Singular-Value-Decomposition- and Dynamic-Mode-Decomposition-Based Reconstruction of Cylinder Wake
This work evaluates the singular value decomposition (SVD) and the dynamic mode decomposition (DMD) for modal decomposition of coupled velocity and thermal fields in the wake of a heated circular cylinder at [Formula: see text] and [Formula: see text]. Both velocity and thermal fields are analyzed by unified transient simulations of flow past a heated circular cylinder performed using the buoyantBoussinesqPimpleFoam solver of OpenFOAM version 2412. SVD and DMD are performed on 51 snapshots of velocity and temperature after transients, and the authors get six modes for velocity and 12 modes for temperature with a 99% cumulative energy truncation criterion by SVD. The relative Frobenius norm of the reconstruction error decreases from 0.0807 to 0.0067 for 1 to 6 modes of velocity and from 0.0022 to 0.0002 for 1 to 12 modes of temperature with SVD giving better performance than DMD. DMD eigenvalue analysis reveals the fundamental shedding frequency at [Formula: see text] and a first harmonic in the velocity field absent in the temperature spectrum, providing quantitative evidence, for [Formula: see text] (water), of Prandtl-number-induced harmonic suppression, in which the harmonic mode present at [Formula: see text] in the velocity field is absent from the temperature spectrum.
Authors
- Anuj Shukla (ORCID: https://orcid.org/0000-0002-3603-2317)
- Sourabh Luthra
Publication Details
- Journal
- Journal of Thermophysics and Heat Transfer
- Published
- 2026-11-15
- DOI
- https://doi.org/10.2514/1.t7461
- Primary Topic
- Model Reduction and Neural Networks
- Type
- article
- Field-Weighted Citation Impact
- 0.00