Video-based diagnosis of a rolling element bearing using a high-speed camera
Feedback on the Survishno 2023 conference contest
Quentin Leclère (INSA Lyon)
Hugo André (Université Jean Monnet Saint-Etienne)
Jérôme Antoni (INSA Lyon)
Arthur Burel (Safran Helicopter Engines, LaMCoS)
Cécile Capdessus (PRISME - Eure-et-Loir Campus)
Marco Cocconcelli (Università Degli Studi di Modena e Reggio Emilia)
Gianluca D'Elia (Università Degli Studi di Modena e Reggio Emilia)
Alessandro Paolo Daga (Politecnico di Torino)
Jean Luc Dion (ISAE-Supméca)
Mohammed El Badaoui (Université Jean Monnet Saint-Etienne, Safran Tech)
Abdallah El Hidali (Safran Aircraft Engines)
Luigi Garibaldi (Politecnico di Torino)
François Girardin (INSA Lyon)
Julien Griffaton (Safran Aircraft Engines)
Konstantinos Gryllias (Katholieke Universiteit Leuven)
Yunhyeok Han (ISAE-Supméca)
Jan Helsen (Flanders Make, Vrije Universiteit Brussel)
Fadi Karkafi (Safran Helicopter Engines, INSA Lyon)
Kayacan Kestel (Vrije Universiteit Brussel)
Layla Kordylas (ISAE-Supméca)
Deepti Kunte (Siemens PLM Software, Katholieke Universiteit Leuven)
Stefania Lo Feudo (ISAE-Supméca)
Adrien Marsick (INSA Lyon)
Douw Marx (Katholieke Universiteit Leuven)
Alex Ricardo Mauricio (Katholieke Universiteit Leuven)
Johann Miranda-Fuentes (École Centrale de Lyon)
Cédric Peeters (Vrije Universiteit Brussel)
Thomas Poupon (Laboratoire Roberval)
Didier Rémond (LaMCoS)
Franck Renaud (ISAE-Supméca)
Julien Roussel (PRISME - Eure-et-Loir Campus)
Jimmy Touzet (Université Jean Monnet Saint-Etienne)
Toby Verwimp (Katholieke Universiteit Leuven)
Luca Viale (Politecnico di Torino)
Mahsa Yazdanianasr (Katholieke Universiteit Leuven)
Rui Zhu (Katholieke Universiteit Leuven)
More Info
expand_more
Abstract
The aim of this paper is to provide a feedback on the signal processing contest organized at the Survishno/Resonance conference held in 2023 in Toulouse, France. The aim of the competition was to demonstrate the possibility of diagnosing a bearing operating at a variable rotation speed using high-speed video data only. To this end, a video of an operating faulty bearing was proposed to registered people a month before the event, with the task of extracting the instantaneous rotation speeds of the various rotating parts, and proposing a methodology for identifying the type of fault (which was only known by the contest organizers). Ten teams of researchers from academia and industry were then formed, and proposed different approaches, the results of which were compared with reference data. The diagnostic task proved difficult, with none of the teams achieving the correct diagnosis of the fault. However, it is shown in this paper that by crossing the results of the different teams, it was possible to achieve the correct diagnosis. A tutorial is proposed at the end of the paper, presenting the application of a complete processing chain from video data to order envelope spectral analysis. Results illustrate the ability to recover bearing fault signatures, and also show the possibility to enhance the diagnosis by taking advantage of the fine tracking of the position of each part of the system offered by the video.