Both LES and k-e turbulence-treatment in CFD-approaches for creation of the database on flare-stacks flame-behaviour

There is an emerging importance to understand and to "map" the reacting complex-flows of the (larges-cale) flare-fires. In this study, the influence of the atmospheric events onto the flame-behaviour, has a start-up in investigation, by watching the development of the temperature-, velocity- and irradiance-fields as combustion-consequence of flared hydro-carbons. Heat radiation of the flame and the gaseous products of such non-premixed combustion (NPC) depend both on wind-condition as well as on the geometrical characteristics of the "torch". In this very study, we focused so far our interest by observing the top (with three pilots) of the API-537-standard-flare. Particularly in this report, we relied on the accuracy of the CFD-investigation tool (proven in earlier numerical experiments) while choosing the best of benefits of the both k-emodel (for treatment of turbulence in case of a steady-state investigation) and LES-approach in case of time-dependent numerical observations. So gained data on influence of the atmospheric changes onto the characteristic of the flare-fires is serving our aim: to have a "closer look" at such large-scale fires being under "disturbing" factors; and - by creating possible data-base - setting so further mosaic stone in a map of the crosstalk between the large-scale (reacting) complex flows, the flare-geometry and atmospheric circumstances where they occur.

BOTH LES AND k-e TURBULENCE-TREATMENT IN CFD-APPROACHES FOR CREATION OF THE DATABASE ON FLARE-STACKS FLAME-BEHAVIOUR M. Muhasilovic1, J. Duhovnik1, M. O. Deville2, K. Ciahotny3, V. Koza3 1 University in Ljubljana, Mechanical Engineering, the LECAD Laboratories, Askerceva st 6 Ljubljana, Slovenia [e-mail] 2 EPFL-Ecole Polytechnique Federale du Lausanne, [e-mail] Station 9 , 1015 Lausanne, Switzerland 3 Institute of Chemical Technology, the VSCHT Prague, Tehnicka st 7 Prague Czech Republic [e-mail] Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza ABSTRACT There is an emerging importance to understand and to "map" the reacting complex-flows of the (largescale) flare-fires. In this study, the influence of the atmospheric events onto the flame-behaviour, has a start-up in investigation, by watching the development of the temperature-, velocity- and irradiancefields as combustion-consequence of flared hydro-carbons. Heat radiation of the flame and the gaseous products of such non-premixed combustion (NPC) depend both on wind-condition as well as on the geometrical characteristics of the "torch". In this very study, we focused so far our interest by observing the top (with three pilots) of the API-537-standard-flare. Particularly in this report, we relied on the accuracy of the CFD-investigation tool (proven in earlier numerical experiments) while choosing the best of benefits of the both k-emodel (for treatment of turbulence in case of a steady-state investigation) and LES-approach in case of time-dependent numerical observations. So gained data on influence of the atmospheric changes onto the characteristic of the flare-fires is serving our aim: to have a "closer look" at such large-scale fires being under "disturbing" factors; and - by creating possible data-base - setting so further mosaic stone in a map of the crosstalk between the large-scale (reacting) complex flows, the flare-geometry and atmospheric circumstances where they occur. 2 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza 1. INVITATION TO RESEARCH The aim set, is to create a database for the IFRF and with it´s outcome to fulfil the "re-bourn" urge in the combustions society to understand more accurately the behaviour of the flare fires. This research takes in account the achievements done earlier describing the efficiency[1] of the combustion of these large-scale fires, and will also reflect the achievements into the experimental results[2]. Great role in the entire research on flaring the gas in petrochemical industry and off-shore facilities[3] plays the environmental[4, 5], security[6], caloric and eventually economic[7, 8] aspect, that have been "distributed" over a longer period of time[9]. Applying the modern CFD-based investigation, as a tool for overall understanding of the characteristics of flare-fires[10, 11] - the determining of their roles and the function of different burner geometry[12, 13], will be established as well. Next to these to-beinvestigated points, it will be investigated the influence of ambient conditions[14, 15] and entrained simple agents for decreasing the heat-irradiance, such is water steam[13] under considerations of physical properties of the flaring gas in the supply-pipe-line: Interaction between pressure changes, combustible gas composition[16] and flame stability. In first steps - in this research, at the beginning, we consulted the real geometry[2, 17] data of a flarestack for preparation and it´s embedding in the computational domain for subsequent numerical evaluation of the mentioned working-conditions[8]: better understanding of the instantaneous irradiative heat flux[13, 18], flame stability, generation and distribution of the post-combustion volatiles. The state-of-art in CFD-combustion society, the various exploring-attempts, done by far in the field of the flare-stability - are not "closing" this research-area, but are inviting to additional investigation, which implies that there is a gap in research-results, for understanding the mechanism of the crosstalk between flare and ambient appearances where it´s combustion is occurring. Therefore the major aim of this research concept is to use the achievements of the efforts done so far and to create a data-base on flame stability of the flare-stacks that are in operation under different ambient conditions. 1.1 Turbulence treatment k-e The flow phenomena within the object of interest are computed by the Reynolds Averaged NavierStokes (RANS) equations, with the turbulence k-e model, representing the major characteristic of the applied CFD-investigation-tool with the FLUENT. Since the investigations have shown that the Mach Number never surpass the order of 0.038, such a flow can be assumed as incompressible[19]. So, assumed as incompressible, the fluid while crossing the reaction front, doesn´t undergo thermal-caused expansion and the reaction makes no impact onto flow-velocity[20]. Further assumption, to have a planar propagation front of combustion in a motionless fluid, leads to the application of the 3 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Boussinesq approximation[21, 22] without external forces[23]. Here, the flow velocity obeys the incompressible Navier-Stokes equation[24] with a temperature-dependent force term[23]. The change of temperature is described by an advection-reaction-diffusion equation. For this incompressible gaseous reactive flow at low velocity, the governing equations of the combustioninduced flow read: ¶v j ¶x j =0 (1) ¶vi ¶ ( vi v j ) 1 ¶p 1 ¶t ij + =+ - giaDT ¶t ¶x j r ¶xi r ¶x j (2) ¶T ¶ ( Tv j ) ¶ + = ¶t ¶x j ¶x j (3) æ l ¶T ö 1 çç ÷÷ + R ( T ) r c ¶ x p j è ø z Here vi denotes the average velocity component, T the mean local temperature, p the pressure, r the density, t the time and xi the space coordinates. The R(T) = ¼ T (1 - T) stands for reaction rate[23] where the reciprocal value of reaction time-scale is represented by z, l is the thermal conductivity, cp is the heat capacity at the constant pressure. Temperature T will be used as expression for reaction-progress-variable as well, whose purpose is to distinguished burned, unburned and partially burned state, providing an easy interpretation of flame propagation. The term - g iaDT denotes buoyancy, treated according to the Boussinesq approximation, where DT is showing the difference between local and reference temperature. The symbol g denotes the gravity and a is the coefficient of thermal expansion. The model for the stress tensor[25], t ij is related to the local strain rate: t ij = ( t ij )N + ( t ij )T (4) where we distinguish between the Newtonian stress ( t ij )N = 2 m Sij featuring molecular viscosity; and the turbulent Reynolds stress ( t ij )T = 2 mT Sij , since the stress rate tensor Sij is defined as: 1 æ ¶v ¶v ö Sij º ç i + j ÷ 2 çè ¶x j ¶xi ÷ø mT = Cm and the turbulent viscosity: k2 e (5) (6) with k the turbulent kinetic energy and  the dissipation rate of turbulent energy. The applied k-e model[26] is a two-equation eddy viscosity model[27, 28] that uses transport equations for these two variables[29]. One of these equations governs the distribution through the field 4 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza of k, the local kinetic energy of the fluctuating motion. The other one yealds the energy dissipation rate e [30]. The energy term Gb = a g i æ k2 ö ¶k k2 - Ñ × ç Cm Ñk ÷ = Cm Pd - e - Gb ¶t e e è ø (7) æ k2 ö ¶e e - Ñ × ç Ce Ñe ÷ = C1kPd - ( C3lv N 2 + C2e ) ¶t e k è ø (8) mt ÑT is modelling the buoyancy effects, where Prt denotes turbulent Pr t Prandtl Number (which is of the order of unity). The constants are given: C1=0.126, C2 =1.92, Cµ=0.09, C=0.07. The combustion - the chemistry development is explained by fast chemistry assumption including the prePDF[31] and in the ideal stoichiometric conditions the reaction runs as follows: ignition - energy C7 H16 + 11( O2 + 3, 76 N 2 ) ¾¾¾¾¾ ® 7CO2 + 8 H 2 O + 41,36 N 2 + DH (9) LES In Large Eddy Simulation, the turbulent motion is decomposed as large- and small-scale motions by filtering. The large-scale flow structures are calculated by solving the differential equations numerically. The effect of small-scale motions will be represented by stress terms similar to Reynolds stresses called subgrid-scale Reynolds stresses to be modelled[21]. The first step of LES is filtering r r which decomposes a variable G ( x , t ) into a large-scale component G ( x , t ) and a small-scale r component (subgrid-scale component) G '( x , t ) , i.e., r r r G ( x , t ) = G( x, t ) + G '( x , t ) r (10) r r The large-scale component, G ( x , t ) is obtained by taking a function F ( x - x ', D ) as the filter r r r r r G ( x , t ) = ò F ( x - x ', D)G '( x , t )dx ' kernel[21] (11) W where Ω is the domain of interest; Δ is the filter width, given by Δ=V1/3; and V is the volume of a computational cell, V=∏i=13Δxi=ΔxΔyΔz , where Δxi is the grid interval along the xi direction. The filter function is r r r r r F ( x - x ', D) = 1/ V ; ( x ' ÎV ) Ù F ( x - x ', D) = 0 ; ( x Ï V ) 5 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA (12) Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Filtering each term in governing equations gives us: ¶v j ¶x j ¶vi ¶ ( vi v j ) 1 ¶p 1 æ ¶t ij ¶t ij ,S + =+ ç + ¶t ¶x j r ¶xi r çè ¶x j ¶x j ¶T ¶ (Tv j ) ¶ + = ¶t ¶x j ¶x j =0 (13) ö ÷÷ - g iaDT ø (14) æ l ¶T ö ¶R j q& + çç ÷÷ + è r c p ¶x j ø ¶x j c p (15) where the overbar denotes the filtered variable. The large-scale eddies are computed directly at the resolved scale. The scales of motions unresolvable on the computational mesh are removed. In most fire simulations, the primary momentum transport and turbulent diffusion are sustained by large-scale eddies. Smaller-scale eddy motions are not accounted for. The subgrid-scale (SGS) motions are represented by an eddy viscosity with the length scale related to the grid size in the computing domain. The time scale is determined by the local resolvable dissipation. The SGS motion is calculated by the Smagorinsky-Lilly model where the unknown SGS Reynolds stresses, t ij ,S are related to the local large-scale rate of strain, Sij , by 1 t ij , S - t kk ,S d ij = 2 mt Sij 3 (16) t ij ,S and Sij are defined, respectively, as 1 - t ij ,S º vi v j - vi v j r (17) 1 æ ¶v ¶v ö Sij º ç i + j ÷ 2 çè ¶x j ¶xi ÷ø The subgrid-scale turbulent viscosity μt is used to provide the role of modelling the dissipative behavior of the unresolved small scales. The eddy viscosity is modelled by mt = r L2S Sij (19) Sij = 2 Sij Sij (20) and where ( LS = min k × d × CS × 3 V ) (21) where k =0.42; d is the distance to the closest wall; Cs is the Smagorinsky constant, lying between 0.1 and 0.23, taken as 0.1[21]. 6 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza 1.2 Numerical approach For our transient simulations, the governing equations of the applied mathematical model are discretised in both space and time[28, 32]. In choosing the numerical method, we rely on the standard of the finite volumes [19, 20, 32, 33] for both of the turbulence-models, we applied. The spatial discretisation of time-dependent equations employed a segregated solution method. Since we took a cell-based computational method, the linearised equations result then in a system of linear equations for each cell in the computational domain, containing the unknown variable at the cell centre as well as the unknown values in surrounding neighbour cells within the computational mesh. Meshing Figure 1.2.1-1: Preparations in Computational Domain - grid around the top of the API-537-standardflare: closer view to the flare-top and denser mesh (API-537 - courtesy of NAFTA[17]) Due to the expected flame behaviour, the computational domain, the space around the last two meters of the tip of a flare-stack, was chosen to be cone-shaped. Grid within the domain was composed out of 1'643'660 cells - more dense in the zone of the combustion (on the tip of the flare as well as around the pilots), having different cell-shapes in asymmetric mesh, throughout the domain - pointing so to the mesh-independency in this approach. 7 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Figure 1.2.1-2: Preparations in Computational Domain: The "cone" shaped area in CFD-approach Figure 1.2.1-3: Top-view of the meshed flare-stack and three pilots surrounding it 8 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza 2. RESULTS Once obtained, our results demonstrate well the CFD-"toolbox"-performances we have chosen. Through running of the begin of this research, the results of such investigation on reactive flows, are offering the next step for further development of technical solutions, that are used at the flare-stacks for both energy-transformation-processes[16, 34] and commercial way, which can do a credit to the general safety, conservation of environment[35] and successful human protection as well[36]. Observing different state-of-the-art shapes of the flare-stacks[4, 17], with this research attempt, we intend to follow the "footsteps" of some further investigators in this area and investigate the influence of the characteristic physical values, such are among others: mass flow rate (of the combustible flaregas)[35], distribution of the iso-surfaces of the heat-irradiance[35] around the burner[34, 35, 37] due to the wind influence for flare-stacks under both high-pressure- and the low-pressure service. Figure 2-1: A side-wind influencing the API-537-Flare, "framed" within the steady-state investigation 9 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Figure 2-2: Another view-angle to the same temperature field coming out of the steady-state investigation: the blue-coloured grid-elements on the far side of the computational domain, signifies the area where-through the "wind" of 5 m/s enters into the domain Figure 2-3 Velocity-field in the 27thsecond of the fire with the concave "wind"-inlet-side (upper sketch) 10 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Figure 2-4: Flare-stack-tip, looking along the "wind": noticed flame-zone with strongest velocityvectors 3. DISCUSSION AND OUTLOOK One of the studies, even in other fields of large-scale combustion, performed earlier by Liu et al. [38] pointed at the important difference between k-e turbulence treatment and the LES-approach in exploration of the same phenomena. We noticed similar reaction after using the mentioned CFDmodules and this in such way where k-e applied model gave general forecast of zones for temperature development, iso-surfaces on heat-irradiance propagation and velocity vector fields - demonstrating here the area of interference between side-wind and buoyancy-influence of the flame. Complementary to this, LES-turbulence treatment, applied in transient CFD-calculations, gave us more precisely the shape of the flame-front during such large-scale combustion and offered so more clearly the areas of time-dependent changes in irradiance, temperature and surface-emissive-power. So after performing these preliminary results, we are aiming to certain "matrix" of results-to-be-gained, based on the different working-conditions of flares - where ambient conditions: wind-direction and velocity will be changed and the composition of the atmosphere as well (changing the moisture of the air from foggy conditions to a rain). Our planned "matrix" will eventually have "additional dimensions" when we undertake those CFD-investigations on the different flare-stack geometries. 11 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza Figure 3-1: the 5m/s-wind influence in the 6th second of the established flare-flame and within it, the temperature-zone of the 1800K (presented here as iso-surface) Figure 3-2: A "cross-panel" for picturing-up the temperature iso-zones is showing here the 5m/swind-influence in the 6th second of the established flare-combustion Simultaneously to our idea to research on gas-flaring and flame-stability, the flare-gas recovering[16] and general application of FGRU[16] (Flare Gas Recovering Units)[34] in the processing industry[11], 12 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza gives further parallel contribution in the investigation of flares, taking in account the differing composition of the flaring gas[16]. Finally, it is the urge to recognize the trend in the results that will be gained by further CFD-based research and then, by following this trend, to enter deeply in the behaviour of the flare-fires exposed to deferent atmospheric influences, creating mentioned database. 4. ACKNOWLEDGEMENT The authors do thank the petrochemical refinery-company "Nafta-Lendava" in Slovenia, for technical support and advices on particular types and function of the flare-stacks. 5. LITERATURE: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. J. Lee J. H. Pohl, R. Payne, B. A. Tichenor Combustion efficiency of flares. Combustion Science Technology, 1986. 50: p. 217. N. Papanikolaou, An experimental investigation of the flow structure and stability limits of jet diffusion flames in a co-flowing oxidizing stream in Department of Mechnanical Engineering. 1998, University of CAlgary: Calgary, CANADA. S. N. Singhal, Offshore Stack-enclosed Gas Flares. Fire Safety Journal, 1989. 15: p. 227. K. E. Abhulimen A. A. Susu, A. B. Adereti, Modeling of Air Pollution Systems with Chemical Reactions: Application to Gas Flares in Nigeria. International Journal for Computational Methods in Engineering Science and Mechanics, 2005. 6(3): p. 201. K. Muirhead and A. P. Cracknell, Identification of gas flares in the North Sea using satellite data. International Journal for Remote Sensing, 1983. 5: p. 199. A.M. Birk, An Experimental Investigation of a Cylindrical Vessel Engulfed in Fire With a Burning Relief Valve Flare Present. Chemical Engineering Communications, 1988. 80: p. 173. S. Mustafiz N. Bjorndalen, M. H. Rahman, M. R. Islam, No-Flare Design: Converting Waste to Value Addition. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2005. 27(4): p. 371. G. C. Oguejiofor, Gas Flaring in Nigeria: Some Aspects for Accelerated Development of SasolChevron GTL Plant at Escravos. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2006. 28(15): p. 1365. A. P. Cracknell K. Muirhead, Identification of gas flares in the North Sea using satellite data. International Journal of Remote Sensing, 1984. 5(1): p. 199. C.-Y. Lin C.-H. Cheng, W. Aung, Predictions of developing flow with buoyancyassisted flow separation in a vertical restangular duct: parabolic versus elliptic model. Numerica l Heat Transfer, 2000. 46: p. 567. J. Akhtar A. Basu, M. H. Rahman, and M. R. Islam, A Review of Separation of Gases Using Membrane Systems. Petroleum Science and Technology, 2004. 22: p. 1343. 13 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. R. Schwartz, New API Standard provides comprehensive information on flares. Hydrocarbon Processing, 2004(May 2004): p. 73. D. Knott K. Leary, R. Thompson, Water-injected flare tips reduce radiated heat, noise. Oil & Gas Journal, 2002(May 2002). N. Papanikolaou G. P. Song, A. A. Muhamad, Flame stability with elliptical nozzles in a crossflow. Combustion Science and Technology, 2004. 176: p. 359. J. Franklin J. Bellovich, C. Baukal, The last fire defence - discussion on the importance of a reliable flare pilot system for plant safety. Hydrocarbon Engineering, 2006(April 2006). N. Tuttle J. Peterson, H. Cooper, C. Baukal, Minimize facility flaring. Hydrocarbon Processing, 2007(June 2007): p. 111. C. Zalik A. Gomzi, On Flare tips and pilots, M. Muhasilovic, Editor. 2009: Ljubljana, Slovenia. J. E. S. Venart M. Tunc, Incident Radiation from a Flare to a Horizontal Cylinder. Fire Safety Journal, 1984-1985. 8: p. 89. J. H. Ferziger M. Peric, Computational Methods for Fluid Mechanics. 2001, Berlin: Springer Verlag. 423. M. Deville M. Muhasilovic, Tunnel-Curvatire´s influence on the propagation of the consequences of large-scale accidental fire - a CFD-investigation. Turkish Journal of Engineering and Environmental Sciences, 2007. 31: p. 391. S. L. Liu P. Z. Gao, W. K. Chow, N. K. Fong, Large eddy simulations for studying tunnel smoke ventilation. Tunneling and Underground Space Technology, 2004. Vol. 19(Issue 6): p. 577 - 586. et. al. H. R. Baum. Gravity-current transport in buildings fires. in International Conference on Fire Research and Engineering. 1995. Orlando, Florida, USA. N. Vladimirova, Model flames in the Boussinesq limit. 2006, ASC / Flash Center, Dept. of Astronomy and Astrophysics, The University of Chicago, IL 60637: Chicago, USA. J. S. M. Li W. K. Chow, Case study: vehicle fire in a cross-harbour tunnel in Hong-Kong. Tunneling and Underground Space Technology, 2001. Vol. 16(Issue 1): p. 23 - 30. T. Jongen T. B. Gatski, Nonlinear eddy viscosity and algebraic stress models for solving complex turbulent flows. Progress in Auerospace Sciences, 2000. 36: p. 655. C. Leupi, Numerical modeling of cohesive sediment transport and bed morphology in estuaries, in La faculte sciences at techniques de língenieur. 2005, Ecole Polytechnique Federale de Lausanne: Lausanne. et. al. W. Zhang, Turbulence statistics in a fire room model by large eddy simulation. Fire Safety Journal, 2002. 37: p. 721. http://www.fluent.com. R. E. Britter P. J. Woodburn, CFD-simulations of a tunnel fire - part one. Fire Safety Journal, 1996. 26: p. 35. M. R. Malin N. C. Markatos, Mathematical modelling of buoyancy-induced smoke flow in enclosures. International Journal of Heat Mass Transfer, 1982. 25: p. 63. O O. Megret, Vauquelin, A model to evaluate tunnel fire characterisrics. Fire Safety Journal, 2000. 34: p. 393. W. Malalasekera H. K. Versteeg, An Introduction to computational fluid dynamics. 1995, London: Longman Group Ltd. 14 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA Both LES and k-e turbulence treatment in CFD-approaches for creation of the database on flare-stacks-flame M. Muhasilovic, J. Duhovnik, M. O. Deville, K. Ciahotny, V. Koza 33. 34. 35. 36. 37. 38. C. Hirsch, Numerical Computation of Internal and External Flows. Vol. I. 1988, Chichester Brisbane Toronto New York: John Wiley & Sons. 515. D. Brennan P. W. Fisher, Minimize flaring with flare gas recovery. Hydrocarbon Processing, 2002(June 2002): p. 83. C. Baukal J. Hong, R. Schwartz, M. Fleifil, Industrial-Scale Flare Testing. Chemical Engineering Progress, 2006(May 2006). Marlon Harding John A. Alderman, Proactive vs. proscriptive fire protection for the offshore industry. Fire protection engineering, 2001(12): p. 23-27. Z. Kodesh R. E. Schwartz, M. Balcar, B. Bergeron, Improve flaring operation. Hydrocarbon Processing, 2002(January 2002): p. 59. P. Z. Gao S. L. Liu, W. K. Chow, N. K. Fong, Large eddy simulations for studying tunnel smoke ventilation. Tunnel and Underground Space Technology, 2004. 19: p. 577. 15 16th International IFRF Members' Conference Combustion and sustainability: new technologies, new fuels, new challenges 8th-10th Jun, 2009 IFRF Members Conferences Boston, USA