Abstract

Abstract
Fault diagnosis and prediction are essential tasks in industrial systems to ensure safe and reliable operation. In the literature, several techniques can and should be used for fault prediction and diagnosis in different industrial systems, e.g. Fault Trees, Artificial Neural Networks, Deep Learning, Signal processing..., etc. These techniques also have limitations. One of the main challenges is the availability and quality of data, which can have a significant impact on the accuracy of fault diagnosis and prediction techniques. In addition, the complexity and size of industrial systems can pose significant computational challenges, requiring the development of efficient algorithms and computing techniques. In this thesis, the proposed method is using Bayesian networks with the process of multi-source information fusion to constitute an approach for fault diagnosis and prediction. The combination of the Bayesian network and the multi-source information fusion allows the development of a complete fault diagnosis and prediction system. This approach can take into account various factors such as sensor data, operational conditions, and historical data to generate accurate and reliable predictions. In addition, this approach can used to identify potential faults before they occur, allowing for proactive maintenance and reduced downtime, and easier decision-making. The results of the industrial case study show that the use of Bayesian networks and multi-source information fusion techniques for fault diagnosis and prediction in industrial systems can improve reliability and efficiency.
Keywords: Diagnostics, Fault prediction, Industrial system, Bayesian network, Multi-source information fusion

Information

Item Type:	Thesis
Divisions:	» Laboratory of Management, Maintenance and Rehabilitation Of Facilities and Urban Infrastructure » Faculty of Science and Technology
ePrint ID:	4470
Date Deposited:	2023-10-12
Further Information:	Google Scholar
URI:	https://www.univ-soukahras.dz/en/publication/article/4470

BibTex