SpringerLink
Log in

An interpretable graph convolutional neural network based fault diagnosis method for building energy systems

  • Research Article
  • Building Systems and Components
  • Published:
Building Simulation Aims and scope Submit manuscript

Abstract

Due to the fast-modeling speed and high accuracy, deep learning has attracted great interest in the field of fault diagnosis in building energy systems in recent years. However, the black-box nature makes deep learning models generally difficult to interpret. In order to compensate for the poor interpretability of deep learning models, this study proposed a fault diagnosis method based on interpretable graph neural network (GNN) suitable for building energy systems. The method is developed by following three main steps: (1) selecting NC-GNN as a fault diagnosis model for building energy systems and proposing a graph generation method applicable to the model, (2) develo** an interpretation method based on InputXGradient for the NC-GNN, which is capable of outputting the importance of the node features and automatically locating the fault related features, (3) visualizing the results of model interpretation and validating by matching with expert knowledge and maintenance experience. Validation was performed using the public ASHRAE RP-1043 chiller fault data. The diagnosis results show that the proposed method has a diagnosis accuracy of over 96%. The interpretation results show that the method is capable of explaining the decision-making process of the model by identifying fault-discriminative features. For almost all seven faults, their fault-discriminative features were correctly identified.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price includes VAT (Germany)

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Abbreviations

b c :

bias

e uv :

edge feature connecting nodes u and v

FI:

feature importance

h tu :

hidden representation of node u at step t

I 0 :

image

M t :

aggregation function

m u :

local structural expression of node u

N(u):

set of neighbor nodes of node u

R :

contribution

S c(I):

class score function

U t :

update function

w c :

weight vector

c :

class

m :

number of nodes

u :

node

t :

step

ASHRAE RP-1043:

American Society of Heating, Refrigerating, and Air-Conditioning Engineers Research Project 1043

ch:

channel

FC:

fully connected layer

G:

graph convolution layer

GC:

graph-level classification

GC-GNN:

graph neural networks based on graph classification

GCN:

graph convolutional network

GNN:

graph neural network

HVAC:

heating, ventilation and air-conditioning

lr:

learning rate

NC:

node-level classification

NC-GNN:

graph neural networks based on node classification

References

  • Ahmad MW, Mourshed M, Mundow D, et al. (2016a). Building energy metering and environmental monitoring.A state-of-theart review and directions for future research. Energy and Buildings, 120: 85–102.

    Article  Google Scholar 

  • Ahmad MW, Mourshed M, Yuce B, et al. (2016b). Computational intelligence techniques for HVAC systems: A review. Building Simulation, 9: 359–398.

    Article  Google Scholar 

  • Apicella A, Isgrò F, Pollastro A, et al. (2023). Adaptive filters in Graph Convolutional Neural Networks. Pattern Recognition, 144: 109867.

    Article  Google Scholar 

  • Beghi A, Brignoli R, Cecchinato L, et al. (2016). Data-driven Fault Detection and Diagnosis for HVAC water chillers. Control Engineering Practice, 53: 79–91.

    Article  Google Scholar 

  • Betkier I, Oszczypała M, Pobożniak J, et al. (2023). PocketFinderGNN: A manufacturing feature recognition software based on Graph Neural Networks (GNNs) using PyTorch Geometric and NetworkX. SoftwareX, 23: 101466.

    Article  Google Scholar 

  • Chen K, Chen S, Zhu X, et al. (2023). Interpretable mechanism mining enhanced deep learning for fault diagnosis of heating, ventilation and air conditioning systems. Building and Environment, 237: 110328.

    Article  Google Scholar 

  • Comstock MC J. E. Braun JE, Bernhard R (1999). Development of analysis tools for the evaluation of fault detection and diagnostics in chillers. Purdue University.

    Google Scholar 

  • Costa A, Keane MM, Torrens JI, et al. (2013). Building operation and energy performance: Monitoring, analysis and optimisation toolkit. Applied Energy, 101: 310–316.

    Article  Google Scholar 

  • Daigavane A, Ravindran B, Aggarwal G (2021). Understanding convolutions on graphs. Distill, https://doi.org/10.23915/distill.00032.

    Google Scholar 

  • Du Z, Chen S, Li P, et al. (2023). Knowledge-extracted deep learning diagnosis and its cloud-based management for multiple faults of chiller. Building and Environment, 235: 110228.

    Article  Google Scholar 

  • Ebrahimifakhar A, Kabirikopaei A, Yuill D (2020). Data-driven fault detection and diagnosis for packaged rooftop units using statistical machine learning classification methods. Energy and Buildings, 225: 110318.

    Article  Google Scholar 

  • Eom YH, Yoo JW, Hong SB, et al. (2019). Refrigerant charge fault detection method of air source heat pump system using convolutional neural network for energy saving. Energy, 187: 115877.

    Article  Google Scholar 

  • Fan C, **ao F, Song M, et al. (2019). A graph mining-based methodology for discovering and visualizing high-level knowledge for building energy management. Applied Energy, 251: 113395.

    Article  Google Scholar 

  • Fan C, Lin Y, Piscitelli MS, et al. (2023). Leveraging graph convolutional networks for semi-supervised fault diagnosis of HVAC systems in data-scarce contexts. Building Simulation, 16: 1499.1517.

    Article  Google Scholar 

  • Gilmer J, Schoenholz SS, Riley PF, et al. (2017). Neural message passing for Quantum chemistry. In: Proceedings of the 34th International Conference on Machine Learning, Sydney, Australia.

    Google Scholar 

  • Han H, Cui X, Fan Y, et al. (2019). Least squares support vector machine (LS-SVM)-based chiller fault diagnosis using fault indicative features. Applied Thermal Engineering, 154: 540–547.

    Article  Google Scholar 

  • Han H, Gu B, Hong Y, Kang J (2011a). Automated FDD of multiple-simultaneous faults (MSF) and the application to building chillers. Energy and Buildings, 43: 2524–2532.

    Article  Google Scholar 

  • Han H, Gu B, Wang T, Li ZR (2011b). Important sensors for chiller fault detection and diagnosis (FDD) from the perspective of feature selection and machine learning. International Journal of Refrigeration, 34: 586–599.

    Article  Google Scholar 

  • Han Y, Li Q, Wang C, et al. (2022). A novel knowledge enhanced graph neural networks for fault diagnosis with application to blast furnace process safety. Process Safety and Environmental Protection, 166: 143–157.

    Article  Google Scholar 

  • Hong Y, Yoon S, Kim Y-S, et al. (2021). System-level virtual sensing method in building energy systems using autoencoder: Under the limited sensors and operational datasets. Applied Energy, 301: 117458.

    Article  Google Scholar 

  • Hu Y, Chen H, **e J, et al. (2012). Chiller sensor fault detection using a self-Adaptive Principal Component Analysis method. Energy and Buildings, 54: 252–258.

    Article  Google Scholar 

  • Jia Y, Wang J, Reza Hosseini M, et al. (2023). Temporal graph attention network for building thermal load prediction. Energy and Buildings, https://doi.org/10.1016/j.enbuild.2023.113507.

    Google Scholar 

  • Jiang W, Luo J (2022). Graph neural network for traffic forecasting: A survey. Expert Systems with Applications, 207: 117921.

    Article  Google Scholar 

  • Katipamula S, Brambley MR (2005a). Methods for fault detection, diagnostics, and prognostics for building systems—A Review, Part I. HVAC&R Research, 11: 3–25.

    Article  Google Scholar 

  • Katipamula S, Brambley MR (2005b). Methods for fault detection, diagnostics, and prognostics for building systems—A review, part II. HVAC&R Research, 11: 169–187.

    Article  Google Scholar 

  • Keramatfar A, Rafiee M, Amirkhani H (2022). Graph Neural Networks: A bibliometrics overview. Machine Learning with Applications, 10: 100401.

    Article  Google Scholar 

  • Kim M, Yoon SH, Domanski PA, et al. (2008). Design of a steady-state detector for fault detection and diagnosis of a residential air conditioner. International Journal of Refrigeration, 31: 790–799.

    Article  Google Scholar 

  • Kindermans PJ, Schütt K, Müller K-R, et al. (2016). Investigating the influence of noise and distractors on the interpretation of neural networks. ar**v: 1611.07270.

    Google Scholar 

  • Kipf TN, Welling M (2016)). Semi-supervised classification with graph convolutional networks. ar**v: 1609.02907.

    Google Scholar 

  • Lee T, Yoon S, Won K (2022). Delta-T-based operational signatures for operation pattern and fault diagnosis of building energy systems. Energy and Buildings, 257: 111769.

    Article  Google Scholar 

  • Li S, Wen J (2014). A model-based fault detection and diagnostic methodology based on PCA method and wavelet transform. Energy and Buildings, 68: 63–71.

    Article  Google Scholar 

  • Li G, Hu Y, Chen H, et al. (2016). An improved fault detection method for incipient centrifugal chiller faults using the PCA-RSVDD algorithm. Energy and Buildings, 116: 104–113.

    Article  Google Scholar 

  • Li G, Hu Y, Chen H, et al. (2017). Data partitioning and association mining for identifying VRF energy consumption patterns under various part loads and refrigerant charge conditions. Applied Energy, 185: 846–861.

    Article  Google Scholar 

  • Li Y, O’Neill Z (2018). A critical review of fault modeling of HVAC systems in buildings. Building Simulation, 11: 953–975.

    Article  Google Scholar 

  • Li D, Zhou Y, Hu G, et al. (2019). Identifying unseen faults for smart buildings by incorporating expert knowledge with data. IEEE Transactions on Automation Science and Engineering, 16: 1412–1425.

    Article  Google Scholar 

  • Li B, Cheng F, Zhang X, et al. (2021a). A novel semi-supervised data-driven method for chiller fault diagnosis with unlabeled data. Applied Energy, 285: 116459.

    Article  Google Scholar 

  • Li G, Yao Q, Fan C, et al. (2021b). An explainable one-dimensional convolutional neural networks based fault diagnosis method for building heating, ventilation and air conditioning systems. Building and Environment, 203: 108057.

    Article  Google Scholar 

  • Li G, Li F, Xu C, et al. (2022a). A spatial-temporal layer-wise relevance propagation method for improving interpretability and prediction accuracy of LSTM building energy prediction. Energy and Buildings, 271: 112317.

    Article  Google Scholar 

  • Li T, Zhou Y, Zhao Y, et al. (2022b). A hierarchical object oriented Bayesian network-based fault diagnosis method for building energy systems. Applied Energy, 306: 118088.

    Article  Google Scholar 

  • Li T, Zhou Z, Li S, et al. (2022c). The emerging graph neural networks for intelligent fault diagnostics and prognostics: A guideline and a benchmark study. Mechanical Systems and Signal Processing, 168: 108653.

    Article  Google Scholar 

  • Li G, Chen L, Fan C, et al. (2023a). Interpretation and explanation of convolutional neural network-based fault diagnosis model at the feature-level for building energy systems. Energy and Buildings, 295: 113326.

    Article  Google Scholar 

  • Li G, Chen L, Liu J, et al. (2023b). Comparative study on deep transfer learning strategies for cross-system and cross-operation-condition building energy systems fault diagnosis. Energy, 263: 125943.

    Article  Google Scholar 

  • Li G, Wang L, Shen L, et al. (2023c). Interpretation of convolutional neural network-based building HVAC fault diagnosis model using improved layer-wise relevance propagation. Energy and Buildings, 286: 112949.

    Article  Google Scholar 

  • Li G, Chen L, Fan C, et al. (2024). Improved convolutional neural network chiller early fault diagnosis by gradient-based feature-level model interpretation and feature learning. Applied Thermal Engineering, 236: 121549.

    Article  Google Scholar 

  • Liang X, Zhu X, Chen S, et al. (2023). Physics-constrained cooperative learning-based reference models for smart management of chillers considering extrapolation scenarios. Applied Energy, 349: 121642.

    Article  Google Scholar 

  • Liu J, Li G, Chen H, et al. (2017). A robust online refrigerant charge fault diagnosis strategy for VRF systems based on virtual sensor technique and PCA-EWMA method. Applied Thermal Engineering, 119: 233–243.

    Article  Google Scholar 

  • Liu J, Shi D, Li G, et al. (2020). Data-driven and association rule mining-based fault diagnosis and action mechanism analysis for building chillers. Energy and Buildings, 216: 109957.

    Article  Google Scholar 

  • Long S, Marjanovic O, Parisio A (2019). Generalised control-oriented modelling framework for multi-energy systems. Applied Energy, 235: 320–331.

    Article  Google Scholar 

  • Lu J, Zhang C, Li J, et al. (2022). Graph convolutional networks-based method for estimating design loads of complex buildings in the preliminary design stage. Applied Energy, 322: 119478.

    Article  Google Scholar 

  • Mirnaghi MS, Haghighat F (2020). Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: A comprehensive review. Energy and Buildings, 229: 110492.

    Article  Google Scholar 

  • Reddy TA (2007). Application of a generic evaluation methodology to assess four different chiller FDD methods (RP-1275). HVAC&R Research, 13: 711–729.

    Article  Google Scholar 

  • Singh V, Mathur J, Bhatia A (2022). A comprehensive review: Fault detection, diagnostics, prognostics, and fault modeling in HVAC systems. International Journal of Refrigeration, 144: 283–295.

    Article  Google Scholar 

  • Sun K, Hong T, Kim J, et al. (2022). Application and evaluation of a pattern-based building energy model calibration method using public building datasets. Building Simulation, 15: 1385–1400.

    Article  Google Scholar 

  • Tang R, Fan C, Zeng F, et al. (2022). Data-driven model predictive control for power demand management and fast demand response of commercial buildings using support vector regression. Building Simulation, 15: 317–331.

    Article  Google Scholar 

  • Tian Y, Wang J, Qi Z, et al. (2023). Calibration method for sensor drifting bias in data center cooling system using Bayesian Inference coupling with Autoencoder. Journal of Building Engineering, 67: 105961.

    Article  Google Scholar 

  • Toub M, Reddy CR, Robinett RD III, et al. (2021). Integration and optimal control of MicroCSP with building HVAC systems: Review and future directions. Energies, 14: 730.

    Article  Google Scholar 

  • Tran DAT, Chen Y, Ao HL, et al. (2016a). An enhanced chiller FDD strategy based on the combination of the LSSVR-DE model and EWMA control charts. International Journal of Refrigeration, 72: 81–96.

    Article  Google Scholar 

  • Tran DAT, Chen Y, et al. (2016b). Comparative investigations on reference models for fault detection and diagnosis in centrifugal chiller systems. Energy and Buildings, 133: 246–256.

    Article  Google Scholar 

  • Vishwanathan SVN, Schraudolph SS, Kondor R, et al. (2010). Graph kernels. Journal of Machine Learning Research, 11: 1201–1242.

    MathSciNet  Google Scholar 

  • Wang P, Yoon S, Wang J, et al. (2019). Automated reviving calibration strategy for virtual in situ sensor calibration in building energy systems: Sensitivity coefficient optimization. Energy and Buildings, 198: 291–304.

    Article  Google Scholar 

  • Wen H, Guo W, Li X (2023). A novel deep clustering network using multi-representation autoencoder and adversarial learning for large cross-domain fault diagnosis of rolling bearings. Expert Systems with Applications, 225: 120066.

    Article  Google Scholar 

  • Wu Z, Pan S, Chen F, et al. (2021). A comprehensive survey on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems, 32: 4–24.

    Article  MathSciNet  Google Scholar 

  • Xu X, **ao F, Wang S (2008). Enhanced chiller sensor fault detection, diagnosis and estimation using wavelet analysis and principal component analysis methods. Applied Thermal Engineering, 28: 226–237.

    Article  Google Scholar 

  • Yan R, Ma Z, Zhao Y, et al. (2016). A decision tree based data-driven diagnostic strategy for air handling units. Energy and Buildings, 133: 37–45.

    Article  Google Scholar 

  • Yan K, Huang J, Shen W, et al. (2020). Unsupervised learning for fault detection and diagnosis of air handling units. Energy and Buildings, 210: 109689.

    Article  Google Scholar 

  • Yang Z, Liu Z, Zhou J, et al. (2023). A graph neural network (GNN) method for assigning gas calorific values to natural gas pipeline networks. Energy, 278: 127875.

    Article  Google Scholar 

  • Ying R, Bourgeois D, You J, et al. (2019)). GNNExplainer: Generating explanations for graph neural networks. ar**v:1903.03894.

    Google Scholar 

  • Yuan H, Yu H, Gui S, et al. (2022). Explainability in graph neural networks: A taxonomic survey. ar**v:2012.15445v3.

    Google Scholar 

  • Zhang R, Hong T (2017). Modeling of HVAC operational faults in building performance simulation. Applied Energy, 202: 178–188.

    Article  Google Scholar 

  • Zhang D, Stewart E, Entezami M, et al. (2020). Intelligent acousticbased fault diagnosis of roller bearings using a deep graph convolutional network. Measurement, 156: 107585.

  • Zhang Y, Yu J (2022). Pruning graph convolutional network-based feature learning for fault diagnosis of industrial processes. Journal of Process Control, 113: 101–113.

    Article  Google Scholar 

  • Zhang C, Tian X, Zhao Y, et al. (2022a). Causal discovery-based external attention in neural networks for accurate and reliable fault detection and diagnosis of building energy systems. Building and Environment, 222: 109357.

    Article  Google Scholar 

  • Zhang H, Li C, Wei Q, et al. (2022b). Fault detection and diagnosis of the air handling unit via combining the feature sparse representation based dynamic SFA and the LSTM network. Energy and Buildings, 269: 112241.

    Article  Google Scholar 

  • Zhang J, **ao F, Li A, et al. (2023). Graph neural network-based spatio-temporal indoor environment prediction and optimal control for central air-conditioning systems. Building and Environment, 242: 110600.

    Article  Google Scholar 

  • Zhao Y, Wang S, **ao F, et al. (2013). A simplified physical model-based fault detection and diagnosis strategy and its customized tool for centrifugal chillers. HVAC&R Research, 19: 283–294.

    Article  Google Scholar 

  • Zhao Y, Li T, Fan C, et al. (2019). A proactive fault detection and diagnosis method for variable-air-volume terminals in building air conditioning systems. Energy and Buildings, 183: 527–537.

    Article  Google Scholar 

  • Zhou Q, Wang S, **ao F (2009). A novel strategy for the fault detection and diagnosis of centrifugal chiller systems. HVAC&R Research, 15: 57–75.

    Article  Google Scholar 

  • Zhou J, Cui G, Hu S, et al. (2020). Graph neural networks: a review of methods and applications. AI Open, 1: 57–81.

    Article  Google Scholar 

  • Zhou Z, Chen H, Li G, et al. (2021). Data-driven fault diagnosis for residential variable refrigerant flow system on imbalanced data environments. International Journal of Refrigeration, 125: 34–43.

    Article  Google Scholar 

Download references

Acknowledgements

This work is jointly supported by the Opening Fund of Key Laboratory of Low-grade Energy Utilization Technologies and Systems (Chongqing University), Ministry of Education of China (LLEUTS202305), the National Natural Science Foundation of China (51906181), the Opening Fund of State Key Laboratory of Green Building in Western China (LSKF202316), the open Foundation of Anhui Province Key Laboratory of Intelligent Building and Building Energy-saving (IBES2022KF11), “The 14th Five Year Plan” Hubei Provincial advantaged characteristic disciplines (groups) project of Wuhan University of Science and Technology (2023D0504), the Wuhan University of Science and Technology Postgraduate Innovation and Entrepreneurship Fund (JCX2022016), and the 2021 Construction Technology Plan Project of Hubei Province (2021-83).

Author information

Authors and Affiliations

  1. School of Urban Construction, Wuhan University of Science and Technology, Wuhan, 430065, China

    Guannan Li, Zhanpeng Yao, Liang Chen, Tao Li & Chengliang Xu

  2. Anhui Province Key Laboratory of Intelligent Building and Building Energy-saving, Anhui Jianzhu University, Hefei, 230601, China

    Guannan Li

  3. State Key Laboratory of Green Building in Western China, **’an University of Architecture & Technology, **’an, 710055, China

    Guannan Li

  4. Key Laboratory of Low-grade Energy Utilization Technologies and Systems (Chongqing University), Ministry of Education of China, Chongqing University, Chongqing, 400044, China

    Guannan Li

  5. Hubei Provincial Engineering Research Centre of Urban Regeneration, Wuhan University of Science and Technology, Wuhan, 430081, China

    Tao Li

Authors
  1. Guannan Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Zhanpeng Yao
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Liang Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Tao Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Chengliang Xu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Guannan Li: original draft, review, editing and funding acquisition. Zhanpeng Yao: manuscript first draft writing and data preparation. Liang Chen: methodology and software. Tao Li: review and editing. Chengliang Xu: review and editing.

Corresponding author

Correspondence to Tao Li.

Ethics declarations

The authors have no competing interests to declare that are relevant to the content of this article.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, G., Yao, Z., Chen, L. et al. An interpretable graph convolutional neural network based fault diagnosis method for building energy systems. Build. Simul. (2024). https://doi.org/10.1007/s12273-024-1125-6

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s12273-024-1125-6

Keywords

Search

Navigation