Advanced search
1 file | 10.00 MB Add to list
Author
Organization
Abstract
Solenoid valves are critical components in many process control systems, as their failure is often a root cause for plant shutdowns. Therefore, the ability to predict the remaining useful life (RUL) of solenoid valves is highly desirable. In this paper, a novel data-driven RUL prediction methodology for solenoid valves is proposed, by training deep neural networks on images constructed from raw current signatures. The performance is compared to shallow machine learning algorithms, trained on features which are obtained using domain expertise on the valves. We show that an ensemble of CNN subnetworks (constructed by the AdaNet algorithm) achieves a predictive performance comparable to the feature -based approaches (p = 0.584 for linear regression, p = 0.321 for Gradient Boosted Regression Trees). By focusing on reducing the modeling effort required for constructing features and fixing neural network architectures, this study offers a new and promising approach for solenoid valve prognostics, and thus even more complex systems in the future.
Keywords
alternating current solenoid valve, convolutional neural network, AdaNet, AutoML, RUL prediction, prognostics

Downloads

  • (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 10.00 MB

Citation

Please use this url to cite or link to this publication:

MLA
Mazaev, Tamir, et al. “Data-Driven Prognostics of Alternating Current Solenoid Valves.” 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020), edited by Jianyu Long et al., IEEE, 2020, pp. 109–15, doi:10.1109/phm-besancon49106.2020.00024.
APA
Mazaev, T., Ompusunggu, A. P., Tod, G., Crevecoeur, G., & Van Hoecke, S. (2020). Data-driven prognostics of alternating current solenoid valves. In J. Long, Z. Pu, & P. Ding (Eds.), 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020) (pp. 109–115). Besancon, France: IEEE. https://doi.org/10.1109/phm-besancon49106.2020.00024
Chicago author-date
Mazaev, Tamir, Agusmian Partogi Ompusunggu, Georges Tod, Guillaume Crevecoeur, and Sofie Van Hoecke. 2020. “Data-Driven Prognostics of Alternating Current Solenoid Valves.” In 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020), edited by Jianyu Long, Zhiqiang Pu, and Ping Ding, 109–15. IEEE. https://doi.org/10.1109/phm-besancon49106.2020.00024.
Chicago author-date (all authors)
Mazaev, Tamir, Agusmian Partogi Ompusunggu, Georges Tod, Guillaume Crevecoeur, and Sofie Van Hoecke. 2020. “Data-Driven Prognostics of Alternating Current Solenoid Valves.” In 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020), ed by. Jianyu Long, Zhiqiang Pu, and Ping Ding, 109–115. IEEE. doi:10.1109/phm-besancon49106.2020.00024.
Vancouver
1.
Mazaev T, Ompusunggu AP, Tod G, Crevecoeur G, Van Hoecke S. Data-driven prognostics of alternating current solenoid valves. In: Long J, Pu Z, Ding P, editors. 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020). IEEE; 2020. p. 109–15.
IEEE
[1]
T. Mazaev, A. P. Ompusunggu, G. Tod, G. Crevecoeur, and S. Van Hoecke, “Data-driven prognostics of alternating current solenoid valves,” in 2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020), Besancon, France, 2020, pp. 109–115.
@inproceedings{8669313,
  abstract     = {{Solenoid valves are critical components in many process control systems, as their failure is often a root cause for plant shutdowns. Therefore, the ability to predict the remaining useful life (RUL) of solenoid valves is highly desirable. In this paper, a novel data-driven RUL prediction methodology for solenoid valves is proposed, by training deep neural networks on images constructed from raw current signatures. The performance is compared to shallow machine learning algorithms, trained on features which are obtained using domain expertise on the valves. We show that an ensemble of CNN subnetworks (constructed by the AdaNet algorithm) achieves a predictive performance comparable to the feature -based approaches (p = 0.584 for linear regression, p = 0.321 for Gradient Boosted Regression Trees). By focusing on reducing the modeling effort required for constructing features and fixing neural network architectures, this study offers a new and promising approach for solenoid valve prognostics, and thus even more complex systems in the future.}},
  author       = {{Mazaev, Tamir and Ompusunggu, Agusmian Partogi and Tod, Georges and Crevecoeur, Guillaume and Van Hoecke, Sofie}},
  booktitle    = {{2020 PROGNOSTICS AND SYSTEM HEALTH MANAGEMENT CONFERENCE (PHM-BESANCON 2020)}},
  editor       = {{Long, Jianyu and Pu, Zhiqiang and Ding, Ping}},
  isbn         = {{9781728156767}},
  issn         = {{2166-563X}},
  keywords     = {{alternating current solenoid valve,convolutional neural network,AdaNet,AutoML,RUL prediction,prognostics}},
  language     = {{eng}},
  location     = {{Besancon, France}},
  pages        = {{109--115}},
  publisher    = {{IEEE}},
  title        = {{Data-driven prognostics of alternating current solenoid valves}},
  url          = {{http://dx.doi.org/10.1109/phm-besancon49106.2020.00024}},
  year         = {{2020}},
}

Altmetric
View in Altmetric
Web of Science
Times cited: