Uncertainty quantification in deep convolutional neural network diagnostics of journal bearings with ovalization fault

Autor(es)

Alves, Diogo Stuani ^[1] ; Daniel, Gregory Bregion ^[1] ; de Castro, Helio Fiori ^[1] ; Machado, Tiago Henrique ^[1] ; Cavalca, Katia Lucchesi ^[1] ; Gecgel, Ozhan ^[2] ; Dias, Joao Paulo ^[3] ; Ekwaro-Osire, Stephen ^[2]

Afiliação do(s) autor(es)

^[1] Univ Estadual Campinas, Sch Mech Engn, Rua Mendeleyev 200, BR-13083970 Campinas, SP - Brazil

^[2] Texas Tech Univ, Dept Mech Engn, 805 Boston Ave, Lubbock, TX 79409 - USA

^[3] Shippensburg Univ Penn, Dept Civil & Mech Engn, Shippensburg, PA 17257 - USA

• Tipo de documento: Artigo Científico
• Fonte: MECHANISM AND MACHINE THEORY; v. 149, JUL 2020.
• Citações Web of Science: 0

Resumo

Bearings play a crucial role in machine longevity and is, at the same time, one of the most critical sources of failure in rotor dynamics. Particularly for journal bearings, it is not completely understood how specific damages may influence the response of the rotating system. Consequently, the identification of hydrodynamic bearing faults is challenging. Most of the literature relies on large amounts of training data collections from physical experiments or from the field, which are high in cost. This paper offers a deep learning approach to identify ovalization faults aiming to develop condition monitoring model-based strategies applied to hydrodynamic journal bearings. Therefore, a numerical model was developed to simulate the ovalization fault conditions in order to build training datasets. Afterwards, a deep convolutional neural network algorithm was trained with the generated datasets and used to predict the faults conditions. Finally, the identification performance was evaluated statistically regarding the true-positive identification by both probability density function and subjective logic. The classification accuracy showed promising results for training the machine learning algorithms with simulated data. (C) 2020 Elsevier Ltd. All rights reserved. (AU)

• Processo FAPESP: 19/00974-1 - Monitoramento da condição e prognóstico de mancais considerando incertezas
• Beneficiário: Katia Lucchesi Cavalca Dedini
• Modalidade de apoio: Auxílio à Pesquisa - Regular
• Processo FAPESP: 15/20363-6 - Identificação e controle tolerantes a falhas em sistemas rotativos
• Beneficiário: Katia Lucchesi Cavalca Dedini
• Modalidade de apoio: Auxílio à Pesquisa - Temático
• Processo FAPESP: 18/21581-5 - Avaliação experimental de um modelo de falha por desgaste em mancais hidrodinâmicos
• Beneficiário: Diogo Stuani Alves
• Modalidade de apoio: Bolsas no Brasil - Pós-Doutorado