Details

Autor(en) / Beteiligte

• Reynaud, J.
• Weiss, P.-E.
• Deck, S.
• Guillen, P.

Titel

A comprehensive framework for high fidelity computations of two-species compressible turbulent flows

Ist Teil von

• Journal of computational physics, 2022-08, Vol.462, p.111222, Article 111222

Ort / Verlag

Cambridge: Elsevier Inc

Erscheinungsjahr

2022

Link zum Volltext

Quelle

Elsevier ScienceDirect Journals Complete

Beschreibungen/Notizen

• •Proposal of a strategy for scale-resolving simulations of bi-species turbulent flows.•Combination of ZDES and an original low-dissipative version of the AUSM scheme.•Assessment of the numerical framework on a coaxial Air/Argon jet configuration.•Application of a white noise/dynamic forcing method to a curvilinear bi-species flow with ZDES mode 3. A strategy for scale-resolving simulations of bi-species turbulent flows is presented. It relies on a combination of Zonal Detached Eddy Simulation (ZDES) and an original low-dissipative version of the AUSM scheme that adapts its dissipation to capture flow discontinuities while ensuring a low numerical dissipation level in resolved turbulence regions. The corresponding discretized equations are thoroughly detailed. This comprehensive numerical framework is evaluated for the coaxial Air/Argon jet configuration investigated experimentally by Clifton and Cutler [1]. A robust method based on a combination of white noise to generate velocity fluctuations at the inlet of attached turbulent boundary layers together with a dynamic forcing method is applied for the first time in a curvilinear bi-species flow framework with ZDES mode 3. It is shown that the taking into account of resolved turbulence in the incoming attached boundary layers provided by this approach improves the prediction of the mixing process in the early stages of the mixing layer. A good agreement with experiment is also observed with a RANS description of attached boundary layers as permitted by ZDES mode 2 (2020) [2] where the switch between RANS and LES zones is set dynamically by the model itself.