A novel semi-supervised data-driven method for chiller fault diagnosis with unlabeled data. (1st March 2021)

• Record Type: Journal Article
• Authors: Li, Bingxu; Cheng, Fanyong; Zhang, Xin; Cui, Can; Cai, Wenjian
• Is Part Of: Applied energy. Volume 285(2021)
• Publication Date: 2021-03-01
• Subjects: Fault diagnosis -- Chiller -- Semi-generative adversarial network -- Unlabeled data -- Semi-supervised learning; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2021.116459
• Languages: English
• ISSNs: 0306-2619

Abstract

Graphical abstract: Highlights: A novel data-driven fault diagnosis method is proposed for chiller systems. The method can learn the information of data distribution from unlabeled data. The diagnosis accuracy can be improved significantly when labeled data is limited. The method fills the gap on how to use unlabeled data in chiller fault diagnosis. Abstract: In practical chiller systems, applying efficient fault diagnosis techniques can significantly reduce energy consumption and improve energy efficiency of buildings. The success of the existing methods for fault diagnosis of chillers relies on the condition that sufficient labeled data are available for training. However, label acquisition is laborious and costly in practice. Usually, the number of labeled data is limited and most data available are unlabeled. Most of the existing methods cannot exploit the information contained in unlabeled data, which significantly limits the improvement of fault diagnosis performance in chiller systems. To make effective use of unlabeled data to further improve fault diagnosis performance and reduce the dependency on labeled data, we proposed a novel semi-supervised data-driven fault diagnosis method for chiller systems based on the semi-generative adversarial network, which incorporates both unlabeled and labeled data into learning process. The semi-generative adversarial network can learn the information of data distribution from unlabeled data and this information can help to significantly improve the diagnostic performance. Experimental results demonstrate the effectiveness of the proposed method. Under the scenario that there are only 80 labeled samples and 16, 000 unlabeled samples, the proposed method can improve the diagnostic accuracy to 84%, while the supervised baseline methods only reach the accuracy of 65% at most. Besides, compared with the supervised learning method based on the neural network, the proposed semi-supervised method can reduce the minimal required number of labeled samples by about 60% when there are enough unlabeled samples.