Thèse soutenue

Jumelage numérique et analyse transitoire pour le diagnostic des défauts dans les systèmes de ventilation

FR  |  
EN
Auteur / Autrice : Mohammad Yakhni
Direction : Anas SakoutHassan AssoumSébastien CauëtMohamad El-Gohary
Type : Thèse de doctorat
Discipline(s) : Image, signal et automatique
Date : Soutenance le 30/11/2023
Etablissement(s) : La Rochelle en cotutelle avec Beirut arab university
Ecole(s) doctorale(s) : École doctorale Euclide (La Rochelle ; 2018-....)
Partenaire(s) de recherche : Laboratoire : Laboratoire des Sciences de l’Ingénieur pour l’Environnement (La Rochelle)
Jury : Président / Présidente : Xavier Brun
Examinateurs / Examinatrices : Anas Sakout, Hassan Assoum, Sébastien Cauët, Mohamad El-Gohary, Xavier Brun, Sohair Rezeka, Amani Raad, Nezha Maamri-Trigeassou
Rapporteurs / Rapporteuses : Xavier Brun, Sohair Rezeka

Résumé

FR  |  
EN

This research discusses condition monitoring in the industry of the future using the digital twinning principle. We begin by creating a digital twin of the actual system, which is a fan motor system used for ventilation in industrial settings. By analyzing the motor current, we identify system faults, without any additional sensors. We study the common system defects. The study is based on transient signal analysis. We developed two approaches to detect defects. The first one is based on Adaptive Notch Filtering and track the frequency of defects at varying operating speeds, enabling their identification and intensity determination. This approach is tested by simulations, demonstrating its effectiveness in online condition monitoring, with some limitations. The second approach involves combining several techniques, including Hilbert transform, generalized demodulation, Vold Kalman filtering, and fast Fourier transform. It aims to enhance fault detection by leveraging the strengths of each technique. We validate this approach through simulations and experimental testing at variable speeds. A notable advance in this research is the incorporation of a self-adjusting mechanism based on real-time operating speed. This innovative feature enables Filtering technique to automatically adapt its parameters, without the need for human intervention. The results show excellent performance and effectiveness in achieving the desired goal. In conclusion, our study showcases how digital twinning and transient analysis can contribute to condition monitoring in the future industry, with a specific focus on detecting system defects through motor current signal analysis.