The transfer learning of uncertainty quantification for industrial plant fault diagnosis system design

Tools

Blair, Jennifer and Amin, O. and Brown, B. D. and McArthur, S. and Forbes, A. and Stephen, B. (2024) The transfer learning of uncertainty quantification for industrial plant fault diagnosis system design. Data-Centric Engineering, 5. e41. ISSN 2632-6736 (https://doi.org/10.1017/dce.2024.54)

[thumbnail of Blair-etal-DCE-2024-The-transfer-learning-of-uncertainty-quantification-for-industrial-plant-fault-diagnosis]
Preview
 Text. Filename: Blair-etal-DCE-2024-The-transfer-learning-of-uncertainty-quantification-for-industrial-plant-fault-diagnosis.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (753kB)| Preview

Abstract

The performance and confidence in fault detection and diagnostic systems can be undermined by data pipelines that feature multiple compounding sources of uncertainty. These issues further inhibit the deployment of databased analytics in industry, where variable data quality and lack of confidence in model outputs is already a barrier to their adoption. The methodology proposed in this paper supports trustworthy data pipeline design and leverages knowledge gained from one fully-observed data pipeline to a similar, under-observed case. The transfer of uncertainties provides insight into uncertainty drivers without repeating the computational or cost overhead of fully redesigning the pipeline. A SHAP-based human-readable explainable AI (XAI) framework was used to rank and explain the impact of each choice in a data pipeline, allowing the decoupling of positive and negative performance drivers to facilitate the successful selection of highly-performing pipelines. This empirical approach is demonstrated on bearing fault classification case studies, using well-understood open-source data.

ORCID iDs

Blair, Jennifer, Amin, O., Brown, B. D., McArthur, S. ORCID logoORCID: https://orcid.org/0000-0003-1312-8874, Forbes, A. and Stephen, B. ORCID logoORCID: https://orcid.org/0000-0001-7502-8129;
CORE (COnnecting REpositories)