Damage quantification using transfer component analysis combined with Gaussian process regression
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Outros Autores: | , , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UNESP |
Texto Completo: | http://dx.doi.org/10.1177/14759217221094500 http://hdl.handle.net/11449/240171 |
Resumo: | Machine learning methods used in Structural Health Monitoring applications still have generalization difficulties among structures, even when structures are nominally and topologically similar. The data sets present divergences between their probability distributions that do not allow the model’s generalization for damage detection. This issue is even more complex in situations where one wants to quantify damage levels through data sets collected from different structures. Transfer learning methods offer a solution to overcome those limitations, using relevant information from a labeled structure (source domain) to assist the analysis of another structure (target domain) under unknown conditions. Therefore, this paper proposes the use of transfer component analysis to mitigate divergences between the model/structure’s features, and the label consistency requirement is applied in combination with a Gaussian process regression model for damage quantification. The effectiveness of the estimated model improves when the labels consistency between domains is achieved, indicating the current damage level in the structure when the regression model achieves its best performance (lowest error). The proposed methodology is applied on the benchmark data of a three-story building structure from the Los Alamos National Laboratory using the knowledge from its numerical model under several conditions, where the complete information of its behavior is available. The results compare the analysis in the original space and after applying the proposed methodology, demonstrating an improvement of the performance in the damage detection and quantification steps. |
id |
UNSP_7d4caecca57ae75e6e344c5e9ec3bfb4 |
---|---|
oai_identifier_str |
oai:repositorio.unesp.br:11449/240171 |
network_acronym_str |
UNSP |
network_name_str |
Repositório Institucional da UNESP |
repository_id_str |
2946 |
spelling |
Damage quantification using transfer component analysis combined with Gaussian process regressiondamage identificationdomain adaptationGaussian process regressionStructural Health Monitoringtransfer component analysisTransfer learningMachine learning methods used in Structural Health Monitoring applications still have generalization difficulties among structures, even when structures are nominally and topologically similar. The data sets present divergences between their probability distributions that do not allow the model’s generalization for damage detection. This issue is even more complex in situations where one wants to quantify damage levels through data sets collected from different structures. Transfer learning methods offer a solution to overcome those limitations, using relevant information from a labeled structure (source domain) to assist the analysis of another structure (target domain) under unknown conditions. Therefore, this paper proposes the use of transfer component analysis to mitigate divergences between the model/structure’s features, and the label consistency requirement is applied in combination with a Gaussian process regression model for damage quantification. The effectiveness of the estimated model improves when the labels consistency between domains is achieved, indicating the current damage level in the structure when the regression model achieves its best performance (lowest error). The proposed methodology is applied on the benchmark data of a three-story building structure from the Los Alamos National Laboratory using the knowledge from its numerical model under several conditions, where the complete information of its behavior is available. The results compare the analysis in the original space and after applying the proposed methodology, demonstrating an improvement of the performance in the damage detection and quantification steps.Departamento de Engenharia Mecânica UNESP - Universidade Estadual PaulistaFaculty of Engineering Lusófona UniversityCERIS Instituto Superior Técnico Universidade de LisboaDepartamento de Engenharia Mecânica Centro Tecnológico UFES - Universidade Federal do Espírito SantoDepartamento de Engenharia Mecânica UNESP - Universidade Estadual PaulistaUniversidade Estadual Paulista (UNESP)Lusófona UniversityUniversidade de LisboaUniversidade Federal do Espírito Santo (UFES)Yano, Marcus Omori [UNESP]Silva, Samuel da [UNESP]Figueiredo, EloiVillani, Luis G Giacon2023-03-01T20:04:33Z2023-03-01T20:04:33Z2022-01-01info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articlehttp://dx.doi.org/10.1177/14759217221094500Structural Health Monitoring.1741-31681475-9217http://hdl.handle.net/11449/24017110.1177/147592172210945002-s2.0-85131174562Scopusreponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengStructural Health Monitoringinfo:eu-repo/semantics/openAccess2023-03-01T20:04:33Zoai:repositorio.unesp.br:11449/240171Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462023-03-01T20:04:33Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false |
dc.title.none.fl_str_mv |
Damage quantification using transfer component analysis combined with Gaussian process regression |
title |
Damage quantification using transfer component analysis combined with Gaussian process regression |
spellingShingle |
Damage quantification using transfer component analysis combined with Gaussian process regression Yano, Marcus Omori [UNESP] damage identification domain adaptation Gaussian process regression Structural Health Monitoring transfer component analysis Transfer learning |
title_short |
Damage quantification using transfer component analysis combined with Gaussian process regression |
title_full |
Damage quantification using transfer component analysis combined with Gaussian process regression |
title_fullStr |
Damage quantification using transfer component analysis combined with Gaussian process regression |
title_full_unstemmed |
Damage quantification using transfer component analysis combined with Gaussian process regression |
title_sort |
Damage quantification using transfer component analysis combined with Gaussian process regression |
author |
Yano, Marcus Omori [UNESP] |
author_facet |
Yano, Marcus Omori [UNESP] Silva, Samuel da [UNESP] Figueiredo, Eloi Villani, Luis G Giacon |
author_role |
author |
author2 |
Silva, Samuel da [UNESP] Figueiredo, Eloi Villani, Luis G Giacon |
author2_role |
author author author |
dc.contributor.none.fl_str_mv |
Universidade Estadual Paulista (UNESP) Lusófona University Universidade de Lisboa Universidade Federal do Espírito Santo (UFES) |
dc.contributor.author.fl_str_mv |
Yano, Marcus Omori [UNESP] Silva, Samuel da [UNESP] Figueiredo, Eloi Villani, Luis G Giacon |
dc.subject.por.fl_str_mv |
damage identification domain adaptation Gaussian process regression Structural Health Monitoring transfer component analysis Transfer learning |
topic |
damage identification domain adaptation Gaussian process regression Structural Health Monitoring transfer component analysis Transfer learning |
description |
Machine learning methods used in Structural Health Monitoring applications still have generalization difficulties among structures, even when structures are nominally and topologically similar. The data sets present divergences between their probability distributions that do not allow the model’s generalization for damage detection. This issue is even more complex in situations where one wants to quantify damage levels through data sets collected from different structures. Transfer learning methods offer a solution to overcome those limitations, using relevant information from a labeled structure (source domain) to assist the analysis of another structure (target domain) under unknown conditions. Therefore, this paper proposes the use of transfer component analysis to mitigate divergences between the model/structure’s features, and the label consistency requirement is applied in combination with a Gaussian process regression model for damage quantification. The effectiveness of the estimated model improves when the labels consistency between domains is achieved, indicating the current damage level in the structure when the regression model achieves its best performance (lowest error). The proposed methodology is applied on the benchmark data of a three-story building structure from the Los Alamos National Laboratory using the knowledge from its numerical model under several conditions, where the complete information of its behavior is available. The results compare the analysis in the original space and after applying the proposed methodology, demonstrating an improvement of the performance in the damage detection and quantification steps. |
publishDate |
2022 |
dc.date.none.fl_str_mv |
2022-01-01 2023-03-01T20:04:33Z 2023-03-01T20:04:33Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://dx.doi.org/10.1177/14759217221094500 Structural Health Monitoring. 1741-3168 1475-9217 http://hdl.handle.net/11449/240171 10.1177/14759217221094500 2-s2.0-85131174562 |
url |
http://dx.doi.org/10.1177/14759217221094500 http://hdl.handle.net/11449/240171 |
identifier_str_mv |
Structural Health Monitoring. 1741-3168 1475-9217 10.1177/14759217221094500 2-s2.0-85131174562 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
Structural Health Monitoring |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.source.none.fl_str_mv |
Scopus reponame:Repositório Institucional da UNESP instname:Universidade Estadual Paulista (UNESP) instacron:UNESP |
instname_str |
Universidade Estadual Paulista (UNESP) |
instacron_str |
UNESP |
institution |
UNESP |
reponame_str |
Repositório Institucional da UNESP |
collection |
Repositório Institucional da UNESP |
repository.name.fl_str_mv |
Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP) |
repository.mail.fl_str_mv |
|
_version_ |
1799964377997115392 |