The effect of varying viscosity in turbulent channel flow

TL;DR

This study investigates how spatially varying viscosity affects turbulence in channel flow using DNS, revealing delays in turbulence structure changes and providing a new benchmark case for such flows.

Contribution

It introduces a novel DNS benchmark for turbulent channel flow with spatially varying viscosity and analyzes turbulence behavior under these conditions.

Findings

Variation in Reynolds number does not cause immediate change in turbulent structures.

A delay is observed in wall shear and friction Reynolds number response.

Results are compared with existing correlations for validation.

Abstract

In this article we examine channel flow subject to spatially varying viscosity in the streamwise direction. The Reynolds number is imposed locally with three different ramps. The setup is reminiscent of transient channel flow, but with a space dependent viscosity rather than a time dependent viscosity. It is also relevant to various applications in nuclear engineering and in particular in test reactors, where the viscosity changes significantly in the streamwise direction, and there is a severe lack of Direct Numerical Simulation (DNS) data to benchmark turbulence models in these conditions. As part of this work we set up a novel benchmark case: the channel is extended in the stream-wise direction up to 20p. The viscosity is kept constant in the first 4p region. This inlet region is used as a cyclic region to obtain a fully developed flow profile at the beginning of the ramping region. In the ramping region the Reynolds number is linearly increased along the channel. The flow is homogenous in the spanwise direction, while it is nonhomogenous in the stream-wise and wall-normal direction. We perform here Direct Numerical Simulation (DNS) with Nek5000, a spectral-element computational fluid dynamics (CFD) code developed at Argonne National Laboratory. In this study, specific focus is given to the investigation of turbulence properties and structures in the near-wall region along the flow direction. Turbulent statistics are collected and investigated. Similarly to transient channel flow, the results show that a variation in the Reynolds across a channel does not cause an immediate change in the size of turbulent structures in the ramp region and a delay is in fact observed in both wall shear and friction Reynolds number. The results from the present study are compared with a correlation available in the literature for the friction velocity and as a function of the Reynolds number.