Refrigerant charge fault detection method of air source heat pump system using convolutional neural network for energy saving

Yong Hwan Eom, Jin Woo Yoo, Sung Bin Hong and Min Soo Kim

Energy, 2019, vol. 187, issue C

Abstract: In heat pumps, refrigerant leakage is one of the frequent faults. Since the systems have the best performance at the optimal charge, it is essential to predict refrigerant charge amount. Hence, the refrigerant charge fault detection (RCFD) methods have been developed by researchers. Due to improvements in computing speed and big-data, data-driven techniques such as artificial neural networks (ANNs) have been highlighted recently. However, most existing ANN-based RCFD methods use low-performance shallow neural networks (SNNs) and require the features extracted by experts’ experiences. Also, they have some critical limitations. First, they cannot provide quantitative information on recharge amount due to a simple classification such as undercharge or overcharge. Second, many ANN-based RCFD methods can be used in one operation mode (cooling or heating mode). To improve the limitations, a novel RCFD strategy based on convolutional neural networks (CNNs) was suggested. Two prediction models using classification and regression can predict the quantitative refrigerant amount in both cooling and heating mode with a single model. The mean accuracy of the CNN-based classification model was 99.9% for the learned cases. Also, the CNN-based regression model showed the excellent prediction performance with root-mean-square (RMS) error of 3.1% including the untrained refrigerant charge amount data.

Keywords: Heat pump system; Refrigerant charge fault detection; Convolutional neural network; Quantitative prediction; Classification; Regression (search for similar items in EconPapers)
Date: 2019
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (20)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S036054421931549X
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:187:y:2019:i:c:s036054421931549x

DOI: 10.1016/j.energy.2019.115877

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().