Reduced models identification from experimental modal analysis of non-self adjoint systems: rotordynamics, active control and vibroacoustics applications - Université de Technologie de Belfort-Montbeliard Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

Reduced models identification from experimental modal analysis of non-self adjoint systems: rotordynamics, active control and vibroacoustics applications

Résumé

In the context of experimental modal analysis of non-self adjoint problems, like rotordynamics, vibroacoustics or active control applications, the nonsymmetry of the system induces specificities that must be considered for proper use of identification techniques. In this paper, the particularities of this kind of problem are addressed in order to be able to efficiently identify the dynamic behaviour. There are many aspects induced by the non-self adjoint character of the behaviour, some of them have been studied in the past years, while some remain unclear. This paper addresses some of these points. The first matter which is detailed is related to the ability of the technique to identify both right and left eigenvectors. The classical LSCF technique is adapted in order to reach this objective. A few adaptations are required, but the extension is quite natural and the efficiency of the technique is not degraded in the context of non self-adjoint problems. The second point is associated to the regularization of inverse problem for matrices identification using the complex eigenvectors. This inverse procedure, which is one of the ways allowing the damping matrix identification, is known to be very sensitive to noise. The technique of properness enforcement, already available in the context of structural dynamics, has been extended to non-self adjoint in order to regularize the problem. The basic idea is to slightly modify the identified complex vectors, in order that they verify the so-called properness condition. A low-cost specific iterative procedure is proposed to reach this objective. A numerical test-case is then presented on a rotordynamics application, and some experimental results are presented on a structural active control application, in order to show the ability of the approach to identify reduced models from experimental modal analysis. In particular, some considerations about proper damping identification are exhibited. Finally, some issues about vibroacoustics applications are considered: in this case, a direct relationship between left and right eigenvectors is available. The technique has then to be slightly adapted in order to take into account the dependency of the variables used in a constrained optimization equation. This leads to a non linear matrix problem, which can be difficult to solve. Some approximate techniques are exhibited and applied on an experimental test-case in order to illustrate their efficiency.
Fichier principal
Vignette du fichier
Ouisse2010.pdf (254.69 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03154010 , version 1 (26-02-2021)

Licence

Paternité

Identifiants

  • HAL Id : hal-03154010 , version 1

Citer

Morvan Ouisse, Emmanuel Foltête. Reduced models identification from experimental modal analysis of non-self adjoint systems: rotordynamics, active control and vibroacoustics applications. ISMA2010 Conference, 2010, Leuven, Belgium. ⟨hal-03154010⟩
28 Consultations
81 Téléchargements

Partager

Gmail Facebook X LinkedIn More