A COMPUTATIONAL MODEL FOR THE RISE AND DISPERSION OF WIND-BLOWN, BUOYANCY-DRIVEN PLUMES .1. NEUTRALLY STRATIFIED ATMOSPHERE

Citation
Xm. Zhang et Af. Ghoniem, A COMPUTATIONAL MODEL FOR THE RISE AND DISPERSION OF WIND-BLOWN, BUOYANCY-DRIVEN PLUMES .1. NEUTRALLY STRATIFIED ATMOSPHERE, Atmospheric environment. Part A, General topics, 27(15), 1993, pp. 2295-2311
Citations number
50
Categorie Soggetti
Environmental Sciences","Metereology & Atmospheric Sciences
ISSN journal
09601686
Volume
27
Issue
15
Year of publication
1993
Pages
2295 - 2311
Database
ISI
SICI code
0960-1686(1993)27:15<2295:ACMFTR>2.0.ZU;2-V
Abstract
A multi-dimensional computational model for the rise and dispersion of a wind-blown, buoyancy-driven plume in a calm, neutrally stratified a tmosphere is presented. Lagrangian numerical techniques, based on the extension of the vortex method to variable density flows, are used to solve the governing equations. The plume rise trajectory and the dispe rsion of its material in the crosswind plane are predicted. It is foun d that the computed trajectory agrees well with the two-thirds power l aw of a buoyancy-dominated plume, modified to include the effect of th e initial plume size. The effect of small-scale atmospheric turbulence , modeled in terms of eddy viscosity, on the plume trajectory is found to be negligible. For all values of buoyancy Reynolds number, the plu me cross-section exhibits a kidney-shaped pattern, as observed in labo ratory and field experiments. This pattern is due to the formation of two counter-rotating vortices which develop as baroclinically generate d vorticity rolls up on both sides of the plume cross-section. Results show that the plume rise can be described in terms of three distinct stages: a short acceleration stage, a long double-vortex stage, and a breakup stage. The induced velocity field and engulfment are dominated by the two large vortices. The effect of a flat terrain on the plume trajectory and dispersion is found to be very small. The equivalent ra dii of plumes with different initial cross-sectional aspect ratios inc rease at almost the same rate. A large aspect-ratio plume rises slower initially and then catches up with smaller aspect-ratio plumes in the breakup stage. The Boussinesq approximation is found to be valid if t he ratio of the density perturbation to the reference density is less than 0.1.