Large eddy simulations of a confined rectangular jet

Date
2007-01-01
Authors
Gokarn, Anup
Major Professor
Advisor
Francine Battaglia
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Mechanical Engineering
Organizational Unit
Journal Issue
Series
Department
Mechanical Engineering
Abstract

Turbulent flows have the ability to transport momentum and mix species at a higher rate than molecular diffusion alone, which is critical in reacting flows. Of importance to this research is the mixing of liquid-phase high Schmidt number flows for applications in the chemical process industry. Computational fluid dynamics (CFD) has the potential to be used as a tool to improve reactor designs and examine the mixing characteristics but requires extensive validation of the computational models. This research presents detailed analysis and validation of a nonreacting turbulent flow for a confined rectangular jet with a co-flowing fluid using CFD. Large eddy simulations (LES) for the incompressible Filtered Navier-Stokes equations are performed on a partially staggered finite difference grid. A second-order central difference scheme and sixth-order compact scheme are employed for the spatial derivatives. A third-order low storage Runge-Kutta method is used for the temporal derivatives. To reduce computational memory and time requirements, message passing interface (MPI) is implemented and an efficient parallel linear equation solver (Aztec) is utilized for solving the elliptical pressure Poisson equation. Solutions for the momentum and non-reacting scalar transport are obtained for a Reynolds number of 20,000 based on the average velocity at inlet and hydraulic diameter. Validation is performed for the LES by comparing one-point and two-point statistics with particle image velocimetry data for the velocity field and planar laser induced fluorescence measurements for the scalar concentration. Such detailed validation with experiments is performed for the very first time. The effect of different parameters such as grid resolution, numerical schemes and subgrid models on the numerical solution are studied. For the scalar transport, numerical schemes that preserve boundedness are tested and implemented. Overall, the LES compared very well with the experiments and recommendations are made to extend the LES work toward reacting flows.

Comments
Description
Keywords
Citation
Source