Understanding the dispersion of pollutants in atmospheric flow is of vital importance for infrastructure planning and urban development as well as forecasting the spread of air-borne hazards in outdoor and indoor environments. To provide analyses in minimal time, empirical approaches, like Gaussian puff methods and single or multizone models, are presently used. Puff methods assume an idealised outdoor emission plume and neglect interaction with complex topologies. Zone models can provide reasonable predictions of dispersion through buildings, but their applicability is questionable for large spaces, where their homogeneous mixing assumption is invalid.
Three-dimensional computational fluid dynamics simulations can overcome the limitations of empirical dispersion models, however the turn-around times for solving the incompressible Navier-Stokes equations on complex meshes are enormous. As an alternative, a novel parallel adaptive lattice Boltzmann method (LBM) for large eddy simulation (LES) of atmospheric flows is currently under development within our AMROC software framework. In this approach, geometries are represented with automatically generated Cartesian meshes that are non-uniformly refined and de-refined at run time.
In this PhD project, you will work on fine-tuning the LBM-LES approach for atmospheric dispersion prediction and supplement the method with the following features: realistic inflow conditions to model the atmospheric boundary layer, representation of complex release sources, efficient import of complex geometries. In addition, you will develop a version of the parallel LBM-LES code (written in C++) that is tailored especially for execution on sets of graphical processing units (GPUs) using CUDA or OpenACC. Together, these developments will lead to a step-change in atmospheric flow prediction. The new capability will be evaluated in close collaboration with subject matter experts at the Defence Science and Technology Laboratory (Dstl) in Porton Down and you will spend at least 3 months in total on site.
This PhD project requires UK citizenship.
If you wish to discuss any details of the project informally, please contact Dr. Ralf Deiterding, Aerodynamics and Flight Mechanics research group, Email: [email protected], Tel: +44 (0) 2380 59 3384