Ensemble of LSTM Networks for Fault Detection, Classification, and Root Cause Identification in Quality Control Line

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Jun 29, 2021
Gurkan Aydemir Adem Avcı Mustafa Kocakulak Tahir Bekiryazıcı

Abstract

Industrial systems with multiple subsystems are monitored via various sensors to control the ongoing process. If the number of monitoring signals collected from these sensors is high and the number of faulty samples is low, then the machine learning methods may fail to provide effective solutions for fault detection and root cause identification. This paper proposes an efficient feature selection model based on the regularized LSTM neural networks, and fault detection and classification using an ensemble of binary LSTM classifiers. The model is verified in PHME Data Challenge 2021 which provides quality-control-pipeline monitoring data.

How to Cite

Aydemir, G., Avcı, A., Kocakulak, M., & Bekiryazıcı, T. (2021). Ensemble of LSTM Networks for Fault Detection, Classification, and Root Cause Identification in Quality Control Line. PHM Society European Conference, 6(1), 6. https://doi.org/10.36001/phme.2021.v6i1.3043
Abstract 491 | PDF Downloads 542

##plugins.themes.bootstrap3.article.details##

Keywords

Prognostics and Health Management, Condition Monitoring, LSTM, Deep learning, Regularized Neural Networks

Section
Data Challenge Winners