351659 Air Quality Impact of the Startup of a Single Olefin Plant

Wednesday, April 2, 2014: 11:15 AM
Newberry (Hilton New Orleans Riverside)
Jian Zhang, Ziyuan Wang, Qiang Xu and Thomas C. Ho, Dan F. Smith Department of Chemical Engineering, Lamar University, Beaumont, TX

Ground-level ozone is one of six common pollutants defined by US EPA (Environmental Protection Agency).  It can cause health effects, including lung diseases.  HRVOC (Highly Reactive Volatile Organic Compounds) emissions from industries are highly responsible for ground-level ozone pollution, especially around industrial district.  Flare emissions from MSS (maintenance, startup and shutdown) & events are major sources of HRVOC emissions.

Most of air quality modeling treats industrial flares as constant emissions, which means the emissions keep constant for the whole scenario days, and the total amount of emissions equal to the total emissions reported to environmental authorities.  However, real flare emissions from MSS and events are time-various.  At certain time, flare emissions could be very huge, and at other time, flare emissions could be very small.  During emission peaks, the ozone formation could be much higher than a flat emission curve.

In this paper, time-various flare emissions of startup procedures are considered for air quality modeling.  Emission data are obtained from dynamic simulation of a typical ethylene plant.  The dynamic simulation results were validated with real plant data and emission data of startup events.  The time-various flare emission data were added into Houston-Galveston-Brazoria (HGB) 2006 episode base case modeling as a point emission source.  Three startup scenarios were simulated, which include a baseline scenario and two flare reduction scenarios.  For each scenario, 120 cases were built to simulate different starting time covering the whole episode days.

Simulation results show that for the baseline scenario, the maximum 8-hour average ozone increment could be 15 ppb at certain starting time.  It means starting a single plant could temporary increase the regional ozone concentration by 15 ppb, which is 20% of current ozone standard (75 ppb).  For the two flare reduction scenarios, the maximum 8-hour average ozone increment could also reach 10 ppb.  Simulation results also show that ozone increment is high when starting time is before 8 am.  Ozone increment is low when starting time is after 12 pm.  For each scenario, air quality simulation gives least impact period, which means ozone impact could be low when starting time is within that period.  For the baseline scenario, the least impact period is only 2 hours.  For the two flare reduction scenarios, the least impact periods are 4 hours and 8 hours.

In conclusion, dynamic simulation and air quality simulation show that the flare emissions from MSS and events are highly time-various, and may not be treated as flat emissions.  Air quality simulations also give least impact periods of startup time, which could help chemical plants to arrange their startup schedule to reduce ozone formation.

Extended Abstract: File Not Uploaded