A Validation of the Emissions Model from EPA's Metal Finishing Facility Risk Screening Tool - MFFRST

A paper presented at the 2000 AESF/EPA conference provided a comparison of MFFRST-predicted emissions from stacks with those reported in the Toxics Release Inventory (TRI) data base. While predictions of some of the organic solvent emissions (TCE, e.g.) were similar to reported quantities, there was apparently a large discrepancy between predicted and reported emissions of metals. MFFRST was predicting emissions two orders of magnitude and more lower than were reported in TRI. This paper continues the testing of the emissions model. Twelve stack test reports from chrome plating facilities in California were used in a model validation exercise. All pertinent parameters (tank size, initial concentrations, current density, ventilation rates and others) were extracted and values assigned. Model predictions of chromium (Cr +6 and total chromium) emissions were compared against measured emissions. Initial testing showed an r 2 correlation of 0.59, between predicted and observed emissions. The model mostly underpredicted emissions by an average of about a factor of 6. This is much less than the underprediction of MFFRST compared to TRI emissions. Causes for this model underprediction will be explored and described in the paper.