Non-dissipative large-eddy simulation of wall-bounded transcritical turbulent flows

TL;DR

This paper develops a non-dissipative large-eddy simulation framework for wall-bounded transcritical turbulent flows, evaluating different models and approaches to improve accuracy in predicting heat transfer and flow characteristics.

Contribution

It introduces a stable, non-dissipative LES framework for transcritical flows and assesses various subgrid-scale and wall models against DNS benchmarks.

Findings

Wall-resolved LES slightly deviates from DNS velocity and temperature profiles.

Nusselt number and skin-friction are well captured at walls.

Wall-modeled approaches improve first-order statistics and heat flux predictions.

Abstract

A posteriori analysis based upon a recently proposed non-dissipative large-eddy simulation framework for transcritical wall-bounded turbulence has been carried out. Due to the complexities arisen in such flows, the discretization requires kinetic-energy- and pressure-equilibrium-preservation schemes to yield stable and non-dissipative scale-resolving simulations. On the basis of this framework, the objectives are to (i) compute wall-resolved and wall-modeled large-eddy simulations of a high-pressure transcritical turbulent channel flow, and (ii) assess the thermofluid performance with respect to a direct numerical simulation. In this regard, three different subgrid-scale stress tensor models have been considered, together with models for the unresolved scales of the filtered pressure transport equation and equation of state. In terms of wall-modeling, models based on the "standard law of the wall" and velocity-temperature coupled approaches have been assessed. The results show that for the wall-resolved approaches, the subgrid-scale stress tensors examined slightly deviate from the time-averaged velocity and temperature reference profiles. In terms of bulk performance, it has been found that the Nusselt number and skin-friction coefficient are relatively well captured at the cold and hot walls, respectively. While heat transfer phenomena are fairly well reproduced, particularly the Prandtl and heat flux trends along wall-normal direction. Additionally, the wall-modeled strategies improve the recovery of first-order statistics, although they do not attain the profile in the log-law region. However, they enhance the wall metrics and the heat flux prediction. It is, thus, concluded that dedicated efforts by the research community are needed to improve the prediction accuracy of existing subgrid-scale and wall models for wall-bounded transcritical turbulence.