ESTIMATION OF VIRTUAL FLARE STACK PARAMETERS FOR A TYPICAL REFINERY FLARE

Abstract

Flares are normal and inescapable part of refinery and petrochemical projects. It is an emergency safety equipment aimed at converting flammable, toxic and corrosive vapors into environmentally acceptable discharges. The two main problems associated with the phenomena are thermal radiation and emissions from flare stack. The study of thermal radiation is necessary to protect the working personnels and expensive equipments. The operation and emissions from flares often create problems (e.g. luminosity, noise, smoke, NOx, SOx) for the surrounding communities. Thus in the present work an effort has been made to predict the safe distance from a flare for the working personnels and ground level concentration (GLC) profiles for SOx and NOx pollutants. Gibb's free energy minimization technique is used to predict the compositions of all possible combustion products. Since equilibrium composition and adiabatic flame temperature are interdependent, so the developed software package gives equilibrium composition and adiabatic flame temperature simultaneously using minimization of Gibb's free energy of mixture of gases for equilibrium composition and enthalpy balance for adiabatic flame temperature estimation. In the process of flaring combustion occurs outside the flare stack. Therefore the actual flare stack characteristics like stack diameter, height, exit velocity, temperature of flue gas etc. can not be used directly for air quality modeling. Therefore to predict the virtual flare parameters for the purpose of dispersion modeling, Thorton model has been iii chosen.Thornton model assumes flame as a frustum of cone and has more realistic representation of flame geometry, suit to predict the virtual flare parameters as compared to other models. Safe distance from flare has been predicted using surface emissive power, transmissivity of air and view factor of the flame surface geometry at the receptor height concept. The 3-D concentration profiles and concentration Vs downwind distance plots have been presented to show vividly the change of concentration around stack with down wind and cross wind distances. The plots show clearly the high concentration zones and sharp concentration profiles which vary with distance. The effect of timing in the 24 hour day on the ambient atmospheric conditions and hence the pollutant GLC's has also been studied into four parts — Morning, Day-time, Evening and Night-time conditions. Complete algorithms and associated mathematical correlations for computational purposes have been developed for efficient programming. C++ codes have been developed in the present work and tested on Pentium-III computers.