Jacqueline Chen, Ed Richardson, Ray Grout, Tianfeng Lu,

1
Reactive Scalar Mixing and Turbulence-Chemistry
Interactions in Turbulent Combustion

• Jacqueline Chen, Ed Richardson, Ray Grout,
  Tianfeng Lu,
• Chung K. Law, and Chun Sang Yoo
• Combustion Research Facility
• Sandia National Laboratories
• Livermore, CA
• jhchen_at_sandia.gov
• Symposium on Turbulence and Combustion
• August 3-4, 2009
• Cornell University
• Ithaca, New York
• Supported by the Division of Chemical Sciences,
  Geosciences, and Biosciences, Office of Basic
  Energy Sciences and the Office of Advanced
  Scientific Computing Research of DOE

2
Changing World of Fuels and Engines

• Fuel streams are rapidly evolving towards
  renewable sources
• Ethanol
• Biodiesel
• New engine technologies
• Direct Injection (DI)?
• Homogeneous Charge Compression Ignition (HCCI)?
• Low-temperature combustion
• High efficiency and low emissions requires mixed
  combustion modes (dilute, high-pressure,
  low-temp.)?
• Sound scientific understanding is necessary to
  develop predictive, validated multi-scale models!

3
Multi-scale Modeling of IC engine processes

• Multi-scale modeling describes IC engine
  processes, from quantum scales up to
  device-level, continuum scales
• Multi-scale Strategy
• Use petascale computing power to peform direct
  simulation at the atomistic and fine-continuum
  scales (4 decades), and develop new
  parameterizations that will enable bootstrapping
  information upscale

4
Petascale High Performance Computing

• Petascale computing for scientific discovery
• DOE INCITE Awards and NSF grants large computing
  allocations

Cray XT5, ORNL 1.6 Pflop 2009
5
Direct Numerical Simulation Code S3D

• Used to perform first-principles-based DNS of
  reacting flows
• Solves compressible reacting Navier-Stokes
  equations
• High-fidelity numerical methods
• Detailed reaction kinetics andmolecular
  transport models
• Multi-physics (sprays, radiation and soot) from
  SciDAC-TSTC
• Ported to all major platforms, scales well
• Particle tracking cabability

DNS provides unique fundamental insight into the
chemistry-turbulence interaction
Engineering CFD codes (RANS, LES)?
Physical models
DNS
6
A Systematic Procedure for Dimension Reduction
Stiffness Removal
Skeletal mechanisms C2H4 30
species nC7H16 100 species
Detailed mechanisms C2H4 70 species nC7H16
500 species
Dimension Reduction With DRG
Time Scale Reduction With QSSA
Reduced mechanisms C2H4 20 species nC7H16 60
species
Minimal diffusive species C2H4 9
groups nC7H16 20 groups
Diffusive Species Bundling
On-the-flyStiffness Removal
Non-stiff reduced mechanisms C2H4 20
species nC7H16 60 species
Sponsored by DOD AFOSR
7
Simulation Benchmark Data for Model Development
and Validation
8
Role of DNS 2 Case Studies

• Stabilization of a Lifted Ethylene-Air Jet Flame
  in Heated Coflow
• Reactive Scalar Mixing in Premixed Methane/Air
  Flames under Intense Turbulence

9
Stabilization of Lifted Ethylene/Air Turbulent
Jet Flames in Heated Coflow

• Chun Sang Yoo and Jacqueline Chen
• Sandia National Laboratories
• Tianfeng Lu
• University of Connecticut
• Chung Law
• Princeton University

10
Motivating Example Diesel Lift-off Stabilization
What is the role of ignition in lifted flame
stabilization?
Chemiluminescence from diesel lift-off
stabilization for 2 diesel, ambient 21 O2,
850K, 35 bar L. Pickett and S. Kook, 2008.
Lift-off distance (mm)
Time (ms)
11
DNS of Lifted Ethylene-air Jet Flame in a Heated
Coflow
Ethylene-air lifted jet flame at Re10000

• 3D slot burner configuration
• Lx ? Ly ? Lz 30 ? 40 ? 6 mm3 with
• 1.28 billion grid points
• High fuel jet velocity (204m/s)
• coflow velocity (20m/s)
• Slot width, H 2.0mm
• Rejet 10,000 ?j 0.15ms
• 6 flow through times
• Cold fuel jet (18 C2H4/ 82 N2) at 550K,
• ?st 0.27
• Reduced C2H4/air chemistry (Lu and Law)
• 22 species 18 global reactions, 206 steps
• Hot coflow air at 1,550K
• Performed on CrayXT4 at ORNL on 30,000 cores and
  7.5 million cpu-hrs
• 240 TB field data, 50TB particle data

OH
HO2
12
Favre Mean and Instantaneous Temperature and
Species Mass Fractions
CH2O
HO2
T
OH
13
Temporal Evolution of Stabilization Point (left
branch)
0.33 ?j
0.77 ?j
Temporal evolution of OH mass fraction isocontour
5e-4 (5 of mean max) at t/?j 0.227 1.160
14
Tracking of Stabilization Point
0.33
0.77
x
-u.n
15
Conceptual Stabilization Mechanism
Su Mungal
Temporal evolution of OH mass fraction isocontour
at t/?j 0.227 1.160
1/ Ignition occurs in lean mixtures with low ? 2/
No self-propagation upstream with mixing
structure 3/ Local extinction occurs by high ? or
flame shortening occurs as the point is
convected downstream 4/ Ignition occurs in
another coherent jet structure
Convective velocity greater than displacement
speed for ?st 0.27
16
Correlation with Large-Eddy Structure
Power spectrum of the stabilization point
fluctuations and the correlation function Ru
oscillations between axial velocity fluctuations
over a transverse separation of 2d1/2 at mean
stabilization height of 6H.

• Strouhal number shows 3 dominant correlation
  function frequencies in the near
• field of planar jet (Thomas and Goldschmidt 1986)
• From the dominant frequencies of the spectra the
  fluctuations of the stabilization point
• appear to be correlated with the passage of
  large-scale flow structures

17
Chemical Explosive Mode and Damköhler Number
Da l/
18
Conclusions

• Lifted ethylene/air jet flame in heated coflow is
  stabilized by autoignition upstream of the high
  temperature flame
• Chemical observables include high levels of HO2,
  and low levels of CH3 and CH2O.
• Autoignition occurs at a preferred mixture
  fraction (fuel-lean) and at low scalar
  dissipation rate.
• Dynamics of stabilization are determined by
  competition between upstream autoignition and
  convective motion of the jet. Ignition is
  characterized by high front speeds (gt 100 m/s)
  and relaxation on longer diffusive timescales to
  laminar propagation speeds (lt 10 m/s).
• Chemical explosive mode analysis (CEMA) useful to
  pinpoint stabilization point and rich and lean
  premixed branches. CEMA confirms occurrence of
  autoignition upstream and on the rich side. Key
  species and reactions were identified.
• Correlation between large-scale jet mixing
  structure and fluctuations in the lift-off
  height.

19
Reactive Species Mixing Rates in Turbulent
Premixed Methane-Air Combustion

• E.S. Richardson, R.W. Grout and J.H. Chen
• Combustion Research Facility
• Sandia National Labs
• R. Sankaran (31D4)
• NCCS, Oak Ridge National Laboratory

20
Introduction

• Industrial need for turbulent reacting flow
  models capable of accurately predicting
  combustion performance and pollutant emission.
• Suitable combustion models typically require
  turbulent mixing timescales/dissipation rates as
  inputs.
• Common practice to use
• Algebraic/transport models for the premixed flame
  progress variable dissipation rate have been
  studied/developed using simple chemistry DNS.
• Analysis of reactive species mixing rates of in
  3D premixed flame simulations with realistic
  transport and chemistry.

21
Dissipation rate equation (almost ) exact
Convection Diffusion
Dissipation
Turbulent transport Gradient source
Dilatation
Turbulence-chemistry interaction (T32)
Reaction

22
Turbulent Premixed Slot Bunsen Flame
Time integration 4th order R-K
Finite difference 8th order
Grid spacing 20µm
Time step 2ns
Case A Case C
Slot width H1.2mm H1.8mm
Jet velocity uj60ms-1 Uj100ms-1
Coflow velocity 15ms-1 25ms-1
Domain size 12Hx12Hx3H 13Hx12hx3H

• Constant Lewis number transport.
• 13 Species, non-stiff CH4-air reaction mechanism.
• Averages formed by integrating over the spanwise
  direction and time.

SL 1.8ms-1
dL 0.3mm
tf 0.17ms
R. Sankaran et al. 31D4. R. Sankaran et al. Proc.
C.I. 2007
23
Flame characteristics
Comparison of tt-1 and tc-1
Case A Case C
Ka(a/SL?k)2 2.3 5.2
Da(SLlt/dLu) 0.23 0.15
attf, x/Lx0.25 2.0 4.75
attf, x/Lx0.50 1.5 3.0
attf, x/Lx0.75 1.25 2.25

• Wrinkled/thickened flamelet regimes.
• Flame-flame interaction/pinch-off.
• tc-1 1/2Cf tt-1

24
Species timescale ratios ti-1/tO2-1 (case C)
x/Lx0.50
x/Lx0.25

• Lewis number
• effects?
• Damhöhler number effects?

Species LeO2/Lei
CO 1.01
OH 1.54
H2 3.72
H 6.35
x/Lx0.50
x/Lx0.75
25
ei transport equation case C
26
Model development

• Algebraic and transport models for ec exist which
  account for dilatation and propagation effects
  seen here focus on flamelet based models for
  intermediate species dissipation rates
• Develop model for tc/ti

• c and Y are available in the PDF method.
• Take conditional dissipation rate from laminar
  flame solutions at attf0.05, 1.5, 3.0.

• Improved estimate for ti
• Implicitly includes dilatation and propagation
  effects acting on the progress variable.

Denotes estimates based on laminar flame
structure.
27
Flame normal species gradients, x/Lx0.5
28
Ratios of species gradients, x/Lx0.5
OHO2
COO2
H2O2
HO2
29
Scalar alignment characteristics
Case A Case C

• Alignment is weak in the thickened preheat
  zone.
• Largest gradients of the intermediate species are
  aligned with progress variable
• 1D approximation reasonable.

30
Model predictions with low-strain flame solution
C
x/Lx0.25
x/Lx0.50

• attf0.05.
• Correct shape.
• Laminar strain rate can be adjusted to give
  quantitatively correct predictions

x/Lx0.50 case A
x/Lx0.75
Symbols DNS No Symbols model
31
Conclusions

• Confirmed roles of propagation and dilatation in
  thin/thickened flame mixing.
• Intermediate species gradients set by
  chemistry/dissipation balance.
• Species time scale ratios explained and modelled
  by laminar flamelet structure.
• Need to develop mixing models which correctly
  reflect the differing mixing rates in flame
  structures.