Monitoring of a Dynamic System Based on Autoencoders

Abstract

Monitoring industrial infrastructures are undergoing a critical transformation with industry 4.0.  Monitoring solutions must follow the system behavior in real time and must adapt to its continuous change. We propose in this paper an autoencoder model-based approach for tracking abnormalities in industrial application. A set of sensors collects data from turbo-compressors and an original two-level machine learning LSTM autoencoder architecture defines a continuous nominal vibration model. Normalized thresholds (ISO 20816) between the model and the system generates a possible abnormal situation to diagnose. Experimental results, including hyper-parameter optimization on large real data and domain expert analysis, show that our proposed solution gives promising results.

Cite

Text

Osmani et al. "Monitoring of a Dynamic System Based on Autoencoders." International Joint Conference on Artificial Intelligence, 2019. doi:10.24963/IJCAI.2019/254

Markdown

[Osmani et al. "Monitoring of a Dynamic System Based on Autoencoders." International Joint Conference on Artificial Intelligence, 2019.](https://mlanthology.org/ijcai/2019/osmani2019ijcai-monitoring/) doi:10.24963/IJCAI.2019/254

BibTeX

@inproceedings{osmani2019ijcai-monitoring,
  title     = {{Monitoring of a Dynamic System Based on Autoencoders}},
  author    = {Osmani, Aomar and Hamidi, Massinissa and Bouhouche, Salah},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year      = {2019},
  pages     = {1836-1843},
  doi       = {10.24963/IJCAI.2019/254},
  url       = {https://mlanthology.org/ijcai/2019/osmani2019ijcai-monitoring/}
}