Transfer Learning Approaches for Wind Turbine Fault Detection using Deep Learning
Wind park operators start to recognize the cost-effectiveness of intelligent maintenance solutions for wind turbines based on the readily available 10-minute SCADA data. In particular, recent advances have shown that deep learning algorithms can enhance the performance and robustness of fault detection algorithms which are fed with such SCADA data. In order to deploy deep learning fault detection algorithms, a large amount of historical data is needed. In case the data is not available for a certain turbine, training the algorithms becomes challenging. The common approaches in this case are referred to as transfer learning or domain adaptation methods, which attempt to allow the transfer of knowledge between different machines.
In this paper we explore the main challenges of domain adaptation for fault detection based on wind turbine SCADA data. We focus on practical use cases, stemming from the commercial need to deploy fault detection algorithms for newly installed turbines, or turbines with little historical data under diverse operating conditions. We analyze different reasons for domain shifts between turbines, which require the development of new domain adaptation approaches beyond the ones familiar for other PHM applications, and present results for several of these challenging cases.
How to Cite
Deep Learning, Domain Adaptation, Fault Detection, Convolutional Neural Networks, Wind Turbines, SCADA Data
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.