CFD-Aided Building Downwash Parameterization and Evaluation Using Wind Tunnel and Field Measurement Databases

Siting hydrocarbon-fueled distributed generation (DG) units, located near population centers and having shorter stacks than central station power plants, requires careful consideration of the impact of the building downwash effect. AERMOD adopted the Plume Rise Model Enhancement (PRIME) and its preprocessor, Building Profile Input Program (BPIP) to account for the downwash effect. Recent studies showed some improvements of BPIP-PRIME using computational fluid dynamics (CFD) and wind tunnel data. However, there are modeling challenges under oblique wind directions, which means the wind direction is not perpendicular to the windward wall of a box-shaped building. In this study, we conducted Embedded Large Eddy Simulations (ELES) of the flow field and the concentration field around a series of box-shaped buildings with a relatively short stack (no more than twice of the building height). The ELES results were evaluated using the USEPA wind tunnel measurement. A training data set was built based on our ELES simulations. We proposed a building downwash parameterization method, named the Mixture Model, which is the aggregation of different Gaussian-based distributions. The new method was evaluated by using the USEPA wind tunnel measurement and the field measurement data of the gas tracer study of roof-vent effluent diffusion at the Millstone Nuclear Power Station. Promising performance made this work capable of improving the building downwash parameterization.