Retour à l'affichage de la conférence
Europe/Paris Français S'authentifier

RESONANCE 2023

→ Europe/Paris
Toulouse

Toulouse

10 Avenue Edouard Belin, 31400 Toulouse FRANCE
Guilhem MICHON (ISAE - DMSM), Leonardo SANCHES
Description

Important Updates and Deadlines

  • Wifi connection at ISAE-Supaero: please find the instructions here.
     
  • Stay up-to-date and make the most of your conference experience by accessing the congress program, pratical informations and real time notifications through a smartphone application.

    Download the "Eventool" app on your Android or Apple device and search for "RESONANCE CONFERENCE 2023".

    We are excited to connect with you through the app!
     
  • Check out the final program of the conference here!
     
  • The Technical Exhibition page has just been updated. Check it out here.
     
  • Registration deadline EXTENDED: June, 23th (00:00 at UTC+2).
    No further extension will be considered!
     
  • Survishno Contest: click here for more information!

Conference Goal

We are pleased to organize the RESONANCE Conference in Toulouse, France, from July 10th to 13th 2023. This conference aims bringing together researchers, scientists and industry professionals to provide a forum with the most recent advances in the fields of structure vibrations, structural health monitoring, and acoustics and the opportunity for interacting in a collegial atmosphere.

 

The RESONANCE Conference aims at gathering the following subevents together at the same time and place:

  • SURVISHNO (SURVeillance, VIbration, SHocks and NOise)
  • JISFA (Industry Days of the French Acoustical Society)
  • JJCAB (Young Researchers in Acoustics, Vibration and Noise)
  • RESONANCE SUMMER CAMP (Research on Nonlinear Vibration Control Enhancement)
  • TECHNICAL EXHIBITION

 

SURVISHNO

SURVISHNO is a joint organization of the conferences Surveillance, VISHNO (Vibration Shocks and Noise) and EVA (Experimental Vibration Analysis).

Participation: 
Submission of Extended Abstract or paper for oral presentation in english

Important dates:
February 1st 2023: Deadline for submission of abstracts 
June 1st 2023: Deadline for submission of complete paper or extended abstract

More details:  here

 

 

 

 

 

 

 

 

 

JISFA

JISFA is a Industrial Day from the French Society of Acoustics. 

Participation:
Invited Industries are sharing their scientific advances in oral presentations 

More details: here

 

 

 

 

 

 

 

 

JJCAB

   JJCAB  

JJCAB is an scientific day organized by the PhDs for the PhDs students.

Participation:
Flash oral presentation and poster presentation

Important dates:
April 15th 2023:   Opening of the abstract submission
May 15th 2023:  Abstract submission deadline

More details: here

 

 

 

 

 

 

 


 

 

Resonance Summer Camp

Resonance Summer Camp aims at gathering PhD researchers to develop collaborative scientific projects in the field on Nonlinear Dynamics.

Participation:
Candidate application

Important dates:
June 5th -16th

More details: here

 

 

 

 

 

 

 

 

Technical Exhibition

Technical Exhibition

Technical Exhibition displays the most recent technologies, products and software for application in acoustic and vibration domain, provided by the companies and corporate groups.

Participation:
Start-up, micro to medium companies and corporate groups

Exhibitors' activities:
Instruments, Measurement systems, Acoustic and vibration control technologies, Software

More details: here

 

 

 

 

 

 

 

 

Conference location

The conference will be held in the amphitheaters at the "Maison de Formation Jacqueline Auriol" (MFJA) and at the ISAE-SUPAERO, Toulouse, France.

 

Registration fees  

  • Early-Bird fees

500 Euros

  • Regular fees

600 Euros

  • Student fees

100 Euros

 

 

 

 

 

 

 

The Conference registration fee includes:

  • Participation in all oral sessions
  • Entrance to the exhibition/poster session area
  • Access to full conference abstracts proceedings
  • Refreshment breaks
  • Tickets for the midday meals
  • One ticket for the reception dinner
  • One ticket for the gala dinner

 

 Follow us on Twitter : 
 
 #RESONANCE2023

Film Destination Toulouse.mp4 Resonance_ISAE_WiFi.pdf
Participants
  • Abdulaziz Shabara
  • Adam Pinto
  • Agnes TRINCAL
  • Alexandre DUCHEVET
  • Alexandre Marois
  • Allan Armougum
  • Amandine Audouy
  • Anabel Fernandez Alvarez
  • Anaïs Servais
  • Anca Stratulat
  • ANDRE Jean Marc
  • Angelotti Giorgio
  • Anke Brock
  • Anthony VACHER
  • Arnaud Violland
  • Asao Kobayashi
  • Aurore Moskwa
  • Aurélie CAMPAGNE
  • Axel Schulte
  • Benjamin Berton
  • Benjamin Hari
  • Benjamin Whitworth
  • Benoit LE BLANC
  • Bernhard Conzelmann
  • Bertille Somon
  • BRAS Philippe
  • Brian Hilburn
  • Bruno Stoufflet
  • Camille Sigoillot
  • Caroline Chanel
  • Caroline Creane
  • Charles Fricaud
  • CHAZAPIS SPYRIDON
  • Chouchane Lisa
  • Christelle Duchossoy
  • Christine GRIS
  • Christoph Vernaleken
  • Christophe Hurter
  • Christophe Lounis
  • Cindie De Faria
  • Claire McDermott Ealding
  • CLAUDINE MELAN
  • Corinne CHABANNE
  • Cyril Camachon
  • Cédric Albinet
  • David Doull
  • Denys Bernard
  • Desislava Georgieva
  • Diana Donath
  • Dominique Fohr
  • Ednei Romano
  • Emilie Jahanpour
  • Emilien Dubois
  • Emmanuelle Escorihuela
  • Engelbert Hartmann
  • ENRIQUE LASO
  • Eric BERNARD
  • Eve Fabre
  • Fabrice Jimenez
  • Fanny ROME
  • Feriel Dacine Ainouz
  • Florence Paris
  • Francesco Frattolillo
  • frederic Beltran
  • Frederic Gouezec
  • FREDERIC LEBOEUF
  • frederic noguera
  • Gaelle MARIS
  • georges sangis
  • Gilles Tuduri
  • Grégoire Petitot
  • Helene LESIRE-CABANIOLS
  • Herimanana ZAFIHARIMALALA
  • Hervé Girod
  • Huu Nghia HOANG
  • Ibtihel DHAYA
  • Irina Alyanova
  • Irina Illina
  • Ismaël Aït Ouali
  • Ivan BURDUN
  • Ivan Kesedzic
  • Jade Guenot
  • Jean Louis Gueneau
  • Jean-François Bonnefon
  • Jean-Paul IMBERT
  • Jean-Samuel de Visme
  • JEREMY GUERIN
  • Joanna KOKOU
  • Jonathan Deniel
  • Jonay Ramon Alaman
  • Julia Behrend
  • Julie Juaneda
  • Julien SENDRÈS
  • Jérôme RANC
  • Jérôme TRUC
  • Kresimir Cosic
  • Laura Begalli
  • Laura Christon
  • Laura Salvan
  • Lauren Dargent
  • Laurent Chaudron
  • Lucas Cavatoni
  • Manuel Mercier
  • Marc Baumgartner
  • Marc GATTI
  • Marcel Hinss
  • Marianne Jarry
  • Marie-Line BERGEONNEAU
  • Marius Dudek
  • Mate Gambiraza
  • Matthieu Mogno
  • Maud Dupuy
  • Maxime Bardou
  • Maxime Péré
  • Maëlis LEFEBVRE
  • Michel LENE
  • Mick Salomone
  • Mickaël Causse
  • Milan Vrdoljak
  • Mohammad Radaei
  • MONICA LAVOYER ESCUDEIRO
  • Monika Litvova
  • Nasrine DAMOUCHE
  • Nelly Chouvy
  • Nicolas Devaux
  • Nicolas Drougard
  • Nikolaos Pernientakis
  • Noha BACHIRI TAOUFIQ
  • Olivier Bartheye
  • Olivier Joubert
  • Patin Bruno
  • Patrick FABIANI
  • Patrick Ringeard
  • Pauline Muñoz
  • PIERRE FOSSIER
  • Priya Doobaree
  • Quentin Laborde
  • Quentin Montardy
  • rafail tsifetakis
  • Raphaël CHIRAC
  • Raphaëlle Roy
  • Rodolphe Bonet
  • Rodolphe Wiese
  • Romain Kiba Bassandi
  • Samy Chikhi
  • Sandro Guidetti
  • Sean Caufield
  • Sebastien Tremblay
  • Simon Schwerd
  • Sinisa Popovic
  • Sonia Dodd
  • SORET Rébaï
  • Soufiane KHAIT
  • Stéphane DURAND
  • Sylvain Yan
  • Sébastien Michel
  • Thibault Lahire
  • Thierry Fautrel
  • Thierry ROZEC
  • Thomak LEDUC
  • Thomas de Groot
  • Théo de la Hogue
  • Udo Indlekofer
  • Valentin Ligier
  • Vesa Keinänen
  • Vincent Tardan
  • Virginie BAYET
  • Vsevolod Peysakhovich
  • Weifei Xie
  • Wided Oueslati
  • yann duval
  • Yann LETORT
  • Zoe Gozzi
Resonance 2023
  • lundi, 10 juillet
    • 08:00 → 08:30
      Café d'accueil 30m MFJA

      MFJA
    • 08:30 → 09:00
      Introduction & Ouverture JJCAB 30m MFJA

      MFJA
    • 09:00 → 09:40
      JJCAB1 MFJA

      MFJA

      • 09:00
        "JJCAB1#1 - Optimisation topologique de patchs viscoélastiques contraints pour aubes fan de turbomachine" 5m

        Au sein d’un turbomoteur, le module Fan est soumis à différents phénomènes vibratoires pouvant conduire, dans des cas extrêmes, à une ruine par fatigue des aubages. Avec le besoin de concevoir des moteurs de plus en plus performants, l’optimisation des géométries d’aube et des architectures moteur accroît la sensibilité à ces phénomènes vibratoires. De nouvelles approches doivent donc être mises en place afin de les maîtriser. Parmi ces innovations, le contrôle vibratoire passif par patchs viscoélastiques est une solution dont l’efficacité a déjà été démontrée dans de nombreuses applications industrielles.
        L'objectif de ce travail est de maitriser cette technologie pour une application sur des aubes fan en rotation. Cet objectif implique d’une part le développement de modèles numériques prédictifs permettant d’estimer l’amortissement apporté par le patch et d’autre part d’optimiser le patch afin de maximiser l’amortissement ajouté sans détériorer les propriétés mécaniques et aérodynamiques de l’aube.
        Pour cela, un modèle Eléments Finis 3D basé sur un calcul de réponse harmonique linéaire autour d’un état précontraint est développé. La précontrainte due au chargement centrifuge est obtenue en résolvant un problème statique prenant en compte les non-linéarités géométriques et matérielles. De plus, les propriétés du patch sont modélisées à travers une loi viscoélastique fractionnaire pour prendre en compte la dépendance fréquentielle et un modèle hyperélastique néo-hookéen pour représenter le comportement non-linéaire. Une approche de réduction adaptée à ce type de problème est mise en œuvre pour réduire le coût du calcul harmonique.
        Un processus d’optimisation topologique basé sur le modèle Eléments Finis de la structure est par ailleurs mis en œuvre. La fonction objectif du processus est définie comme la maximisation de certains amortissements modaux sous une contrainte de masse, afin d’optimiser l’amortissement ajouté sur un ensemble de modes choisis dans les conditions de fonctionnement voulues.

        Orateur: Matteo Couet
      • 09:05
        "JJCAB1#2 - Étude de l’utilisation des PINNs et de la réduction de modèle en dynamique des structures pour une intégration dans un dispositif immersif" 5m

        Dans le cadre du projet ANR JENII, ce travail de recherche s'intéresse au développement de jumeaux numériques de structures mécaniques assemblées utilisables au sein de dispositifs immersifs et interactifs (réalité virtuelle) pour des applications d'enseignement. Ces derniers doivent être capables de refléter la physique du système réel (tels que des structures aéronautiques) afin d’illustrer des notions importantes de conception mécanique pour des apprenants lors de travaux pratiques. L’objectif de cette étude est de fournir ces résultats expérimentaux simulés crédibles en temps interactif (déformations, accélérations…). Cela se traduit donc par la résolution efficace du problème mécanique haute dimension, potentiellement non-linéaire, pour plusieurs configurations de chargements (intensité, localisation) et/ou liées à des variations de paramètres de conception (géométrie, conditions aux limites). La méthode développée doit notamment être robuste face à des problèmes sous sollicitation dynamique (choc) qui font intervenir le régime transitoire du système étudié.

        La réduction de modèle paramétrique (POD, PGD) et notamment l’écriture sous forme de variables séparées de la solution semble être une réponse à cette problématique malgré des limites dans la reconstruction de la réponse transitoire des structures. L’utilisation du Deep Learning est également envisagée avec le paradigme PINN (Physic-Informed Neural Network) qui allie les avancées technologiques du Machine Learning avec les équations de la mécanique. De plus, ce formalisme présente également l’avantage de pouvoir intégrer directement des données expérimentales pour le recalage et l’ajout d’incertitudes dans le modèle entraîné. Cependant, le passage à l’échelle pour la résolution de problème dynamique 3D ainsi que la création d’un modèle paramétrique forment un verrou scientifique.

        Cette présentation résume donc les recherches effectuées dans ces deux champs à travers des exemples académiques représentatifs pour aboutir à des premières conclusions.

        Orateur: Florian DUPONT
      • 09:10
        "JJCAB1#3 - Prédiction de l’amortissement de structures metalliques remplies de matériau de viscoélastique" 5m

        La signature accoustique des navires et des sous-marins provient, pour une large part, des vibrations structurelles induites par les équipements internes ou l’écoulement. L’optimisation accoustique des structures navales repose donc sur la minimisation des vibrations alors que d’autres contraintes d’architecture plus globales incitent à une réduction de masse. Une des voies de recherche pour atteindre ce double objectif consiste à évider les strucutres et à les remplir d’un matériau moins dense capable d’amortir les vibrations. L’objectif de la thèse est d’évaluer quantitativement et de justifier le gain de performance vibro-acoustique apporté par certains matériaux de remplissage.

        Une solution pour quantifier l’impact d’un matériau de remplissage sur l’amortissement vibratoire est d’étudier les caractéristiques modales de la structure remplie. Cependant le comportement viscoélastique des matériaux utilisés induit une dépendance en fréquence qui complexifie la résolution du problème aux valeurs propres. De plus, la méthode MSE (Modal Strain Energy) est rendue inefficace par la présence de modes internes liés à la géométrie utilisée.

        Trois méthodes numériques, sont donc implémentées et comparées. La résolution directe du problème, utilisée comme solution de référence, et la méthode dite de Base Réduite permettent de tracer des fonctions de réponses en fréquences à partir desquelles il est possible de déterminer les caractéristiques modales de la structure. La dernière méthode est un modèle à variables internes utilisé pour calculer directement les modes complexes de la structure. Les caractéristiques modales issues des calculs sont comparées à des résultats expérimentaux afin de valider la modélisation du matériau de remplissage. Les modèles pourront ensuite être utilisés pour faire de l’identification inverse ou de l’optimisation des propriétés matériaux.

        Orateur: Matthieu Marion
      • 09:15
        "JJCAB1#4 - Simulation numérique du bruit d’alimentation des moteurs électriques par MLI" 5m

        Les moteurs électriques peuvent fonctionner à vitesse variable grâce à un convertisseur qui transforme la tension continue délivrée par une batterie en une tension alternative par Modulation en Largeur d’Impulsion (MLI). Cette technique de découpe repose sur la comparaison entre un signal pseudo-sinusoïdal et une porteuse à haute fréquence, source d’harmoniques perçues comme très gênantes.

        Un des objectifs de la thèse consiste à concevoir un modèle de simulation numérique du bruit électromagnétique des moteurs électriques, afin de mieux appréhender les problématiques acoustiques liées à la MLI. L’enjeu est de réaliser ce modèle dans le domaine temporel, pour rendre possible l’étude des stratégies de découpe aléatoires, et faciliter les études de perception. Un modèle Simulink a été développé pour simuler le bruit rayonné par le stator à partir des signaux d’alimentation en tension. Il se base sur une formulation analytique de la force électromagnétique générée par l’interaction entre le champ magnétique du rotor et les 3 phases de courant du stator. Afin de valider le modèle, un banc d’essai a été adapté à l’étude. Piloté par un script Python, il permet d’alimenter un moteur par des stratégies de découpe à la fois déterministes et aléatoires. Ainsi, une comparaison directe est possible entre les signaux générés par Simulink et les enregistrements réalisés sur le banc d’essai.

        A ce jour, une étude perceptive est en préparation pour évaluer la qualité des signaux simulés par rapport aux signaux réels. Le modèle de simulation permettra dans la suite des travaux de faire un lien entre la nature des stratégies de découpe et le caractère désagréable du bruit. L’objectif final étant de proposer une nouvelle stratégie de découpe, présentant un bon compromis entre qualité sonore et rendement du convertisseur.

        Orateur: Salomé Wanty
      • 09:20
        "JJCAB1#5 - Modelling of a resonant cavity with sonic black holes using patch transfer functions" 5m

        Noise control in room acoustics is of paramount importance for the comfort of dwelling inhabitants, automobile passengers, and rail and aircraft pilots and passengers. The broadband dissipation of noise is a challenging task, especially at low frequencies, and many options have been proposed in the past decades. A relatively recent one is that of acoustic black holes (ABHs). An ABH is a passive and lightweight device for the control of noise and vibrations. It basically consists of a retarding wave guideline that slows down impinging waves and concentrates them at the ABH center, where energy gets dissipated by means of viscoelastic materials. Essentially, there are two types of ABHs: vibrational black holes (VBHs), which focus on the reduction of flexural waves in plates and beams; and sonic black holes (SBHs), which consist of specific duct terminations that dissipate propagating acoustic waves and avoid reflections. The present study focuses on SBHs, which have been less investigated than VBHs. In particular, the influence of an array of SBHs on the acoustic behavior of an acoustic room has not been assessed, while is of interest in many practical applications (buildings, automobiles, etc.). The acoustic field inside a SBH is complex, so building e.g., a finite element model (FEM) that considers a room with a wall of embedded SBHs is computationally prohibitive. Substructuring techniques can circumvent this problem. In the present work, we study the interaction between an array of SBHs and a rectangular acoustic cavity using the transfer matrix method (TMM) to characterize the SBHs and the patch transfer functions (PTF) approach to couple the SBHs to the acoustic cavity. A significant reduction in computational cost is observed with respect to FEM. The influence of various SBH design parameters on the cavity response can then be studied.

        Orateur: Pedro Cerântola
      • 09:25
        "JJCAB1#6 - Optimisation de la méthode des champs virtuels pour l'identification de chargements dynamiques et de paramètres structuraux" 5m

        Les méthodes inverses vibroacoustiques utilisent la réponse vibratoire de la structure pour identifier une charge dynamique ou un paramètre structure. Dans cette étude, deux méthodes seront utilisées, la Méthode des Champs Virtuels (MCV) et la Résolution Inverse (RI) et ses variantes. Le cas d'une poutre d'Euler-Bernoulli simplement supportée et sollicitée en flexion pure est considéré.
        La Résolution Inverse utilise un schéma aux différences finies afin de discrétiser l’équation d’équilibre local et d'identifier l’inconnue. L'amplification du bruit de mesure due à l'application du schéma nécessite l'utilisation d’opérations de filtrage et fenêtrage. On parle alors de Résolution Inverse Filtrée Fenêtrée. La Résolution Inverse Corrigée (RIC), une variante de la RI, apporte une correction qui permet d'appliquer la méthode dans les hautes fréquences. La correction se fait via un facteur appliqué au schéma aux différences finies.
        La Méthode des Champs Virtuels se base sur le Principe des Travaux Virtuels (PTV) une forme faible de l'équilibre local. L'application de la MCV consiste à choisir une fonction test appelée déplacement virtuel afin de résoudre le PTV. La MCV peut être appliquée sur un segment virtuel dont la longueur est inférieure à la longueur de la poutre. Le segment virtuel balaye l'ensemble de la poutre et pour chacune de ses positions, le PTV est résolu et l'inconnue identifiée.
        Dans cette étude, il est proposé d'optimiser la MCV afin de pouvoir l'appliquer dans le domaine des hautes fréquences. Cette optimisation est réalisée en utilisant une approche similaire à celle de la RIC et consiste à déterminer une longueur optimale du segment virtuel.
        L’étude introduira d’abord les méthodes, puis le processus d'optimisation de la MCV sera développé. La méthode optimisée sera ensuite testée sur des données simulées dans le but d'identifier un chargement dynamique ou un paramètre structure.

        Orateur: Nicolas Madinier
      • 09:30
        "JJCAB1#7 - A multi order synchro-squeezing transform approach for instantaneous angular speed estimation" 5m

        Under non-stationary conditions, estimating the Instantaneous Angular Speed (IAS) of rotating machines from vibration measurements is a practical way for encoder-free condition monitoring. The Phase Demodulation (PD) and the Multi Order Probabilistic Approach (MOPA) are arguably the most popular techniques used in this context. Although they both offer a straightforward approach to determine accurately the IAS of a rotating machine, they still present some limitations that can hinder the estimation accuracy and its overall reliability. On the one hand, PD relies on the presence of a single harmonic with a high signal-to-noise ratio (SNR) for the full duration of the measured vibration signal, which is sometimes not satisfied and, consequently, limiting its applicability for a generic speed estimation scheme. On the other hand, MOPA determines the IAS via probability density functions that are constructed from the Time-Frequency Representation (TFR) of the signal followed by an a priori continuity concept that degrades the estimation resolution due to Gaussian smoothing. In this paper, we propose a novel Multi Order Synchro-squeezing Transform (MOST) approach for estimating the IAS in low SNR conditions without compromising the angular resolution. Therefore, it addresses the limitations of PD and MOPA by providing a more accurate method for estimating the IAS under non-stationary conditions. Firstly, the proposed method provides a preliminary estimation of the instantaneous speed for frequency reallocation in the signal's TFR; then it defines a normalized threshold to preserve the corresponding energy in the TFR at each bin of time. Secondly, the probability density functions are constructed using multiple harmonics, benefitting from the repetitive profile of the IAS. Finally, an automatic setting is proposed to keep the useful information of given harmonics and neglect the noisy ones. The proposed method is thoroughly investigated by assessing its performance on a vibration signal of a turbojet engine.

        Orateur: Fadi Karkafi
    • 09:40 → 10:10
      Session Poster + Café 30m MFJA

      MFJA
    • 10:10 → 10:50
      JJCAB2 MFJA

      MFJA
      • 10:10
        "JJCAB2#1 - Bearing degradation indicator using characteristic frequencies applied on non-stationary vibration signals" 5m

        Rolling-element bearings are wear components in rotating machinery. Usually, their condition is monitored with signal-processing tools tailored for vibration-based surveillance. From cyclostationary theory, angle-time analysis, or sparsity-based approaches, numerous methods have been developed relying on the properties of incipient fault signals. The scalar indicators coming from these techniques are well-suited for detection and localisation, but face difficulties in prognosis. The main issue is that the released vibration energy may not correlate with the severity of the fault.

        There is a need to develop scalar indicators dedicated to gravity estimation. Our belief is that the geometry modification of the fault will induce a change in the kinematics of the rolling-element bearing. Far from the ideal perfect-rolling assumption, the relationship between the races' rotation and that of the fundamental train is the fruit of complex interactions. What is usually summarized in the contact angle in the traditional fault frequency equations conceals various effects of axial-radial load ratio, skidding between the elements, and imperfect transmission. As the fault extends, these interactions may change with a direct impact on the fault characteristic frequencies. Bertoni and Andre [1] proposed to monitor these characteristic frequencies with the BeaFEM method to estimate the most probable contact angle. They showed on an academic test-bench run-to-failure experiment that monitoring the fault frequencies could be a valuable degradation indicator. The validation was done on stationary conditions with a main focus on instantaneous angular speed measurements. However, under non-stationary conditions, the equivalent contact angle is sensitive to various factors such as loading, angular acceleration, or lubrication.

        An extension of the method is presented to prove the relevance of the concept on an industrial case with non-stationary conditions. The effect of load and rotation speed is assessed and the method adjusted to emphasize only the contribution of the fault.

        Orateur: Adrien MARSICK
      • 10:15
        "JJCAB2#2 - Resynchronization of sequential measurements using the Maximally-Coherent Reference technique" 5m

        Remote sensing of a physical field generated by a small number of sources is limited by the size of the available array of sensors and by the array density. These limitations may lead to asynchronously measuring the field by sequentially moving a small prototype array around its facets, however at the cost of losing the phase information between every array position. Resynchronization using references, i.e. fixed sensors, can be used for phase retrieval, given that these references are of a number that is at least equal to the stochastic dimension of the field, and are not inter-correlated to the degree that hinder them from spanning the source signals' subspace. However, when the number of references largely exceeds the number of sources, the reference cross spectral matrix becomes ill-conditioned leading to the irrelevance of the least squares solution (LSS). Although the truncated singular value decomposition (TSVD) was successfully applied in the literature to solve this problem, its validity is restricted only to the case of scalar noise on the references. In this poster, a solution based on finding a set of virtual references that is maximally correlated with the field measurements, named the Maximally-Coherent Reference Technique (MCR), accompanied with a technique for estimating the number of sources, are proposed for resynchronization. The method is validated both numerically and by using real acoustical data, and its results was compared to those of the LSS and the TSVD when employed for the same purpose.

        Orateur: Muhammad Nabil Albezzawy
      • 10:20
        "JJCAB2#3 - Conception d'indicateurs optimaux pour la détection précoce de défauts à l'aide d'un test de rapport de vraisemblance généralisé" 5m

        Cette étude propose une méthodologie de construction d'indicateurs d’état de santé pour la surveillance vibratoire des machines tournantes en vue d'une détection précoce des défaillances. Ces indicateurs sont optimaux pour maximiser la probabilité de détection compte tenu d'un taux constant de fausses alarmes. Deux fonctions de densité de probabilité modélisant les états sain et défectueux des signaux de vibration sont exploitées pour générer des indicateurs de santé à l'aide d'un test de rapport de vraisemblance généralisé. Le point clé est la formulation d'un cadre permettant d'exprimer les indicateurs de santé en tant que fonction des dérivées partielles de la PDF de l'état sain et d'une fonction de modulation. La fonction de modulation permet de détecter de légers écarts par rapport à l'état sain des signaux. En outre, il est démontré que des choix spécifiques de fonctions de modulation permettent de récupérer des indicateurs conventionnels tels que l'aplatissement, l'asymétrie, les normes Lp/Lq, la néguentropie de l'enveloppe quadratique, etc. Comme les indicateurs sont asymptotiquement distribués sous la forme d'une distribution de chi-2, un seuil statistique peut être estimé pour évaluer l'état de la machine par rapport à son état sain. La performance des seuils est démontrée sur des signaux de vibration simulés et expérimentaux. Il est démontré que la transition de l'état sain à l'état défectueux de la machine peut être détectée par le seuil. Une conclusion importante de cette étude est que de nombreux indicateurs de santé conventionnels sont optimaux si et seulement si l'état de santé des machines est gaussien. La méthodologie proposée montre comment concevoir des indicateurs pour des descriptions non gaussiennes de l'état de santé et peut ouvrir la voie au développement de nombreux autres indicateurs de santé en sélectionnant soigneusement la PDF de l'état de santé et la fonction de modulation.

        Orateur: Kayacan Kestel
      • 10:25
        "JJCAB2#4 - Stochastic assessment of electric powertrain whining noise under early-stage design uncertainties" 5m

        Despite the advantage of being quieter than traditional internal combustion engine vehicles, electric vehicles are often distinguished by high-frequency tonal components, which can be perceived as unpleasant to the occupants. To ensure optimal acoustic comfort in electric vehicles, it is important to analyze the NVH behavior of e-powertrains during the early stages of the design process which poses inherent uncertainties, such as varying operating conditions, partial knowledge of design parameters, dispersion in measurement data, etc. To effectively address these uncertainties, it is necessary to use fast and comprehensive stochastic models during the design phase.

        In this work, a probabilistic framework is presented to estimate the electric powertrain’s interior whining noises considering the structure-borne contribution. Hence, two different stochastic metamodels are developed for efficient quantification and propagation of uncertainties from electric motor stage to powertrain mounting system. Multivariate Bayesian regression models help to incorporate prior knowledge on the uncertain parameters and generate the respective posterior distributions using Markov chains Monte Carlo (MCMC) techniques. The data is generated through weakly-coupled multi-physical domains estimated using semi-analytical approaches and combined with measured vehicle transfer functions. Importantly, the validation of each domain is conducted separately to ensure accurate representation. The results obtained from the developed probabilistic framework will aid in the early design stages by guiding engineers in making informed decisions to optimize NVH performance.

        Orateur: Vinay PRAKASH
      • 10:30
        "JJCAB2#5 - Surfaces et interfaces métaporoélastique pour la mitigation conjointe des énergies acoustiques et élastiques" 5m

        Les matériaux poreux sont utilisés depuis longtemps pour leur efficacité en absorption acoustique. Pourtant, dans leurs usages il existe des contraintes d’épaisseur qui limitent leur performance d’absorption en basses fréquences à cause de leurs mécanismes de pertes. Le développement de surfaces métaporeuses est donc un enjeu pour la réduction de ces matériaux absorbants, différentes structures présentant une absorption large-bande et sub-longueur d’onde sont étudiées dans la littérature. Pourtant, la compréhension du comportement modal de ce type de système reste limitée. En effet, les méthodes permettant de calculer les relations de dispersion complexes de structures (méta)poroélastiques impliquent généralement une recherche de racines complexes (Müller, etc.) à partir d’un déterminant analytique, méthode coûteuse en calculs et n’assurant pas de trouver toutes les racines. Pour cela, une méthode de collocation spectrale, déjà appliquée pour des guides élastiques peut être étendue pour l’étude de matériaux poroélastiques, dont les ondes guidées sont alors étudiées. En discrétisant une couche de matériau le long de l’épaisseur, un problème aux valeurs propres est obtenu dont les valeurs propres sont des nombres d’ondes complexe, donnant des courbes de dispersion en $(\omega, k), k \in \mathbb{C}$ de ce type de système. Un changement de variable permet de considérer la radiation des ondes guidées dans un fluide. La comparaison avec des résultats analytique montre une concordance des résultats. Ainsi, cette méthode peut être étendue à une structure multicouche comme un bicouche élastique-poroélastique. Ces travaux ont pour objectif de complexifier à terme la structure étudiée avec l’ajout d’inclusions périodiques.

        Orateur: Mathieu Maréchal
      • 10:35
        "JJCAB2#6 - Modeling and optimization of acoustic metamaterials for cabin noise treatment" 5m

        The objective of our research is to develop an acoustic fairing for UAVs and other VTOL aircraft by using an acoustic metamaterials developed at Capgemini, which have been tested for noise insulation in new workspaces. These metamaterials will be used to create specific liners for the fairings and designed with a porous matrix that is printed using the Fused Deposition Modeling (FDM) method. The matrix includes resonant elements strategically placed to achieve the desired acoustic properties Various printing configurations will be explored, including different patterns and fill densities, to optimize performance within the targeted frequency ranges. Tests will be conducted to compare the noise emissions of each acoustic fairing configuration. This will allow us to assess the effectiveness of the different metamaterial designs and identify the configurations that provide the best noise reduction. By using these advanced materials and conducting rigorous testing, we aim to develop an effective fairing solution that can significantly mitigate noise emissions for UAVs and VTOL aircraft.

        Orateur: Anas Lachheb
    • 10:50 → 11:20
      Session Poster + Café 30m MFJA

      MFJA
    • 11:20 → 12:00
      JJCAB3 MFJA

      MFJA
      • 11:20
        "JJCAB3#1 - Conception et contrôle vibratoire d'une imprimante 3D à câbles" 5m

        La fabrication additive (FA) a connu un essor spectaculaire ces dix dernières années. Ce procédé permet de réaliser en petites séries des pièces à haute valeur ajoutée et de formes complexes. Le potentiel de la FA est très important, notamment dans les domaines de l'aéronautique et du médical. Par ailleurs, l'application de la technologie des Robots Parallèles à Câbles (RPC) pour l'impression 3D a vu le jour ces dernières années. Les RPC se caractérisent par le fait que les câbles ne peuvent travailler qu'en traction.

        L'objectif du projet CABTIVE (suite du projet CABFAB (2018-2021), qui a montré l'intérêt d'une augmentation de la précharge des câbles et du contrôle actif de vibrations en embarquant des actionneurs piézoélectriques sur les câbles) est d'accroître la précision des RPC pour l'impression 3D de pièces de tailles moyennes, afin de garantir la qualité géométrique des pièces produites.

        L'intérêt est donc ici d'utiliser un modèle dynamique afin d'établir les interactions entre les commandes moteurs de génération de trajectoire et les algorithmes de contrôle actif via les actionneurs piézoélectriques embarqués. On se concentrera notamment sur leur nature, leur nombre et leur position, tout en mettant en place un couplage des lois de commande entre génération de trajectoire et contrôle actif. Pour le contrôle actif, la question de la mesure de la tension des câbles sera à prendre en compte dans les boucles de contrôle, nécessairement colocalisées (actionnement et mesure au même point).

        Afin de tester expérimentalement les concepts introduits, un RPC plan sera mis au point, en se concentrant notamment sur le fonctionnement synergique des contrôles robustes. On cherchera principalement à optimiser la rapidité, la précision et la stabilité dans des mouvements simples.

        Orateur: Thibault GARCIA
      • 11:25
        "JJCAB3#2 - Détectabilité des défauts de roulements dans les transmissions mécaniques par engrenages: Conditions de fonctionnement, capteurs et indicateurs optimaux" 5m

        Dans une transmission mécanique, les roulements assurant le guidage des arbres en rotation sont des pièces importantes à surveiller, à hauteur des efforts qu’ils reprennent et de la gravité que leur défaillance peut avoir. Dans les secteurs aéronautique et ferroviaire entre autres, il est crucial de repérer les défauts de roulements au plus proche de leur apparition, ce qui implique d’avoir recours à une méthode de maintenance prédictive, surveillance d’un système en fonctionnement à l’aide de données capteurs. Les données utilisées incluent des signaux d’accélération vibratoire mesurés sur le boîtier de la transmission, ainsi que des signaux de vitesse angulaire instantanée (IAS) mesurés directement sur l’arbre.
        L’objectif de cette thèse est d’optimiser la méthode de détection de défauts de roulement en simulant les interactions mécaniques par lesquelles défaut donné perturbe les signaux. Pour ce faire, un modèle numérique du comportement dynamique d’une transmission est réalisé : sur la base d’un code simulant les contacts et les effets dynamiques entre les arbres, roulements et engrenages (Thèse N.Thibault, 2022), l’étude des interactions entre le rotor et le stator est envisagée en couplant ce modèle dynamique avec un modèle vibratoire par éléments finis du carter de la transmission.
        Une spécificité de la méthodologie utilisée est l’approche angulaire, qui consiste à échantillonner les signaux en fonction de l’angle parcouru par l’arbre, et à repérer des défauts dans le domaine des fréquences angulaires, exprimées en évènements par tour. Les connaissances apportées par la simulation sur l’interaction rotor-stator permettront de déterminer des excitations types à retrouver sur les signaux d’une transmission en fonctionnement, et donc d’améliorer la précision de la détection de défauts de roulements.

        Orateur: Arthur Burel
      • 11:30
        "JJCAB3#3 - Méthode d’assimilation de données sur des structures en bois de grande hauteur" 5m

        La construction de bâtiment est un enjeu actuel important dus à notre démographie croissante mais également d’un point de vue écologique. Pour concilier ce désaccord, il est intéressant de construire des bâtiments avec du bois plutôt que du béton. Cependant, les structures en bois ont un comportement plus flexibles que celles en béton, notamment lorsque leur hauteur est conséquente. Cette propriété entraîne l’inconfort de ses usagers, d’où l’intérêt de mieux connaître leur dynamique afin de savoir parer ce désagrément.
        Pour cela, ma thèse s’applique à créer un jumeau numérique de ces structures à partir de méthodes d’assimilation de données : les filtres de Kalman. Ces méthodes permettent d’estimer un état du système à partir de mesures bruitées ou incomplètes. Afin de ne pas être intrusif par rapport à la structure, les mesures ne seront pas issues d’une analyse modale expérimentale mais d’expériences par caméras rapides qui extrairont les déplacements de la structure. Ces observations, complétées par un modèle physique du système, permettront d’obtenir les paramètres modaux et les propriétés dynamiques de la structure. Une prévision de l’évolution du système peut se dessiner grâce à l’attribut du jumeau numérique. Ma présentation montrera l’utilisation des méthodes d’assimilation de données sur une structure simple.

        Orateur: Layla KORDYLAS
      • 11:35
        "JJCAB3#4 - Digital twin for predicting energy consumption for heating in non-residential buildings" 5m

        The following work aims to develop an identification and prediction tool for forecasting energy consumption for environmental heating in non-residential buildings. The proposed method is based on the design of a digital twin of the real system under consideration. This twin consists of a virtual model (analytical state model) combined with Kalman filters in a learning process.
        The state model presents several unknown parameters which refer to specific properties of the system. The training phase allows to gain insight into the unknown properties of the system. Once the parameters are identified, the state model is disconnected from Kalman filters and employed to run simulations to predict the energy consumption for heating. This forecasting offers significant potential to optimize energy consumption by improving the heating system scheduling.
        The methodology is detailed with reference to an industrial use-case in the automotive manufacturing sector: the CRF bodyshop in Turin. The target process is the environmental heating of the industrial building, which accounts for a major part of total energy demand of the plant. The thermal dynamics of the building is described through a multi-zone RC circuit thanks to the electrical network analogy. The entire warehouse is represented by means of a few thermal resistors and capacitors, that are equivalent to multiple thermal elements and physical effects. Heating system heat gains and solar heat gains are included in the model. The virtual model is trained by Kalman filters through measurements of indoor temperature, outdoor temperature and solar radiation, assuming that the heating system scheduling and technical data are known.

        Orateur: Antonio Baldassarre
      • 11:40
        "JJCAB3#5 - Reconstitution non-supervisée des séquences d’activation des actionneurs issus d’un système de production" 5m

        En vue d’améliorer la durabilité énergétique des usines, de nombreuses initiatives voient le jour, allant du jumeau numérique à la maintenance prédictive. Pour ces applications, les séquences d’activation des actionneurs d’une machine de production sont nécessaires à l’obtention d’indicateurs de performance (par opération et par actionneur). En pratique, ces séquences peuvent être synthétisées à partir de capteurs non-intrusifs. Nous nous intéressons notamment aux accéléromètres et capteurs de courant, capables de capter plusieurs sources simultanément.
        Les méthodes sous-espace sont couramment utilisées dans les problèmes de séparation de sources, celles-ci se heurtent néanmoins à deux difficultés: l’obtention de séquences d’activation binaires, et l’estimation précise du nombre de sources. Dans cette thèse, nous proposons une approche de séparation de sources par segmentation et sous-segmentation palliant ces deux difficultés. Cette méthode s’appuie sur la représentation temps-fréquence dans le domaine complexe (transformée de Fourier à court terme), et se limite aux signaux stationnaires et aléatoires.

        Orateur: Romain Delabeye
      • 11:45
        "JJCAB3#6 - Video-based vibration analysis of medium and high wood constructions" 5m

        With the aim to analyze the vibration of civil and large-scale mechanical structures, video-based displacement measurement has emerged as an attractive technique thanks to its advantages, such as non-contact, full-field measurement, and low interference to occupants. Furthermore, there is a growing demand for wooden buildings since timber is considered a sustainable alternative to concrete and steel. In this context, the thesis focuses on developing a computer vision (CV)-based measurement protocol to measure the vibrations of high-rise timber buildings and large structures. As the first step of this thesis, a video-based displacement measurement using a camera network was conducted on a tower crane. It's because the measurement condition based on the CV for a tower crane is close to that for a high-rise timber building. For CV-based 3D displacement calculation, a procedure is applied that consists of (1) Data acquisition with a camera network, (2) Camera synchronization using a coded light signal that we developed, (3) Camera calibration with two steps, (4) Motion tracking based on Digital Image Correlation (DIC) and Line Segment Tracking that we developed for the robustness, (5) Structural 3D displacement calculation with non-linear triangulation. As a result, 3D displacement of the tower crane was measured based on CV. Furthermore, the 3D displacement results extracted from videos are well-superposed with data acquired by the contact-based sensor (the accelerometer).

        Orateur: Yunhyeok HAN
      • 11:50
        "JJCAB3#7 - A experimental parametric study of brake squeal" 5m

        Many systems have vibration properties, modes, and non-linear limit cycles, that are significantly affected by environmental properties such as temperature, pressure, rotation speed. Brake systems can have a mode coupling instability, known as squeal, for certain parameter combinations. It has also been previously observed that the limit cycle frequency evolves periodically with wheel spin, further motivating the need to study parametric variations.
        An experimental campaign was designed to characterize parametric evolution of limit cycle properties with pressure, wheel spin, and to a lesser extent the testbed torsion mode and temperature. Short term Fourier transforms clearly indicate parametric variations but have limitations in their ability to track properties varying in time. A multi-output harmonic balance signal model varying slowly in time is estimated using a process mostly based in synchronous demodulation.
        The relevance of this model for the analysis limit cycle sensitivity is illustrated for different test runs showing intermittent squeal, transition between multiple instabilities, evolution of limit cycle properties with wheel position, pressure, … The analysis of principal coordinates in the process demonstrates the changes in shapes for different conditions.

        Orateur: Guilherme Malacrida Alves
    • 12:00 → 12:30
      Session Poster + Café 30m MFJA

      MFJA
    • 12:30 → 14:00
      Pause Déjeuner 1h 30m MFJA

      MFJA
    • 14:00 → 14:30
      Diffusion des videos laboratoires 30m MFJA

      MFJA
    • 14:30 → 15:05
      JJCAB4 MFJA

      MFJA
      • 14:30
        "JJCAB4#1 - Acoustic Liners Intergrated Inside Blades of Unmanned Air Vehicle Rotors" 5m

        This investigation proposes a new concept of rotor-liner integration with the objective of reduce noise. The liner concept to be integrated in the lades of the rotor blades is called LEONAR (long elastic open neck resonator). This type of liner provides good absorption in the low frequency range and at the same time allows to have cavity thicknesses up to one-thirtieth of the wavelength. A model for the rotor-liner is proposed, based on linear acoustic theory integrating a tailored Green's function for the presence of the liner. Prototypes of a first rotor-liner configuration have been made and tested in an anechoic room. The results permitted to validate the proposed concept and at the same time the limitations of the physical model.

        Orateur: Santiago Montoya-Ospina
      • 14:35
        "JJCAB4#2 - Conception de Panneaux d’Habillage Légers Type Aéronef à Fort Indice d’Affaiblissement Acoustique" 5m

        L’habillage interne des aéronefs est souvent composé de matériaux multicouches de type panneaux sandwichs. Ils se composent d’un cœur léger en nid d’abeille emprisonné entre deux parois rigides. Cette configuration permet aux panneaux, malgré leur faible masse et leur faible épaisseur, de résister aux efforts de flexion mais les prive malheureusement d’une bonne isolation acoustique à cause de leur légèreté.
        L’objectif de la thèse est alors d’améliorer le confort acoustique en cabine d’aéronef en associant plusieurs concepts vibro-acoustiques, à savoir l’intégration d’une distribution fractale massique ainsi que des résonateurs acoustiques au sein des cellules du nid d’abeille.
        Un schéma fractal est constitué d’un motif qui se répète périodiquement à plusieurs échelles. Il a été montré que, pour 10 % de masse en plus, des masselottes disposées selon un schéma fractal dans le panneau sandwich agissait sur son schéma vibratoire en créant des réflexions multiples sur les inclusions. Elles génèrent ainsi des bandes de fréquences interdites répétitives en fréquence, tout en atténuant le rayonnement que nous allons modifier localement grâce à des matériaux absorbants de type LEONAR (Long Elastic Open Neck Acoustic Resonator). Ceux-ci sont constitués de parois perforées prolongées par des tubes qui leur permettent, malgré leur faible épaisseur, d’absorber en basse fréquence.
        L’étude consiste à développer dans un premier temps un modèle numérique vibro-acoustique permettant d’optimiser les caractéristiques des deux concepts au sein d’un panneau d’habillage de type avion, à réaliser des tests en laboratoire (analyse modale, intensité de structure, indice d’affaiblissement acoustique) pour valider les phénomènes physiques souhaités et enfin à démontrer l’applicabilité sur des configurations avions en vue d’un essai à l’échelle 1 sur le démonstrateur ZAL d’Airbus Hambourg.

        Orateur: Tiffany Raharimanana
      • 14:40
        "JJCAB4#3 - Conception d'un filtre absolu d'onde mécanique à l'aide d'un réseau périodique 3D de treillis de barre." 5m

        Les structures périodiques ont été l'objet de nombreuses études ces dernières années en raison de leur capacité à filtrer les ondes sur une large plage de fréquences. Cette recherche se concentre sur la conception d'une cellule unitaire 3D constituée d'un treillis de barres (structure lattice) capable de filtrer tous les types d'ondes mécaniques (compression, flexion, torsion) dans une plage de fréquences donnée. Les structures lattice offrent une alternative aux structures conventionnelles grâce à leur grande robustesse statique et à leur faible poids. Par exemple, l'âme en nid d'abeille, qui peut être considérée comme une structure lattice, est largement utilisée dans les structures sandwich en raison de ses propriétés mécaniques statiques.
        De manière similaire aux cristaux phononiques, les structures en treillis peuvent être conçues de manière périodique pour produire des bandes de fréquences absolues où aucune onde mécanique ne peut se propager à travers le réseau. Dans cette étude, nous proposons la conception d'une cellule unitaire modélisée en utilisant la méthode des éléments spectraux (SEM) couplée à un algorithme génétique. La méthode des éléments spectraux est un puissant outil de calcul pour les systèmes en treillis de poutres, car contrairement à la méthode des éléments finis, un seul élément suffit à converger vers la solution exacte, ce qui réduit le temps de calcul et permet une optimisation plus économique.
        En résumé, notre étude présente une méthode innovante pour concevoir une structure lattice capable de générer des bandes de fréquences interdites pour tous les types d'ondes mécaniques. Cette approche utilise des techniques avancées telles que la méthode des éléments spectraux (SEM) et l'algorithme génétique afin d'atteindre une efficacité maximale dans la conception de ces dernières.

        Orateur: Théo BONNEVAL
      • 14:45
        "JJCAB4#4 - Contrôle des instabilités dynamiques non-linéaires à l'aide de la prédiction" 5m

        Les systèmes mécaniques et fluide-structure peuvent présenter des résonances et des instabilités, notamment le flottement aéroélastique, une oscillation structurelle instable due au couplage entre la structure de l'aéronef et l'écoulement du fluide. Cette interaction fluide-structure peut entraîner de l’endommagement structurel, immédiat ou dû à la fatigue. Pour contrôler ces instabilités, de nouvelles stratégies ont été développées, notamment l'utilisation d'absorbeurs passifs non linéaires. Cependant, la conception et l'optimisation de ces dispositifs nécessitent une étude approfondie des régimes post-flottement définis par des diagrammes de bifurcation. Dans ce contexte, une nouvelle approche basée sur une technique de prédiction est introduite. Cette technique de prédiction est sans modèle et utilise des mesures de la réponse du système collectées uniquement dans la phase pré-bifurcation. Elle permet de prédire le point critique et la dynamique post-bifurcation des systèmes non linéaires. Cette approche a montré son potentiel à prédire le régime non linéaire des systèmes aéroélastiques avec des mesures basées sur le régime stable. Cependant, cette technique de prédiction présente certaines limitations. Dans ces travaux, nous proposons le développement d'une technique de prédiction basée sur la physique pour améliorer la prédiction des bifurcations et des oscillations à plus grande amplitude. Nous mettons également en évidence l'utilisation de cette technique pour optimiser la conception d'absorbeurs passifs non linéaires afin de contrôler le flottement aéroélastique. Pour valider les résultats, des simulations numériques et des expériences en soufflerie sont réalisées. Cette approche de prédiction basée sur des mesures pré-bifurcation offre de nouvelles perspectives pour la conception et l'optimisation des absorbeurs passifs non linéaires dans le contrôle du flottement aéroélastique. Les résultats préliminaires obtenus montrent le potentiel de cette technique pour prédire l’instabilité et la capacité de l’absorbeur non linéaire à contrôler le comportement aéroélastique d’une structure aéroélastique pour des conditions de vol plus sûres et une enveloppe de vol élargie.

        Orateur: Jesus García Pérez
      • 14:50
        "JJCAB4#5 - Stratégie d’identification de l'influence dynamique des éléments non-structuraux sur les structures aérospatiales" 5m

        Les structures aérospatiales sont exposées à des vibrations de forte amplitude au cours de leur vol, affectant à la fois la structure principale et les éléments non-structuraux (ENS). Des études ont montré que les ENS peuvent amortir de manière significative la structure globale et expliquer les différences entre les modèles numériques et les résultats expérimentaux dans les vibrations de grande amplitude. Les ENS, tels que les faisceaux de câbles, représentent près de 10 à 30 % de la masse totale des engins spatiaux, mais leur comportement dynamique n'est pas pris en compte dans le modèle global. Le développement d'un couplage dynamique entre les ENS et la structure principale peut améliorer les prévisions du comportement dynamique de l'ensemble de la structure et réduire les coûts totaux, permettant ainsi une réduction globale de la masse.

        Cependant, le développement d'un modèle prédictif des ENS basé sur les seules définitions de la géométrie et des matériaux est trop complexe pour comprendre pleinement leur impact sur le comportement dynamique des structures aérospatiales. Pour résoudre ce problème, il est possible d'analyser les fonctions de réponse en fréquence des ENS, mais cette méthode n'est pas toujours adaptée à certains types d'éléments, tels que ceux des lanceurs, qui ne sont pas toujours accessibles individuellement.

        Dans cette étude, la structure principale est modélisée à l'aide d'un modèle éléments finis et couplée à des ENS inconnus dont l'identification du modèle est basée sur la comparaison des données numériques et expérimentales des fonctions de réponse en fréquence de la structure avec et sans ENS. L’enjeu principal de ces travaux de recherche est de développer une stratégie adaptée pour identifier l'effet dynamique des ENS sur la structure principale, en combinant des données expérimentales et un modèle d'éléments finis.

        Orateur: Lisa Fournier
      • 14:55
        "JJCAB4#6 - Modélisation numérique d’un instrument de musique à vent et comparaison à la mesure expérimentale" 5m

        La fabrication des saxophones a longtemps reposé sur un savoir-faire artisanal, associé à une connaissance empirique de l'acoustique qui régit leur fonctionnement. La conception de ces instruments chez Henri Selmer Paris est aujourd'hui principalement basée sur une grandeur physique : l'impédance d'entrée. Différentes approches, expérimentales et numériques, permettent d'accéder à cette grandeur. L'utilisation de méthodes purement numériques est actuellement limitée par la précision (modèles semi-analytiques), ou par le temps de calcul (résolution numérique des EDP) non compatible avec des délais industriels. Les méthodes expérimentales ou hybrides (combinant mesures et calculs) sont encore compétitives et donc privilégiées pour le développement. L'enjeu de cette thèse est de proposer une méthode numérique alliant précision et temps limité, pour accéder aux champs d’intérêt (pression et vitesse acoustiques) dans l'instrument réel, avec la précision requise par l'industrie musicale.

        Notre objectif a été de développer un modèle FEM 3D, performant (parallèle), pour modéliser précisément (quelques cents sur les résonances) le comportement acoustique de la partie passive du saxophone, sous conditions de jeu réelles. Différentes stratégies de modélisation (physique et numérique), spécifiques aux instruments à vent, ont permis de réduire le temps de calculs d’un ordre de grandeur, et s’appuient sur trois aspects principaux : d’une part, la modélisation des pertes visco-thermiques aux parois au juste coût ; d’autre part, la réduction de la taille des systèmes linéaires considérés (>1M degrés de liberté) ; et enfin, des stratégies de réutilisation et de réduction de modèle limitant l’effet des multiples résolutions imposées par la large plage fréquentielle d'intérêt pour le facteur. Le modèle numérique est validé par comparaison avec des mesures expérimentales, hors en conditions de jeu (pour l’instant). Un système de mesure d'impédance permettant des acquisitions sous écoulement, de débit et température contrôlés, a été spécifiquement développé et validé pour reproduire les conditions de jeu.

        Orateur: Marie Jeanneteau
    • 15:05 → 15:35
      Session Poster + Café 30m MFJA

      MFJA
    • 15:35 → 16:10
      JJCAB5 MFJA

      MFJA
      • 15:35
        "JJCAB5#1 - Characterization of the dynamic behavior of laminated electric rotors" 5m

        The rotors of traction motors have a direct impact on reliability and vibrational stability of the whole system. The rotors of Alstom electric motors are composed of a shaft onto which magnetic sheets are assembled. The sheets are held in position by a shrink fit on the shaft and are held under axial pressure by clamping plates. The vibrational behavior of this complex assembly, due to its nature and method of assembly, is difficult to predict to date. Therefore, mastering the mechanical properties of these complex assemblies and the development of appropriate dynamic models are a major challenge for the Alstom company. The vibration of rotors has a direct impact on the following issues: the design of rotating parts under torsional excitation; the design of rotors with respect to increasingly stringent acoustic criteria; the design of rotors for increases in rotational speed that induce new vibrational risks. The objectives of the thesis are focused on the mechanical modeling of rotating parts in the presence of significant unknowns. In particular, the thesis should allow: a finer characterization of the mechanical behavior of the laminated sheet assembly as a function of residual mounting pre-tension through testing of sub-assemblies and elementary numerical models; validation of the approach within the framework of a simplified academic case (including, in particular, a calculation/test correlation and the consideration of non-linearities); the establishment of a methodology for predicting the vibrational behavior of the laminated rotor based on mounting conditions; validation of the approach within the framework of an application to synchronous and asynchronous machine rotors; integration of this methodology into a rotor optimization approach. The integration of unknowns into such complex simulations, incorporating realistic behaviors of technological components, is both a scientific challenge for the laboratory and a strong methodological contribution for industrial design offices

        Orateur: Nour Abuhemeida
      • 15:40
        "JJCAB5#2 - Contribution à l'amélioration de la modélisation multi-corps de turbomachines" 5m

        Le prototypage virtuel est désormais largement utilisé en ingénierie mécanique dans les phases de préconception et de conception pour l'optimisation des performances, ainsi que dans les phases d'exploitation et de maintenance pour la détection des dommages et le diagnostic rapide des problèmes. Les simulations dynamiques multi-corps flexibles sont efficaces pour représenter le comportement de structures complexes avec des non-linéarités locales d'amortissement et de rigidité. Cependant, les intégrations temporelles peuvent tout de même représenter une charge de calcul élevée selon le choix des liaisons entre les composants et la bande de fréquence étudiée. Dans la plupart des logiciels commerciaux multi-corps, le comportement des composants élastiques linéaires est pris en compte via des superéléments de Craig-Bampton. Cette présentation abordera la mise en œuvre de méthodes de sous-structuration pour réduire davantage l’ordre des superéléments. Un cas d’application industriel de montée en régime d’un moteur d’avion permettra de comparer les méthodes éprouvées en termes de gain de temps et de perte de précision sur la réponse dynamique.

        Orateur: Océane TOPENOT
      • 15:45
        "JJCAB5#3 - Conception d'une suspension multi-axiale à verrouillage pour le contrôle des vibrations" 5m

        La maitrise des niveaux vibratoires est fondamentale dans le développement de matériels optroniques ou d’autodirecteurs. Les environnements mécaniques et thermiques dans lesquels évoluent les équipements sont très sévères. Depuis quelques années, le bureau d’études de Thales LAS OME s’est spécialisé dans la conception de suspension afin d’isoler les équipements optroniques. Ces suspensions doivent répondre à deux objectifs : limitation du transfert des vibrations vers les équipements optroniques en phase de vol et limitation du débattement lors de sollicitations transitoires de forte intensité. Cet objectif est ambitieux étant donné l’intensité des sollicitations mécaniques et les contraintes environnementales. Les travaux de thèse consistent à développer des suspensions efficaces pour les différents régimes de sollicitation. Le travail de thèse propose d’investiguer une nouvelle stratégie de suspension dite « à verrouillage », présentant un comportement à rigidité élevée lorsque soumise à un évènement transitoire, et présentant un comportement de suspension classique le reste du temps. Le verrouillage devra être multiaxial, en garantissant un blocage dans au moins deux directions de déplacement. Deux voies technologiques sont envisagées :
        • Le développement d’une suspension à verrouillage thermique (1ère partie de thèse)
        • Le développement d’une suspension basée sur un métamatériau architecturé (2nd partie de thèse)

        Mots Clés : Verrouillage thermique – Verrouillage Métamatériau – Dynamique des structures – Choc – Conception – Polymère à mémoire de forme

        Orateur: Johann Roberjot
      • 15:50
        "JJCAB5#4 - Analyse numérique du comportement vibratoire d'une machine électrique synchrone à aimants permanents" 5m

        Dans le but de concevoir des machines électriques synchrones à aimants permanents les plus silencieuses possibles, nous proposons un modèle magnéto-mécanique. Ce modèle permet de prédire la dynamique du système en condition de fonctionnement, en incluant un couplage fort entre les efforts électromagnétiques appliqués sur le rotor et le stator ainsi que la dynamique de l'arbre moteur. Les pressions de Maxwell sont calculées de manière analytique par la méthode des sous-domaines, tandis que la dynamique du rotor est basée sur un modèle en éléments finis 1D. Le couplage entre ces deux physiques est résolu en dynamique à l'aide d'une méthode d'intégration temporelle à pas de temps décalés. Ces simulations visent d'abord à mettre en évidence des régimes en couplage fort et leur influence sur le contenu harmonique. Un second objectif concerne l'influence des excentricités statiques et dynamiques sur le contenu spectral, au moyen d'une prédiction quantitative des ordres d'excitation spatiaux. Enfin, une analyse de la convergence et de la stabilité des résultats issus du schéma d'intégration temporelle est menée afin de garantir la robustesse et la stabilité des signaux prédits par ce modèle couplé.

        Orateur: Thomas Poupon
      • 15:55
        "JJCAB5#5 - Modal analysis of in-duct fan tonal noise at varying shaft speed with a Bayesian approach" 5m

        A modal analysis of tonal fan noise is applied to deceleration measurements collected during the European project TurboNoiseBB at the Anecom test facility. The azimuthal modal content of the three first Blade Passing Frequencies (BPF) is obtained through an iterative inverse Bayesian approach. Particularly the second BPF showed an interesting pattern of rotor-stator interaction modes with peaks and troughs throughout the deceleration. This pattern suggests an influence of the sensor array configuration over the modal content. A full azimuthal and radial mode decomposition showed that correlated modes associated with high mutual coherence of the sensor array induce an amplitude inversion of dominant modes.

        Orateur: Johann Miranda-Fuentes
      • 16:00
        "JJCAB5#6 - Geometrical, Flexural and Vibrational Characterization of Saxophone Reeds" 5m

        The sound production by the saxophone, a single-reed wind instrument relies mainly on the control of the vibration of a plate by the player’s lower lip and the air blown into the instrument. This vibrating plate is securely attached to the mouthpiece and is known as a reed. It is one of the principal components responsible for generating notes played by musicians. Traditionally, saxophone reeds are made from natural cane, particularly Arundo-Donax. However, these reeds may have certain disadvantages such as limited durability and anisotropy bucur_traditional_2019. To address these inconsistencies, synthetic reeds have emerged as an alternative.

        In this study, we have investigated eight different reeds, including both synthetic and natural ones and, compared their design, material and vibroacoustic properties. We employed microtomography to capture the 3D view of each reed and redesigned it using CAD software. A bending test was conducted to test the property of the reed material. Finally, vibroacoustic tests were performed to effectively analyze the dynamic properties of the reeds, and modal analyses were carried out to identify the vibrational modes and their corresponding damping factors. Ultimately, these reeds were compared based on their respective design, rigidity, and vibroacoustic properties. The long-term aim of this study is to facilitate the fabrication of reeds aiming to aide musicians in customizing their reeds that can align with their specific needs.

        Orateur: Sashi Kiran MADUGULA
    • 16:10 → 16:40
      Session Poster + Café 30m MFJA

      MFJA
    • 17:00 → 18:30
      Visite Laboratoires 1h 30m MFJA / IRT / ICA

      MFJA / IRT / ICA
    • 19:30 → 23:00
      JJCAB Diner de Gala 3h 30m Sapiens La cité

      Sapiens La cité

      55 Av. Louis Breguet Bâtiment La Cité, 31400 Toulouse
  • mardi, 11 juillet
    • 09:00 → 09:40
      Welcome Ceremony (Survishno & JISFA)
    • 09:40 → 10:20
      Keynote / Dr. Julien Caillet, Acoustics - ETGV, Airbus Helicopters 40m

      Title:
      eVTOL noise acceptance : a short overview from the perspective of a helicopter manufacturer.

      Abstract:
      New technologies such as distributed electric propulsion and autonomy offer new opportunities for VTOL designs and operations. They are also new challenges for engineers, since noise may be a major barrier for a large deployment of such new vehicles. The presentation will focus on eVTOL noise and will include discussion on the noise sources, acceptance, and the potential solutions to reduce noise for communities and passengers. Dr. Julien Caillet is an acoustic expert at Airbus Helicopters, also in charge of noise certification activities. He had a PhD on helicopter interior noise diagnosis and modelling in 2006. He has over 20 years of acoustics engineering and project management experience with expertise in developing low noise technology for helicopters. He currently supports UAM acoustics activities within Airbus Group and coordinates research projects on Helicopter & eVTOL Low noise design and acceptance.

      Orateurs: Frank Simon, Julien Caillet (ETGV, Airbus Helicopters)
    • 10:20 → 10:40
      Coffee break 20m
    • 10:40 → 12:40
      JISFA 1 / Aircraft Exterior Noise
      Président de session: Frank Simon
      • 10:40
        JISFA1#1 - Engine noise: panorama of future architectures 20m

        As a key brick of aviation decarbonization challenge, Airbus is curently exploring many engine architectures. This presentation will give an overview of the various engines currently studied, concentrating on noise aspects.

        Orateur: Stefan Moal (AIRBUS, Toulouse)
      • 11:00
        JISFA1#2 - Combustion noise and engine noise: current computational capabilities and fuel expected issues 20m

        Combustion noise is becoming an increasing subject of interest and has been brought forward with the uprise of premixed combustors. This combustor technology which has entered in production and comes with lower pollutant emissions of new engines, also comes with a sensitivity to the engine acoustics and the potential triggering of new and unwanted noises heard outside the engine. Indeed, such combustion dependent features can now be heard outside because of all the progresses built around engine noise mechanisms which are today better controlled leaving the place for combustion noise to express when it could not with older engines. This question of combustion generated noise is furthermore expected to re-emerged especially with the foreseen use of new fuels like hydrogen. In the present talk, recent advances on this problem and obtained within CERFACS will be detailed and discussed based on high-end fully unsteady numerical predictions and their exploitation for a better understanding of combustion noise.

        Orateur: Laurent Gicquel (CERFACS)
      • 11:20
        JISFA1#3 - To be defined 20m
      • 11:40
        JISFA1#4 - Reverse problems and Analytical Sensitivity analysis for landing gear noise calibration and prediction 20m

        Reverse problems are used in order to calibrate landing gears noise spectra. The idea consists to minimize a least square function where the SMITH semi-empirical spectrum is optimized with respect to the experimental measured data. Hence, the SMITH “empirical” parameters are the control parameters for these acoustics optimization problems. The number of these optimization parameters can quickly increase when it concerns real landing gears with hundreds of components. An analytical sensitivity analysis is carried out using SMITH analytical formulations in order to identify the most sensitive and influencing control parameters and to reduce significantly the big number of optimization parameters. The optimized SMITH numerical coefficients can be efficiently used to predict the noise of news landing gear configurations.

        Orateur: Paul Hanappier (Safran Landing Systems - Vélizy)
      • 12:00
        JISFA1#5 - Noise of realistic slats with tracks 20m

        Aircraft noise needs to be mitigated to be compliant with certification requirements and operational restrictions imposed by airports. Aircraft noise comprises engine noise and airframe noise, which itself consists mainly of landing gear noise, slat noise and flap noise. The presentation is focused on noise produced by slats, which significantly contribute to aircraft noise at approach.

        Recent works have shown the importance of slat tracks which are structural supports connecting the slat to the wing. These devices are indeed located in a region of high-speed cross-flow generating self-noise and wakes, which impact the slat cove and the wing leading edge causing additional noise sources.

        The solutions to reduce slat track noise aim to improve the aerodynamic shape of the track with add-on fairings and to reduce the impact of the wakes on the slat cove.
        Results from numerical simulations and from experiments will be shown to illustrate the presentation.

        Orateur: Fleury Vincent
      • 12:20
        JISFA1#6 - Round table about JISFA 1 20m

        Additional Q&A and open discussion about the session topic.

    • 10:40 → 12:40
      Survishno 1 / Model & System identification
      Président de session: Simon Chesné
      • 10:40
        "SYSID#1 - Comparative study of the first bending modes of a cantilever beam from a video measurement" 20m

        Today, there is a real interest in using video cameras for operational modal analysis or preventive maintenance applications rather than using traditional sensors. Indeed, the video camera allows a non-intrusive diagnosis and allows a synchronous measurement of displacement and deformation, all with few cables and equipment. The potential of video-based vibration monitoring is all the more interesting because it could potentially be achieved in the future by using any smartphone. The objective of this work is to provide a new methodology of operational modal analysis adapted to the bending vibrations of the beams from video measurements. The major advantage of video measurements is the spatial resolution of the observations: a pixel can be seen as a sensor in its own right as long as its spatial position can be estimated throughout the acquisition. Comparative measurements with a scanning laser Doppler vibrometer will be performed on a more limited number of points. This work presents in detail the steps prior to the comparison, which are particularly influential on the conclusions concerning the estimation of the Modal Assurance Criterion (MAC). Indeed, it is necessary to compare what is comparable: the video camera gives information for each pixel while the scanning laser Doppler vibrometer gives information on a pre-defined mesh. It is therefore important to be able to match pixels to points on the mesh by performing spatial synchronization operations. A comparative study is implemented on a test bench presenting a cantilever beam and observed with a high-speed camera (more than 2000 frames per second). In order to study a maximum of modes and to know if the camera is able to observe the first bending modes despite their low amplitude, a random excitation (white noise) excites the beam. The parameters of the first bending modes estimated by the video camera are compared to those obtained by the scanning laser Doppler vibrometer.

        Orateur: Touzet Jimmy
        Comparative study of the first bending modes of a cantilever beam from a video measurement
      • 11:00
        "SYSID#2 - Harmonic Modal Analysis Using Hydroelectric Runner Steady-State Strain Gauge Measurements" 20m

        Hydroelectric turbine runners' fatigue life is reduced by more recurrent off-peak operations induced by the integration of wind and solar energy sources to the power grid. This increase in fatigue degradation entails the need for a better comprehension of runners' dynamic loadings and responses. This needed knowledge includes a more accurate prediction of runners' modal characteristics and solicitations. It allows better-suited design of turbines to the fluctuating demand and improves fatigue analysis tools, life estimation and diagnosis tools. Analysis of Francis runners' operational vibrations has recently shown that non-trivial rotor-casing interactions (NTRCI) can give rise to resonance. Those NTRCI, induced by the interaction of turbine casing non-uniformities and natural modes of the runner, generate a richer harmonic response than typical rotor-stator interactions (RSI). It is proposed in the literature that the amplitude of each harmonic observed in the runner's forced response resulting from NTRCI, and RSI, is modulated according to its proximity to specific natural modes. The synchronous excitation from the casing is defined as periodic and independent of those modes, allowing the decomposition of the excitation in distinct Fourier coefficients. This brings an opportunity for a new, nodal diameter-specific, steady-state harmonic-based modal analysis method of the runner. This paper proposes a method to be best suited near the best efficiency point of operation, where harmonic excitations dominate stochastic excitations, non-ideal conditions for usual operational modal analysis algorithms. The proposed Bayesian inference-based methodology uses Francis runner strain gauge measurements and a harmonic forced response model to evaluate the modal parameters of the runner and the associated uncertainties. The probabilistic identification tool relies on a prediction error model using a Metropolis-Within-Gibbs sampling algorithm and a physical model-based likelihood to infer the excitation's Fourier coefficients, the modal parameters of the runner and a homoscedastic error. The modal identification method was implemented on a model simulating NTRCI and RSI harmonics observed in steady-state strain gauge measurements. The method was shown efficient to identify the modal parameters and the harmonic excitation of the model. In this paper, the algorithm's robustness is evaluated on real Francis runner prototype strain gauge measurements in steady-state operating conditions.

        Orateur: Morin Nicolas
        Harmonic Modal Analysis Using Hydroelectric Runner Steady-State Strain Gauge Measurements
      • 11:20
        "SYSID#3 - Identification of the Payne effect in a viscoelastic material coupling Bayesian identification and Digital Twin" 20m

        The Payne or Fletcher-Gent effect is a particular behavior occurring in viscoelastic materials containing fillers. It induces a nonlinear dependency of the viscoelastic storage modulus on the amplitude of the applied strain. This amplitude dependency must be taken into account in engineering applications as it does change the elastodynamic behavior of the overall structure integrating the material. A methodology is developed in the proposed work to identify this nonlinear effect. The approach first starts with the identification by Bayesian inference of the frequency and damping properties of a viscoelastic sample. The input data are obtained from a modified Oberst test, conducted for different displacement load amplitudes. A digital twin of the experimental set-up is then used to evaluate the stiffness behavior, and in particular the Young's modulus of the material, at the measured frequencies, and to deduce the strain level within the sample. This work shows that the Payne effect, which leads to the decrease with the amplitude of the Young's modulus, can thus be estimated in terms of mean and standard deviation by combining experiments and numerical simulations.

        Orateur: Jaboviste Kévin
        Identification of the Payne effect in a viscoelastic material coupling Bayesian identification and Digital Twin
      • 11:40
        "SYSID#4 - Parametric modal testing using slow but continuous variation of operating conditions. Illustration on a contact bench." 20m

        Several systems have vibration properties, modes and non-linear limit cycles, that are significantly affected by environmental properties such as temperature, pressure, rotation speed, ... The classical testing strategy is to select a number of fixed operating conditions, wait for the system to stabilize at each condition and extract the vibration characteristics once stabilized. A lot of test configurations may thus be required to finely characterize the influence of environmental parameters on the system dynamics. The novel point of view taken here is to consider that in a number of applications, one can generate a slow variation of the operating condition that is sufficiently below the frequencies of interest to allow continuous monitoring of vibration properties. The application chosen as illustration of the process is a test bench seeking to characterize contact stiffness properties in a brake subassembly focusing on piston/backplate and piston/chamber contacts. A first rough estimate of the relation between modal frequency and pressure changes is obtained by traditional modal tests. Then pressure and excitation frequency are linked together in a feedforward iterative testing process to track phase resonance, thus giving access to modal property (resonance frequency, damping and shape) evolution at all intermediate pressures. Demodulation is used to address leakage issues in the signal processing and is shown to give appropriate results for both the excitation fundamental and its harmonics. The notion of harmonic modulation is finally introduced: it is used to analyze the fluctuations of the response within a given cycle and thus provides an understanding of spatial variations of non-linearities. To allow detailed shape characterization, a 3D vibrometer scan is used and the optimization of the scanning procedure is discussed.

        Orateur: Malacrida Alves Guilherme
        Parametric modal testing using slow but continuous variation of operating conditions. Illustration on a contact bench.
      • 12:00
        "SYSID#5 - Estimation of the piezoelectric factor in nonlinear transducers" 20m

        This study focuses on the characterisation of non-linear and piezoelectric systems. The behaviour of a structure displaying a geometric elastic nonlinearity and coupled to piezoelectric elements is explored. The elaboration of piezoelectric transducers is based on the maximization of their coupling factor to ensure optimal performances. The measurement of the coupling factor is thus a key step in the design of a non-linear piezoelectric transducer. In the case of linear transducers, several equivalent definitions based on mechanical parameters are used to estimate the coupling. The differences between the short-circuit (SC) and open-circuit (OC) behaviours are exploited to assess the electromechanical conversion taking place. The validity of the equations based on natural frequencies, stiffness and energies are questioned in the non-linear case. The definition based on energy transfer inside the resonator could be extended to evaluate its coupling. For a non-linear and piezoelectric ensemble, it is supposed that the coupling will vary with the amplitude of the device. The coupling would not only be function of the equivalent parameters but also of the usage range. An experimental study is performed an on a clamped-clamped beam covered with piezoelectric patches. A first methodology based on a mechanical quasi-static cycling in open and closed circuit is proposed to estimate the coupling for a given solicitation. A loading and unloading cycle is performed while switching the electric condition between the two steps. Cycles were performed for increasing solicitation levels to visualize their evolution. This study can be performed by prescribing either the displacement or the force. The equivalences between these two commands and the order of the OC and SC steps are further questioned. A numerical model is developed to better apprehend the portion of mechanical and electrical energies to considerate, in order to understand the impact of the geometric non-linearity on the piezoelectric elements. A second dynamical experiment is conducted to extend the equations based on the OC and SC frequencies of the system. The piezoelectric beam is mounted on a vibrating to assess the evolution of its backbone with the amplitude solicitation. The two results are then confronted.

        Orateur: Peyrouse Floriane
        Estimation of the piezoelectric factor in nonlinear transducers
      • 12:20
        "SYSID#6 - Design of a Resonant Plate for Pyroshock Testing based on Shape and Size Heuristic Optimization" 20m

        Resonant plate testing is a common laboratory test method for qualifying space equipment which needs to withstand strong high-frequency shocks, usually activated by pyrotechnic devices and transmitted to the structure. Given their impulsive nature, shocks could critically damage onboard equipment and jeopardize the success of the mission. NASA-STD-7003A international standards are usually adopted to establish the requirements, in terms of a Shock Response Spectrum (SRS), for the qualification of space equipment according to the launch vehicle characteristics. To foster repeatability and safety in laboratories, the most common test facilities exploit the launch of an impacting object (e.g., hammers, dropping masses, pistons, or bullets) on a resonant plate on which the component under test is mounted. In this work, a numerical model able to completely simulate a pyroshock test is used to perform a shape and size heuristic optimization of a resonant plate to match the required SRS. Limiting the energy inputs, the performances of regular polygonal rather than irregular quadrilateral shapes and sizes are investigated and compared. The algorithm features an embedded Computer-Aided Design (CAD) modeler, a Finite Element (FE) solver, and a Genetic Algorithm (GA) optimizer, ensuring accuracy and flexibility in predicting the behavior of a resonant plate with a complex shape. The optimized design of a resonant plate permits improvements in both the SRS accuracy and the time and cost efficiency of its tuning.

        Orateur: Viale Luca
        Design of a Resonant Plate for Pyroshock Testing based on Shape and Size Heuristic Optimization
    • 10:40 → 12:40
      Survishno 2 / Fleet Monitoring
      Président de session: Konstantinos Gryllias
      • 10:40
        "FLEET#1 - Fleet Condition Monitoring in non-stationary conditions" 20m

        Machine condition monitoring is a challenging task. This paper presents a technique on the domain of rotating machine fleet condition monitoring for non-stationary operating with emphasis on gears and bearings as the most critical mechanical components. The machines are working on non-stationary conditions, on factual applications. This condition is causing an effect on the vibration signal as the operating frequencies are varying. As a result, the signal cannot be used as a condition estimator on its raw form. Additionally, the signal could contain more than one mechanical part interference, as well as casing or other resonances. While we present the factor of non-stationary conditions and signal decomposing, current studies do not follow this path. Most researchers require a large historical data as they use artificial intelligence on their try to be less unproductive. More than that, on this study a comparison between the whole fleet and a single machine will be presented, compared to cluster methods or machine to machine comparison. The proposed solution contains the decomposition of the vibration signals, received from the main mechanical components of each individual, in order to separate the internal forces that contain information on kinematic parts and are therefore potentially symptomatic of faults in gears and bearings, to be used for Machine Health Monitoring. This decomposed signal of each individual defines a stochastic process in the fleet provide different realizations. The deviating machine is spotted within a monitoring framework by controlling if any new measurement can be accepted as a realization of this stochastic process. The originality of this method stands on the equalization scheme which is able to remove the effect of speed-varying transfer functions, so as to normalize the signal with respect to the structural fingerprint of each individual part of the fleet. All the above, were applied in parallel with the experimental procedure, on a set of machines, acting as a fleet. Vibration signals were taken from three points of each mechanism. On the paper, the experimental procedure is presented in depth as well as the evaluation of signals and the complete methodology of the condition monitoring.

        Orateur: Achilleos Achilleas
        Fleet Condition Monitoring in non-stationary conditions
      • 11:00
        "FLEET#2 - Effective Identification of Cyclic Excitation and Resonance in Non-stationary Gearbox Vibration Monitoring" 20m

        Since as rotating machinery, the gearbox is an important source of vibration and the identification of signals and parameters of some critical components of the gearbox is necessary for the monitoring of the system. Inside this mechanism, the rotation of elements with periodic discrete geometries (bearings, gears, turbines, ...) is the origin for potential excitations which are commonly cyclic. These excitations are therefore characterized by frequencies that evolve with rotation speed Reconstructing them is equivalent to an inverse problem or an identification problem that can be treated utilizing the Least-Squares Complex Frequency (LSCF) estimator and its Poly-reference implementation to detect the temporal events. The main problem is to find the right external excitation conditions to reveal behaviors that are either resonances (a well-known linear problem and treated by identification) or internal excitations (less known and less well-treated). The difficulty here lies effectively in non-stationary operating conditions for good exploitation of time and angular Fourier Transform framework. In this work, a 1-stage gearbox is numerically developed to simulate various working conditions and generate signals which are necessary for the identification. The cyclic excitations generated by rotating elements with periodic discrete geometries pose a challenge in reconstructing their frequencies, as they vary with rotation speed but are definitely valuable excitations for FRF identification of structural parts. The presence of resonances and internal excitations further complicates the identification process, particularly under non-stationary operating conditions. To address this in this simulation, a careful balance is struck between the length of the temporal signal and the rate at which the rotational speed increases. This ensures effective separation of time-related and angle-related phenomena within the signals, facilitating accurate estimation of the Frequency Response Function (FRF) and resonance identification. Then by utilizing a meticulously induced swept frequency and appropriate working conditions, it becomes possible to mitigate temporal phenomena and achieve precise identification of cyclic phenomena.

        Orateur: Ahani Mojtaba
        Effective Identification of Cyclic Excitation and Resonance in Non-stationary Gearbox Vibration Monitoring
      • 11:20
        "FLEET#3 - Automated domain adaptation for bearings fault detection and classification" 20m

        Effective fault classification of mechanical components based on vibration signals requires robust domain adaptation methods, particularly in scenarios where complete data on all fault types is unavailable. The challenge lies in monitoring the health of components with variable characteristics, such as speed, load, and torque, where the lack of information on their current state renders target domain labels unavailable. Maximum Mean Discrepancy (MMD) is a widely used loss function in domain adaptation; however, its popular Gaussian kernel necessitates prior knowledge of the target domain data, posing a limitation in domain adaptation problems. To address this challenge, this paper proposes a novel approach that aims to achieve two key goals: (1) accurate fault identification and (2) automatic tuning of the MMD Gaussian kernel to adapt to different domains. The proposed approach consists of two stages. Firstly, a static representation of the standard deviation is identified using the Pascal Triangle that allows the reconstruction of a Gaussian kernel accurately. In the second stage, a dynamic parameter is computed by considering the difference between the source and target features distribution, extracted from the desired layers of the considered model. The proposed approach is evaluated by the classification of bearings health state, using the Case Western Reserve University (CWRU) signals as the source dataset and Jiangnan University (JNU) signals as the target one. The model employed in this evaluation is a 2D Convolutional Neural Network (CNN) model designed to process 2D reshaped signals. The desired layers for adaptation were identified as the fully connected layers of the network. The results demonstrate that the proposed method outperforms traditional MMD with a manually tuned Gaussian kernel, as well as other domain adaptation methods such as Correlation Alignment (CORAL) and Wasserstein distance that do not require any parameter tuning, making it a valuable contribution to the field of intelligent fault classification based on vibration signals.

        Orateur: Karkafi Fadi
        Automated domain adaptation for bearings fault detection and classification
      • 11:40
        "FLEET#4 - Results on Experimental Data Analysis of Independent Cart Systems in Non-Stationary Conditions" 20m

        This paper presents an analysis of the vibration data obtained from an independent cart system. The study focuses on comparing the vibration data collected with and without the presence of outer race faults, while the 12 roller cart/mover maneuvers at different speeds. Vibration data was captured using accelerometers positioned along the system and acquired through the Beckhoff PLC and National Instrumentation data acquisition systems. The data acquisition process involved running the mover under no-load conditions with and without an outer race fault, at speeds of 500 mm/s and 3000 mm/s. The analysis revealed valuable insights into the behavior and characteristics of the signals under different conditions. Results indicated that envelope analysis in the frequency domain is effective in identifying fundamental frequencies related to the system's dynamics. The calculated frequencies associated with the rotational speed of the mover and the outer race frequency provided a further understanding of the system's behavior.Despite some frequency smearing due to the the non-stationary conditions, fault frequencies were identifiable. The maximum peak hold plots demonstrated that at low speeds, the fault-related frequencies were more susceptible to noise interference, while at higher speeds, the fault frequencies became more prominent and distinguishable. This study contributes to the overall understanding and diagnosis of potential faults within independent cart systems, with further analysis planned to precisely identify the fault frequencies.

        Orateur: Jabbar Abdul
        Results on Experimental Data Analysis of Independent Cart Systems in Non-Stationary Conditions
      • 12:00
        "FLEET#5 - Performance study of DTW-based spike measurement anomaly detection in sensors on real world tests." 20m

        Anomaly detection in sensor data plays a crucial role in various applications, including predictive maintenance, quality control, and prototype testing. The objective is to identify patterns that deviate from the expected results in the time series. This can be achieved through statistical analysis that exposes outliers or using domain knowledge to understand the expected values. Additionally, machine learning models can learn normal sensor behavior and detect anomalies. In the automotive industry, improving the efficiency and effectiveness of prototype testing is essential to reducing errors. This is a complex challenge due to harsh testing conditions and the complexity of some tests, such as the need to monitor over 50 channels simultaneously during an NVH vehicle validation test. Furthermore, typical vehicle tests must be performed on specific test tracks or benches, with limited access time, making it imperative to avoid repeating tests due to a defective acquisition. This study focuses solely on one specific type of anomaly, known as €œSpikes€, which are sharp outlier values with no correlation to surrounding samples. In addition, is important to highlight that its only of interest any issue regarding the acquisition chain (sensors, cables, recording device), the object under test and the specific test procedures or maneuvers are not considered. Therefore, it is assumed that this anomaly detection algorithm must be capable of detecting spikes disregarding the application. The study presents a novel Dynamic Time Warping (DTW) technique for detecting these spikes in real-time, multi-channel acquisitions during automotive testing. The technique has been validated on a 24-signal dataset, consisting of 12 anomalous and 12 non-anomalous signals with varying dynamic ranges, patterns, lengths, and sensor types. The results show the tool's accuracy in avoiding false positives, such as mistaking spikes for other physical impulses during the test, like tires squeaking, or any other physical impulse coming from the engine or other sub-component of the vehicle under test. Moreover, the study demonstrates the ability of the Dynamic Time Warping technique to effectively detect spikes in sensor data, leading to improved efficiency and effectiveness in prototype testing in the automotive industry.

        Orateur: Cornelis Bram
        Performance study of DTW-based spike measurement anomaly detection in sensors on real world tests.
      • 12:20
        "FLEET#6 - Investigating the usage of Proxy-A-Distance as a measure of dataset shift detection and quantification in an automotive booming noise classification setting" 20m

        "Machine learning methods, which can be effective tools for NVH end-of-line testing applications, are typically data-demanding. Techniques like transfer learning and data augmentation have been proposed to overcome this need. For effective transfer learning, it is imperative to compare the source and target datasets and assess the disparity between the two.

        In this paper, we study the applicability of Proxy-A-distance for dataset shift detection and quantification. We use the case-study of booming noise detection in automotive end-of-line quality testing with simulated class distribution and mixture component shifts for this investigation. The Proxy-A-distance method works with the help of a domain classifier and emerges as a straightforward and reliable procedure for shift detection. Furthermore, its gradual monotonic rise with increasing dataset divergence makes it suitable for shift quantification. However, it is important to note that there seems to be a low correlation between Proxy-A-distance and transferability, which warrants further exploration."

        Orateur: Kunte Deepti
        Investigating the usage of Proxy-A-Distance as a measure of dataset shift detection and quantification in an automotive booming noise classification setting
    • 10:40 → 12:40
      Survishno 3 / Signal Processing
      Président de session: Gianluca D'Elia
      • 10:40
        "SIGPRO#1 - Towards a self-evaluation of the reconstruction of the instantaneous phase of a vibration signal" 20m

        The cyclo-stationary nature of rotating machine signals is an essential lever for most diagnostic strategies. For this lever to be operated, the signal must be analysed with a detailed knowledge of the machine's instantaneous frequency. In a context where many monitoring systems do not have access to an adequate angular sensor, the extraction of the instantaneous frequency of the machine from the vibration signal alone has been the object of particular attention from the community. Thus, multiple tools have been developed in different contexts/environments, often with reduced efficiency as soon as the study case moves away from the one for which the tool was created. This work focuses on the proposal of a key performance index (KPI) evaluating the quality of the reconstructed instantaneous quantity. Such a tool can be useful for determining the most suitable instantaneous frequency extraction method for the case under study, for assisting the expert in the possible parameterisation of the method chosen, or simply for discriminating between the signals from which it is relevant to perform a cyclostationary diagnosis. The KPI is only based on the analysis of the estimation of the instantaneous phase of the signal. While the instantaneous phase is traditionally used to diagnose balancing problems, this quantity also has the advantage of being able to superpose (spacially synchronize) signals obtained at different stages of the system's life. While the adaptation of instantaneous frequency extraction tools to estimate the instantaneous phase does not pose any major difficulty, the construction of the KPI is subject to a number of cautions. Several KPI proposals will be presented in this paper, and successively tested on a simulated and then real measurement campaign. The signals studied are those transmitted by inductive sensors observing the shaft orbit of several power plant turbines. The effectiveness of the KPIs will be assessed by tachometers installed on these machines. The indicators will then be compared with another campaign using the signal emitted by an accelerometric sensor installed on a mechanical transmission involving an epicyclic gear train.

        Orateur: Andre Hugo
        Towards a self-evaluation of the reconstruction of the instantaneous phase of a vibration signal
      • 11:00
        "SIGPRO#2 - Time and angle analysis of Instantaneous Angular Speed signal: impact on average velocity and order spectrums" 20m

        To this day, monitoring of rotating machinery by vibration signals has often relied on time-domain signal analysis, for bearing or gear faults to be recognized by the frequency at which they create their very specific impact. However, time domain analysis is not ideal for it implies assuming that the shaft rotates at a constant speed. Even if in certain conditions the rotating speed may be forced around a constant value for the need of the test, the assumption is not valid as long as the defects we are looking for have an impact on the rotating speed. To overcome this problem, an emerging approach consists in measuring the signal of Instantaneous Angular Speed (IAS) directly on the shaft, by the means of an angular coder: the €œElapse-Time€ method gives the angle-time relationship which allows one to get rid of the constant speed assumption, and analyze the system based on its true velocity. Knowing the true velocity also allows signal processing in the domain of angular frequencies, particularly useful for fault detection. In this paper, a one-degree-of-freedom rotating system is simulated under different periodic excitations, in order to extract its IAS signal. A study of the corresponding spectrums in terms of time-domain and angle-domain frequencies shows a shift in peak locations from one domain to the other. This interrogates the constant speed assumption used for order spectrum analysis, and shows a distinction between angle-averaged and time-averaged velocity.

        Orateur: Burel Arthur
        Time and angle analysis of Instantaneous Angular Speed signal: impact on average velocity and order spectrums
      • 11:20
        "SIGPRO#3 - Enhancing the Performance of the Multi-Order Probabilistic Approach in Angular Speed Estimation through Adaptive Window Selection" 20m

        The Multi-Order Probabilistic Approach (MOPA) is a method for estimating the instantaneous angular speed from vibration signals with high precision. However, determining the optimal choice of inputs can be challenging as the method requires many input parameters, including the window size selection used to generate the spectrogram for computing the probability density functions. This study presents a new approach that utilises statistical indicators to evaluate the information content of various window sizes and selects the optimal size for each signal segment. The algorithm is tested using simulated sensor data and real data from offshore wind turbines, and the results are compared to benchmark parameters such as the Root Mean Squared Error (RMSE). The results show that the proposed approach provides similar accuracy in estimating angular speed without the necessity to pre-select a window size for the spectrogram. Overall, this study demonstrates that using an adaptive window based on statistical indicators can improve the flexibility of MOPA as well as reduce the time required to detect the appropriate window for each data set.

        Orateur: Protopapadakis Georgios
        Enhancing the Performance of the Multi-Order Probabilistic Approach in Angular Speed Estimation through Adaptive Window Selection
      • 11:40
        "SIGPRO#4 - Exploring the potential of sparse spectral estimation for vibration analysis" 20m

        Recently, sparse signal processing algorithms have attracted much interest in the field of vibration-based machine monitoring as they leverage the assumption that most vibrations are sparse in at least one domain. While a large body of research focuses on the application of sparsity for gear and bearing fault detection in the time and envelope domain, this paper investigates the performance of a sparse data-adaptive spectral estimator as a precursor tool for other vibration signal pre-processing techniques. The Sparse Iterative Covariance-based Estimator or SPICE is a hyperparameter-free sparse spectral estimation method that has some promising statistical properties when compared to the standard Fourier transform. These characteristics make SPICE an interesting candidate for the analysis of complex vibration data coming from industrial machinery. This paper examines the efficacy of SPICE for vibration analysis by employing it for instantaneous angular speed estimation through integration with the multi-order probabilistic approach. Its advantages and shortcomings are explored and compared to the standard short-time Fourier transform as well as the short-time iterative adaptive approach. This new SPICE-based speed estimation approach is validated on realistic simulated data. The performance of the proposed methodology is then showcased on two complex wind turbine gearbox vibration data sets.

        Orateur: Peeters Cédric
        Exploring the potential of sparse spectral estimation for vibration analysis
      • 12:00
        "SIGPRO#5 - Design considerations for smartphone camera-based rotational speed measurement" 20m

        "Condition monitoring of rotating machinery gains importance in order to optimally schedule maintenance and to guarantee operation safety and production efficiency. Varying speed conditions are common in rotating machinery operations, but pose a challenge for vibration analysis. Nevertheless, a direct measurement or an indirect estimation of the rotational speed can simplify the monitoring process. Recent studies have shown that a smartphone's low-cost camera can serve as a rotational speed measurement tool, even though it has a low frame rate, by exploiting the rolling shutter effect. In rolling shutter cameras the pixel lines of an image are not sampled simultaneously, but sequentially at a sample rate that is much higher than the frame rate. This rolling shutter frequency results in a very high Nyquist limit allowing for the measurement of high rotational speeds.
        A recent study proposed a theory and the corresponding methodology to measure the rotational speed of a shaft using a smartphone's rolling shutter camera pointed towards the curved surface of the shaft around which a wide zebra tape was glued. Due to the rolling shutter effect, the stripes are captured as inclined lines of which the angle depends on the rotational speed. The goal of this paper is to investigate and validate the influence of various design parameters, e.g. the dimensions of the zebra tape, on the rotational speed measurement. The design considerations are applied on an in-house experimental drivetrain running at shaft speeds in the range from 5 to 50 revolutions per second and their effect is evaluated in order to conclude with general design rules for smartphone camera-based rotational speed measurement."

        Orateur: Verwimp Toby
        Design considerations for smartphone camera-based rotational speed measurement
      • 12:20
        "SIGPRO#6 - Extraction of the acoustic modal content of a turbofan engine in non stationary conditions using order analysis" 20m

        The analysis of acoustic fields in turbofan engines is a challenging task, due to the geometrical complexity of the ducts, of the multiple noise generation mechanisms and the poor signal-to-noise ratio (SNR) related to the strong flow noise. Modal analysis can be carried out using different kinds of microphone arrays. It is proposed in this work to conduct the analysis on non-stationary data using order extraction based on angular resampling of signals. A modal analysis of tonal fan noise is applied to deceleration measurements collected during the European project TurboNoiseBB at the Anecom test facility. Orders are extracted from time domain microphone signals so as to extract the instantaneous envelope and phase. The instantaneous modal content is then extracted using modal identification. The azimuthal mode content of the three first Blade Passing Frequencies (BPF) is obtained through an iterative Bayesian inverse approach. Particularly the modal spectrum for the second BPF shows an interesting pattern with peaks and troughs throughout the deceleration. It suggests that detected modes are influenced by the azimuthal distribution of probes. Results are compared to steady state measurements at different operating points obtained from the same test rig.

        Orateur: Miranda-Fuentes Johann
        Extraction of the acoustic modal content of a turbofan engine in non stationary conditions using order analysis
    • 12:40 → 14:00
      Lunch Break 1h 20m
    • 14:00 → 16:00
      JISFA 2 / Environmental Noise and Perceptive approaches
      Président de session: Adrien Pelat (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613)
      • 14:00
        JISFA2#1 - Applications of experimental processes to large set-up (videos) 20m
        Orateurs: Adrien Pellat, Frank Simon
      • 14:20
        JISFA2#2 - Short term annoyance due to aircraft flyovers 20m

        Noise annoyance can be defined in two different ways, depending on whether the sound is listened actively or whether this noise is heard passively. In the first case, the noise can produce an immediate reaction of unpleasantness, linked to its sound signature. In the second case, the noise is not listened to, but suffered. The noise annoyance can then correspond to the short-term disturbance of an activity and, on a longer term, to the feeling of annoyance towards the sound environment.
        In this first lecture, the researches which focus on unpleasantness will be presented, as well as those which focus on activity disturbances. These researches are generally based on laboratory experiments. Five important perceptual dimensions can be observed concerning the sound characterization an aircraft flyover, and a sixth one which is less important. The first corresponds to the loundness, and the second is linked to the tonalities which emerge between 3 and 5 kHz and on which the Doppler effect is clearly heard. The third corresponds to the duration of the flyover and the fourth to the fluctuations heard on the sound rise of the flyover. As for the fifth, it corresponds to the very close harmonics often heard when the aircraft takes off. This effect, called "buzz saw" gives a very rough aspect at the beginning of the sound. The sixth factor, perceptually less important than the others, is the low pitch broadband noise, often associated with jet noise and engine combustion sources. When the flyovers are repeated and participants no longer actively listen to the aircraft, the previous perceptual dimensions lose their importance. Some people build their annoyance feeling on the number of flyovers, others on the maximum noise level, finally others integrate these two elements to focus on the equivalent noise level.

        Orateur: Catherine Lavandier (ETIS laboratory, CY Cergy Paris University)
      • 14:40
        JISFA2#3 - Long term annoyance and its impact on health 20m

        In this second presentation, researches that focus on the overall feeling of residents towards their sound environment will be presented. This corresponds to the long-term annoyance. It can only be studied in the field, during surveys. Numerous studies show that this does not simply depend on the exposure of populations to noise, which explains only a third of the annoyance expressed by the residents. The annoyance also depends on so-called non-acoustic factors. An individual sensitivity to noise in general is one of the important non-acoustic factors, which also explains around a third of the annoyance reported. The last third could be explained by the attitude of residents towards their strategies to cope with noise problems. These three factors, one being directly linked to the sound environment and the other two linked to non-acoustic parameters, share the explanation for the long-term annoyance. Long term annoyance is one of the critical health outcomes among cardiovascular disease, effects on sleep, cognitive impairment, hearing impairment and tinnitus. The WHO environmental noise guideline for Europe considers that the risk for noise to have an impact on the health starts when 10% of the population is highly annoyed. According to the WHO, average aircraft noise exposure above about 45 dB is associated with adverse health effects, because 10% of the population is highly annoyed when living at LDEN 45 dB. This value is much lesser than the average exposure for road or railway noise exposure (53 dB). This value is still debated with new method of analysis or with new data around French airports.

        Orateur: Catherine Lavandier (ETIS laboratory, CY Cergy Paris University)
      • 15:00
        JISFA2#4 - Uncomfort due to noise and vibration in aircrafts 20m

        The comfort of aircraft passengers depends on many factors, including the noise and vibrations passengers are submitted to. Levels can be very high (in helicopters in particular) or more reasonable (large airplanes). In some cases (e.g. engines), the same source is responsible for both sensations. Noise assessment procedures use classical indicators (A-weighted level, possibly penalised according to tonal emergence). For vibrations, the ISO 2631 standard is used. This presentation will show examples of the use of these indicators in helicopters and aircraft. For noise, the work has shown that, as in many cases, loudness metrics are able to predict discomfort better than the overall A-weighted level. For vibration, the indicator proposed by the ISO standard can be slightly improved by considering amplitude modulations that may be present in engine vibration signals. Finally, a combined noise/vibration discomfort model will be proposed.

        Orateur: Etienne Parizet (Laboratoire Vibrations Acoustique, INSA-Lyon)
      • 15:20
        JISFA2#5 - Aircraft noise certification: evolutions & trends 20m

        Focusing on external certification of aircrafts: current procedures, test means evolutions, future regulations at ICAO level.

        Orateur: Stefan Moal (AIRBUS, Toulouse)
      • 15:40
        JISFA2#6 - Round Table JISFA2 20m

        Additional Q&A and open discussion about the session topic.

    • 14:00 → 16:00
      Survishno 4 / Dynamic modelling 1
      Président de session: M. Ouisse (SUPMICROTECH-ENSMM, Univ. Franche-Comté, CNRS, institut FEMTO-ST)
      • 14:00
        "DYN1#1 - A model-based approach for the NVH performance improvement of Soft Close Actuators for automotive applications" 20m

        The aim of this paper is to describe a numerical vibro-acoustic methodology, experimentally assessed, for the estimation of the overall vibratory and acoustic level of a Soft Close Actuator (SCA), for automotive applications. The process is carried out in order to develop a digital twin enable to represent the Noise, Vibration and Harshness (NVH) behaviour of the real mechanism. The vibro-acoustic model is the combination of three sub-models: a multibody (MB) model, a structural finite-element (SFE) model and an acoustical boundary-element (ABE) model. The MB model is used to obtain the reaction forces on the case of the actuator during working conditions. Reaction forces are employed as an input for the further SFE dynamic model to evaluate the dynamic response of the SCA's case, which is the only meshed part. The dynamic response is exploited to set-up an ABE model which allows to estimate the noise generation in terms of acoustic properties. The numerical simulation results are validated using experimental data acquired on a real SCA. The developed model is a powerful tool for the improvement of NVH performance of the analysed actuator.

        Orateur: Natali Caterina
        A model-based approach for the NVH performance improvement of Soft Close Actuators for automotive applications
      • 14:20
        "DYN1#2 - Optimization of the Energy Input and Output Parameters for Pyroshock Testing" 20m

        During the different phases of a mission, spacecraft equipment is subject to high-frequency shocks, commonly known as pyroshock since they are caused by explosive materials. The requirements for space qualification are usually expressed in terms of Shock Response Spectrum (SRS) acceleration €“ indicating the damage that a shock could potentially cause to a structure €“ and depend on the characteristics of the launch vehicles. Therefore, to simulate real shock load conditions in a repeatable and safe manner, hammers, projectiles, or generic impacting objects are commonly launched against a fixture (such as an anvil, a Hopkinson bar, or a resonant plate) supporting the component under test. This work exploits a numerical model entirely developed in the frequency domain which has been built to optimize the energy input and output parameters of a test facility. The development of the proposed model in the frequency domain contributes to considerable advantages, such as the reduction of calculation times. Based on the SRS requirements, the mass and speed of the bullet, the impact location, and the position of the object under test (mounted over a triangular plate) are comprehensively investigated by the proposed algorithm. Given the notable influence of these parameters and the complexity of their interactions, the proposed analysis considerably simplifies the tuning process, reduces calibration time, and, in general, avoids costly trial-and-error repetitions.

        Orateur: Viale Luca
        Optimization of the Energy Input and Output Parameters for Pyroshock Testing
      • 14:40
        "DYN1#3 - Numerical analysis of the dynamic behavior of rotor shafts in permanent magnet synchronous machines" 20m

        In order to design permanent magnets synchronous machines as quiet as possible, we propose a predictive magneto-mechanical vibration model. This model allows for describing the dynamics of the system under operating conditions by including a strong magneto-mechanical coupling between the electromagnetic forces and the dynamics of the motor shaft. The Maxwell pressures are calculated analytically using the subdomain method, while the rotor dynamics are based on a 1D beam model. The coupling between these two physical phenomena is solved using time integration. This numerical study aims first at highlighting the strong coupling regimes and their influence on the harmonic content of the displacement. The second objective is concerned with the study of the influence of static eccentricity on the spectral content of the dynamic response.

        Orateur: Poupon Thomas
        Numerical analysis of the dynamic behavior of rotor shafts in permanent magnet synchronous machines
      • 15:00
        "DYN1#4 - Control of an acoustic mode by a digitally created Nonlinear Electroacoustic Absorber at low excitation levels: Analytical and Experimental results" 20m

        Noise reduction becomes important to raise acoustical comfort and to prevent the population from additional disorders. Acoustic passive devices such as acoustic foams and passive resonators have their limitations. Active Noise Control has a high energy consumption and works well only in local zones. The concept of Impedance Control is very efficient around its resonant frequency. The Impedance Control concept lies on a loudspeaker collocated to microphones, and equipped of a processor, named as Electroacoustic Absorber. By calculating and injecting the electrical current into the loudspeaker coil based on the sensed pressure, one can be able to provide a desired linear behavior for any loudspeaker. This linear behavior is generally experimentally implemented through the Infinite Input Response (IIR) method to control the absorption and reflection conditions on the loudspeaker membrane. The IIR method is limited to linear regimes. Passive nonlinear systems are efficient for large frequency widths, both in transient and stationary regimes, but usually activated at high excitation amplitudes. As a result, our work focuses on the creation of a Nonlinear Electroacoustic Absorber (NEA) at low excitation amplitudes. This work is based on an innovative method, allowing to digitally program any nonlinear desired behavior of the loudspeaker for all excitation amplitudes. It is a Real Time Integration (RTI) method, where the current is calculated at each time step, based on the measured pressure. To show the advantages of this device, a NEA is coupled with an acoustic mode of a tube at a normal incidence. Multiples nonlinear behaviors are shown at low excitation amplitudes, and compared with linear control devices. Analytical and Experimental results will be presented and discussed.

        Orateur: Morell Maxime
        Control of an acoustic mode by a digitally created Nonlinear Electroacoustic Absorber at low excitation levels: Analytical and Experimental results
      • 15:20
        "DYN1#5 - Exploring the Impact of Defect Geometry on Bearing Dynamic Behavior Using Spall and Indentation Models" 20m

        Dynamic modeling of a bearing is an effective tool that is used to research the bearing behavior under various conditions. Models yield clean results that include data that cannot be obtained through direct experimentation, such as internal forces and the positions of rolling elements. By simulating a faulty bearing, it is possible to thoroughly examine the effect of the defect's shape on the bearing's dynamic behavior. The model enables the exploration of correlations between the severity of the defect and observable outcomes (such as outer ring vibrations). These insights are essential for developing diagnostic techniques to assess the bearing's condition during operation.  This study introduces two types of defect geometries incorporated into a previously validated deep-groove dynamic model. The first is an improved spall model that represents the entrance and exit of the spall as linear-like slopes, and the second is a model of indentation defect. A comparison of the defect geometries indicated that both indentation and spall exhibit a similar pattern of dynamic behavior. The study's results demonstrate that under specific conditions, the slope of the spall's exit may affect the duration of the ball-defect interaction and the maximum acceleration of the outer ring. However, when only the intermediate range of exit slopes was examined the effect on the bearing dynamics was negligible. This could be explained by the relatively large size of the ball compared to the defect depth that causes it to strike the same location at the spall's exit. Conversely, the slope of the spall's entrance has no influence. This can be attributed to the free fall of the ball into the spall, which prevents the interaction of the ball with the entrance.

        Orateur: Schwarcz Ori
        Exploring the Impact of Defect Geometry on Bearing Dynamic Behavior Using Spall and Indentation Models
      • 15:40
        "DYN1#6 - Unknown load torque estimation on rotary drivetrains with exploitation of angular periodicity in an extended Kalman filter" 20m

        Accurate knowledge of the torsional vibrations is key for condition monitoring, control and design optimization of mechatronic drivetrains. An often applied solution is to combine the knowledge of the system dynamics in the form of a physics-based model with an informative but limited set of measurements in a stochastic estimation algorithm. A critical difficulty in this estimation arises from the external torques acting on the drivetrain. For many rotary drivetrains, these external torques show cyclic behavior in function of the rotational position to some extent. This contribution presents a strategy to exploit such behavior in an augmented extended Kalman filter. The strategy is experimentally validated on a mechatronic drivetrain setup, consisting of a back-to-back induction motor with a cardan axle connection. The root-mean-square error of the estimated load torque with regard to a validation torque sensor is used to compare the results with a conventional estimator. A significant reduction in root-mean-square error is obtained for the new strategy as compared to the conventional augmented Kalman filter for a number of validation experiments. The amount of error reduction is shown to be dependent on the relative contribution of the cyclic term to the overall unknown torque.

        Orateur: Van Der Veken Thijs
        Unknown load torque estimation on rotary drivetrains with exploitation of angular periodicity in an extended Kalman filter
    • 14:00 → 16:00
      Survishno 5 / Fault Diagnostics & prognostics 1
      Président de session: Hugo André
      • 14:00
        "FDP1#1 - A federated learning approach for rolling bearing fault diagnosis on data sources with imbalanced class distribution" 20m

        Rotating machinery fault diagnosis is a field of intensive research, attracting the last years a particular interest for data-driven methodologies such as Machine Learning and Deep Learning. In order to build such models, the general assumption is that a sufficient number of healthy and fault samples, collected under various working conditions, are available for model training. This assumption is often not realistic in real industry. This limitation could be avoided by exploiting data sets collected at multiple industrial partners, but this is in practice not easily feasible since companies prefer not to share their data for privacy reasons. Federated Learning (FL) is an emerging machine learning approach proposed to train a global model without sharing data among users. In this context an FL methodology for fault classification based on Convolutional Neural Networks (CNN) is proposed in this paper. Local models are trained on local data sets, each one owned by a single client, i.e. an industrial participant, and are then aggregated at a server level. In the baseline Federated Learning approach, local models optimization fails to increase global accuracy with model aggregation in case there is significant statistical heterogeneity in the data distributions among clients. Thus the aim of this paper is the proposal of an enhanced strategy that accounts for adaptive local updates and the comparison of its performance with state-ofthe-art techniques. Each participant computes the local stochastic gradients within an adaptive interval, set by the server at the aggregation step, when the models are loaded by the participants at the end of each communication round. The improved method is applied for bearing fault diagnosis and its effectiveness and accuracy are evaluated in the case of imbalanced class distribution in rolling bearing fault local data sets, i.e. considering a scenario where fault types are non-independent and identically distributed (non-i.i.d.) among clients. This case is addressed in literature to be one of the main challenges in FL and is of practical interest since skewed data sets are common in real-world factories.

        Orateur: De Fabritiis Fabrizio
        A federated learning approach for rolling bearing fault diagnosis on data sources with imbalanced class distribution
      • 14:20
        "FDP1#2 - CANCELLED - Cyclic monitoring of the Remaining Useful Life RUL for the Bearing fault prognosis" 20m

        Over the last decades, Prognostics has played a dominant role in preventive maintenance in manufacturing. It usually involves estimating the Remaining Useful Life (RUL) or the Time to Failure (TTF) of mechanical systems. In Prognosis, the analysis could often be purely data-driven (Trend Analysis). It requires a vast data set and offers the double benefit of being both applicable in many systems and being relatively precise. The published works demonstrate that the Trend analysis tends to be the most adopted method in prognosis, most notably for bearing fault. Nonetheless, other learning techniques such as Bayesian Network, Convolutional Neuronal Network or the Support Vector Machine (SVM) [1] are also applied to predict the evolution of the bearing's degradation and forecast the exact time of its failure in hours and minutes. In this paper, we implemented the « Threshold Data » approach, considering the limited amount of run-to-failure data and the fact that the Bearings' features are suitable for a degradation model creation [2]. Moreover, this study sheds light on Cyclic Prediction Method, intending to prove that cyclic monitoring could also estimate the RUL for the bearing fault with the integrity of data, and precisely track the degradation. Firstly, different filters are compared: Simple Moving Average (SMA), Cumulative Moving Average (CMA) and Exponential Moving Average (EMA). After that, both the Principal Component Analysis (PCA) and Model Fitting are deployed in order to construct a degradation model and fit the exponential function to the last n data. Finally, using the selected indicators, we managed to estimate the RUL of the bearing cyclically, thus exhibiting accurate predictions throughout each phase of its life till failure. Keywords: Prognosis, Preventive Maintenance, Remaining Useful Life (RUL), Cyclic monitoring. References:  Dong S., Luo T. (2013), €œBearing degradation process prediction based on the PCA and optimized LS-SVM model,€ Meas. J. Int. Meas. Confed, 46, 9, pp. 3143€“3152 Hai Q., Jay L., Jing L. (2006), €œWavelet Filter-based Weak Signature Detection Method and its Application on Roller Bearing Prognostics.€ Journal of Sound and Vibration, 289, pp.1066-1090

        Orateur: Benyagoub Abderrahmane
        Cyclic monitoring of the Remaining Useful Life RUL for the Bearing fault prognosis
      • 14:40
        "FDP1#3 - Modelling and diagnosis of a crack of a bearing inner ring" 20m

        Cracking of the inner ring of a bearing can occur for various reasons: poor handling during assembly, thermal stresses during start-up, etc. This results in a loss of tightness between the ring and the shaft with various consequences: fretting corrosion, heating of the inner ring leading to a reduction in the operating clearance and then to the blocking of the bearing, or even the rotation of the shaft in the bearing bore. The collateral damage can therefore be very significant. The difficulty in diagnosing a crack on the inner ring of a bearing is that it will generate a vibratory signature similar to that of other defects located on the inner ring (marking, corrosion, grooves due to the passage of leakage currents, etc.) but with very different consequences. With the experience accumulated at DYNAE in particular on paper machines, some veteran experts are able to identify a cracked ring simply by listening to the measured signal. However, a theoretical formalization of this technique seems desirable in order to achieve a reliable and semi-automated detection. This requires the modelling of the phenomena induced by a cracking and their associated vibratory signature. The method proposed here is based on the concept of Modal Kurtosis, as an improved version of the Spectral Kurtosis allowing to identify the frequency range in which shocks are the most readable, by considering the physical parameters associated with the excitation of the natural modes. The analysis of the envelope signal obtained after filtering is then performed in the spectral and time domains, by estimating the extent of the bearing load zone, the number of shocks per revolution due to rolling elements passing over the crack, and the attenuation factor of the spectral envelope of the modal response. The latter allows to identify another source of shocks, in particular a rotational shock due to a clearance. These processing allow an expert to orientate the diagnosis either towards a ring crack with loss of tightening, or towards other types of defects affecting the inner ring (spalling, grooves). Industrial case studies are presented to illustrate the technique.

        Orateur: Combet Francois
        Modelling and diagnosis of a crack of a bearing inner ring
      • 15:00
        "FDP1#4 - Study of bearing fault detectability on a rotating machine by vibro-acoustic characterisation as a function of a noisy surrounding machine" 20m

        In a context study linked to the preventive maintenance of industrial machinery, a vibro-acoustic study aimed at identifying the defect presence in cylindrical bearings was carried out under noisy environmental conditions. The simultaneous acquisition of acoustic and vibratory signals providing information on the condition of the bearings was carried out using microphonic and accelerometric sensors positioned in close and direct contact with an engine test bench. Three types of roller bearing were studied: one with no fault and two others with different numbers of faults. The signals acquired were analysed using the RMS, Kurtosis and Talaf indicators and discussed as a function of the rotating machine speed and the signal-to-noise ratio in the time domain. The acoustic and vibration indicators are compared and analysed to assess their relevance for fault detection.

        Orateur: Attal Emmanuel
        Study of bearing fault detectability on a rotating machine by vibro-acoustic characterisation as a function of a noisy surrounding machine
      • 15:20
        "FDP1#5 - On the detection of rolling contact fatigue in large bearings using roller embedded sensors." 20m

        Sensor Roller (SR) system is a prototype technology, developed by SKF, that uses sensors embedded in bearing rollers to collect data on load, roller accelerations & rotation speed, and temperature. This data is then transmitted wirelessly to operations and maintenance staff, allowing them to monitor the state and performance of machinery in real time. The system is currently being used in customer pilots for design validation and is being tested in various types and sizes of bearings. There is a desire to further analyse the data collected to improve the performance and reliability of new bearings. Towards this approach, a study was conducted to determine if SR can be used to detect surface damage. Early detection of damage is important to prevent problems such as increased friction, decreased efficiency, and, therefore, reduced lifespan of the machinery. Gathering the data for this study involved artificially damaging an SKF Nautilus bearing raceway and testing it at the SKF Sven Wingquist Test Centre (SWTC), where the testing rigs are designed to reproduce all the loads that this bearing may experience in operation. The Nautilus bearing, mainly used as a main shaft bearing for offshore wind turbines, has an outside diameter of up to four meters and a weight of up to 14 tonnes, and an expected bearing life of more than 20 years. A testing protocol of various machine speeds and loads was initiated with the SR. The damage detection algorithm is comprised of mapping multiple records of data around the bearing race by performing signal processing techniques on the SR data. The results of the study show that it is possible to detect damage on a bearing raceway using a SR.

        Orateur: Baggerohr Stephan
        On the detection of rolling contact fatigue in large bearings using roller embedded sensors.
      • 15:40
        FDP1#6 - 20m
    • 14:00 → 16:00
      Survishno 6 / Condition monitoring 1
      Président de session: Jérôme Antoni
      • 14:00
        COMO1#1 - Acoustic Monitoring of Rolling Element Bearings using a Sparse Microphone Array 20m

        Acoustic monitoring of rolling element bearings in industrial environments can provide a non-contact solution for early detection of bearing failures and prevention of costly downtime. Nevertheless, extracting the bearing signature of interest from the other contributions in the acoustic signals remains a key challenge. A possible approach to improve the Signal-to-Noise Ratio (SNR) of bearing signatures is to exploit spatial information, obtained using acoustic imaging techniques. However, Conventional microphone arrays are rarely used for bearing condition monitoring in an industrial setting due to their extensive number of microphones, which leads to high complexity and cost. In this paper, we propose a sparse microphone array, exploiting the modulation characteristics of the bearing signatures. The process begins with the extraction of the envelope from the raw signals, followed by frequency domain conventional beamforming on the envelope signals to generate acoustic imaging maps of the bearing signals. These maps are then used to optimize an indicator within a predefined region, effectively separating the bearing signal. The validity of this methodology is confirmed through numerical simulations of point sources in a free field environment, using simple amplitude modulated signals and signals generated with a phenomenological rolling element bearing model. Finally, experimental results collected from an in-house test rig demonstrate the robustness and accuracy of the proposed methodology in industrial environments.

        Orateur: Wu Xian
        Acoustic Monitoring of Rolling Element Bearings using a Sparse Microphone Array
      • 14:20
        COMO1#2 - Automatic Processing of Air Gap Monitoring Signals in Hydro-Generators 20m

        In run-of-the-river hydroelectric power plants, compact turbines are typically used to minimize disturbance to the river flow and preserve the ecosystem. However, compact turbines are often designed at the cost of a small air gap which can lead to rotor-stator collisions. Such accidents lead to a temporary halt in production, can cause collateral damage and are very expensive to repair. Air gap monitoring is therefore the best solution to predict and prevent such collisions.  Due to recent industry needs, we consider the case of stator deformation with dynamic rotor eccentricity. Specifically, we opt for a set of capacitive sensors attached to the rotor to monitor the air gap (wireless transmission) and a set of displacement sensors to monitor the rotor eccentricity (wired transmission). Each set is associated with a keyphasor. However, if the literature in the field of air gap monitoring is rich in sensor instrumentation, it is rather poor in signal processing, especially for the case of rotor deformation. In this work, we take a first step to fill this gap and propose a solution to automatically process the recorded signals.  After resampling the signals in the angular domain, a first classic step is to consider the cyclo-stationary properties of the recorded signalÌ€s. The first order component provides a denoised signal while the second component can quantify its relative strength. Next, we present two main results. (1) We show how to compute the stator profile by correcting the eccentricity of the measured air gap, which is useful for long-term maintenance. (2) For each point of the stator, we calculate the critical air gap that corresponds to its closest position to the rotor which is essential to protect the turbine in real time. Furthermore, this solution allows to considerably reduce the size of the data, thus facilitating its storage. Finally, we propose visualizations that can be interpreted by experts in the field who have no prior knowledge of the techniques used. In particular, we display the rotor trajectory and the 3D stator deformation map while displaying the acyclic noise strength to provide insights on the reliability of the results.

        Orateur: Faure--Giovagnoli Pierre
        452975.pdf
      • 14:40
        COMO1#3 - Bearing diagnostics and speed estimation via smartphone standalone data 20m

        Condition monitoring of rolling element bearings is a point of interest for early damage detection and prediction to avoid unexpected rotating machinery breakdown. Rolling element bearing related signals are often acquired with accelerometers, as vibration signals usually carry sensitive early information related to the bearing damages. However the accelerometers need to be physically mounted in contact near the bearing to be monitored. Acoustics signals from microphones provide an alternative solution as they acquire signals from multiple sources and bearings, without needing to be in contact with the bearing housing. Accurate speed estimation is also a necessity for detection of the bearing damages, as the bearing related frequencies are dependent of the shaft speed. Encoders are then needed to be added during design phase and mounted to the drivetrain, or a zebra tape and optical tachometer are glued and mounted on one shaft of the drivetrain to properly extract the speed signal. The goal of this paper is to propose using a smartphone instead of expensive measurement equipment and data acquisition systems can result in a fast and cost-effective methodology to monitor the health status of rolling element bearings. The shaft speed extraction is based on video images acquired by the smartphone camera. This methodology exploits the deformation of the video image due to the rolling shutter effect of the smartphone camera. The bearing damage detection is based on the audio of the smartphone video. The audio is captured in stereo by a dual plug-in microphone. The speed estimation via video and signal processing methodologies proposed for bearing diagnostics are applied on real data from an experimental drivetrain different cases of damaged bearings running at shaft speeds in the range from 5 to 40 revolutions per second.

        Orateur: Mauricio Alexandre
        454073.pdf
      • 15:00
        COMO1#4 - Bearing diagnostics based on a Spectral combination of Hjorth's parameters 20m

        In the realm of gear and bearing diagnostics a plethora of condition indicators and signal decompositions has been introduced since 1980s, with the aim of extracting useful information concerning the machine state of health. Nowadays, new emphasis has been given to signal decomposition methods that aims to return components which maximize a particular criterion or signal properties such as kurtosis or negative entropy. Usually, this information is given in a frequency/frequency resolution plane defining the well-known "kurtogram" and "infogram". In this paper, a novel signal decomposition called €œdetectogram€ is presented, based on the maximization of the Detectivity, a proper combination of the well-established Hjorth's parameters (namely activity, mobility and complexity) . The effectiveness and sensitivity of such a signal decomposition to bearing fault diagnostics has been investigated based on both synthetic and real bearing vibration signals. In particular, the bearing data of the Politecnico di Torino test rig are taken into account, which deal with various working conditions and different damage levels. The presented signal decomposition method has been compared to the well-known kurtogram and infogram, showing good performances in extracting the frequency band which carries the most useful diagnosis information, and sensibility to the different fault severities. Moreover, the proposed Spectral Detectivity has been shown to be insensitive to single transient impulse, overcoming therefore possible pitfalls in the diagnosis of bearing and gear faults.

        Orateur: D'elia Gianluca
        454033.pdf
      • 15:20
        COMO1#5 - Data-driven Interpretable Optimized Weights Derived from A Sparsity Measure Framework and Their Applications in Machine Condition Monitoring 20m

        Machine condition monitoring aims to acquire information about machine's health status and prevent pending machine failures, by analyzing collected sensor data. Since most machine fault signatures will introduce repetitive impulsive transients into collected signals (e.g., vibration signals), sparsity measures that can effectively detect such signatures receive much research attention and are eagerly studied. A sum of weighted normalized square envelope spectrum (SWNSES) is a new framework for generalization of the sparsity measures, and optimized weights derived from this framework have been mathematically proved to be fully interpretable and demonstrate eminent properties for machine condition monitoring. What's more, by extending the SWNSES to a sum of weighted normalized Fourier spectrum (SWNFS), the corresponding optimized weights are capable of identifying fault frequency components and healthy frequency components, respectively. This paper intends to present several optimized-weights-based new technologies and their applications in machine condition monitoring. If square envelope spectrum in fast kurtogram is replaced by SWNSES-based optimized weights, a more robust OSESgram is obtained to locate informative bands; based signal processing methods such as variational mode decomposition and wavelet packet transform, a new indicator could be obtained to precisely quantify and extract machine fault components. Moreover, by introducing a new threshold approach, a new signal decomposition method named difference mode decomposition (DMD) could be designed to decompose a mixed signal into reference signals and concerned signals. Both simulation and real-world experimental vibration signals validated the effectiveness and superiority of the proposed optimized-weights-based new technologies, and the optimized weights are useful in boosting the development of machine condition monitoring.

        Orateur: Bingchang Hou
        452685.pdf
      • 15:40
        COMO1#6 - Information Fusion of Infrared Images and Acoustic Signals for Bearing Fault Diagnosis of Rotating Machinery 20m

        To minimize the operation and maintenance costs of rotating machinery, damaging conditions should be detected in the early stages. In recent years, infrared (IR) imaging and acoustic-based condition monitoring methods have gained attention for this purpose. To solve the problems when one single type of data, either acrostic signals or IR images, cannot be individually used to assess the severity of faults in rotating machinery, in this paper we proposed an information fusion method to improve the utilization of multisource sensor systems. Fusing the extracted information from both sources is performed with the aid of a fuzzy inference system. The effectiveness of the proposed techniques is demonstrated through the data that has been captured by an inspection mobile robot for monitoring conveyor belt idlers at a mining site.

        Orateur: Siami Mohammad
        454023.pdf
    • 16:00 → 16:20
      Coffee break 20m
    • 16:20 → 17:00
      Exhibitor session
      Président de session: Simon CHESNE (LaMCoS INSA-Lyon)
    • 17:00 → 18:00
      Round Table - IROQUIA

      De 2005 à 2010, la démarche IROQUA avait pour objectif de réduire de 50% le bruit perçu des aéronefs à l’horizon 2020. Pour ce faire, l’ensemble des acteurs du domaine aéroacoustique (avionneurs, motoristes, laboratoires universitaires, CNRS, ONERA…) s’était réunis dans le cadre de Journées thématiques annuelles et au travers de projets de recherche ambitieux financés par la FRAE.
      Ces projets portaient aussi bien sur le bruit des trains d’atterrissage et dispositifs hypersustentateurs, que sur les bruits de combustion et jet, ainsi que sur les stratégies de contrôle actif et le comportement des liners acoustiques des entrées d’air des réacteurs.
      De 2010 à 2016, d’autres acteurs tels que Aéroports De Paris, Air France, la DSNA et la FNAM ont intégré le Comité Directeur IROQUA aux côtés des partenaires initiaux. En outre des équipes de recherche universitaires s’intéressant aux questions d’impact et de perception se sont joints au réseau. Cette 2ème convention a permis de monter plusieurs projets financés par la FRAE et l’ANR notamment dans les domaines du bruit de couche limite, de la combustion et de la psychoacoustique.
      En 2019, avait été prise la décision de ressusciter les journées IROQUA, sans pour autant aboutir à une renaissance du réseau.
      Dès son origine, IROQUA se voulait être « un espace de dialogue destiné à améliorer la connaissance mutuelle et la fluidité des échanges entre les différentes équipes impliquées dans les recherches sur la réduction du bruit aéronautique » et « un atelier où ces équipes pourraient concourir à monter des projets relatifs à ce thème ».
      L’objectif de cette table ronde est de rappeler le contexte des précédentes initiatives IROQUA et de proposer un nouveau cadre IROQUA en relation avec les groupes thématiques GAHA (Groupe Aéro et Hydro-Acoustique) et GVB (Groupe Vibro acoustique et Contrôle du Bruit) de la SFA.
      Participants de la table ronde: Présentateurs JISFA Industriels et SFA

      Président de session: Frank Simon (ONERA)
    • 18:00 → 19:30
      Visite ISAE 1h 30m
  • mercredi, 12 juillet
    • 09:20 → 10:00
      Keynote / Dr. Wenyi Wang - Aerospace Division, Defense Science and Technology Group, Australia 40m

      Title:
      Tackling a challenging problem: diagnosis of helicopter planetary gear rim crack – with a historical reflection of DSTG’s gearbox diagnostics work.

      Abstract:
      This keynote will address a worldwide challenging problem - finding solutions to detecting the planetary gear rim cracking in helicopter main rotor gearboxes. This failure mode caused two fatal helicopter crashes in recent years, and these is no reliable means of detecting the gear fault. Defences Science and Technology Group (DSTG) generated a dataset in an accelerated rim crack propagation test with the Bell-206B helicopter main gearbox as the test article, and shared the dataset with the international HUMS/PHM community. With this dataset, DSTG conducted a benchmark analysis using the conventional signal processing and machine learning techniques. This keynote will present the results of this analysis with a historical reflection of DSTG's work of gearbox diagnostics in the last four decades.

      Orateur: Jérôme Antoni (Laboratoire Vibration Acoustique, INSA Lyon)
    • 10:00 → 10:20
      Start-up Session
      Président de session: Quentin Leclère (Laboratoire Vibration Acoustique, INSA Lyon)
    • 10:20 → 10:40
      Coffee break 20m
    • 10:40 → 12:40
      JISFA 3 / Aircraft Interior Noise
      Président de session: Frank Simon (ONERA)
      • 10:40
        JISFA3#1 - Main Gearbox noise in helicopter cabins and dedicated noise control treatments 20m

        The presentation will introduce Main Gear box noise generation mechanism and transfer path to helicopter cabins.
        From this introduction, various noise control solutions dedicated to this typical issue in helicopter cabins will be presented, providing an overview of typical treatments used to improve comfort of passengers.

        Orateur: Julien Caillet (ETGV, Airbus Helicopters)
      • 11:00
        JISFA3#2 - Modeling the wall pressure of a turbulent flow to predict the noise radiated by a structure 20m

        The internal noise inside an aircraft induced by a turbulent flow results from a coupling between aerodynamics and vibroacoustics through the turbulent wall pressure (TWP). By applying instationary forces on a structure, this pressure initiates a vibratory motion which will generate noise by acoustic radiation. Modeling these phenomena is equivalent to solving a filtering problem. The input is the TWP spectrum in the frequency/wave-number domain while the output is the acoustic pressure radiated by the structure.
        The link between these quantities is realized through a modal decomposition. A mode will contribute significantly to the radiation at a given frequency only if its resonance or radiation frequency and the wave-numbers associated are aligned with the characteristics of the TWP spectrum. Therefore, modelling this spectrum is a key step in the determination of the internal noise. The most common model was established by Corcos and is based on empirical observations. The only mechanism involved is the convection and the boundary layer profile is averaged. These hypothesis limit the frequency range scope of this model and those developed latter on the same basis.
        In order to improve the estimation of the internal noise, a more realistic model of the TWP spectrum has been developed. It is based on the resolution of Poisson’s equation. In this model, the full profile of the boundary layer is used and four mechanisms are involved, each one acting as a filter in the final formulation through four characteristic lengths based on convection, turbulence, altitude and thickness of the boundary layer.
        With this model, new modes can contribute to the radiation and sometimes much more significantly than those selected by Corcos’s model. In particular, the importance of the low part of the boundary layer has been highlighted.

        Orateurs: Sylvain Morilhat (DMPE - ONERA - Université de Toulouse), Frank Simon (DMPE - ONERA - Université de Toulouse), François Chedevergne (DMPE - ONERA - Université de Toulouse), Bertrand Aupoix (DMPE - ONERA - Université de Toulouse)
      • 11:20
        JISFA3#3 - Engine-to-cabin airborne transmission 20m

        Engine noise plays a variable role in cabin comfort, depending on the engine type. This presentation will focus on engine-to-cabin airborne transmission, prediction process and validation tests.

        Orateur: Stefan Moal (AIRBUS, Toulouse)
      • 11:40
        JISFA3#4 - A multi-port scattering matrix formalism for the acoustic prediction in duct networks 20m

        Duct acoustic network modelling is commonly carried out using the transfer matrix formalism which is limited to the low frequency range. The aim of this work is to extend it to higher frequencies by taking into account the multi-mode acoustic propagation. The first step is to compute, via finite element discretisation (FEM), the multi-port multi-modal scattering matrix of each element. The second step is to transform it into a scattering matrix for the acoustic power, relying on assumptions which are often used for the study of medium-to-high frequency broadband noise. The method is applied to typical elements such as expansion chamber mufflers and air conditioning veins. In all cases, the power-flow model is compared to the FEM solution in terms of Transmission Losses. It is concluded that this simplified model is a reliable tool for the analysis of complex networks encountered in HVAC duct networks.

        Orateurs: Cyril Calmettes (Airbus Helicopters), Emmanuel Perrey-Debain (Université Technologique de Compiègne, Laboratoire Roberval (FRE UTC - CNRS 2012)), Emmanuel Lefrançois (Université Technologique de Compiègne, Laboratoire Roberval (FRE UTC - CNRS 2012)), Julien Caillet (Airbus Helicopters)
      • 12:00
        JISFA3#5 - Processing and analysis of flight test internal / external microphone measurements 20m

        The identification and ranking of sound sources contributing to aircraft interior noise is not an easy task, due to the complexity of the involved physical phenomena. The turbulent boundary layer developed on the fuselage is recognized as one of the dominant sources, but there is also engine noise, airframe noise and noise from the air-conditioning system. When implementing flight tests, microphones can be installed on the fuselage outside the cabin to characterize external sources (jet and fan noise), but these suffer from heavy contamination by pressure fluctuations induced by the boundary layer. In this presentation, we'll look at how to take advantage of synchronous internal-external microphone measurements, whether for denoising external microphones, or assessing the contributions of different sources of internal noise.

        Orateurs: Quentin Leclère (Laboratoire Vibration Acoustique, INSA Lyon), Jérôme Antoni (Laboratoire Vibration Acoustique, INSA Lyon), E. Julliard (Acoustics Department, Airbus Operations S.A.S)
      • 12:20
        JISFA3#6 - Round Table JISFA3 20m

        Additional Q&A and open discussion about the session topic.

    • 10:40 → 12:40
      Survishino 7 / Dynamic Modelling 2
      Président de session: Sadoulet-Reboul Emeline
      • 10:40
        "DYN2#1 - Non-Linear data-driven model for a solar tracker aeroelastic simulation" 20m

        The solar tracker industry has evolved dramatically in recent years with huge PV-farm demanding for challenging structural design. In such framework it is essential to consider aeroelastic wind loads due to flow unsteadiness and vortex shedding. Dynamic loads estimation is therefore not easy: this is due to the fact that, being the tracking systems equipped with a torsional flexible constraints, complex fluid-structure interaction phenomena, briefly called flutter, can be triggered. The latter modify the load scenario and the structural damping in an apparently unpredictable way, generating a dangerous self-excited load on the system, that can be unhealthy for the structure integrity. The aeroelastic behaviour of the panel could be described through dimensionless parameters (the so-called "Scanlan Derivatives"), that usually are function of the reduced wind speed and that can be estimated starting from damping and frequency under freely excited oscillation, basing on the classic vibration's theory. Several experiments have been performed in the €œR. Balli€ Wind Tunnel at the University of Perugia to estimate such parameters in different wind condition using a scaled model of the solar tracker. Results demonstrate how these aeroelastic parameters are function of wind speed and of the angle of attack; this behavior affect dramatically the final critical wind speed (the speed at which flutter occurs). Basing on the experimental campaign a numerical model has been developed in order to reproduce the non-linear aeroelastic behavior of the structure. A comparison of the results of the linear and the data-driven non-linear model were discovered to be a very useful tool in order to understand the complex dynamic behavior of the structure: it has been found that under certain operating conditions the non-linear model provides an average amplitude error five times lower than the value obtained from the linear model.Only this type of analysis can give useful insight to improve the challenging design of huge solar tracker structures.

        Orateur: Castellani Francesco
        Non-Linear data-driven model for a solar tracker aeroelastic simulation
      • 11:00
        "DYN2#2 - Phase transitions in a resonating free-piston engine generator" 20m

        Free-piston engine generators (FPEGs) offer significant advantages over conventional combustion generators, in particular, higher efficiencies, compactness, and lighter weight. These attributes give higher gravimetric and volumetric energy densities, compared to alternatives. Moreover, an FPEG allows unlimited scalability to high power, and also, electrically controlled variable compression ratio, which offers fuel flexibility, including potential to run on zero carbon fuels. By fitting a very stiff resilient member to an FPEG, a nearly-linear mass-elastic system can be created capable of resonance. Resonating FPEGs offer significant advantages over FPEGs that use a bounce chamber. These advantages include more precise control, a significant reduction in electrical machine currents, and a corresponding reduction in electrical power losses. Resonance using a square wave external excitation source, has also been used to start an FPEG fitted with a bounce chamber.  In operation, a resonating FPEG is actually a nonlinear, self-excited system, where the nonlinearity stems from the adiabatic gas compression and expansion processes. Self-excitation stems from the occurrence of combustion which largely depends on the piston displacement. A significant difficulty starting any type of FPEG by exploiting resonance, is first to recognise that the phase of the fundamental harmonic component of any linear system response, is Ï€/2 radians out-of-phase from the external excitation. In operation however, under self-excited resonant conditions, the phase difference between the dynamic response and the excitation, is significantly different from Ï€/2. The first question therefore is: were the system to be entirely linear, in operation, what would the resonant phase difference be between the self-excitation and the response? The second question is: if the phase difference is significantly different from Ï€/2, how is it possible to make the transition from externally excited resonance (used for starting) to self-exited steady-state resonant operation. The third question is how does nonlinearity affect the phase differences in resonant starting and resonant steady-state operation? This paper addresses these three questions by modelling and simulation. The overall objective is to provide clear guidance on how to transition between resonant starting and subsequent self-excited operation of a resonating free-piston engine generator.

        Orateur: Dunne Julian
        Phase transitions in a resonating free-piston engine generator
      • 11:20
        "DYN2#3 - Rubber part characterisation for rotordynamics analysis." 20m

        Abstract. This paper presents a method to determine stiffness dynamic characteristics of rubber parts which are used for vibration isolation. The study was motivated by the need of such characteristics to model vibroflot's dynamic behaviour. Vibroflots are used by ground improvement companies to deeply densify sandy soils by vibrations in order to make stable future infrastructures. A classical vibroflot is a slender structure hung from a crane with several extension tubes and cables. The vibroflot is composed of a non-rotating tube casing in contact with the soil, containing a mass unbalanced rotor mounted on rolling bearings driven by an asynchronous electrical motor that produces orbital vibration and therefore the soil compaction. In order to avoid the vibrations to propagate to the extension tubes two big rubber coupling parts are used. The method presented here is based on five steps. The first step aims at determining material properties and at presenting the hyper-elastic and viscoelastic models used [1, 2]. Then real dynamic tests are carried out on one type of rubber coupling, at different frequencies and dynamic amplitudes, using a hydraulic shaker capable to develop a force up to 62 kN. Then Finite Elements simulations are conducted to validate material models and parameters on the first geometry. Once the results are satisfactory, simulations are made on the second rubber coupling made of the same material. Finally, both rubber parts stiffness properties are known for different sets of parameters and can be introduced into the global Finite Elements multi-rotors model for investigating operational runs.   Keywords: Rotordynamics, Rubber characterization, Vibration isolation, Vibroflot.   References 1. G. A. Holzapfel, Nonlinear Solid Mechanics: A Continuum Approach for Engineering, Chichester, England: Wiley, 2000. 2. G. Petitet and M. Barquins, Matériaux caoutchouteux, Lausanne, Switzerland: PPUR, 2008.

        Orateur: Tezenas Du Montcel Florian
        Rubber part characterisation for rotordynamics analysis.
      • 11:40
        "DYN2#4 - Determining Loads for Down-scaled Testing of Wind Turbine Pitch Bearings using an Augmented Kalman Filter and Reference Measurements" 20m

        This article proposes an approach to determine the loads to apply during down-scaled testing of wind turbine pitch bearings. These bearings present unique challenges and failure modes due to their large size, their limited range of motion, and their loads. Because of these particularities, learnings from more common bearing applications cannot be relied upon. On the other hand, accelerated full-size testing of this type of bearings is lengthy, up to 5 months for one bearing, and expensive, due to the specialized infrastructure required, the assembly costs of the test-bench and the manufacturing cost of the prototype bearing itself. For these reasons, a trend is emerging in the wind industry to down-scale tests with the goal of using already existing, more affordable test-benches to design and test new larger bearings. This raises the question of what loads to apply on the down-scaled test-bench bearing in order to reproduce the fatigue life and defect formation of the larger reference bearing. Manufacturers currently use proprietary methods to determine the loads for down-scaled testing as there is no established standard. Given a specific failure mode, the goal of this work is that the down-scaled testing reproduces stresses at key locations around that failure mode. This ensures that the down-scaled testing is representative of the operation of the reference bearing. First, these key stresses are computed using design loads and a model of the reference bearing. Then, a load-estimation problem is solved around the model of the scaled bearing to determine the loads to apply on this scaled bearing to replicate those key stresses. The resulting inverse problem is solved using an Augmented Kalman Filter, which has proven successful for this type of load-estimation problem. The novelty is that this estimator uses as measurements those key stresses, which, while relevant, cannot be physically measured. Finally, this is validated in simulation by showing that the down-scaled testing replicates the stresses at the chosen key locations. This article demonstrates the proposed approach on the industrial case of testing a 3.4MW wind turbine pitch bearing for ring structural failure using a 1.5MW bearing test-bench.

        Orateur: Dwek Nathan
        Determining Loads for Down-scaled Testing of Wind Turbine Pitch Bearings using an Augmented Kalman Filter and Reference Measurements
      • 12:00
        "DYN2#5 - Identification of vibration damping in 3D-printed lattice structures" 20m

        As metallic 3D-printing opens the design space for lightweight structural applications, it also raises challenges in terms of low level vibration damping. Reducing the number of parts is avoiding some classical damping phenomena due to friction mechanisms in mechanical assemblies. In addition, metallic materials do not exhibit intrinsic viscous properties that could significantly contribute to structural damping. In this work, the damping properties of lattice-based structures and possible design solutions offered by Powder Bed Fusion are investigated. In particular, auxetic behaviour, passive energy transfer, and viscoelastic shear layer could be considered. The approach consists in identifying promising concepts from numerical simulations, and validate them experimentally. Underlying questions of predictive capability of damping simulations and identification of damping properties from experiments are addressed.

        Orateur: Tanays Rémy
        Identification of vibration damping in 3D-printed lattice structures
      • 12:20
        "DYN2#6 - Stochastic Digital Twin of a Composite Plate for Predicting Lamb Wave Propagation" 20m

        This work presents a framework for stochastic updating for verifying and validating a finite element (FE)-based model - the Digital-Twin - of a composite plate, considering temperature influence on Lamb wave propagation. It begins with a deterministic updating procedure to find optimal mechanical properties, followed by a stochastic updating procedure to obtain probability density functions for meaningful parameters. The stochastic updating procedure is divided into two steps: a sensitivity analysis using Sobol Indices and a Bayesian inference process using Markov-chain Monte Carlo (MCMC) simulations and the Metropolis-Hastings sampling algorithm. To reduce the computational time required for the MCMC process, the work proposes using a surrogate model based on artificial neural networks (ANNs). The ANN can be trained using parallelized Monte Carlo simulations, in contrast to the sequential nature of the MCMC process. This approach reduced the time required for updating rounds by 450 times in the studied case without compromising the accuracy of the resulting probability density functions for model parameters.

        Orateur: Teloli Rafael
        Stochastic Digital Twin of a Composite Plate for Predicting Lamb Wave Propagation
    • 10:40 → 12:40
      Survishno 8 / Fault Diagnostics and prognostics 2
      Président de session: Renata Klein
      • 10:40
        "FDP2#1 - Impact of low ring waviness orders on Hybrid ball bearing under high speeds" 20m

        Deep groove ball bearings (DGBBs) have a crucial role in various rotating equipment applications, particularly in high-performance systems such as electric vehicle Powerpacks, where noise and vibration levels are of utmost importance. The roundness and surface shapes of the ring races significantly influence the performance and service life of DGBBs. But the fact remains that in the bearing industry, it is well-known and widely accepted that contact surfaces between bearing components are never perfectly smooth, even with the most advanced grinding and honing processes. Despite manufacturers' efforts to achieve high levels of precision, there are always inherent surface irregularities and deviations on the ring races and ball surfaces that generate vibrations due to uneven ball load distribution. To address this issue, DGBB manufacturers have developed vibration classification systems that help to classify the vibration levels of DGBBs. These classification systems are based on international standards, such as ISO and ANSI, and they provide a standardized way of measuring and classifying the vibration levels of DGBBs. The vibration classification system for DGBBs is based on the vibration velocity and is measured in the micron scale (μm/s), that means that those defects at manufacturing stage still of very small amplitudes and are in general not harmful under applicative conditions. However, excessive surface irregularities, such as deep scratches, indentations, or waviness orders with high amplitudes, lead to increased friction and localized stresses that affect bearing performance and may result in premature failure. In this article, inner and outer ring low waviness orders are studied on Hybrid Deep Groove Ball Bearings (DGBB). Both numerical model and, experimental tests were carried out to study the effect of rings waviness on hybrid bearing under high rotation speed.

        Orateur: Berrada Salma
        Impact of low ring waviness orders on Hybrid ball bearing under high speeds
      • 11:00
        "FDP2#2 - Orthogonal nonnegative matrix factorization as informative frequency band selector" 20m

        One of the most common representations of acquired vibration signals from a faulty machine is the time-frequency representation in the form of a spectrogram matrix. Because the magnitude part of the spectrogram matrix consists only of non-negative elements, it can be decomposed using non-negative matrix factorization (NMF) into a base matrix and weight matrix, which represent the frequency and time content of the signal, respectively. The frequency features of the base matrix can be used as filters to detect local damage in bearings by filtering the original signal with these filters. However, classical NMF provides filters that cover all frequency bands with different amplitudes. Unfortunately, such filters cover both informative and non-informative frequency bands, second ones corresponds to the noise. To solve this problem, the NMF can be enhanced by using orthogonal non-negative matrix factorization (ONMF), which imposes orthogonality constraints onto the NMF model. The orthogonality constrained applied to NMF improves the quality of clustering properties of NMF, which corresponds to better detecting of informative frequency bands. Additionally, the orthogonality constraints make the decomposition more sparse, which translates into zero amplitude at the non-informative frequency band related to the noise. Hence, using ONMF we can obtain a more selective filter which filters out only the most relevant information from the signal. The ONMF works for both signals with Gaussian and non-Gaussian noises. The analyzed signals come from a test rig with faulty bearings (Gaussian noise) and belt conveyor (non-Gaussian noise).

        Orateur: Gabor Mateusz
        Orthogonal nonnegative matrix factorization as informative frequency band selector
      • 11:20
        "FDP2#3 - Noise robust gearbox defect diagnosis" 20m

        "Vibratory analysis focused on the detection of periodically impulsive signatures caused by tooth crack of a gearbox in early stage is a great interest in industrial applications. This problem has received great attention in the last decades. A wide variety of model-based approaches requiring advanced methods for processing and analyzing the vibration signal for early fault detection have been proposed. For the vibration signal processing about gearbox fault diagnosis, the noise caused by data acquisition systems and other machine elements, must be properly filtered from the recorded vibration signal. A novel gearbox fault diagnosis method taking into account modeling errors of vibration signal, additive noises and the knowledge of the default frequency is proposed.
        A new model of faulty vibration signal from those usually used for gearbox fault diagnosis is introduced. Then an optimal estimator, combining simultaneously the noise filtering of the recorded signal and the estimation of fault signal by a demodulation approach based on the least squares optimization, is determined. The new approach has been applied to faulty vibration signal and shown its robustness to measurement noise and its reliability for fault diagnosis.
        The new approach has been applied to faulty vibration signal and shown its robustness to measurement noise and its reliability for fault diagnosis.
        The problem of finding the optimal estimation of fault signal from a new model vibration signal has been addressed under constant speed conditions. It integrates an optimal strategy for noise filtering. The experimental results showed the robustness of the proposed gear fault diagnosis technique."

        Orateur: Sekko Edgard
        FDP23.pdf
      • 11:40
        "FDP2#4 - Modeling and identifying non-stationary long-term historical condition monitoring data in the presence of noise with non-Gaussian characteristics" 20m

        "The advancement of condition monitoring systems has led to an increase in the utilization of long-term monitoring data for diagnostics and prognostics. The effective utilization of such data, collected over extended periods ranging from months to years, is a critical factor in enabling accurate diagnosis and prognosis.
        However, most of these industrial applications operate under time-varying conditions, making the diagnosis and prognosis approach to condition-based maintenance (CBM) complicated. Furthermore, many machines work in harsh environments, such as mining machines, wind turbines, helicopters, etc. The data acquired from these machines is often affected by noise with non-Gaussian characteristics. Therefore, it is a necessary task to analyse real data and introduce a proper model that could consider time-varying parameters and the effects of non-Gaussian noise.
        In this research, we first conducted a short literature review on the state of the art in long-term data modelling, focussing on statistical-based models. Then we analysed a well-known benchmark data set collected from the high-speed shaft of a wind turbine. Finally, we present the results for this data set and draw conclusions based on our findings."

        Orateur: Shiri Hamid
        Modeling and identifying non-stationary long-term historical condition monitoring data in the presence of noise with non-Gaussian characteristics
      • 12:00
        "FDP2#5 - Robust and automatic diagnosis of rotating machine faults by long-term spectral analysis" 20m

        "Although it is now recognised that the most effective methods for diagnosing rotating machines are based on the analysis of the cyclo-stationary content of the vibration signal, they can nevertheless be classified into two main families. The first group includes temporal indicators combining statistical moments of the signals, possibly filtered around the orders of interest. These methods, both simple and global, are effective when the fault is prevalent in the signal. The second method consists of those which analyse the spectra (1st or 2nd CS orders) of the vibration signal more directly. These methods often allow earlier detection, but require precise knowledge of the frequency channels likely to carry the defects.
        In contrast to these approaches, this paper propose to study the distribution of the frequency channels whose amplitude presents the strongest macroscopic evolution. As could be done via the analysis of a spectrum cascade, we determine the channels showing a suspicious trend in a robust way, by classifying them according to a rank statistic. For each frequency channel thus identified, we evaluate the distance that separates it from the harmonic families corresponding to the characteristic frequencies of the fault. This distance provides a variable whose distribution we study to diagnose the machine.
        This new strategy is innovative from several points of view. Firstly, it combines the qualities of the two families of pre-existing methods, benefiting simultaneously from the precision of a spectral observation and the simplicity of a global method, without the need to know a priori the frequencies carrying the fault. Furthermore, this new strategy differs from others in that it is intrinsically necessary to study a measurement campaign rather than a single signal. This apparent limitation is alliviated by the increase in our storage and processing capacities, and favors the hope of combining the diagnostic (signal-oriented) and prognostic (data-oriented) approaches in the same analysis step.
        The approach will be presented through an ageing campaign of aeronautical epicyclic gear trains. Although such gear faults are linked to well-known characteristic frequencies, their signature happens to be carried by a specific combination of these frequencies, whose orders are unpredictable apriori."

        Orateur: Andre Hugo
        Robust and automatic diagnosis of rotating machine faults by long-term spectral analysis
      • 12:20
        FDP2#6 - 20m
    • 10:40 → 12:40
      Survishno 9 / Condition monitoring 2
      Président de session: Marco Cocconcelli
      • 10:40
        "COMO2#1 - Non-Gaussian noise in rotating machines: sources, impact to local damage detection procedures and possible solutions" 20m

        Local damage detection in rotating machines is one of the most important research problem in condition monitoring. There are many available solutions that are based on properties of informative signal (Signal of Interest, SOI) such as impulsiveness and/or periodicity (cyclo-stationarity), amplitude modulation in time domain, etc. As signal is non-stationary (due to impulsive nature it is wideband in frequency domain), the time frequency domain is often selected as a basis for further analysis. All these techniques are very effective even for really poor signal-to-noise ratios (or in other words in case of a weak SOI signal that could be considered as early stage of fault development). Unfortunately, the situation is changing dramatically if noise associated to almost all real signals is not Gaussian. Such a case is not so obvious for many researchers, especially if they work only with test rig data (a non-Gaussian noise is not present there). In the paper, we will present several examples of various machines that may produce strongly non-Gaussian signals. Non-Gaussian background noise is related to technological processes performed by these machines. It could be also related to measurement and data transmission systems as reported in work of [1]. Another possible situation is just random, single excitation (shock) that may appear during measurement [2]. A Machine located nearby or a specific design of mechanical systems could be also a source of non-Gaussian noise in case of acoustic measurement [3] How does non-Gaussian noise may affect diagnostic procedures? Obviously it depends on the case, but in general, most of procedures are based on mean removal, normalized variance, etc. Non-Gaussianity of background noise may imply lack of variance (not defined for heavy tailed distribution). It may affect synchronous averaging [4]. Detection of impulsive SOI cannot be based on impulsiveness criteria anymore (kurtosis will react to high amplitude spikes, not to cyclic impulses). It could be surprising, but for heavy tailed noise the autocovariance function might not be allowed to use (variance may not exist). There are some solutions related to outliers removal, robust estimators, alternative measures of dependence. They will be illustrated in the presentation.

        Orateur: Radoslaw Zimroz
        450760.pdf
      • 11:00
        "COMO2#2 - Bearing degradation indicator using characteristic frequencies applied on non-stationary vibration signals" 20m

        Rolling-element bearings are wear components in rotating machinery. Usually, their condition is monitored with signal-processing tools tailored for vibration-based surveillance. From cyclostationary theory, angle-time analysis, or sparsity-based approaches, numerous methods have been developed relying on the properties of incipient fault signals. The scalar indicators coming from these techniques are well-suited for detection and localisation, but face difficulties in prognosis. The main issue is that the released vibration energy may not correlate with the severity of the fault. There is a need to develop scalar indicators dedicated to gravity estimation. Our belief is that the geometry modification of the fault will induce a change in the kinematics of the rolling-element bearing. Far from the ideal perfect-rolling assumption, the relationship between the races' rotation and that of the fundamental train is the fruit of complex interactions. What is usually summarized in the contact angle in the traditional fault frequency equations conceals various effects of axial-radial load ratio, skidding between the elements, and imperfect transmission. As the fault extends, these interactions may change with a direct impact on the fault characteristic frequencies. Bertoni and Andre [1] proposed to monitor these characteristic frequencies with the BeaFEM method to estimate the most probable contact angle. They showed on an academic test-bench run-to-failure experiment that monitoring the fault frequencies could be a valuable degradation indicator. The validation was done on stationary conditions with a main focus on instantaneous angular speed measurements. However, under non-stationary conditions, the equivalent contact angle is sensitive to various factors such as loading, angular acceleration, or lubrication. An extension of the method is presented to prove the relevance of the concept on an industrial case with non-stationary conditions. The role of apparent contact angle as a scalar accounting for the kinematics is clarified. The effect of load and rotation speed is assessed and the method adjusted to emphasize only the contribution of the fault. The method is applied to industrial vibration signals from a damaged wind turbine operating over a wide spectrum of speed and load conditions. This original approach provides a tailored tool for optimal maintenance decisions. [1]https://doi.org/10.1016/j.ymssp.2022.109891

        Orateur: Marsick Adrien
        452328.pdf
      • 11:20
        "COMO2#3 - Optimal filter design for rotating machinery fault detection under time-varying speed conditions" 20m

        Critical assets found in the mining and power generation industries need to operate reliably. The harsh operating environments can lead to accelerated degradation of the components, which can lead to unexpected failures and long downtimes that impede production. Vibration-based condition monitoring methods are used to detect changes in the condition of critical components, and to identify and potentially trend the damaged components and potential damage modes. This information can support maintenance decisions. Damage often manifests as weak components in the vibration signals and is masked by dominant signal components attributed to gear mesh excitations and the operating environment. Furthermore, rotating machinery found in the power generation industry (e.g., wind turbines) and mining industry (e.g., bucket wheel excavators) operate under time-varying operating conditions. The speed and load variations add additional modulation to the signals that impede fault detection. Hence, methods are required to enhance damage components for early fault detection and improved fault characterisation under time-varying operating conditions. Blind deconvolution methods are used to recover the sources that are convolved with the impulse response function of the system and are corrupted by other extraneous sources from the raw vibration signal. Optimal filtering methods are implemented by identifying a measure that is maximised by the source of interest, e.g. if the source is non-Gaussian, then a feature that measures the deviation from Gaussianity can be maximised. Features such as kurtosis, correlated kurtosis, ICS2, Hoyer index, L2/L1, and negentropy have been used with great success under constant-speed operating conditions, but their application under varying operating conditions is still rather limited. In general, the optimal filtering problem is solved by formulating an optimisation problem by defining relevant loss and constraint functions. Various frameworks are available to solve the optimization problem, including classical gradient-based methods when gradients are readily available. The optimisation problem provides flexibility in the form of the loss and constraint functions to identify novel features with which to perform optimal filtering. In this study, optimal filtering using these features is performed on gearbox data that were generated under time-varying speed conditions.

        Orateur: Schmidt Stephan
        450408.pdf
      • 11:40
        "COMO2#4 - The design of optimal indicators for early fault detection using a generalized likelihood ratio test" 20m

        This study elaborates on a methodology to design health indicators for vibration-based condition monitoring of rotating machines for early fault detection. These indicators are optimal in maximizing the probability of detection given a constant rate of false alarms. Two probability density functions (PDF) modeling the vibration signals' healthy and faulty states are exploited to generate health indicators using a generalized likelihood ratio test. The key point is formulating a framework to express the health indicators as the function of the partial derivatives of the PDF of the healthy state and a modulation function. The modulation function allows the detection of slight deviations from the healthy state of the vibration signals. Furthermore, it is shown that specific choices of the modulation functions recover conventional health indicators such as kurtosis, skewness, Lp/Lq-norms, the negentropy of the squared envelope, etc. Since the indicators are asymptotically distributed as a chi-squared distribution, a statistical threshold can be estimated to assess the machine's state with respect to its healthy state. The proposed approach is formulated in this study in detail. The performance of the thresholds is demonstrated on simulated and experimental vibration signals. It is shown that the transition from the healthy to the faulty state of the machine can be detected via the threshold. An important conclusion of this study is that many conventional health indicators are optimal if and only if the healthy state of the machines is Gaussian. The proposed methodology shows how to design indicators for non-Gaussian descriptions of the healthy state and can pave the way for the development of many other health indicators by carefully selecting the PDF of the healthy state and the modulation function.

        Orateur: Kestel Kayacan
        452087.pdf
      • 12:00
        "COMO2#5 - Combined bearing faults detection using the Multiple Improved Envelope Spectra via Feature Optimization gram (MIESFO-gram) in complex systems" 20m

        Bearing diagnostics is a growing field of study, with a focus on complex machinery that includes a variety of bearing and gear components operating under varying conditions (e.g. speed and load). Meanwhile, some components with weak signatures may remain hidden while others with intensive defects are detected. Therefore, the ability to detect combined faults in the machinery, having different cyclic frequencies is critical. Envelope Analysis is a popular method for bearing diagnostics, in which a filter is typically applied around an excited frequency band and the signal is enveloped, yielding the Squared Envelope Spectrum. However, as several damaged bearings may excite not only different but also several frequency bands simultaneously, band-pass filtering around only one frequency band may not be sufficient to detect all bearing faults in the machine. This will be more challenging when an operational condition like speed is changing over time which causes the fault-related cyclic frequencies to be smeared and/or hidden. Furthermore, different carrier frequencies may be excited under different speed conditions. Recently, IESFOgram has been proposed utilizing Targeted and Blind features. The method has been essentially developed based on either the Cyclic Spectral Correlation or the Cyclic Spectral Coherence, in order to select the optimal frequency band and extract the corresponding Improved Envelope Spectrum. In the IESFOgram with the Targeted features, the possible existing fault frequencies are fed into the algorithm as inputs. However, the IESFOgram with Blind features aims to find the most-occupied frequency band with the cyclic frequencies. On the other hand, when there are more than one bearing faults exciting different natural frequencies, selecting only the single most dominant carrier may prove insufficient to detect other damages present in the signals. In this paper, an extension of the Blind IESFOgram is introduced with the aim of finding all possible unique frequency bands occupied by cyclic frequencies. The method is applied and evaluated on simulated and experimental data with different types of faults under steady and varying speed conditions in a complex system. Finally, the results are compared with the conventional Targeted and Blind IESFOgram.

        Orateur: Yazdanianasr Mahsa
        449060.pdf
      • 12:20
        "COMO2#6 - Temporal companding for the evaluation of a rotor rotation speed during strong transients" 20m

        This paper presents a new autonomous vibration monitoring system for monitoring the mechanical health of aircraft engines. The first section of the paper describes the new global monitoring system. The constraints of the embedded part of the system are detailed. One of the most important constraints is the need to estimate the rotation speed of a rotor, from a vibratory signal, even when the rotation speed varies strongly. The second section of the paper presents a solution to facilitate the estimation of the rotational speed when the speed varies strongly. The solution is based on a temporal companding (compressing and expanding) algorithm, which is described in detail. In the last section, the performance and limitations of the algorithm are evaluated on some data acquired on an aircraft engine in flight.

        Orateur: Griffaton Julien
        447222.pdf
    • 12:40 → 14:00
      Lunch Break 1h 20m
    • 14:00 → 16:00
      JISFA 4 / Passive and Active Control of Aeronautical Noise
      Président de session: Adrien Pelat (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613)
      • 14:00
        JISFA4#1 - Absorption and transmission of boundary layer noise through micro-perforated panel structures: measurements and modellings 20m

        Reducing the airframe noise under a low-speed flow is a challenging problem in the acoustic design of automotive and aircraft cabins as it should comply with low drag, space- and mass-saving constraints. Moreover, the use of lighter and stiffer structures with low critical frequencies transmit more efficiently mid-frequency noise components towards the cabin, such as those induced by a turbulent boundary layer (TBL) of air. Backed or unbacked micro-perforated panels (MPP) are potential lightweight and low-drag solutions that may enhance the absorption and decrease the transmission of flow-induced noise. This work describes experimental and modelling studies that show the effect of MPP partitions, either flush-mounted or over a cavity floor, on the wall-pressure fluctuations induced by a low-speed TBL.
        Measurements have been performed in a low-speed wind-tunnel to determine the frequency-averaged power flow injected by TBL wall-pressures into flush-mounted MPP partitions. They were assessed against modal simulations of the power absorbed and transmitted by flexible MPP partitions forced by a TBL. At low frequencies, a large part of the power injected by the aerodynamic pressures is transmitted through the apertures with inefficient back-scattering properties. As frequency increases, the absorption steeply decays due to the MPP apertures that efficiently back-scatter small-scale turbulence into sound. Absorption peaks are observed at the MPP aerodynamic coincidence frequency and at the Helmholtz resonance frequency.
        Numerical and experimental studies have also been carried out to assess the effect of microperforating the base wall of shallow cavities in transitional and closed flow regimes to reduce their tonal and broadband noise components under a low-speed TBL. Measurements showed that the first peak levels observed towards the leading edge of the cavity were reduced by up to 8 dB whereas flow-induced noise was generated towards the trailing edge. This trends were confirmed by Lattice-Boltzmann simulations that predict an attenuation of the dominant peaks at the floor and at the mouth of the cavity. It was found that the dissipation of energy occurs at the regions of maximum velocity fluctuations induced by well established outflow conditions within and at the inlet-outlet of the base wall apertures.

        Orateurs: Cédric Maury (Aix Marseille Univ. CNRS Centrale Marseille, Laboratory of Mechanics and Acoustics (LMA UMR 7031) ), Teresa Bravo (Instituto de Tecnologías Físicas y de la Información (ITEFI) , Consejo Superior de Investigaciones Científicas (CSIC) ), Daniel Mazzoni (Aix Marseille Univ. CNRS Centrale Marseille, Institut de Recherche sur les Phénomènes Hors Equilibres (IRPHE UMR 7342) )
      • 14:20
        JISFA4#2 - Overview of Onera acoustic active control activities in helicopter cabin 20m

        Within a helicopter cabin, passengers are in close proximity to disturbing sources that contribute to interior noise: main and tail rotors, engines, main gearbox and aerodynamic turbulence. These sources generate bending vibrations of the entire tail boom, induced vibrations in the cabin at blade passing frequencies, transient vibrations of rotor blades and structure borne noise induced by gear meshing within gearboxes. Conventional passive systems, as trim panels or passive anti-resonance isolation systems, are still the main way to control the cabin noise. Nevertheless, (semi-)active control techniques have been the subject of numerous studies for decades, especially to improve the performance of passive concepts in low frequencies. Unfortunately, many turned out to be unreliable or completely inapplicable because problems of robustness, time convergence of algorithms, but also a difficult identification of optimal locations for actuators and sensors or a high added mass and electrical power. The objectives of this paper is to describe in details the research projects in active noise control carried out at Onera since the 1990s, i.e. from numerical modeling and laboratory tests to flight tests in helicopter cabin. This overview will feed into the reflection on the renewal of active vibro-acoustic control within different laboratories.

        Orateur: Frank Simon (DMPE - ONERA - Université de Toulouse)
      • 14:40
        JISFA4#3 - Acoustic Absorption in business jets cabin 20m

        The acoustic design of a philharmonic auditorium involves controlling the propagation path between the instruments and the listeners. Today, this control has become a standard practice and is applied more broadly, even extending to the definition of the sound environment in business aircraft. Absorption needs to be strategically distributed based on its cost, mass, size, and effectiveness on specific sources (aerodynamics, systems, passengers, etc.).

        Before modeling the distribution of absorption, an assessment is necessary. The perception and measurements of in-situ reverberation times indicate that a cockpit does not have the same properties as a cabin. Thus, a primary analysis objective is to quantify and differentiate this observation. By increasing precision, it is possible to distinguish between different cabin configurations based on the materials used.

        To bridge the gap between overall observations and the role of each component, a link must be established between elementary measurements taken with a "Kundt tube" or in an "alpha" chamber, and the effect of bursting balloons in different areas of the aircraft. A testing pyramid must be implemented, taking into account the specificities and constraints of each level. The measurement tools and analysis methods must consider the mechanical properties of material samples with a diameter of 30 mm, as well as the geometry of full-scale prototypes of interior elements or local in-situ measurements.

        Once the measurement data is gathered, it is necessary to verify that the calculation tools allow for restitution and extrapolation. Typically, and in coherence with the testing pyramid, there are several levels and types of modeling: synthesis, energy-based propagation methods, "BEm," volumetric finite elements depending on the degree of homogenization, etc.

        Once the observations and requirements have been identified, and the foundations of measurement and simulation techniques have been established, the discussion can focus on control, i.e., the design of an optimized interior. This involves considering the localization/application (sound traps, silencers, wall treatments, etc.) and the elements used (meta-)materials and specific features (perforations, gaps, assemblies, elemental mesh architectures, etc.).

        Orateur: Yann Revalor (Bruit interne, Dassault Aviation )
      • 15:00
        JISFA4#4 - Multimodal characterisation of acoustic liners using MAINE Flow facility 20m

        MAINE Flow (for Multimodal Acoustic ImpedaNce Eduction with Flow) is a large-scale duct that allows investigating the acoustic properties of liners with flow and acoustic conditions typically found in nacelles of aircraft engines: the incident acoustic level can go up to 150 dB and the flow velocity up to Mach 0.6. Compared to other large-scale experiments, this facility permits a precise control of the modal content and amplitude, as well as the measurement of the scattering matrix of the liner. In this presentation, we will describe the several methods that have been developed to permit these measurements, including inverse and direct impedance eduction. Results on different liner technologies will be displayed and compared against measurements of the same acoustic treatments using smaller facilities. Finally, flow and modal content effects will be studied and discussed.

        Orateurs: T. Humbert (Laboratoire d'Acoustique de l'Université du Mans (LAUM), UMR CNRS 6613, Institut d'Acoustique - Graduate School (IA-GS), Le Mans Université), J. Golliard (CTTM)
      • 15:20
        JISFA4#5 - Smart Acoustic Lining for UHBR Engines 20m

        The SALUTE project aims at evaluating performance of electroacoustic metasurface, employing a surface array of controlled electroacoustic actuators, for smart acoustic lining under grazing turbulent flow for UHBR Engines. Theoretical and numerical investigations have been carried out for designing innovative concepts for complex aero-acoustic characterization in an engine mock-up. A specific focus was placed in the realization of prototypes for evaluating the metacomposite liner performances in 3D liners close to real engine implementation, its process complexity and robustness. 
This talk presents the concept development from theory to technological realization and characterization by produced numerical tools. The experimental results obtained with the liners in acoustic flow duct facilities have been realized in the PHARE facilities of Ecole Centrale de Lyon. Different configurations of liners have been tested using flow conditions of the engine: a passive liner used as reference and a 3D active liner based on an array of electroacoustic absorbers. These tests combine acoustics and aerodynamics measurements to characterize the aeroacoustics flow conditions, the membrane behavior, the achieved acoustic impedance and the resulting insertion loss.
The SALUTE project has received funding from Clean Sky 2 Joint Undertaking under the European Union’s Horizon 2020 research and innovation program under grant agreement N° 821093.

        Orateurs: K. Billon (Univ. Lyon, Ecole Centrale de Lyon, LTDS UMR 5513), M. Gillet (SUPMICROTECH-ENSMM, Univ. Franche-Comté, CNRS, institut FEMTO-ST), E. Salze (Univ. Lyon, Ecole Centrale de Lyon, LMFA UMR 5509), M. Volery (Signal Processing Laboratory LTS2, Ecole Polytechnique Fédérale de Lausanne), E. De Bono (SUPMICROTECH-ENSMM, Univ. Franche-Comté, CNRS, institut FEMTO-ST), M. Ouisse (SUPMICROTECH-ENSMM, Univ. Franche-Comté, CNRS, institut FEMTO-ST), H. Lissek (Signal Processing Laboratory LTS2, Ecole Polytechnique Fédérale de Lausanne), M. Collet (Univ. Lyon, Ecole Centrale de Lyon, LTDS UMR 5513), J. Mardjono (Safran Aircraft Engines)
      • 15:40
        JISFA4#6 - Round Table JISFA4 20m
    • 14:00 → 16:00
      Survishino 10 / Data driven condition monitoring 1
      Président de session: Eltabach Mario
      • 14:00
        "DATA1#1 - Vibration based milling diagnostics using Artificial Intelligence" 20m

        In industrial machining processes, tool failures may result in losses in surface and dimensional accuracy of a finished part, or possible damage to both the work piece and the machine. Consequently, tool condition monitoring has become essential to achieve high-quality machining as well as cost-effective production. Moreover, cutting tool degradation may vary considerably under different operation conditions and materials behaviour. Therefore real time identification of the tool state during machining, before it reaches its failure stage, is critical. In this study the vibrations of the cutting tool and of the workpiece material are online measured. Features are then calculated in time domain from the raw signals. Transient zones, when the cutting tool enters or exits the material, can be considered or not. All the calculated features are normalized and stored in a table. It is then necessary to make a dimensional reduction of that feature table in order to avoid overfitting and to reduce the computing time of the learning algorithm. In this study, 54 milling experiments were conducted from which features were calculated and then split into two groups: 80% for the machine learning model training, 20% for the test phase. The first part of this study proposes an analysis on the impact of the features on the robustness of the models, and a second part focuses on a real-time data driven prognostics and health management (PHM) approach for tool condition monitoring, based on supervised machine learning techniques (i.e. the model training needs labelled data). The fusion of decision coming from several machine learning algorithms (kNN, decision trees (DT) and random forest (RF)) is then used to predict the tool quality in real time. All parameters and configurations of the algorithms are optimized in order to maximize the real time diagnosis accuracy. The experimental results show that our proposed approach achieves good accuracy and real time performances in dry milling operations. Results of our study are implemented in real tool wear diagnosis, and thus give new opportunities toward realizing Industry 4.0

        Orateur: Knittel Dominique
        Vibration based milling diagnostics using Artificial Intelligence
      • 14:20
        "DATA1#2 - CANCELLED - Multi-Source Information Fusion Fault Diagnosis for Rotating Machinery using Signal and Data Processing" 20m

        Machine fault diagnosis is crucial in industrial systems to enhance reliability, lifetime, and service availability. Intelligent fault diagnosis (IFD) using artificial intelligence (AI) techniques has emerged as a promising approach for automating machine health assessment and reducing labor costs. One approach to improve fault diagnosis accuracy, which has become a highly relevant research topic, is to use multi-data fusion, which combines information from multiple sources to make a more informed decision. However, there is a lack of research focusing on the detection of combined machinery faults from multiple sensors. Indeed, when combined (and emerging) faults happen in different parts of the rotating machines their features are deeply dependent and the separation of characteristics becomes complex, while multi-sensor information can provide more comprehensive fault features to deal with the diagnosis and identification of multiple combined faults. This paper presents a comprehensive methodology for diagnosing combined faults using data fusion and machine learning techniques. The proposed approach leverages multiple types of sensor data, including vibration, current, temperature, and acoustic data, sensors to provide a comprehensive picture of the machine's health. Our proposed methodology incorporates an ensemble learning approach and time-domain features to improve diagnostic accuracy. The proposed approach is tested on a publicly available dataset of rotating machinery with multiple faults. The results indicate that the method is viable and achieves good accuracy and efficiency. Keywords€” Machinery Fault diagnosis, Combined fault diagnosis, multi-sensor data fusion, Data and Signal processing, Broken rotor bar, Bearing fault

        Orateur: Makrouf Iman
        Multi-Source Information Fusion Fault Diagnosis for Rotating Machinery using Signal and Data Processing
      • 14:40
        "DATA1#3 - Wind turbine drivetrain fault detection using physics-informed multivariate deep learning" 20m

        Vibration analysis is a prevalent technique in the predictive maintenance of wind turbines. It is an effective method for early fault detection and enables the creation of cost-effective maintenance strategies. Commonly used vibration analysis methods in the literature rely on signal processing techniques such as time and frequency domain approaches. However, the signal processing techniques require manual interpretation by domain experts. It is important to note that different indicators exhibit sensitivity to specific faults. Manual analysis of indicators can be avoided by fusing them to derive high-level wind turbine health status. It enables the learning of complex non-linear relationships among the indicators. This research focuses on a multivariate deep learning model, i.e., autoencoder, which fuses different signal processing indicators to provide a single high-level health status. The proposed model is a normal behaviour model that learns the indicator's normal behaviour and labels faults if it observes deviation from the normal behaviour. The proposed fusion method of indicators is robust compared to individual indicator models as it learns complex non-linear relationships among indicators. The proposed method is tested for fleet-level fault detection both with and without fine-tuning for a specific wind turbine. Furthermore, it decreases the time required for wind farm health prognosis analysis and computation. Various autoencoder architectures have been compared, including simple feedforward neural networks, convolutional neural networks, and recurrent neural networks. The proposed method is demonstrated using real-life, high-frequency condition monitoring data from offshore wind turbines over several years, including wind turbines observed faults. The method's effectiveness and performance were demonstrated through analysis of planetary stage, generator, and high-speed stage failure cases.

        Orateur: Jamil Faras
        Wind turbine drivetrain fault detection using physics-informed multivariate deep learning
      • 15:00
        "DATA1#4 - Anomaly Detection in Aircraft Engine Vibration Using Deep Convolutional Autoencoder" 20m

        The useful life of aircraft engines depends on their operating environment (polluted areas, harsh climate, etc.). Detecting signs of degradation and aging can be difficult due to background noise measured on vibrational signals. Statistical methods such as threshold-based monitoring may not be reliable enough. This paper presents a promising method based on learning normal behavior on a population of engines considered to be healthy, such as newly produced engines. The learning is done by calculating spectrograms of the vibrational signals, normalizing them and treating them as images, then using a convolutional autoencoder to learn normal behavior. This model can be used during shop visits to detect early degradation by comparing vibrational signals of in-use engines to the learned standard. Keywords: vibrational signals, background noise, normal behavior, convolutional autoencoder.

        Orateur: El Hidali Abdallah
        Anomaly Detection in Aircraft Engine Vibration Using Deep Convolutional Autoencoder
      • 15:20
        "DATA1#5 - Airplane turbulence detection with hybrid deep learning model" 20m

        Atmospheric turbulence has a significant impact on airplane motions and can induce excessive stress and fatigue damage. The identification of turbulence in aircraft service phase is of particular interest to estimate actual structural fatigue or to define an improved maintenance plan. This article describes a new model to detect turbulence from in-board instrumentation solely. Equations expressing the relationship between the load factors in the 3 directions and the forces applied to the airplane are derived from a rigid body 6 degrees of freedom model. Lift, drag and lateral force coefficients, required to compute the aerodynamic forces, are predicted by multi layer perceptrons. The architecture of the model makes it possible to train the multi layer perceptrons in an unsupervised way with common deep learning techniques, using only sensors commonly present on airplanes. After the training phase, the model is able to predict the lift, drag and lateral force coefficients for various configurations. The error between the measured load factors and the load factors predicted by the model is used to identify the presence of turbulence. The performances of the model to predict the lift and drag coefficient is first evaluated on simulated data. The turbulence detection is then evaluated on a dataset composed of hundreds of commercial flights as well as on simulated data.

        Orateur: Dampeyrou Charles
        Airplane turbulence detection with hybrid deep learning model
      • 15:40
        "DATA1#6 - Ball bearing diagnosis using a homogenous hybrid database in a supervised machine learning" 20m

        Digital twins (DT) are often described as a virtual and dynamic representation of a system. They guarantee interaction between physical and virtual spaces. In the context of maintenance 4.0, the lack of historical data can be caused by an impossible instrumentation for complex systems. To face it, DT offers the possibility to simulate several operating modes which can serve for a diagnostic. This operation can be made by using machine learning algorithm (MLA) through a diagnosis by classification. But the challenge is to identify the best use of both data historical and simulated on a hybridisation database to make the most reliable diagnosis. In this paper, a digital twin combining a discrete element model (DEM) and a finite element model (FEM) is developed to generate data with an outer race default signature. These generated data with five sizes of defaults are also measured on the test bench. According to a percentage, historical data are used to build the homogenous hybrid database. Two MLAs (Support Vector Machines and K-Nearest Neighbours) are used to perform a classification by training the homogenous hybrid database and the test is realised by using the rest of historical data. The results of this approach show a better reliability than existing methods on the tested datasets also it's allowed to evaluate the contribution of historical data in homogenous hybridisation process.

        Orateur: Sow Souleymane
        Ball bearing diagnosis using a homogenous hybrid database in a supervised machine learning
    • 14:00 → 16:00
      Survishino 11 / Passive control of Vibrations
      Président de session: Régis Dufour
      • 14:00
        "PACO#1 - Vibration response of a machine structure filled with high-damping material" 20m

        This work focuses on the use of a polymer concrete as a filling material of structural parts to damp vibrations in machine tools. The goal of the overall project is to improve the dynamic behavior of CNC lathe systems by limiting at most the vibrations triggered in operative conditions, in order to guarantee a high productivity of the machine tool while keeping the machining quality of the final products. The task is very challenging since the vibrations transmitted to the tool-workpiece area through the structures of both the lathe and its joined bar-feeder are mainly associated with the bars' critical velocities, which are extremely variable in a wide range. Indeed, the latter depend on operative factors that are not constant for all the working conditions of the system, e.g. material and geometry of the bar to be machined, working angular speed, instantaneous free-length of the bar supported by the bar-feeder... As a consequence, designing the machine structures in order to avoid resonances in the large broadband of the excitations is basically impossible. Hence, a successful strategy could be levering the structural components' damping in order to limit the vibration levels in a wide frequency band. To this purpose, polymer concretes, mineral casts, and metal foams can be used as filler of machine tools and automatic machines beds. The prototype of a new base €“ filled with a commercial polymer concrete €“ of an automatic bar-feeder was developed and many laboratory tests were performed. Vibration signals were acquired for different working conditions (varying both the bars and the spindle velocity of the connected lathe). The comparison of the dynamic responses of the original machine (no filler) and the new variant highlights an overall favorable effect of the high-damping material properties, though a slight worsening appeared in a few circumstances (likely due to the additional mass of the filler that lowers natural frequencies thus making the vibration amplitudes increase). This work illustrates the design of the new prototype, the performed experiments, the results of the signal analyses, and some concluding remarks on possible future developments for the presented application.

        Orateur: Martini Alberto
        Vibration response of a machine structure filled with high-damping material
      • 14:20
        "PACO#2 - Nonlinear damper approaches to flutter mitigation in highly-flexible wings" 20m

        The High-Altitude, Long Endurance (HALE) aircraft is an innovative concept with great potential in telecommunications and surveillance applications due to its ability to function as a "pseudo-satellite" within the stratosphere. Its unique features are accomplished in practice through numerous design requirements that ensure maximal efficiency. Regarding the airframe in particular, these lead to light wings with extremely high aspect ratios. The resulting -very flexible- structure is consequently sensitive to dynamic instabilities emerging from aeroelastic effects (i.e., flutter), which produce large-amplitude oscillations and drastically diminish the aircraft's flight envelope. Hence, accurate prediction of flutter and the ability to mitigate it are crucial to enable reliable HALE aircraft design and operation. In the present work, we explore passive control strategies for flexible wing flutter through the nonlinear effects introduced by an added damper system. To this end, an original aeroelastic model of the wing is introduced, whose structural component provides a degree of accuracy and complexity that lies at a midpoint between linear and geometrically-exact nonlinear beam models. The damper is modelled as a single-degree-of-freedom secondary subsystem attached to the wing through a linear spring, and the equations of motion of the full system are implemented numerically. Mitigation performance is assessed with regard to flutter speed, instability range and vibration amplitude. Parametric studies are conducted for varying stiffness, span-wise location and dissipation mechanism of the damper. The latter of these aspects is given special attention, and the respective performances of viscous and nonlinear damping are compared. Overall, our results show that dynamic stability of the wing can definitely be improved through the use of a well-tuned damper. However, the choice of design parameters leading to optimal performance is far from obvious. Furthermore, the simplistic mechanism proposed herein is subjected to evident practical limitations, and thus this work concludes with a discussion on potential extentions and considerations for implementation.

        Orateur: Alcorta Roberto
        Nonlinear damper approaches to flutter mitigation in highly-flexible wings
      • 14:40
        "PACO#3 - Nonlinear passive control of galloping of overhead transmission lines: design and numerical verifications" 20m

        In the field of power delivery, the vibration of overhead conductors affects the sustainability of transmission lines. The passive control of galloping of conductors is the subject of this study. Galloping oscillations are caused by ice accretion on a cable that creates an aerodynamic instability. The consequences of galloping on transmission lines are electrical outages, fatigue failure of cables, and the impacts between cables. The study focuses on the nonlinear passive control of a single conductor. The use of a nonlinear absorber called as nonlinear energy sink (NES) with a piecewise linear restoring forcing function for galloping mitigation on a single span of a suspended cable is proposed. An analytical model of a single conductor span coupled to a NES is developed. The fluid-structure interaction, i.e. the interactions of the wind and the ice-accreted cable, is modeled by a parametric excitation supposing the quasi-steady theory. A complexification technique accompanied by the time multiple scale method is used to determine the slow and fast dynamics of the system. The bifurcation diagrams are analytically determined and compared with results obtained from the numerical integration of the governing equations of the system. The system with harmonic excitation is modeled using the finite element (FE) method with the software Code_Aster. An equivalence between harmonic and parametric excitation is addressed to compare the results from the FE model and the analytical developments. The influence of the parameters of the nonlinear absorber: clearance, stiffness, and damping coefficient on galloping mitigation is studied.

        Orateur: Leroux Matthieu
        Nonlinear passive control of galloping of overhead transmission lines: design and numerical verifications
      • 15:00
        "PACO#4 - Understanding vibroimpact damping through a numerical energy based approach" 20m

        Vibroimpact damping classically consists in the integration in a main vibrating structure of oscillating masses. The energy is dissipated thanks to the impact interactions with the main vibrating structure. An assumption is that an energy transfer towards high frequencies occurs thanks to nonlinear effects, leading to a faster and more efficient energy dissipation compared to lower frequencies. This hypothesis is investigated within the framework of the work presented, in the case of a beam equipped with a vibroimpact absorber, assuming conservative impacts. A numerical study is developed to solve the equations of motion and derive energy balance with negligible error. Then, the energy transfers in the case of a harmonic excitation centered on not only the first, but also the second, and third mode are considered demonstrating the effectiveness of the vibro-impact absorber in each case. Additionally, the way the energy is dissipated is explained and quantitatively assessed. At least, the results obtained for the three first modes are compared.

        Orateur: Sadoulet-Reboul Emeline
        Understanding vibroimpact damping through a numerical energy based approach
      • 15:20
        "PACO#5 - A passive nonlinear absorber for controlling pathological tremors of human arm" 20m

        Mechanical vibration absorbers are applied in different domains to protect structures and machines from undesired vibrations. One of the precise applications of such control systems is reducing tremors of the human arm caused by some neurological disorders such as essential tremors. In recent studies, active and linear passive controllers are designed, in the form of a wearable bracelets, to reduce the shaking in the upper limbs of patients. Active controllers require a power source to operate, and their components are a matter of concern regarding safety. Passive linear absorbers need accurate calibration of their operating frequency to be efficient and require the usage of multi-absorbers to solve this problem. Passive nonlinear absorbers are proposed to overcome the shortcomings of linear absorbers because they can be designed to have a broadened frequency bandwidth. In the current work, the arms of the upper limbs are modeled as links connected by joints, and the passive tissues of the muscles as springs and dampers. A voluntary moment is considered to represent the torque exerted by the human to place his arms at a desired position. The involuntary moment is used to excite the system and to reflect the activity of the muscles generating the tremor. Tremor signals at the upper limbs, of a patient with essential tremor, are measured to obtain the translational and angular responses. These measurements are used to calibrate the modeled dynamical system of the upper limbs and to deliver a reliable numerical response very close to the measured one. The system is coupled to a passive nonlinear absorber and the multiscale method is used to study the behavior of the system at the fast and slow timescales. It allows us to obtain the invariant manifold of the system and to detect the equilibrium and singular points needed for the design of a nonlinear energy sink (NES). A non-smooth, piece-wise linear NES is used for its design. The response of the system obtained analytically is validated by those obtained from numerical simulations. The designed NES is fabricated and the preliminary prototype is obtained.

        Orateur: Gebai Sarah
        A passive nonlinear absorber for controlling pathological tremors of human arm
      • 15:40
        PACO#6 - Control based continuation of autonomous system — Stabilization mechanisms 20m

        Experimental analysis of nonlinear dynamical system is always a complicated task because these system can exhibit bifurcations, multi-stability or unstable behavior. When a mathematical system is available, numerical continuation algorithm constitute a powerful tool to investigate the so-called bifurcation diagram. Sieber and Krauskopf introduced the idea of control based continuation, obviating the need of a mathematical model so that it can be used to obtain the bifurcation diagram from a physical experiment [1]. Control based continuation methods rely on a feedback controller to stabilize the possibly unstable response of the system under test. This feedback controller must be non-invasive to ensure that the solution of the controlled system is also a solution of the uncontrolled system.
        Since then, control based continuation received a growing interest and has been mostly used to track branches of periodic solutions of non-autonomous (i.e. forced) systems [4, 3, 2]. Although it has already been applied on various experiments, only few papers deal with the detailed analysis of the underlying stabi- lization mechanism [5]. Applying these methods to autonomous systems presents an additional difficulty since the frequency of the limit cycle is also an unknown. Numerical continuation algorithms solve this problem by appending the system with a phase condition.
        The objectives of this study are twofold. (i) design a controller able to stabilize limit cycle oscillations, (ii) investigate the underlying stabilization mechanism of the designed controller.

        Orateur: Etienne Gourc
        gourc.pdf
    • 14:00 → 16:00
      Survishno 12 / Condition monitoring 3
      Président de session: Castellani Francesco
      • 14:00
        "COMO3#1 - Experimental study on condition indicators for severity estimation of growing spall in bearings." 20m

        Rolling element bearings are essential components for the proper functioning of many types of rotating equipment. Diagnosing faults in bearings has traditionally been done using signal processing techniques inspired by physics, where acceleration signals are analyzed using time-frequency analysis methods. One of the key challenges in classifying the spall severity in practical applications is that changes in acceleration signatures, which are related to the size of the spall, are hard to detect due to low signal-to-noise ratios (SNRs).  The objective of this research is to study and characterize the effect of spall propagation on acceleration signatures to classify and identify the spall severity. To overcome the challenge of low SNRs, we focus on changes in signal trends rather than events in single measurements. Experiments were conducted to gather data from endurance tests with growing faults on the outer ring of cylindrical roller bearings. The data collected includes measurements of acceleration and load at various rotational speeds.  One benefit of conducting endurance tests is that they allow for the natural propagation of spall, however, the extent of spall severity during the test remains uncertain. To overcome this, a spall size estimate is used derived from the load-cell signals, which is validated by means of visual inspections. Although a load-cell is not available in practical applications it is used in our research as the €œground truth€ to validate the acceleration-based algorithms.  A new condition indicator (CI) for classifying spall severity is proposed. This CI was derived through analysis of CIs trends, extracted from order domain signatures. The new CIs enable the identification of several stages of spall propagation prior to reaching the critical size, where beyond asset operation is no longer acceptable. The effectiveness of this new CI was demonstrated using four different endurance tests.

        Orateur: Bublil Tal
        448825.pdf
      • 14:20
        "COMO3#2 - Wavelet-based high order spectrum for local damage diagnosis of gears under different operating conditions" 20m

        Gears play an important role in transmission systems, allowing for high performance in terms of load capacity and efficiency. Common gear faults such as cracked teeth and pitted teeth, can occur as a result of contact fatigue, excessive load, or sudden impact. Starting from an initial stage, their steady growth can lead to irreparable damage and unexpected breakdowns that result in economic losses. Therefore, local tooth damage diagnosis of gears using advanced monitoring techniques is extremely important for the normal operation of drivelines and transmissions. The presence of local tooth damage on gear tooth produces transient impact in the vibration signals, which exhibit non-stationary and non-linear characteristics. Taking into account its ability to characterize the phase coupling between signal components caused by non-linearity, wavelet-based high order spectrum (e.g. wavelet bispectrum/bicoherence) is considered to be effective to attain reliable fault-related features. Among others, wavelet bicoherence technology has been successfully applied to detect the artificially created gear faults under steady speed and load. However, in case the operating condition changes, the effectiveness of this method in detecting gear faults is still unclear. Additionally, there is no mature idea of selecting informative bifrequency bands and extracting instantaneous diagnostic features after applying wavelet bicoherence. This may constraint the widespread application of wavelet-based high order spectrum in gear fault diagnosis. This paper presents a novel strategy for selecting informative bifrequency bands and extracting instantaneous diagnostic features in the time-bifrequency domain. The performance of the proposed methodology is evaluated and extended to cases involving healthy and faulty gears operating under different speeds and loads. To validate the effectiveness of the methodology, a publicly available dataset is utilized, which includes gears with various crack severity as well as different speed and load operating conditions.

        Orateur: Zhu Rui
        454663.pdf
      • 14:40
        "COMO3#3 - Identification of non-informative noise component in time-frequency representations. Application to vibration-based local damage detection" 20m

        In this presentation we highlight the importance of the background noise properties in the context of the vibration-based local damage detection. We assume the model is a mixture of signal of interest (SOI) and the noise. In the case, when the background noise has Gaussian characteristics, the classical methods for local damage detection can be applied. In this case, the most common approaches are based on the measuring of impulsiveness of the vibration signal or cyclostationary analysis. In both cases, the used methods often are applied to the signal in time-frequency representation. However, for many cases in real environment the assumption of Gaussian distribution of the noise is not satisfied and one may expect the large impulses that influence the noise characteristics. It should be noted, the non-Gaussian distribution of the signal may occur for many machines. In that case the impulsiveness criteria fail and the cyclostationary analysis seems to be more useful approach. Since, most of the methods used in the cyclostationary analysis are based on the autocovariance function, we indicate here the important role of the finite variance of the signal. In theory, if the variance is infinite, then the autocovariance is also not defined. We highlight, the problem considered here is much more general than the problem of testing the noise distribution. We present a new approach for the assessment of the noise probabilistic properties. The methodology is applied for the time-frequency representation of the signal. The problem is illustrated for the simulated signals from non-Gaussian distributions and real signals from various machines.

        Orateur: Wylomanska Agnieszka
        445870.pdf
      • 15:00
        "COMO3#4 - Robust estimators of autocorrelation function in application to local damage detection for non-Gaussian signals" 20m

        One of the most common approaches for local damage detection is the cyclostationary analysis. The indicators of cyclostationarity are based on the classical estimation of autocorrelation function (ACF), called sample ACF. It can be applied for the underlying signal in time, time-frequency or frequency-frequency domains. The indicators based on sample ACF are very efficient in case when the informative signal is disturbed by Gaussian- (or close to Gaussian)-distributed noise. However, in case when the background noise has strong non-Gaussian behavior, the sample ACF may fail as it is sensitive to large impulses related to the non-Gaussian characteristics of the noise. Thus, in this presentation we discuss the new approach based on the robust versions of sample ACF. By robust sample ACF we mean the algorithms less sensitive to large observations that estimate theoretical ACF. By relatively simple replacement of the classical statistic by its robust versions, one may decrease the influence the non-Gaussian distribution and identify the cyclostationary behavior also in this case. In the literature there are considered various statistics used as robust versions of sample ACF but they were never used in condition monitoring. In this presentation we demonstrate the general methodology of new cyclostationary indicators that, similar as the classical approach, can give the information in different domains. The results are demonstrated for three selected robust estimators of ACF and two different non-Gaussian distributions of the background noise. The simulation studies are supported by applications of the introduced methodology to real vibration signal.

        Orateur: Zulawinski Wojciech
        446450.pdf
      • 15:20
        "COMO3#5 - Algorithm for the detection of faults in rolling element bearings running under tacholess and variable rotating speed conditions" 20m

        The goal of this research was to implement a new algorithm for the automatic detection of faults in rolling element bearings, in such a way that it does not depend on the estimation of the shaft rotational speed. The idea was that the algorithm could be applied under variable and unknown rotational speed conditions. The proposed algorithm was based on the detection of phase-relationships between spectral components that emerge when an amplitude modulation appears in the gathered vibration. To do that, a mode decomposition procedure, as well as the Hilbert transform, was applied in order to estimate a detection coefficient, which served as indicator of the presence of the modulation produced by the faults. Lock-in amplifiers were used in order to calculate such an indicator. The effectiveness of the method was validated through experiments preformed by using simulation and real signals. It was proven that the application of the proposed method can lead to an effective detection of the modulation featuring the existence of rolling element bearing faults.

        Orateur: Hernandez Fidel
        441874.pdf
      • 15:40
        COMO3#6 - 20m
    • 16:00 → 16:20
      Coffee break 20m
    • 16:20 → 17:00
      Exhibitor session
      Président de session: Simon CHESNE (LaMCoS INSA-Lyon)
    • 17:00 → 17:40
      Survishno Contest Session
      Président de session: Quentin Leclère (Laboratoire Vibration Acoustique, INSA Lyon)
    • 17:40 → 18:40
      Round Table / Adapting Dynamics and Acoustics Professions and Training for Future Industrial Needs
      Présidents de session: Adrien Pelat (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613), Dupont Jean-Baptiste
    • 19:30 → 23:30
      Congress Social Event - JJCAB prices & SFA industrial medal - Congress Diner 4h L'Envol des Pionniers

      L'Envol des Pionniers

      6 Rue Jacqueline Auriol, 31400 Toulouse
  • jeudi, 13 juillet
    • 09:20 → 10:00
      Keynote / Prof. Régis Dufour - INSA Lyon, LaMCoS, UMR CNRS 5259 40m

      Title:
      Nonlinear on-board rotordynamics

      Abstract:
      Unperfect balancing, pedestal motion, rotor-stator interaction, multi-field physics coupling, are capable of triggering nonlinear phenomena that can threaten the integrity of rotating machines. It is therefore essential to predict their nonlinear dynamic behavior in order to be able to avoid the occurrence of such phenomena. The presentation focuses on theoretical and experimental investigations for establishing Finite Element models as reliable as possible. A particular attention is paid on a 6-DOF shaker, unique in the French university, which permits carrying out the tests. The illustrations address several industrial applications dealing mainly with on-board rotors subject to multi-excitations.

      Orateurs: Régis Dufour, Guilhem MICHON (ISAE - DMSM)
    • 10:00 → 10:20
      Resonance Summer Camp Session
      Président de session: Guilhem MICHON (ISAE - DMSM)
    • 10:20 → 10:40
      Coffee break 20m
    • 10:40 → 12:40
      JISFA 5 / Noise and Vibration of Spacecrafts
      Président de session: Leonardo Sanches
      • 10:40
        JISFA5#1 - Some experimental developments on supersonic jet noise reduction by water injection 20m

        In order to protect the payload and the rocket structure from severe vibrations, water injection is widely used on launch pads to reduce the high noise level radiated by booster exhausts at lift-off. If this technique appears to be effective, it remains that the physics of the phenomena associated with the noise reduction of hot supersonic turbulent jets is poorly known, making the optimization of the technique and the development of predictive models particularly difficult.
        Experiments on the MARTEL test facility are then carried out on jets up to Mach 3 and 2100 K in an environment acoustically representative of the launcher lift-off. These researches undertaken in a long-standing cooperation with the CNES aim to understand the noise reduction mechanisms and design the ground infrastructure of the French space centre in Kourou.
        Two main phenomena lead to jet the noise attenuation. The first is associated with the decrease of the mean jet velocity by momentum and heat transfer between the liquid and the gaseous phases. The second concerns the modification of the mixing layer turbulence and subsequently of the shock cells structure.
        In addition, several related phenomena have to be considered to fully address the issue: aerodynamic and acoustic interactions between the jet and the ground structure, with possible feedback loops; water diffusion, fragmentation and vaporization into the mixing layer; parasitic noise due to the impact between the gaseous flow and the water droplets; acoustic waves propagation in the two-phase media. The coupling between these different mechanisms remains partially understood and numerical models are currently not faithfully predictive.
        This lecture will present a review of experimental developments in progress on the MARTEL test bench. Current experiments focus on: i) aero-acoustic scaling laws with water injection depending on jet morphology, temperature, Mach number and nozzle diameter; ii) a two-phase optical metrology development based on a simultaneous measurement of phosphorescence and fluorescence particles; and iii) some jet interaction tests with simplified launch pad structure (flat plate or plate with a hole).

        Orateurs: Patrick Berterretche (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Gildas Lalizel (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Romain Bouju (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Pierre Beltra (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Florian Moreau (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Eva Dorignac (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers), Hadrien Lambare (Centre National d’Etudes Spatiales), Yves Gervais (Institut PPRIME, UPR CNRS 3346 / ISAE-ENSMA / University of Poitiers)
      • 11:00
        JISFA5#2 - Large eddy simulations of launcher lift-off noise and comparisons to experiments 20m

        During the lift-off phase of a space launcher, rocket motors generate harsh acoustic environment that is a concern for the payload and surrounding structures. Hot supersonic jets contribute to the emitted noise from both their own noise production mechanisms and their interactions with launch pad components, such as the launch table and flame trenches.
        The present work describes the results of computations performed by ONERA to predict the lift-off noise from reduced scale models, carried out by CNES at the MARTEL facility operated by Pprime in Poitiers.
        Numerical computations involve two in-house codes: the flow solver CEDRE, used in LES mode to accurately predict the noise sources, and the Euler code SPACE to propagate nonlinearly the generated noise to the far field. Results are discussed and compared with experimental measurements.

        Orateurs: Julien Troyes (DMPE - ONERA - Université de Toulouse), François Vuillot (DMPE - ONERA - Université Paris Saclay (COmUE))
      • 11:20
        JISFA5#3 - Forced and shock vibration reduction in a honeycomb sandwich panel using a Vibro-Impacted Acoustic Black Hole attenuator 20m

        An Acoustic Black Hole (ABH) is a passive technique for vibration reduction without adding mass. It usually consists of a thickness gradient following a parabolic thickness profile, inserted in a thin structure such as a beam or plate. This geometric inhomogeneity, combined with the addition of a viscoelastic thin layer, leads to a localization effect and a particularly efficient damping of bending waves. There is a wide variety of designs with attractive damping performance of vibration levels above a threshold frequency depending on the size of the ABH. In addition, it is possible to significantly improve low frequency performance by introducing contact nonlinearities able to transfer vibrational energy from low frequencies to high frequencies where the resulting device is highly efficient. In the framework of a CNES Research & Development program, an ABH disk with vibro-impactors, designed as an additional attenuator attached to a structure to be damped is proposed. The key points of the design and the performances are studied by means of numerical simulations combining a spatial finite element model and a conservative time integration scheme. A demonstrator is also studied experimentally in both the linear and non-linear regimes (without and with vibro-impactors, respectively). The performance analysis, conducted from the vibrational responses to broadband noise and shock excitations, show promising results for possible further applications.

        Orateurs: Haiqin Li (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613), Patrick O’Donoghue (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613), Adrien Pelat (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613), François Gautier (Laboratoire d’Acoustique de l’Université du Mans, UMR CNRS 6613), Cyril Touzé (Institute of Mechanical Sciences and Industrial Applications, Unité mixte CNRS - ENSTA Paris - EDF – CEA), Florent Masson (ACOEM)
      • 11:40
        JISFA5#4 - Influence of non-structural elements on satellite dynamics 20m

        Satellite structures face a significant challenge in dissipating vibratory energy during launch loads. Currently, modal damping is commonly used to address this issue, although it is quite far from representing the actual physical dissipation mechanism. Various sources contribute to dissipation, including intrinsic material properties, joint interfaces, and fluid-structure coupling. However, existing simulations still fall short in accurately capturing these dissipation sources, leading to overlooked or underestimated effects. Satellite structures also incorporate non-structural components like electrical harnesses, typically treated as distributed masses in simulations due to their relatively small contribution to the total mass. Nonetheless, research has demonstrated that these substructures can actually contribute significantly to dissipation despite their low stiffness, acting as oscillators that recover vibratory energy from the main structure. In light of this, the present study focuses on investigating the nonlinear dynamic coupling between a satellite structure and its cable harness. To achieve this, a numerical model is developed by coupling a finite element model of the main structure with an analytical cable harness model based on beam theory, previously established in other research. An experimental setup is proposed, involving an aluminum honeycomb panel mounted on an electrodynamic shaker. By comparing the dynamic response measured through accelerometers attached beneath the panel, with and without the cable harness, the impact of the harness is observed. Finally, the setup is simulated numerically, and the predictions are compared to the numerical results and discussed.

        Orateur: Xavier Cadiot (Airbus Defence and Space)
      • 12:00
        JISFA5#6 - Round Table JISFA5 40m
    • 10:40 → 12:00
      Survishino 13 / Data driven condition monitoring 2
      Président de session: Wylomanska Agnieszka
      • 10:40
        "DATA2#1 - Detection of machine mechanical faults using vibrations and deep autoencoders" 20m

        Deep learning models represent a new learning paradigm in artificial intelligence (AI). Recent breakthroughs in image analysis and voice recognition have generated enormous interest in many other areas such as the diagnosis of rotating machinery providing voluminous data during the life of these machines. A gear endurance test bench was designed to create a database of vibration signals ranging from a healthy state to a degraded state. The aim of our work is to test common methods and the use of unsupervised deep learning in the detection of deviation from the normal operating state of the machine. After a description of the bench, we present the progress of the test over one year. The data collected consists of a representative set of vibration signals corresponding to different operating environments (speed, temperature, etc.). Then we present the experimental study using deep auto-encoding networks and we compare with usual methods. The principle of the chosen approach is to train the autoencoder with the sane data, so that it should learn to reconstruct only this type of data. When a new sample of data is supplied to the network, the reconstruction error is calculated and the objective to try to achieve is to obtain a low error for the healthy data and an error which begins to increase as the defect develops. We our paper with the work perspectives.

        Orateur: Eltabach Mario
        Detection of machine mechanical faults using vibrations and deep autoencoders
      • 11:00
        "DATA2#2 - Wind turbine gearbox condition monitoring through a multi-scale data-driven approach" 20m

        Since wind is expected to play a crucial role on the worldwide electricity production scenario, the reliability of the turbines is attracting attention from industry as well from the scientific community. New techniques for efficient condition monitoring of the key components can be fundamental in order to optimize the performance and the maintenance of a large fleet of turbines. The gearbox represents the most critical component, as it is responsible for a large part of the wind turbine downtime during its overall life. Anyway, monitoring for wind turbine gears is challenging due to the non-stationarity of the operation and the lack of noise-free vibration measurements.  In the present work, a new approach for long to medium term efficient monitoring of wind turbine drivetrains has been developed basing on real-world data. An incipient fault on the drivetrain of a turbine has been used as a test case for developing a new approach based on the use of multi-scale data sources. On one side SCADA (Supervisory Control And Data Acquisition) data have been used for a general monitoring of the state of the machine's component while, on the other hand, high multi-resolution data from triggered events collected by a CMS (Condition Monitoring System) were used to refine the diagnosis and prognosis of the fault.    Even if triggered events are difficult to be used when classifying a target machine with a healthy reference, the results demonstrate that the use of CMS multi-scale high resolution data can be much more effective in the fault diagnosis. In the present work, the one class-SVM (Support Vector Method) was used for novelty detection. Its effectiveness in unveiling the faulty machines is observed to be strongly boosted by the application of the dimensionality reduction through the PCA (Principal Component Analysis). The approach, when using all the available time scales, is able to detect the incoming fault also several years in advance and can therefore be proposed as a more robust approach with respect to the classical data-driven regression normal behaviour model developed with continuously available SCADA data.

        Orateur: Castellani Francesco
        Wind turbine gearbox condition monitoring through a multi-scale data-driven approach
      • 11:20
        "DATA2#3 - Vibration-based unsupervised detection of common faults in rotating machinery under varying operating speeds" 20m

        Automated condition monitoring (CM) of Rotating Machinery (RM) has several benefits for modern industry including enhanced safety and reduced maintenance costs. Vibration signals may be easily and inexpensively collected from a RM during its normal operation and include rich information about the RM dynamics the proper exploitation of which may lead to effective CM. However, RM operate oftentimes in harsh industrial environments under varying conditions which affect the vibration signals in a similar manner as potential faults. Thus, the detection of incipient faults without imprints to the time domain and controversial effects in the frequency domain due to their similarity with the different operating conditions necessitates robust methods eliminating false alarms and RM downtime. Although there are numerous of studies on the CM of RM, the vast majority focuses on fault classification using a large volume of labeled data from the faulty RM under all possible operating conditions for supervised training, which is impractical for industry. The detection of common incipient faults in RM under different operating speeds is investigated in this study via two unsupervised Machine Learning type methods using a reasonable amount of data from the healthy RM for their training. The first utilizes Functional Pooled (FP) AutoRegression (AR) models with a scalar operating parameter corresponding to the rotating speed of the RM for the interpretation of its healthy dynamics. The second interprets the healthy dynamics using a cloud of AR models within a Multiple Model framework. Vibration signals are captured by a single accelerometer mounted on a RM consisting of two electric motors coupled via a claw clutch and operating at a wide range of different speeds. The methods' performance is systematically assessed based on hundreds of experiments with the healthy RM, as well as with a slight unbalance, minor wear at the base of a single claw clutch tooth, and mechanical looseness at one of the four RM mounting bolts under a wide range of rotating speeds, not necessarily used in the methods training. The results indicate the superiority of the FP based method that achieves remarkable detection of all considered faults under any rotating speed.

        Orateur: Bourdalos Dimitrios M.
        Vibration-based unsupervised detection of common faults in rotating machinery under varying operating speeds
      • 11:40
        DATA2#4 20m
    • 10:40 → 12:40
      Survishno 14 / NVH of eco-efficient powertrains
      Président de session: Didier Remond
      • 10:40
        "EDRIVE#1 - Multi-Physic Analysis Of Power Electronic Control Parameters In A Simulation Framework" 20m

        Incorporating sound quality markers into optimization objectives is a new approach that aims to design more efficient and high-quality powertrains. In this paper, a digital framework, which combines different simulation environments, is proposed with the intent of investigating the influence of pulse width modulation strategies on vibroacoustic response and energy consumption of an electric drive unit. Different switching schemes and frequencies are relatively compared within this process. A high frequency, non-linear, 1D model of an electric drive unit is realized to estimate the phase currents and rotor position of a permanent magnet synchronous machine, for certain working conditions. These results are employed in a 2D electromagnetic FE model to estimate the magnetic forces in the air gap. The structural modes of the stator are calculated by means of finite element method. The vibration modes are used together with the magnetic forces generated previously to calculate a modal-based forced response. A vibroacoustic simulation generates the response up to 12.5 kHz for certain target positions. Finally, the acoustic spectral information is rendered in time domain and a sound quality assessment is performed on the synthesized signals using psychoacoustic metrics.

        Orateur: Salamone Nicolo
        Multi-Physic Analysis Of Power Electronic Control Parameters In A Simulation Framework
      • 11:00
        "EDRIVE#2 - NVH optimization of electric motors: optimization under constraints and uncertainty" 20m

        The optimization methodology described in this paper aioms at minimizing the noise and vibrations of electric motors. It relies on a numerical workflow involving electromagnetic, dynamic and acoustic models. Afterwards, a deterministic optimization method is described and applied to automotive traction motors. A robust enhancement of the optimization is then presented. It aims at considering the effect of manufacturing tolerances, materials properties dispersion and control uncertainties on the vibratory and acoustic levels to minimize, so that the electric motors are really silent when manufactured and operated. An application case of this robust optimization based on an automotive motor is then detailled.

        Orateur: Dupont Jean-Baptiste
        NVH optimization of electric motors: optimization under constraints and uncertainty
      • 11:20
        "EDRIVE#3 - Resynchronization of tonal acoustic field in multi-pass non-stationary microphone array measurements" 20m

        The geometrical limitations of microphone arrays as well as the unavailability of multiple channel data acquisition systems or their incurred costs fundamentally challenge the sound source identification in acoustic imaging. This can be solved by sequentially scanning the object of study with a moving array with or without fixed references and requires the sound field to be stationary. In rotating machines, the generated vibroacoustic phenomena during a run up or coast down are non-stationary, and conventional resynchronization techniques cannot be directly implemented. To solve this problem, this study proposes a novel approach called Angular Speed Resynchronization (ASR). The orders from several sequential measurements are first extracted using time domain order tracking techniques, then synchronized into a single data set over which order based beamforming (OBBF) is performed. OBBF is an acoustic imaging technique that aids in the identification of non-stationary tonal sources. Given that the radiated sound field is determined for a given angular speed, the complex envelops in the data set are realigned with respect to speed as if they were obtained from the same measurement with a full array. The proposed method is evaluated using an experimental data set collected from an electric motor. The approach demonstrates the capability of OBBF to accommodate with non-synchronous measurements.

        Orateur: Kottakota Kalasagarreddi
        Resynchronization of tonal acoustic field in multi-pass non-stationary microphone array measurements
      • 11:40
        "EDRIVE#4 - Virtual measurements for exterior vibroacoustic problems using experimental modal models" 20m

        "This paper explores the application of experimentally derived vibroacoustic state-space models in virtual sensor schemes. These models are obtained through Experimental modal analysis (EMA). The paper demonstrates how both in-band and out-of-band contributions can be integrated into a unified state-space model. Additionally, the study illustrates how various measurement types, such as acceleration or acoustic pressures, can be derived from the temporal evolution of the experimental state space model. This allows the state estimator, specifically a Kalman filter, to effectively merge data from both the acoustical and structural domains with the numerical process model. To validate the framework, an industrial test setup is utilized, where the complexity
        prohibits the use of first-principle techniques for modelling. The results indicate that the state estimation outperforms forward simulation for excitation signals that were not initially incorporated into the model."

        Orateur: Staiger Julian
        Virtual measurements for exterior vibroacoustic problems using experimental modal models
      • 12:00
        "EDRIVE#5 - Resynchronization of sequential measurements using the Maximally-Coherent Reference technique" 20m

        Remote sensing of a physical field generated by a small number of sources is limited by the size of the available array of sensors and by the array density. These limitations may lead to asynchronously measuring the field by sequentially moving a small prototype array around its facets, however at the cost of losing the phase information between every array position. Resynchronization using references, i.e. fixed sensors, can be used for phase retrieval, given that these references are of a number that is at least equal to the stochastic dimension of the field, and are not inter-correlated to the degree that hinder them from spanning the source signals' subspace. However, when the number of references largely exceeds the number of sources, the reference cross spectral matrix becomes ill-conditioned leading to the irrelevance of the least squares solution (LSS). Although the truncated singular value decomposition (TSVD) was successfully applied in the literature to solve this problem, its validity is restricted only to the case of scalar noise on the references. In this paper, a solution based on finding a set of virtual references that is maximally correlated with the field measurements, named the Maximally-Coherent Reference Technique (MCR), accompanied with a technique for estimating the number of sources, are proposed for resynchronization. The method is validated both numerically and by using real acoustical data from an e-motor, and its results was compared to those of the LSS and the TSVD when employed for the same purpose.

        Orateur: Albezzawy Muhammad Nabil
        454298.pdf
      • 12:20
        EDRIVE#6 - 20m
    • 10:40 → 12:40
      Survishno 15 / Structural Health Monitoring
      Président de session: Fabien Chevillotte
      • 10:40
        "SHM#1 - From Lab to tail boom: the challenges to develop a SHM system in an industrial context" 20m

        To prevent any accident, a schedule driven maintenance is mandatory to review the many critical parts of a helicopter. As a result, ratio of maintenance man hours to flight hours is often near 4. Despite the time of meticulous review, 14% of the accidents in the aeronautical industry are linked to maintenance issues or clumsiness. The global cost of maintenance is high due to the wasted time of inactivity, the human resources involved and the risk of over-damaging during disassembly. The transition to a condition-based maintenance requires a reliable monitoring system. For many years, new monitoring techniques have been developed to estimate the health of the structures and shorten maintenance time. The structures are monitored in place without any disassembly to limit handling accident. The rear transmission shafts of helicopters are critical and need to be monitored. A promising technique, the Gapped Smoothing Method (GSM), was developed by Ratcliffe and Bagaria in 1998 and allows to localize damages in beam-type structures. This paper points out the potential of the GSM to monitor helicopter shafts without the need a pristine shaft. The method is based on the modal response of the shaft. The modal curvature of the beam, computed from the measured displacements, is compared to the expected theoretical curvature. This value is compute from the beam vibration theory and is approximated with a third order polynomial. Results from this work are in agreement with those presented by Ratcliffe and Bagaria. This method was applied to a more representative structure of a rear drive shaft, differing from the original rectangular section beam. Different tests highlighted the limitations of the GSM which could not be observed in the case presented in the original article. New perspectives of research will be proposed to consider the use of the GSM in an industrial context.

        Orateur: Bottois Paul
        From Lab to tail boom: the challenges to develop a SHM system in an industrial context
      • 11:00
        "SHM#2 - On the time reversal method for fault diagnosis on a beam: some preliminary results" 20m

        Controlling the health state of mechanical and structural systems for early replacements of some damaged elements, is an important issue for prevention of losses in terms of human, environmental and also economical. This control is leading to extension of usability durations of systems. The structural health monitoring covers large categories of techniques ranging from vibration based to ultrasonic or magnetic field and radiographic methods. All of these techniques analyze changes on mechanical/physical system parameters under excitation sources and then relate them to the state of the systems. In the current work, we use the time reversal (TR) method which can be categorized in the ultrasonic techniques for condition monitoring of materials and systems. The method is based on the symmetry of solutions of the wave equation with respect to the time in a lossless medium. One or several beams with different damage conditions are considered and the TR method is applied to the systems for detections and quantifications of damages. The illustration of the exploited method will be based on numerical and finite element modellings of the system accompanied by some experimental results (depending to the progress of the experimentation and availabilities of data) on beams with different damage states.

        Orateur: Huguet Mélissandre
        On the time reversal method for fault diagnosis on a beam: some preliminary results
      • 11:20
        "SHM#3 - Structural Health Monitoring using time reversal techniques in acoustic domains" 20m

        Since it has been theorized by Matthias Fink, time reversal approaches have been applied in various domain such as medicine (lithotripsy), submarine communication and electromagnetic communication. Those applications are using rather high frequency signals and the focalization properties of time reversal techniques. On the other hand, applications of time reversal techniques in structure health monitoring and especially using vibration frequencies lower than 10kHz are scarce. However, time reversal techniques are well suited for monitoring of complex structures using a few transducers. This paper presents developments made in the frame of the ViDeNS project (Vibration DEvice Network on Structures). First, transducers able to act both as sensors and emitters, have been designed. In a second phase, autonomous and lightweight devices were developed to send and acquire vibratory signals. This equipment allowed outdoor experiments on large structures (pipe network, railways, concrete slab ...). The experiments driven tend to prove that time reversal techniques can detect flaws in structures such as added masses and holes with size lower than the limit induced by the low frequency range in wave propagation theory. Theoretical studies have been led in order to explain those phenomena with modal approaches and assess the sensitivity of the technique. This theoretical approach has been used to determine indicators that could be used in the frame of structural health monitoring. Finally knowing potential indicators, statistical studies on the experimental campaigns have been achieved in order to fit flaw detection threshold. As prospective studies, AI algorithms could be applied on both full temporal signals and potential indicators. To conclude, we have shown that very few transducers allow the monitoring of structures larger than 20 meters long with a good sensitivity to centimetric flaws. Further calculation helped us to determine sensitive indicators opening new perspectives for structure health monitoring applications.

        Orateur: Le Gall Thomas
        Structural Health Monitoring using time reversal techniques in acoustic domains
      • 11:40
        "SHM#4 - Operational modal analysis for scour monitoring of bridge piers" 20m

        Scour is the erosive process of removing sediment and soil around bridge piers in rivers, which can pose significant risks to bridge failures. Therefore, it is essential to monitor this phenomenon, which is difficult because of the various structural, geotechnical and hydraulic actions that are involved. An innovative and effective way to do it seems to take advantage of the structural behavior of bridge piers, which is linked to the scour depth. Indeed, modal characteristics of bridge piers, are very sensitive to change of their boundary conditions including those due to the scour. This study presents some results obtained from vibration-based monitoring of bridge piers subject to scour. As most civil engineering structures, the real-time monitoring concerns output-only systems under environmental and traffic loads. This work aims to propose a method for the operational modal analysis (OMA) of a bridge pier. In this study, the OMA is applied to a real case study in a French motorway bridge. Sensors are placed under the bridge deck and on the top of its pier. The method mixes a frequency domain decomposition (FDD) of the power spectral density matrix with algorithms from unsupervised machine learning (clustering). Natural frequencies of the bridge pier for a period of more than one year are identified and their sensitivities to the water level and the temperature are evaluated in a statistical way. Based on changes of natural frequencies, features which are potentially due to scour are detected.

        Orateur: Belmokhtar Mohamed
        Operational modal analysis for scour monitoring of bridge piers
      • 12:00
        "SHM#5 - A study on efficient approaches for modeling Lamb wave propagation in joint metal plates with embedded defect" 20m

        Ultrasonic Guided Wave (UGW), particularly Lamb waves propagating in plate-like media, has gained significant popularity among NDE methods for damage detection and structural health monitoring (SHM) due to its low attenuation and high frequency. Structure boundaries, discontinuities, and damages reflect or diffract Lamb waves during their travel through the structure. The UGW propagation characteristics of a structural waveguide provide information about the structure's health, defects, and locations. An SHM framework can use this function to evaluate a structure by comparing its signal to a database of potential responses and identifying damage through pattern recognition [1]. This research aims to search for and present an efficient modelling approach based on the existing analytical and numerical methods for UGW propagation through damaged and scattering plates and verify the results with a high-precision full 3D elastic finite element model (FEM). In the case study, we investigate the propagation of Lamb wave through a two-piece metal plate made of steel and aluminum with an embedded circular hole. To achieve the best accuracy and efficiency, several analytical [2,3,4], semi-analytical and hybrid [5], and ray tracing approaches [6,7] will be examined, discussed, and selectively applied to the case study to model the transducer-excited ultrasonic field, the steady-state and transient propagation of UGW, and scattering at defects and boundaries. The quantitative validation of the received UGW signals is performed at different sensor locations in the case study. As the outcome of this study, we will provide an efficient and robust approach for modelling ultrasound propagation in damaged and scattering plates, which will be potentially useful in developing a digital twin to structural health monitoring. References (1) https://doi.org/10.1088/1742-6596/305/1/012123 (2) https://doi.org/10.1109/58.393096 (3) https://doi.org/10.1121/10.0013426 (4) https://doi.org/10.3390/app12105078 (5) https://doi.org/10.12989/sss.2014.13.4.587 (6) https://doi.org/10.1016/0963-8695%2892%2990002-X (7) https://doi.org/10.1016/S0041-624X(99)00068-2

        Orateur: Mardanshahi Ali
        A study on efficient approaches for modeling Lamb wave propagation in joint metal plates with embedded defect
      • 12:20
        "SHM#6 - Effects Of Multi-Axis Random Vibration Environments On Fatigue-Life And Durability Predictions" 20m

        One of the main features of a product is its capability to withstand harsh environments that may compromise its durability over time. Therefore, screening laboratory tests are usually performed in the early stages of the product development process to predict the fatigue life beforehand. In this context, random vibration testing has become one of the most frequently employed procedures to ensure the durability and suitability of a product during his working life. Nowadays, the most performed tests are still single-axis shaker tests, due to cost of the equipment and for their reduced complexity compared to a multi-axial tests. However, real working environments almost always present a multi-axial loading condition. As a consequence, neglecting this aspect may lead to large errors in the estimation of the component durability and cause failures that can endanger equipment and people during the product lifetime. International standards propose to excite the unit under test with single-axis excitations along different directions sequentially, in order to mimic a multi-axial vibration environment by means of single-axis testing procedures. In this scenario, this work presents a testing campaign where sequential single-axis testing procedures are studied and compared with multi-axial vibration environments. Tests were run by taking advantage of the multi-axis shaker table available at the University of Ferrara, which is capable of exciting the unit under test along three independent translational degrees of freedom (DOFs). In particular, a cantilever beam is studied in order to assess the fatigue behaviour and the durability of the specimen under three different types of loading: 3 DOFs multi-axis uncorrelated vibration, a first sequential single-axis vibration and a second sequential single-axis vibration with inverted excitation sequence. Finally, the criticalities of the matter are analysed, exposing the inadequacy of single-axis testing to validate components subjected to multiaxial vibration environments.

        Orateur: Proner Enrico
        Effects Of Multi-Axis Random Vibration Environments On Fatigue-Life And Durability Predictions
    • 12:00 → 12:40
      Survishno 16 / Acoustic Treatment
      Président de session: Quentin Leclère (Laboratoire Vibration Acoustique, INSA Lyon)
    • 12:40 → 13:00
      Closing Ceremony
    • 13:00 → 14:20
      Lunch Break 1h 20m
Fonctionne avec Indico
Your browser is out of date!

Update your browser to view this website correctly. Update my browser now

×