(515e) Proactive Emission Source Detection and Evaluation with Air Quality Monitoring Network

Chemical facilities, where toxic or hazardous chemicals are manufactured and housed, present a high potential of dangerous accidents such as explosion or toxic gas leakage.  These are usually caused by various factors of equipment failure, operation, or extreme conditions such as earthquake, tsunami, hurricane, transportation accidents, or terrorist attack.  In certain accidental emission events, the emission source or severity are indeterminate and immeasurable due to event complexity or the simple fact that the accidental areas are too dangerous to approach.  Lack of necessary information on emission source and severity may lead to insufficient decision makings that keep emergency rescue from timely and effective responses.  Therefore the inverse determination of the emission source and evaluation of its severity is of great significance to all stake holders.   

In this study, a methodology is developed for proactive emission source detection and evaluation in emergent situations.  All the location of emission points have been assumed to be known because chemical facility positions are usually fixed and GPS technology is currently available for positioning chemical transportation vehicles.  This method has included three stages.  In the first stage of modeling constructing, the monitor data is collected from existent AQMN (air quality monitoring network).  The changing trend of monitor result will be correlated with the normal emission condition under meteorological conditions.  In the second stage of modeling optimization, the dispersion model and least square approximation are used for abnormal AQMN measurements diagnosis, which will help determine the possible emission positions.  In the third stage of emission source evaluation, the air pollutant dispersion model is further refined with available abnormal emission measurements.  The fine-tuned model is used to provide the dynamic prediction of future consequence to current or potential impacted areas so that timely response plan can be proactively taken.

The whole procedure is to setup a mechanism with risk early-warning and rapid response functions for dangerous chemical emissions from chemical plants, transportation vehicles, or accidental chemical releases.  The developed methodology involves a complex inverse optimization problem.  An effective solving strategy is also introduced.