ARTIFICIAL NEURAL NETWORK BASED FAULT DETECTION AND CLASSIFICATION METHOD FOR AIR CONDITIONERS

Abstract

Air Conditioners (AC)s are devices that balance the air exchange and humidity rate as well as provide heating/cooling functions in order to keep the temperature of the environment within the desired conditions and needs. In this study, a new fault detection and classification method for AC is proposed. The method is based on the fact that power consumptions of appliances imply significant information about the appliances’ health. Hence, according to the proposed method, power profiles of considered AC are created during its operations. Artificial Neural Network (ANN) configuration of AC is specifically designed and trained by created power profiles. Trained ANN is used to detect and classify faults in the present power profile before major malfunctions occur. Taking action against detected faults helps prevent increased power consumption and serious security issues. Performance and efficiency of the Method designed for classification and detection of errors is between 95.1% - 97.01%

Keywords

• Air Conditioner
• Fault detection
• Fault classification
• Power profile
• Artificial Neural Network

References

1. [1] https://www.apple.com/lae/ios/home/. Access date: 26.02.2019.
2. [2] https://developer.amazon.com/en-US/alexa. Access date: 20.02.2022.
3. [3] https://www.smartthings.com/. Access date: 26.05.2021.
4. [4] https://xiaomi-mi.com/mi-smart-home/. Access date: 26.02.2022.
5. [5] https://www.apple.com/lae/ios/health/. Access date: 26.04.2021.
6. [6] Gupta A, Gupta HP, Biswas B and Dutta T. An unseen fault classification approach for smart appliances using ongoing multivariate time series. IEEE Transactions on Industrial Informatics 2020; 17.6 3731-3738.
7. [7] Prist M, Monteriù A, Freddi A, Pallotta E, Ciabattoni L, Cicconi P, ... & Longhi S. Machine learning-as-a-service for consumer electronics fault diagnosis: A comparison between Matlab and Azure ML. In 2020 IEEE International Conference on Consumer Electronics (ICCE) (pp. 1-5), IEEE, 2020.
8. [8] Fernandes S, Antunes M, Santiago AR, Barraca JP, Gomes D and Aguiar RL. Forecasting appliances failures: A machine-learning approach to predictive maintenance. Information 2020; 11(4) 208.

9. [9] Yang H, Yang Z, Yang H and Xie Y. Fault detection for air conditioner using PCANet. In 2019 Chinese Control Conference (CCC), pp. 3363-3366, IEEE, 2019.
10. [10] Xu X, Chen T, and Minami M. Intelligent fault prediction system based on internet of things. Computers and Mathematics with Applications, 2012;64(5), 833-839,.
11. [11] Hosseini SS, Agbossou K, Kelouwani S, Cardenas A and Henao N. A practical approach to residential appliances on-line anomaly detection: A case study of standard and smart refrigerators. IEEE Access 8, 2020; 57905-57922.
12. [12] Okut H. Bayesian regularized neural networks for small n big p data. Artificial neural networks-models and applications, 2016; 28-48.
13. [13] Da Silva IN, Spatti DH, Flauzino RA, Liboni LHB and Franco S. Artificial Neural Networks. A Practical Course, Springer, 2017.
14. [14] Toplak H, Uzmay I and Yildirim S, An artificial neural network application to fault detection of a rotor bearing system. Industrial Lubrication and Tribology, 2006;58/1,32-44.
15. [15] Vyas NS and Satishkumar D. Artificial neural network design for fault identification in a rotor-bearing system. Mechanism and Machine Theory, 2001;36, 157-175.
16. [16] Beale MH, Hagen MT and Demuth HB, MATLAB Deep Learning Toolbox User's Guide, The Mathworks Inc, 2020.
17. [17] Gouravaraju S, Narayan J, Sauer RA and Gautam SS, A Bayesian regularization-backpropagation neural network model for peeling computations, The Journal of Adhesion 2021; 97/13, 1234-1254.