(199e) Response Surface Modeling of Flare Performance and Characterization of the Incipient Smoke Point

Current EPA regulations (40CFR60.18) require smokeless flaring, which motivates flare operators to over-steam or over-air to suppress smoke at the expense of combustion efficiency (CE). It is also well known that incipient smoke point (ISP) is a good indicator for good combustion, but the phenomena is neither well understood nor scientifically defined. Further, many factors affect soot emission and unburned/ produced volatile organic compounds (VOCs) emissions. This leads to a question of how to operate the flares in order to achieve the optimal over-all environmental performance. In this study, the flare test data from 1983 to 2014 with soot emission, destruction and removal efficiency (DRE), and CE data (including 1983/1984 EPA, 2010 TCEQ, 2009/2010 Marathon Detroit/Texas City, and 2014 Carleton University) were analyzed.  All CE data were corrected for soot emissions. Easy to use response surface models were developed to predict DRE/CE and soot yield for steam and air assisted flares burning propylene, propane, natural gas, methane and ethylene. Soot yield is expressed in pounds per million British Thermal Unit (lb/MMBTU). The input variables include vent gas combustibility (net heating value (NHV) /lower flammability limit(LFL)/combustible concentration(C)), vent gas species (olefins, alkynes, aromatics, hydrogen, nitrogen, hydrogen to carbon ratio, carbon number), and operating parameters (steam/air assists, stack diameter, and exit velocity). Vent gas NHV/LFL/C can be combined with steam and air assists at combustion zone as NHV_CZ/LFL_CZ/C_CZ to simplify the models for easy visualization and simple control. The combination of stack diameter and exit velocity is then used to obtain the Richardson Number (gd/v^2).  NHV_CZ/LFL_CZ/ C_CZcan be redefined to include a portion of upper steam used. To estimate the desirable range of controllable operating variables, inverse response models were developed in which the operating parameters were expressed as a function of performance variables. The goal is to estimate the desirable operating set point with a high CE (>96.5%) and a low soot yield (0.01lb/MMBTU) simultaneously. Test cases were used in characterization of ISP with plume opacity and soot yield. Plume opacity was estimated using the Beer – Lambert law in terms of soot concentration, flame diameter, and black carbon absorptivity.  The average opacity was found to be 10% and 40% for steam and air assisted flares at the ISP. Linear relations were established between soot yield and opacity normalized to flame diameter with goodness of fit 0.98 and 0.99 for air and steam assisted flares. These relations can be used to link the visible emission scale = 5 at the ISP (in Marathon's 2009/2010 data) to opacity = 10% for steam-assisted flares and opacity = 40% for air-assisted flares. As a result, visible emission rating data can be converted to soot yields in the case of Marathon's flare tests data.

 Keywords: Propylene, Propane, Methane, Ethylene, Natural gas, soot emissions, Flare performance, Richardson number, Lower flammability limit, Net heating value at combustion zone, response surface models.