A CNN-based Explainable Fault Diagnosis Model for Gearboxes in Rotating Machinery

Daoguang Yang, Hamid Reza Karimi, Len Gelman

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Convolutional neural networks(CNN) as a class of deep neural networks are attracting remarkable attention due to their powerful feature extraction capability in various areas such as gearbox fault diagnosis in rotary machinery. Although the identification performance of the CNN has demonstrated a superiority over traditional approaches, it is difficult to explain which parts of the inputs to the CNN are learned by this black box model. Hence, understanding the relationship between the inputs and the deep learning models will help to establish connection to the physical meaning of the fault diagnosis, contributing to the broad acceptance of deep learning models as a trustworthy complement to physical-based reasoning by human experts. In this paper, using Gradient-weighted Class Activation Mapping++ (Grad-CAM++) as the interpreter, the CNNs trained by two-dimensional time-frequency domain signal are interpreted by Grad-CAM++ to show the attention part in these signals by the CNN model.

Original languageEnglish
Title of host publication18th International Conference on Condition Monitoring and Asset Management
Subtitle of host publicationCM 2022 The Future of Condition Monitoring
PublisherBritish Institute of Non-Destructive Testing
Number of pages7
ISBN (Electronic)9781713862277
Publication statusPublished - 7 Jun 2022
Event18th International Conference on Condition Monitoring and Asset Management: The Future of Condition Monitoring - London, United Kingdom
Duration: 7 Jun 20229 Jun 2022
Conference number: 18


Conference18th International Conference on Condition Monitoring and Asset Management
Abbreviated titleCM 2022
Country/TerritoryUnited Kingdom

Cite this