In a traditional regulatory modeling analysis, permit applicant compiles site operational and emissions data in the front end. Combining with terrain, meteorological, land use, and population data, applicant runs AERMOD and obtains the predicted concentrations of air contaminants at any locations of interest. However, with many current-day fast-paced projects, especially those green-field sites still undergoing design processes, it is typically difficult to have all confirmed information readily available to set up a single all-inclusive model run. Applicant’s desire to permit maximum production and allow various operating scenarios also adds additional complexities to any modeling analyses.
With the advancement of computing powers and technologies, it is now feasible to address all possible operating scenarios together with ambiguous source locations and parameters, to provide a defendable modeling demonstration to support the air permit applications for most complicated projects. By identifying design options that will work with permitting much earlier in project planning, it will greatly benefit the permitting process to align with demanding construction schedules.
Multiple modeling scenarios call for the needs of automated creation of AERMOD run files, especially when number of scenarios can be huge. Last year, we introduced a few customizable tools written in Python programming language for pre-processing input data, creating AERMOD run files, and summarizing results from hundreds or thousands of model runs. This presentation will further introduce the advancement of these efforts. Additionally, we will explore the possibilities of introducing AI and machine learning to assist with even more streamlined AERMOD model setup, scenario analysis, and results reporting.
.jpg)
