Exploring STAR-CCM+ Capabilities, Enhancements and Practices for - - PowerPoint PPT Presentation

Exploring STAR-CCM+ Capabilities, Enhancements and Practices for Aerospace Combustion

Niveditha Krishnamoorthy CD-adapco

Overvie view of mode deling ing capabi bility lity Ap Applications ations, , Practice tices s and En Enhan hance ceme ments nts

– Gas Turbines – Rocket Nozzles – Scramjets

Sum umma mary

Outli line

Density – and Pressure- based flow solvers Incompressible through hypersonic flow regimes Full range of turbulence models

– RANS, LES and DES

Multi-physics

– Combustion – CHT – Fluid-Structure Interaction – Aeroacoustics

Overvie view w – (1)

Combus ustion tion and nd Emissions

• ns Modeling

ing

– Models for different flame types: Premixed, non-premixed and partially premixed – Multi-component liquids, solids and gases – Homogeneous and Heterogeneous chemistry – Global, tabulated chemistry, reduced/detailed chemistry – Emission models for Soot, and Nox

• Soot

– Two –equation semi-empirical model – Method of moments

• Nox

– Fuel – Prompt – Thermal

Overvie view w – (2)

Mass Fraction of CO (LES run)

Par Particip cipati ting ng Medi dia a Radiat ation ion Model del

Overvie view w –(3)

• Radiative Transport Equation (RTE) solved using the Discrete Ordinates Method (DOM)
• S2 to S16 Quadratures available

Properties of the medium:

• Gray Thermal Radiation
• User defined
• Weighted Sum of Gray Gases
• Multiband Thermal Radiation
• Properties in wavelength bands
• Particle Radiation
• Scattering: isotropic, gray
• Absorption: gray

Applicati ication,

• n, Practices

ctices and Enhanceme ncements nts Ga Gas Tur urbin bine Combu mbust stor

• rs

Gas Turbine Combustor Emissions Fuel Flexibility, Flame Stability Thermo-acoustic Instability Mechanical Durability Cost

• UHC
• Soot
• Nox
• CO
• Flame shape
• Flame location
• Flash-back/ blow-off
• Gaseous/liquid Fuels
• Liner temperature
• Component temperature

System Level Combustion Chemistry Heat Transfer Fluid Dynamics Unit Level

• Flow and mixing
• Swirlers
• Bluff bodies
• Fuel formulation
• Operating conditions
• Chemical kinetics
• Thermodynamics
• Conduction
• Convection
• Radiation

Liquid Droplet Combustion

– Droplet Evaporation

• Quasi-steady
• User defined

– Droplet Break-up

• Primary atomization

– Linear Instability Sheet Atomization (LISA)

• Secondary break-up

– Kelvin Helmholtz-Rayleigh Taylor (KHRT) – Taylor Analogy (TAB) – Stochastic Break-up (SSD)

– Droplet Wall-impingement

• Bai-Gosman
• Satoh

– Collision Detection Model

• No Time Counter (NTC)
• O’Rourke

– Two-way Coupling

– Turbulence Dispersion

• Random Walk Technique

Combustion Chemistry: Models

Global al Chemistr try

– Global multi-step reactions can be calibrated for specific operating conditions – Calibration carried out using HEEDs and DARS-Basic [ Freely propagating Flame]

Gas Phase se Comb mbustion ustion Modelin ling 𝜕 = 𝑩𝑓−𝐹𝐵/𝑆𝑈 𝐷 𝒐 𝐸 𝒏 𝐷 + 𝐸 → 𝐹 + 𝐺

Sample mple Reac action: ion: Reac acti tion

• n Rate:

e:

Variables we can vary

• A : Pre-exponential Factor
• n, m : FORD (forward reaction rate exponents)

Calibration to match flame speed Blue = Literature Red = uncalibrated Green = Calibrated Purple = Hand Calibrated

5 Step Mechanism chanism

Tabul ulat ated ed Chem emist stry

– Presumed Probability Distribution Models (PPDF Flamelet)

• Update PPDF Species once per time step without affecting results
• Can delete species not needed for post-processing: affects table size and look-up

times

• Vectorial table retrieval and more efficient table interpolation

– Flamelet Generated Manifold (FGM)

• Flexible definition of progress variable
• Transports progress variable and its variance. Better representation of state-space

than the traditional progress variable model

• Inclusion of heat loss effects in the table always ensures species and enthalpy are

consistent

Gas Phase se Comb mbust ustion ion enhancem emen ents ts

Performanc

• rmance

e Improvem ement nts: s: Large rge Cases ses (LES ES)

• Flow Solver improvements in v9.04

Case 1 Case 2 Case 1 Case 2

• Combustion solver improvements in

v9.02 (40-50% speedup)

• Flow and Lagrangian solver

improvements in v9.04 (20-25% speedup from v9.02)

Accura urate temperatu perature re distr tribu ibuti tions

• ns are

requi uire red d for emiss ssion ion predic dicti tions

• ns and

to assess ess com

• mpone

ponent nt life Heat t transf nsfer er in solid id compone ponents nts requi uire res adequate e desc scription ription of

– Conduction – Convection – Radiation

Liner ners s can eith ther er be mode delled led as shells lls (1 (1-D D heat at conduct nduction) ion) or with h 3-D heat at conduc nducti tion

• n

Heat Transfer

Impor mport t CAD AD

– Single or multiple injectors – Liners can have all details like dilution holes, effusion holes – All solid components can be included: splash plate, dome etc.

Repair ir the e CAD AD to get et a closed sed geom

• metr

try Ex Extract act flui uid d doma main in and all the e so solid d compon ponents ents from

• m the

e closed ed CAD AD Refine ine prism m layer r mesh esher er for CHT compon ponen ents ts suc uch that t Y+ is close se to 1

Genera eral l Meshing hing Procedure edure for CHT

Need to ensure there are no intersecting parts or gaps between components Fully conformal mesh possible Quality of mesh can be checked by running mesh diagnostic report Checks for:

– Face Validity – Cell quality – Volume change statistics – Cell and boundary skewness angle

Meshing ing

1:1 matching Conformal mesh

Lagrang rangia ian up update e can be done ne once ce ever ery y time me-st step Dynami amic c load d balanc ncing ing for Lagrangi angian an spray y helps lps with th speed eed up up Up Update e species cies and radiat ation ion once ce per time me-st step ep (for

• r LES

ES run uns) s)

Other er useful ful sett ettings ings

Applicati ication,

• n, Practices

ctices and Enhanceme ncements nts Rocket t No Nozz zzles les

Comb mbusti ustion

• n Chamb

mber er

– Multicomponent Droplet combustion – Real Gas Equation of State – Equilibrium and finite rate chemistry models

Nozzl zzle e Flow/P w/Plu lume me Study udy

– Coupled Solver – Equilibrium or finite rate chemistry

Heat t Transf nsfer er (CHT) T)

– Base heating – Film Cooling

Applica icati tion

• n Areas

as

Combustion with Real Gas Model

Fuel: Methane at 275.8 K Oxidizer: Oxygen at 105.9 K Solver: 3D, Steady, k-omega SST Redlick Kwong EOS, Coupled Implicit with solution driver and convergence accelerator (CCA) Non-premixed, non-adiabatic PPDF

Coupled Implicit, Axisymmetric Steady, SST K-Omega turbulence Detailed chemistry: 11 species DARS-CFD Approximation options:

– In-situ Adaptive Tabulation

• Populates source terms as the

simulation progresses for subsequent look-up

• Speeds computational time once the

table is populated

– Equilibrium Time-Scale

• Quick approximate solution for

detailed chemistry calculations

• Assumes chemical composition

relaxes to local equilibrium composition at time-scale determined by flow and chemistry

Reacti cting ng Nozz zzle le Flow

One stream or part of one-stream is inert, its reactivity neglected Sole effect of inert stream is to dilute reacting products Transport equation solved for corresponding mixture fractions Species mass fractions computed as linear combination from inert stream and reacting streams Faster table generation and

• interpolation. Smaller table size

Inert t Stream ream model el for PPDF DF Combust mbustion

Yi=Zinert*Yi_inert+(1-Zinert)*Yi_reaction

Applicati ication,

• n, Practices

ctices and Enhanceme ncements nts Scramj amjets ts

Densi nsity ty-bas based ed Solver er

– Coupled, implicit formulation with AMG acceleration – TVD reconstruction

• AUSM+ or Roe inviscid flux schemes
• MUSCL + Venkata limiter

– Advanced initialization and convergence control – Real gas models

• Redlich Kwong
• Soave-Redlich Kwong
• Modified Soave-Redlich Kwong
• Peng Robinson
• EBU, PPDF combustion models

High Speed Reacting Flows

Advanc nced ed Initi tializ lizat ation ion

– Grid sequencing option – Fully implicit newton-type solution algorithm – Controllable number of coarse levels

Continuity ntinuity Convergence ergence Accel elera erator

• r (CCA)

A)

– Used for high speed flows where convergence for mass flow is slow – Solves pressure correction equation using density based Riemann Flux discretization – Overall and individual cell mass imbalances are minimized at each iteration – Option available for Coupled Implicit Solver.

Advanced Initialization and Convergence Control

Supersonic Combustion

• H2 Fue

ueled led NAS ASA A SCHOLA OLA direct ect-co connect nnect Scramj amjet t engine gine

• Valida

date agains nst t experime eriment nt and NAS ASA A VULCAN CAN code de Mesh: h: 1.4M Hex-dominant 10 Prism Layers Solver: er: Density based solver Steady,k-w SST, AUSM+FVS Non-adiabatic PPDF

Supersonic Combustion (2)

Global bal Chem emist stry

– Single or multi-step – Variants of eddy break-up model

• Standard
• Hybrid
• Combined time-scale
• Kinetics only

Tabul ulat ated ed Chem emist stry

– PPDF Equilibrium

Det etail iled ed Chemis emistr try

– DARS-CFD stiff chemistry solver – Use Equilibrium Time-Scale approximation for initial guess – Then switch to finite rate chemistry

• Laminar flame concept
• Eddy dissipation concept

Comb mbustion ustion Modelin ling g in Scramje amjets ts

Dual al Mode de Scra ramje jet

Meshing hing

– Utilize extruded (directed) mesh as much as possible in long, non-complex, ductwork (isolator, combustor, etc.) – Use directional reordering of mesh in streamwise direction

Coup uple led Flow and En Energy rgy

– Coupled inviscid flux scheme: AUSM+ – Coupled Energy: Enable Enthalpy Formulation

Boun undary y Condi nditi tions

• ns

– Specify fuel inlets as mass flow. Ramp flow rate over 1000 iterations – Pressure outlets can have a small area of extrusion w/ free slip wall

Genera eral l Tips ps for Scram amjets ts

Solver er Sett ettings gs for Scram amje jets ts

Coupl pled ed Imp mplici cit

– CFL = 5.0 – Ramp CFL from 0.1 over first 300 iterations

AMG G Linear near Solv lver er

– Max Cycles = 10 – V-Cycle:

• Pre-Sweeps = 1
• Post-Sweeps = 3
• Max. Levels = 50

Grid Sequenc uencing ng Initia tializa lization tion

– 10 Levels, 150-250 iter./level, Tolerance = 0.005, CFL = 5.0

Exper ert t Driver

– CFL Ramp:End Iteration = 250 –

• Min. Explicit Relaxation = 0.35

–

• Max. Explicit Relaxation = 0.75

– Target AMG Cycles = 6

Continuit tinuity y Converg ergenc ence e Ac Accel celerat erator

• r

– URF = 0.6 (Ramp from 0.03 over first 100 iterations) – Enhanced Mass-Imbalance Calculations Enabled – AMG Solver Convergence Tolerance = 0.05, V-Cycle: 1 Pre-Sweep, 1 Post-Sweep, Max levels = 50

Mul ulti-ph phys ysics ics simula ulati tion

• ns

s involvin

• lving

g high h speed ed flows, ws, reacti ction

• ns

s and heat at transf nsfer er possibl ssible e us using g STAR AR-CC CCM+ Continued ntinued effor

• rts

ts on making ing the e code de fa faster er for com

• mple

lex x flows ws Best t practices ices esta tabl blis ished hed for: r:

– Meshing – Physics set-up – Solver settings – Initial and boundary condition specifications In each area of application across industry sectors

Summ mmar ary