: dc.descriptionFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)-
Descrição: dc.descriptionVersão final do editor-
Descrição: dc.descriptionFAPESP: 2022/16156-9-
Formato: dc.formatapplication/pdf-
Idioma: dc.languageen-
Relação: dc.relationProceedings of the XX International Symposium on Dynamic Problems of Mechanics-
Direitos: dc.rightsinfo:eu-repo/semantics/openAccess-
Palavras-chave: dc.subjectBolted joints-
Palavras-chave: dc.subjectHysteresis mechanisms-
Palavras-chave: dc.subjectModel calibration-
Palavras-chave: dc.subjectNeural differential equation-
Título: dc.titleNeural parameter calibration for reliable hysteresis prediction in bolted joint assemblies-
Título: dc.titleCalibração neural de parâmetros para previsão confiável de histerese em juntas aparafusadas-
Tipo de arquivo: dc.typelivro digital-
Aparece nas coleções:Repositório Institucional - Unesp

Não existem arquivos associados a este item.