Comparison of emission models with computational fluid dynamic simulation and a proposed improved model.

Understanding source behavior is important in controlling exposure to airborne contaminants. Industrial hygienists are often asked to infer emission information from room concentration data. This is not easily done, but models that make simplifying assumptions regarding contaminant transport are frequently used. The errors resulting from these assumptions are not yet well understood. This study compares emission estimates from the single-zone completely mixed (CM-1), two-zone completely mixed (CM-2), and uniform diffusivity (UD) models with the emissions set as boundary conditions in computational fluid dynamic (CFD) simulations of a workplace. The room airflow and concentration fields were computed using Fluent 4. These numerical experiments were factorial combinations of three source locations, five receptor locations, three dilution airflow rates, and two generation rate profiles, constant and time-varying. The aim was to compute plausible concentration fields, not to simulate exactly the processes in a real workroom. Thus, error is defined here as the difference between model and CFD predictions. For the steady-state case the UD model had the lowest error. When the source near-field contained the breathing zone receptor, the CM-2 model was applied. Then, in decreasing agreement with CFD were UD, CM-2, and CM-1. Averaging over all source and receptor locations (CM-2 applied for only one), in decreasing order of agreement with CFD were UD, CM-1, and CM-2. Source and receptor location had large effects on emission estimates using the CM-1 model and some effect using the UD model. A location-specific mixing factor (location factor) derived from steady-state concentration gradients was used to build a more accurate time-dependent emission model, CM-L. Total mass emitted from a time-varying source was modeled most accurately by CM-L, followed by CM-1 and CM-2.