PSI - Issue 80

ScienceDirect Structural Integrity Procedia 00 (2022) 000 – 000 Structural Integrity Procedia 00 (2022) 000 – 000 Available online at www.sciencedirect.com Available online at www.sciencedirect.com ScienceDirect Available online at www.sciencedirect.com ScienceDirect

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

Procedia Structural Integrity 80 (2026) 411–417

© 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Ferri Aliabadi Abstract This study presents a prototype attention domain adaptation framework for explainable bearing fault diagnosis under varying operating conditions. The method aligns class centroids and dynamically weights target samples based on similarity and confidence. It achieves high accuracy with limited labeled data. The framework was evaluated on two benchmark datasets, Paderborn and CWRU, both reaching 100 percent classification accuracy. Interpretability is supported through prototype-to-instance similarity maps and visualizations of feature alignment. © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Professor Ferri Aliabadi Keywords: bearing fault diagnosis; prototype attention; domain adaptation; vibration analysis; explainable artificial intelligence 1. Introduction Rotor systems play a key role in mechanical, electrical, and aerodynamic energy transmission. They are susceptible to imbalance, misalignment, shaft cracks, and bearing damage due to high-speed operation. Bearings carry heavy loads from upstream components such as turbine blades, and even small defects can escalate quickly if left undetected. Failures may occur in the inner race, outer race, or rolling elements, and multiple defects may arise together, making early diagnosis more difficult (Zhao et al., 2023). Among sensing techniques, vibration analysis is the most effective Abstract This study presents a prototype attention domain adaptation framework for explainable bearing fault diagnosis under varying operating conditions. The method aligns class centroids and dynamically weights target samples based on similarity and confidence. It achieves high accuracy with limited labeled data. The framework was evaluated on two benchmark datasets, Paderborn and CWRU, both reaching 100 percent classification accuracy. Interpretability is supported through prototype-to-instance similarity maps and visualizations of feature alignment. © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Professor Ferri Aliabadi Keywords: bearing fault diagnosis; prototype attention; domain adaptation; vibration analysis; explainable artificial intelligence 1. Introduction Rotor systems play a key role in mechanical, electrical, and aerodynamic energy transmission. They are susceptible to imbalance, misalignment, shaft cracks, and bearing damage due to high-speed operation. Bearings carry heavy loads from upstream components such as turbine blades, and even small defects can escalate quickly if left undetected. Failures may occur in the inner race, outer race, or rolling elements, and multiple defects may arise together, making early diagnosis more difficult (Zhao et al., 2023). Among sensing techniques, vibration analysis is the most effective Fracture, Damage and Structural Health Monitoring Prototype ‐ attention domain adaptation for explainable bearing fault diagnosis Nima Rezazadeh a, *, Francesco Caputo a , Antonio Aversano a , Giuseppe Lamanna a , Alessandro De Luca a , Donato Perfetto a Fracture, Damage and Structural Health Monitoring Prototype ‐ attention domain adaptation for explainable bearing fault diagnosis Nima Rezazadeh a, *, Francesco Caputo a , Antonio Aversano a , Giuseppe Lamanna a , Alessandro De Luca a , Donato Perfetto a a University of Campania Luigi Vanvitelli, Department of Engineering – via Roma 29, 81031 Aversa (Italy) a University of Campania Luigi Vanvitelli, Department of Engineering – via Roma 29, 81031 Aversa (Italy)

* Corresponding author. Tel.: +0-000-000-0000 ; fax: +0-000-000-0000 . E-mail address: nima.rezazadeh@unicampania.it * Corresponding author. Tel.: +0-000-000-0000 ; fax: +0-000-000-0000 . E-mail address: nima.rezazadeh@unicampania.it

2452-3216 © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Professor Ferri Aliabadi 2452-3216 © 2023 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Professor Ferri Aliabadi

2452-3216 © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of Ferri Aliabadi 10.1016/j.prostr.2026.02.039