To main content

A sustainable deep learning framework for fault detection in 6G Industry 4.0 heterogeneous data environments

Abstract

The integration of 5G and Beyond 5G (B5G)/6G in Machine-to-Machine (M2M) communications, is making Industry 4.0 smarter. However, the goal of having a sustainable self-monitored industry has not been reached yet. State-of-the-art deep learning-based Fault Detection algorithms cannot handle heterogeneous data, meaning that more than one fault detection computational device has to be used for each data format, in addition to the inability to take advantage of the combination of all the information available in different formats to derive more accurate conclusions. Moreover, these algorithms rely on inefficient hyper-parameters tuning strategies. In this paper, we propose an Advanced Deep Learning framework for Fault Diagnosis in Industry 4.0 (ADL-FDI4), which combines Long Short Term Memory (LSTM), Convolutional Neural Networks (CNN) and graph CNN (GNN), to handle heterogeneous data. Furthermore, our novel framework uses a Branch-and-Bound procedure to guide the learning process. Our experimental results show that ADL-FDI4 outperforms the state-of-the-art solutions in terms of detection rate and running time, and for that, it consumes less energy. In addition to handling heterogeneous data, which implies that one computational device is sufficient to handle all data formats.
Read the publication

Category

Academic article

Language

English

Author(s)

  • Tinhinane Mezair
  • Youcef Djenouri
  • Asma Belhadi
  • Gautam Srivastava
  • Jerry Chun-Wei Lin

Affiliation

  • SINTEF Digital / Mathematics and Cybernetics
  • Kristiania University of Applied Sciences
  • Western Norway University of Applied Sciences
  • Algeria
  • China Medical University
  • Brandon University

Year

2022

Published in

Computer Communications

ISSN

0140-3664

Volume

187

Page(s)

164 - 171

View this publication at Norwegian Research Information Repository