Interpretation of convolutional neural network-based building HVAC fault diagnosis model using improved layer-wise relevance propagation

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

16 Scopus Citations
View graph of relations

Author(s)

  • Guannan Li
  • Luhan Wang
  • Limei Shen
  • Liang Chen
  • Hengda Cheng
  • And 2 others
  • Chengliang Xu
  • Fan Li

Detail(s)

Original languageEnglish
Article number112949
Journal / PublicationEnergy and Buildings
Volume286
Online published2 Mar 2023
Publication statusPublished - 1 May 2023

Abstract

Convolutional neural networks (CNNs) have been widely utili sed for fault diagnosis (FD) in building heating, ventilation, and air conditioning (HVAC) systems. Despite achieving high accuracy in many HVAC FD tasks, misdiagnosis still occurs. As a black-box model, the CNN FD model and its diagnostic mechanism and decision-making process are opaque, making it difficult for HVAC operators and managers to trust it. To address this, this study proposes an improved Layer-wise Relevance Propagation (ImLRP) method for interpreting CNN FD models in HVACs.The proposed method addresses the issue of preserving positive/negative information from HVAC inputs by adopting a Softsign activation function in the CNN. The feature-matching issue is addressed by excluding pooling layers from the CNN. ImLRP evaluates the contribution of each neuron in the network to the output decision by assigning a relevance score to each neuron in each layer during the backpropagation of the feedforward transmission process. The relevance score difference, a new metric, is used to obtain the net impact of HVAC faults. The proposed method was validated using RP-1043 chiller fault experiment data, which showed a CNN FD accuracy of 96%. Both correct-diagnosis and misdiagnosis were interpreted at the feature variable level, and the study also discussed the influence of the CNN model parameter, ImLRP parameter, and the relevance score difference on the results. © 2023 Elsevier B.V. All rights reserved.

Research Area(s)

  • Building heating, ventilation, and air conditioning (HVAC) system, Fault diagnosis, Convolutional neural networks (CNNs), Layer-wise relevance propagation (LRP), Model interpretation, Quantitative evaluation, STRATEGY, METHODOLOGY

Citation Format(s)

Interpretation of convolutional neural network-based building HVAC fault diagnosis model using improved layer-wise relevance propagation. / Li, Guannan; Wang, Luhan; Shen, Limei et al.
In: Energy and Buildings, Vol. 286, 112949, 01.05.2023.

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review