Fault detection and diagnosis for Air Handling Unit based on multiscale convolutional neural networks

• Published in: Energy and buildings 2021-04, Vol.236, p.110795, Article 110795
• Main Authors: Cheng, Fanyong, Cai, Wenjian, Zhang, Xin, Liao, Huanyue, Cui, Can
• Format: Article
• Language: English
• Subjects: Air conditioning; Air Handling Unit; Artificial neural networks; Convolutional neural networks; Diagnosis; Fault detection; Fault diagnosis; Feature extraction; HVAC system; Kernels; Multiscale signals; Neural networks; Signal monitoring
• ISSN: 0378-7788; 1872-6178
• DOI: 10.1016/j.enbuild.2021.110795

[Display omitted] •A novel method is proposed to improve diagnosis performance of AHU system with multiscale signals.•The proposed model eliminates the drawbacks of complicated feature engineering requirements.•The proposed model can effectively extract highly discriminative features from multiscale signals.•The proposed model can obtain better diagnosis performance by an end-to-end learning. This paper proposes a novel fault detection and diagnosis (FDD) method using multiscale convolutional neural networks (MCNNs) for Air Handling Unit (AHU) in Heating, Ventilation, and Air conditioning (HVAC) system. In existing works, it is challenging to achieve high diagnosis performance on multiscale monitoring signals from AHU system since the feature extraction methods in these works are not powerful enough. Although the single-scale convolutional neural networks (CNNs) have been adopted to improve the ability of feature extraction in FDD, it remains difficult to extract strong discriminative feature from multiscale monitoring signals only using single-scale kernels. In this paper, a novel MCNNs-based FDD method is proposed with three different scale kernels to improve the ability of feature extraction and the end-to-end learning strategy is adopted to optimize the model of MCNNs. With strong representation ability, the proposed method can capture highly discriminative features, which can help to improve the diagnostic performance of AHU. The proposed method is compared with other five commonly used methods using the measured data from our AHU experiment platform. The comparison results demonstrate that the proposed MCNNs-based FDD method outperforms other FDD methods.