Abstr:Unsteady near-field plume

Free Flows

Underlying Flow Regime 1-07

Abstract

This document examines the Underlying Flow Regime of a turbulent, non-reacting plume flowing into a quiescent and unstratified environment. Only the unsteady flow behaviour in the near-field is considered, within the region extending up to around five diameters downstream from the source. Figure 1 shows a CFD simulation of such a plume. The fully-developed, far-field behaviour of turbulent plumes is examined in a related UFR (UFR 1-06).

Turbulent buoyant plumes are a feature of many important scientific and engineering applications including flows generated by fires, smokestacks, cooling towers, and large geothermal events, such as volcanoes. The source of the buoyancy may be provided by temperature differences in the fluid or can be related to two fluids of different density mixing together.

Medium to large scale plumes are characterised by the repetitive shedding of coherent vortical structures at a well-defined frequency, a phenomenon known as “puffing”. A number of empirical correlations for the puffing frequency of plumes have been developed, based on the Richardson number, which are described in this UFR.

A brief review is provided of near-field plume experiments and CFD studies. Three CFD studies are examined in greater detail, those by DesJardin et al. [1], Tieszen et al. [2] and Xin [3]. These have all simulated the recent helium plume experiments of O‘Hern et al. [4]. The studies have each used slightly different numerical modelling approaches, although all are based on Large-Eddy Simulation (LES). The plume experiments O‘Hern et al. [4] are particularly well-suited for model evaluation as they involved simultaneous measurement of velocities and mass fraction, allowing both Reynolds and Favre-averaged quantities to be determined.

Based on the three CFD studies, best practice advice is provided for industrial CFD practitioners on some key modelling issues involved in simulating unsteady buoyant plumes.

LES is less mature than RANS turbulence modelling and a number of uncertainties remain when using LES for industrial flow predictions, such as the appropriate grid resolution and the choice of numerical schemes. Some guidance is given on these issues and suggestions are provided for where future work could contribute to providing improved quality and trust in the simulation of plumes.

Figure 1 CFD Simulation of the near-field of a buoyant helium plume. The vorticity isosurface is coloured with the magnitude of the gravitational torque. From DesJardin et al. [1]

Acknowledgements

The author would like to thank Sheldon Tieszen (Sandia), Paul DesJardin (University at Buffalo, NY), Yibing Xin (FM Global), Wolfgang Rodi (Karlsruhe) and Chris Lea (Lea CFD) for their help and assistance in preparing this review.