Publication type: | Conference paper |
Type of review: | Peer review (publication) |
Title: | Fully unsupervised fault detection in solar power plants using physics-informed deep learning |
Authors: | Zgraggen, Jannik Guo, Yuyan Notaristefano, Antonio Goren Huber, Lilach |
et. al: | No |
DOI: | 10.3850/978-981-18-8071-1_P652-cd |
Proceedings: | Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023) |
Editors of the parent work: | Brito, Mário P. Aven, Terje Baraldi, Piero Čepin, Marko Zio, Enrico |
Page(s): | 1737 |
Pages to: | 1745 |
Conference details: | 33rd European Safety and Reliability Conference (ESREL), Southampton, United Kingdom, 3-7 September 2023 |
Issue Date: | Sep-2023 |
Publisher / Ed. Institution: | Research Publishing |
Publisher / Ed. Institution: | Singapore |
ISBN: | 978-981-18-8071-1 |
Language: | English |
Subjects: | Deep learning; Physics informed machine learning; Solar power; Anomaly detection; Unsupervised machine learning; Fault detection; Predictive maintenance; Data contamination; Data refinement |
Subject (DDC): | 006: Special computer methods 620: Engineering |
Abstract: | Machine learning algorithms for anomaly detection often assume training with historical data gathered under normal conditions, and detect anomalies based on large residuals at inference time. In real-world applications, labelled anomaly-free data is most often unavailable. In fact, a common situation is that the training data is contaminated with an unknown fraction of anomalies or faults of the same type we aim to detect. In this case, training residual-based models with the contaminated data often leads to increased missed detections and/or false alarms. While this challenge is rather common, in particular in technical fault detection setups, it is only rarely addressed in the scientific literature. In this paper we address this problem by introducing a data refinement algorithm that is capable of cleaning the contaminated training data in a fully unsupervised manner, and apply the algorithm to a problem of fault detection in grid-scale solar power plants. The data refinement framework is based on an original physics informed deep learning classification algorithm that would require healthy data as its input, in order to generate from it synthetic faulty data and train a binary classifier. We show that in order to achieve high fault detection performance, it is essential to avoid contamination of the original healthy data with unlabelled faults. To this end, we introduce an algorithm that isolates the healthy data in a fully unsupervised manner prior to training the binary classifier. We test our algorithm with field data from an operational solar power plant which includes contamination of unlabelled faulty data and demonstrate its high performance. In addition, we demonstrate the robustness of the proposed refinement method against an increasing fraction of faults in the training data. |
URI: | https://digitalcollection.zhaw.ch/handle/11475/29066 |
Fulltext version: | Published version |
License (according to publishing contract): | Licence according to publishing contract |
Departement: | School of Engineering |
Organisational Unit: | Institute of Data Analysis and Process Design (IDP) |
Published as part of the ZHAW project: | Intelligente Diagnostik von Leistungseinbussen in Solarkraftwerken |
Appears in collections: | Publikationen School of Engineering |
Files in This Item:
There are no files associated with this item.
Show full item record
Zgraggen, J., Guo, Y., Notaristefano, A., & Goren Huber, L. (2023). Fully unsupervised fault detection in solar power plants using physics-informed deep learning [Conference paper]. In M. P. Brito, T. Aven, P. Baraldi, M. Čepin, & E. Zio (Eds.), Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023) (pp. 1737–1745). Research Publishing. https://doi.org/10.3850/978-981-18-8071-1_P652-cd
Zgraggen, J. et al. (2023) ‘Fully unsupervised fault detection in solar power plants using physics-informed deep learning’, in M.P. Brito et al. (eds) Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023). Singapore: Research Publishing, pp. 1737–1745. Available at: https://doi.org/10.3850/978-981-18-8071-1_P652-cd.
J. Zgraggen, Y. Guo, A. Notaristefano, and L. Goren Huber, “Fully unsupervised fault detection in solar power plants using physics-informed deep learning,” in Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023), Sep. 2023, pp. 1737–1745. doi: 10.3850/978-981-18-8071-1_P652-cd.
ZGRAGGEN, Jannik, Yuyan GUO, Antonio NOTARISTEFANO und Lilach GOREN HUBER, 2023. Fully unsupervised fault detection in solar power plants using physics-informed deep learning. In: Mário P. BRITO, Terje AVEN, Piero BARALDI, Marko ČEPIN und Enrico ZIO (Hrsg.), Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023). Conference paper. Singapore: Research Publishing. September 2023. S. 1737–1745. ISBN 978-981-18-8071-1
Zgraggen, Jannik, Yuyan Guo, Antonio Notaristefano, and Lilach Goren Huber. 2023. “Fully Unsupervised Fault Detection in Solar Power Plants Using Physics-Informed Deep Learning.” Conference paper. In Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023), edited by Mário P. Brito, Terje Aven, Piero Baraldi, Marko Čepin, and Enrico Zio, 1737–45. Singapore: Research Publishing. https://doi.org/10.3850/978-981-18-8071-1_P652-cd.
Zgraggen, Jannik, et al. “Fully Unsupervised Fault Detection in Solar Power Plants Using Physics-Informed Deep Learning.” Proceedings of the 33rd European Safety and Reliability Conference (ESREL 2023), edited by Mário P. Brito et al., Research Publishing, 2023, pp. 1737–45, https://doi.org/10.3850/978-981-18-8071-1_P652-cd.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.