4A.1
Evaluation of PRIME and ISC3 versus wind tunnel observations
R. L. Petersen, Cermak Peterka Petersen, Inc., Fort Collins, CO; and B. C. Cochran and J. J. Carter
At present, the ISC3 model is used for estimating concentration levels when building wakes affect pollutant dispersion. A new model developed by EPRI, referred to as PRIME, is in the process of being recommended as an "EPA approved model" for building downwash situations. This model has been designed to simulate stacks located at various distances upwind or downwind of structures and has improved dispersion and plume rise algorithms. The purpose of this study was to evaluate the relative improvement in concentration predictions for the PRIME model versus the ISC3 model for various simple building configurations. To meet this objective, a series of wind tunnel tests were conducted using individual buildings with height to width to depth ratios of 1:2:1. This is the building shape used by EPA to develop the downwash algorithms in ISC3 and used to conduct Equivalent Building Dimension evaluations. A cube shaped building was also assessed. Ground level concentrations due to an elevated point source release were obtained for various stack height to building height ratios with the building at various orientations and heights. Plots of model predictions versus wind tunnel observations were then developed and the normalized bias and root-mean-square error between ISC3 and PRIME model predictions and wind tunnel observations were computed. Two methods were used to determine the ISC3 and PRIME model inputs based on the wind tunnel simulation. The statistical evaluation showed that ISC3 (using a typical wind tunnel scaling method) provided the best estimates for the 1:2:1 building shape while PRIME provided the best estimates for cubical shape. Using exact wind tunnel scaling, PRIME model performance did not change significantly while ISC3 model performance changed to the point where PRIME provided the best agreement for all configurations. Furthermore, the evaluation pointed out a potential problem in that PRIME tended to underpredict concentrations close to the source where the maximum concentrations occurred due to building downwash effects.
Session 4A, Dispersion about buildings and obstacles (Parallel with Session 4B)
Monday, 10 January 2000, 1:30 PM-3:00 PM
Next paper