 |
Analysing discrete and continuous spectrum and dimension reduction for thermal fields
1 : LAboratoire de Modélisation Pluridisciplinaire et Simulations
Université de Perpignan Via Domitia
2 : Institut de Mathématiques de Toulouse UMR5219
Université Toulouse Capitole, Institut National des Sciences Appliquées - Toulouse, Université Toulouse - Jean Jaurès, Université Toulouse III - Paul Sabatier, Centre National de la Recherche Scientifique
3 : Laboratoire Jean Alexandre Dieudonné
* : Auteur correspondant
Université Nice Sophia Antipolis (1965 - 2019), Centre National de la Recherche Scientifique, Université Côte d'Azur
The aim of this work is to compare three methods of data reduction in the
context of heat transfer. This follows the well-known practice of observing unsteady
phenomena according to space or time energetic arguments. Especially, the present study
focuses on the efficiency of the Proper Orthogonal Decomposition (POD), Spectral Proper
Orthogonal Decomposition (SPOD) and Principal Components Analysis in the Frequency
domain (FPCA). In several previous works, both POD and SPOD have been proposed in
the context of fluid mechanics while FPCA is been newly applied to this domain. Thus, in
this work we provide a discussion on the contribution of the FPCA to deal with multiscale physics.
