Modeling CO and Estimating Combustion Efficiency with Flamelet-Generated Manifolds in Flares

Paper from the AFRC 2014 conference titled Modeling CO and Estimating Combustion Efficiency with Flamelet-Generated Manifolds in Flares by A. Jatale.

The goal of this study is to use CFD simulations to accurately estimate the combustion efficiency and CO emission for flares. Combustion engineers face serious challenge in accurate prediction of efficiency and pollutant emissions from combustion systems. Flares, is one such technology which carries additional obstacles in measuring the experimental data. A wind tunnel flare setup was simulated using a Flamelet Generated Manifold (FGM) model in ANSYS R15. This model does not make any assumption on flame structure and the realized trajectories are approximated by scalar evaluation similar to laminar flame. Therefore, it can be successfully used to model ignition, slow chemistry, and quenching that are important in estimation of CO and other pollutant concentrations. Also in this model, 1D premixed flamelets are solved in reaction-progress space, giving better estimation for combustion efficiencies. The simulation uses LES turbulent model and were performed on an optimum size mesh. The experimental data from flare tests performed at the CANMET wind tunnel flare facility1, Ottawa, Canada, was used to compare the simulation results. The simulation results not only show a good match with combustion efficiency data, also the prediction of CO concentration is comparable with experimental data. The effect of crosswind velocity on efficiencies can also be observed in the simulation results.

1 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Modeling CO and Estimating Combustion Efficiency with Flamelet-Generated Manifolds in Flares Anchal Jatale ANSYS Inc. AFRC 2014 INDUSTRIAL COMBUSTION SYMPOSIUM Sept 7-10, 2014 Stefano Orsino ANSYS Inc. Philip Smith University of Utah 2 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • Important control practice for destroying unwanted hydrocarbons and other combustible material released in the atmosphere • Complete Combustion is the goal • CH4 etc. are 20-25 times more effective in trapping heat compare to CO2 • Combustion efficiency is a most common measure of effectiveness of flaring systems • Operating conditions have significant effect on flare flame stability, efficiency and emissions • Important to predict this effect to control emissions ( stringent regulation/taxes) Motivation Most important thing to ensure a proper mixing of fuel and air 3 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • No single or few parameters characterizes the combustion behavior of flare flames. - Complexity of nonlinear mixing, reaction and heat transfer - Motivates the need to accurately measure it from operating flares so that the effect of different designs and operations can be quantified. Why little known about Combustion Efficiency ? 4 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Measurement of Combustion Efficiency • Difficult to measure as needs simultaneous knowledge of the composition and velocity to obtain mass flux surrounding the flare - Measurable combustion species gets diluted in the surrounding atmosphere - Effect of high temperature and radiant heat on test equipment - Irregular nature of flames due to external winds and intrinsic turbulence - Lack of suitable sampling locations due to flare/flame height especially for elevated flares 5 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential CFD Simulations • Industrial flare simulation can provide important information on combustion efficiency and pollutant emissions, • It also provides operational parameter sensitivities for design or operation that can not be measured. • It is cost effective and one can test loads of what if scenarios which are not possible to measure in reality. 6 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • Difficult to compute accurately through traditional computational fluid dynamics (CFD) simulation tools that are based on Reynolds-Averaged Navier-Stokes (RANS) approaches codes - Wide range in length and timescales in the mixing and reacting processes - The large-scale mixing due to vortical coherent structures in these flames is not readily reduced to steady-state CFD calculations with RANS - By time-averaging the equations, unsteady information such as instantaneous mixing and flame shape cannot be captured - Capturing ignition and extinction phenomena CFD Simulations 7 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • Various turbulence models which will be able to resolves all the important characteristics of the flare flames - Large Eddy Simulation (LES) - Detached Eddy Simulations (DES): Unsteady RANS + LES - Scale- Adaptive Simulation (SAS): Improved Unsteady RANS - Embedded Large Eddy Simulation ( ELES) • Various Combustion models are available to capture the chemistry properly. ( Non premixed , premixed and partially premixed models) Fluent as CFD tool 8 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • Fluent's premixed FGM assumes that the thermo-chemical trajectories in a turbulent flame are similar to the trajectories in a laminar flame ‐ No assumption of flame structure ‐ Can model premixed flames, flame lift, ignition, quenching and extinction • Diffusion FGM model is present - For predominantly diffusion turbulent flames - Calculate detailed chemistry a-priori in opposed flow laminar diffusion flames and parameterize by f and c Combustion Model: Diffusion Flamelet Generated Manifold (Diffusion FGM) 9 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Combustion Model: Diffusion Flamelet Generated Manifold (Diffusion FGM) • The traditional Diffusion FLAMELET based Partially Premixed Model defaults to an Equilibrium solution for a pure or pre-dominantly premixed case. Hence cannot predict super-equilibrium species like CO and OH accurately. • The Diffusion FGM model, solves the 1D premixed system (with full convection & diffusion) in the Progress Variable Space and constructs a species-temperature manifold based on the Premixed FLAMELET solution parameterized by the progress variable. • Because equilibrium assumption is not made, it is expected that species like CO and OH can be predicted more accurately. 10 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Diffusion FGM example • LES of Sandia flame D 11 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Cabra CH4 Lifted Flame Details are available at: http://www.me.berkeley.edu/c al/vcb/data/VCMAData.html Proceedings of ASME 2013 Gas Turbine India Conference December 5-6, 2013, Bangalore, Karnataka, India SLFM PFGM Flame lift-off • PFGM Captures the flame lift-off - Measured lift-off distance = 35Dn - Predicted lift-off distance = 33Dn 12 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Case Study: Wind Tunnel Flare Cross Wind Flue Gas 13 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Mesh ~8.5 million cells 14 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Turbulence Model: LES Subgrid-scale Model: WALE Combustion Model: Partially Premixed diffusion FGM Turbulence-Chemistry Interaction: Finite-Rate Variance Method: Algebraic Models used 15 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Fuel Stream @300 K Flow Rate: 0.00556 Kg/s Mole % : CH4 - 95.35 CO2 - 0. 62 C2H6 - 2.1 C3H8 - 0.13 N2 - 1.8 Oxidizer @ 287 K Mole % : N2- 78.9 O2- 21.1 Inlet and Boundary Conditions 16 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential • First Establish mixture fraction variance field by running the case for 10 time steps ( time-step size=0.05 s) • Whole domain was patched with progress variable =1 • The simulation was ran till 2 second of run time. This give enough time to get a developed flow • Time step size was reduced to 0.0005 sec and simulation was run for 2 more second of run time. • Data was sampled from t=4 sec to t= 8 sec Simulation Strategy 17 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Flare at different crosswind velocity 6 m/s 12 m/s Buoyancy effects after certain distance at lower crosswind velocity 18 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Case Study: Wind Tunnel Flare A part of the Temperature profile at lower crosswind in the wind tunnel 19 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential CFD Results Vs Experimental Values 20 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Combustion Efficiency With increase in crosswind velocity efficiency goes down 21 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Effect of Turbulence Models on Combustion Efficiency Unsteady RANS predict almost complete combustion 22 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Flare structure wrt to turbulence model LES U-RANS SAS 23 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential CO Mass Percentage 24 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Effect of turbulence model on other species CH4 CO2 O2 25 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential Diffusion Flamelet Generated Manifold combustion model along with LES turbulent model provides a good combination to tackle the difficult combustion systems like flares Summary 26 © 2014 ANSYS, Inc. August 22, 2014 ANSYS Confidential