Reacting cting Flow Modeling ling in n STAR-CC -CCM+ Rajesh sh Rawat at
Latest est Addit itions ions (v 7.02/v v 7.04) Eulerian ian Multi-phase phase Reacti ction on Model el Soot ot Model el – Moment Methods PPDF Flamel elet – Multi-stream model Comple plex x Chemi mistr stry Model (DARS ARS-CFD) CFD) – EDC (additional modifications) – Dynamic Load Balancing – ISAT Further her enhancem ncemen ents ts and testi ting ng for LES Best st Practi tices ces for LES Misc – Non-transported Passive Scalars – Permeable fixed flux or fixed species Boundary Conditions
Lagrangian rangian Multi tiphas phase e Capabilit bilites es Par Particle cle radiation iation Mul ulti-co component ponent drop ople let t evapora porati tion on Orig igin in Genera neral l Par Particle cle Reacti tions ons Axis is eros osion ion mode dels Inner er Angle le Inject jector or Model dels Outer er Angle le – Part, Point, Hollow Cone and Surface Injectors Primar ary y Atomi mizati zation on Model del – Pressure Swirl Injector (Lisa model) Secondar condary Break akup Model del – KHRT – Reitz-Diwakar – TAB
Par Particle ticle React action ion Mo Model el Par Particle cle Reacti ction on Model del is available ble through ough Lagrangi rangian an Mul ultiphase phase frame mewor ork of STAR AR-CC CCM+ M+ Two o Reacti ction on Models dels are available: ble: Par Particle cle Devolatil atiliza izati tion on and Par Particle cle Comb mbusti ustion on Can simulat ate – Particle Devolatilization, Sublimation, Melting – Particle Combustion, Oxidation, Pyrolysis
Particle Combustion In partic icle le combustion mbustion modeli odeling, ng, a solid d particle cle reacts ts with th a gas-pha phase se species cies to form rm solid id and/or /or gas-phase phase produc ducts ts Three ree different rent meth ethod ods s are available ble for speci cify fying ing the e react ction ion rates: es: – First Order Combined Rate – Half Order Combined Rate – User-Defined Rate The e combine mbined d rate consid nsiders gas-speci pecies es diffusi fusion on to the e particl icle. e. The e Diffusion usion Coef efficient cient can be speci cified ied by the us user as scalar ar profile. le.
Example Reactions Par Particle cle Devolati tiliza izati tion on Model del -- -- CaCO 3 CaO + CO 2 Par Particle cle Comb ombusti ustion on Model del -- -- 2Ca (s) + O 2 (g) (g) 2CaO (s) -- -- NaOH (s) + HF (g) NaF F (s) + H 2 O (g) -- -- Fe Fe 2 O 3 (s) + 3H 2 (g) (g) 2Fe e (s) + 3H 2 O O
LES • All ll Yplu lus treat eatmen ment • Seco cond nd order der im implic licit it tim ime diff ifferenc rencing ing • Both CD and d BCD • Non-reflec reflecting ing bounda undary y condit ndition ion • Synt nthet hetic ic turbulen bulence ce for in inflo low BC BC • Reac acti ting ng Flo low • Thick icken ened ed Fla lame me Model del • Alg lgebraic braic Varian riance ce and d SDR 7
Sandia dia D Flame ame Mesh count: 4.1 M Polyhedrals in the vicinity of inlet Extruded Polyhedrals elsewhere Volumetric control in the flame region Approximately 0.18 mm of mesh size in radial direction
Temperat perature ure – Insta tantan ntaneou eous s & Mean
Movie e - Temperatu perature re
Comparis parison on with th Exp xperi erimen ment t – C Cent nterl erlin ine e Axial ial Veloc ocity ity RMS Axia ial l Vel Mean an Axia ial l Vel
Compari parison on with th Exp xperi eriment ment – Cent nterl erlin ine e Mixt xture ure Fr
LES S Best st Prac actices tices
Librar rary Based ed Models ls PVM/Flamelet Library H298 Source Premixed Flame Emissions source Term (CFM) terms (Nox/soot) Transport Equations Properties STAR-CCM+
The progress gress varia iable ble model el A precomputed table is used to store the effects of detailed • chemistry on combustion: selected species, source term to chemical enthalpy, mixture NASA polynomials, and molecular weight The combustion state is described using a progress variable • based on chemical enthalpy h298 is the chemical enthalpy for the mixture, while • h298|c= 0 and h298|c= 1 are the chemical enthalpies at the initial conditions and equilibrium state, respectively. 15
Premix emixed ed Gas Turbine bine PVM DARS-CFD with GRI-Mech
Temperat perature ure Compar arison ison
Soot Model Validation
Centerline Temperature Profile
Centerline Soot Profile
Radial Profiles for Soot (x = 0.347)
Soot Distributions
Multi ti-Stream tream PPD PDF
Temperat perature ure & Lagrangian rangian Trac acking king
Comple plex x Chemistr emistry • Can read ad Chemk emkin in format rmat and no limit t on num umber ber of species cies • Online ne tabulati tion on us using g ISAT is available ble – Factor of 2-5 speedup is commonly observed • Dynami amic c load d balanc ncing ing is available ble to achieve scalabil bilit ity y for r chemis emistr try y calcula ulati tion on with h large e num umbe ber r of process cessor ors. s. • DAR ARS-Ba Basi sic c provide ides s tool ol to reduc duce e the e chemis emistr try y that t can be imp mpor orted d in STAR AR-CC CCM+ M+ for fur urth ther er speedup eedup for comple plex chemis emistr try y calcula ulati tion ons. s.
Dy Dyna nami mic Load Balancing ancing Same non-premi premixed ed flame me but with more det etailed ailed mechanism hanism with h 53 speci ecies es tested ed for transient sient run Presen ented: ed: compl mplex x chemi mistr stry CPU time after 200 iterati tion ons Transient run with load balancing 130 Hrs 140 120 100 60 Hrs 80 Time (hours) 8p 16p 32p 60 25 Hrs 40 20 0 Dynamic balancing Transient run
The EDC model el The total space is divided into two zones • Fine structures All the chemical reactions takes place Represents the smallest turbulence scales Bulk structure No reactions occur The sub-grid chemistry scales don't need to be resolved at a • cell level. 27
Fine e stru ructures ctures Governing equations for the species and enthalpy • * : quantities in the fine structure < > : the cell mean values. The residence time: how long the species remain in the fine • structures Mass fraction of fine structures in the cell: • 28
Temperature Profiles
CH4 Profiles
SCR SCR Flow ow Direc ection ion is from om left to o righ ght • Solid id Cone ne Spray ray with h 70 o , not ot much h • turbule urbulent nt disper persion ion Therm ermolysi olysis s cons nsum umes Urea ea quit ite rapidl apidly • Conv nvers rsion ion Efficien iciency & U Unif iform rmity ity Index ndex of • NH3 and nd H2O O can n be deduc educed ed from rom this is analys alysis is. This is can n help lp opt ptimiz imize injecti jection n strat rateg egy for or • UWS upstream eam of SCR system em.
Resu sults lts – NOx Reduct uction n Compari parison on Two-Step Model Detailed Surface Chemistry
Conclusions lusions Euler eria ian Multi-Phas Phase e wit ith Rea eacti tions ons LES ef effec ecti tive e but expe pensiv ive Fi Finit ite-rate e kin inet etic ics – Library-based – Direct chemistry coupling Spe peed edup up – Load balancing – Clustering – ISAT
Recommend
More recommend