Data Scarcity in Fault Detection for Solar Tracking Systems: the Power of Physics-Informed Artificial Intelligence
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Combining physical and domain knowledge in artificial intelligence (AI) models has been gaining attention in various fields and applications.
Applications in machine prognostics and health management (PHM) are natural candidates for such hybrid approaches. In particular, they can be efficiently exploited for high fidelity anomaly detection in technical and industrial systems. A natural way for hybridization is using physical models to generate representative data for the training of AI models. Depending on the level of domain knowledge availability, data augmentation can compensate for scarcity of real data from the field. This is particularly attractive for anomaly detection tasks, in which data from the abnormal regimes is limited by definition. On top of this inherent data limitation, many real-world systems suffer from data limitations even within the normal regimes.
In this paper we propose a physics-informed deep learning algorithm for fault detection in grid scale photovoltaic power plants. We focus on a common data scarce scenario that emerges from a low asset monitoring granularity: instead of monitoring the power production of each solar string, the power output is monitored only at combiner-box or even inverter level (monitoring a large number of strings with a single sensor). As a result, the signatures of single local faults can become very subtle and challenging to detect. We show that in this case a physics-informed AI approach significantly outperforms the alternative of a purely data-driven anomaly detection model. This enables high fidelity fault detection in larger solar power plants, that are often limited in the granularity of their condition monitoring data.
How to Cite
##plugins.themes.bootstrap3.article.details##
physics informed deep learning, anomaly detection, fault detection, artificial intelligence, solar power plant, data scarcity, data augmentation
Chao, M. A., Kulkarni, C., Goebel, K., & Fink, O. (2019).
Hybrid deep fault detection and isolation: Combining deep neural networks and system performance models. arXiv preprint arXiv:1908.01529.
Chao, M. A., Kulkarni, C., Goebel, K., & Fink, O. (2022).
Fusing physics-based and deep learning models for prognostics. Reliability Engineering & System Safety, 217, 107961.
Chen, Z., Chen, Y., Wu, L., Cheng, S., & Lin, P. (2019). Deep residual network based fault detection and diagnosis of photovoltaic arrays using current-voltage curves and ambient conditions. Energy Conversion and Management, 198, 111793.
Chine, W., Mellit, A., Lughi, V., Malek, A., Sulligoi, G., & Pavan, A. M. (2016). A novel fault diagnosis technique for photovoltaic systems based on artificial neural networks. Renewable Energy, 90, 501–512. Daliento, S., Chouder, A., Guerriero, P., Pavan, A. M., Mellit, A., Moeini, R., & Tricoli, P. (2017). Monitoring, diagnosis, and power forecasting for photovoltaic fields: A review. International Journal of Photoenergy, 2017. Frank, S., Heaney, M., Jin, X., Robertson, J., Cheung, H., Elmore, R., & Henze, G. (2016). Hybrid modelbased and data-driven fault detection and diagnostics for commercial buildings (Tech. Rep.). National Renewable Energy Lab.(NREL), Golden, CO (United States). Gao, W., & Wai, R.-J. (2020). A novel fault identification method for photovoltaic array via convolutional neural network and residual gated recurrent unit. IEEE access, 8, 159493–159510. Huber, L. G., Palm´e, T., & Chao, M. A. (2023). Physicsinformed machine learning for predictive maintenance: applied use-cases. In 2023 10th ieee swiss conference on data science (sds) (pp. 66–72). Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422440. Kohtz, S., Xu, Y., Zheng, Z., & Wang, P. (2022). Physics-informed machine learning model for battery state of health prognostics using partial charging segments. Mechanical Systems and Signal Processing, 172, 109002. Li, B., Delpha, C., Diallo, D., & Migan-Dubois, A. (2021). Application of artificial neural networks to photovoltaic fault detection and diagnosis: A review. Renewable and Sustainable Energy Reviews, 138, 110512. Li, W., Zhang, J., Ringbeck, F., J¨ost, D., Zhang, L., Wei, Z., & Sauer, D. U. (2021). Physics-informed neural networks for electrode-level state estimation in lithiumion batteries. Journal of Power Sources, 506, 230034. Mansouri, M., Trabelsi, M., Nounou, H., & Nounou, M. (2021). Deep learning based fault diagnosis of photo-
voltaic systems: A comprehensive review and enhancement prospects. IEEE Access. Mellit, A., Tina, G. M., & Kalogirou, S. A. (2018). Fault detection and diagnosis methods for photovoltaic systems: A review. Renewable and Sustainable Energy Reviews, 91, 1–17. Pillai, D. S., & Rajasekar, N. (2018). A comprehensive review on protection challenges and fault diagnosis in pv systems. Renewable and Sustainable Energy Reviews, 91, 18–40. Racharla, S., & Rajan, K. (2017). Solar tracking system–a review. International journal of sustainable engineering, 10(2), 72–81. Rai, A., & Mitra, M. (2021). A hybrid physics-assisted machine-learning-based damage detection using lamb wave. S¯adhan¯a, 46(2), 64. Rausch, R. T., Goebel, K. F., Eklund, N. H., & Brunell,
B. J. (2005). Integrated In-Flight Fault Detection and Accommodation: A Model-Based Study. In Volume 1: Turbo expo 2005 (pp. 561–569). ASME. doi:
10.1115/GT2005-68300
Triki-Lahiani, A., Abdelghani, A. B.-B., & Slama-Belkhodja,
I. (2018). Fault detection and monitoring systems for photovoltaic installations: A review. Renewable and Sustainable Energy Reviews, 82, 2680–2692. Wu, Y., Sicard, B., & Gadsden, S. A. (2024). A review of physics-informed machine learning methods with applications to condition monitoring and anomaly detection. arXiv preprint arXiv:2401.11860. Zgraggen, J., Guo, Y., Notaristefano, A., & Goren Huber,
L. (2022). Physics informed deep learning for tracker fault detection in photovoltaic power plants. In 14th annual conference of the prognostics and health management society, nashville, usa, 1-4 november 2022 (Vol. 14). Zgraggen, J., Guo, Y., Notaristefano, A., & Goren Huber,
L. (2023). Fully unsupervised fault detection in solar power plants using physics-informed deep learning. In 33rd european safety and reliability conference (esrel), southampton, united kingdom, 3-7 september 2023 (pp. 1737–1745).
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.