dual-pump cars and oh plif pi: andrew cutler

National Center for Hypersonic Combined Cycle Propulsion

1

Dual-Pump CARS and OH PLIF

PI: Andrew CutlerPhD Students: Gaetano Magnotti, Luca Cantu, Emanuela Gallo

The George Washington University

CollaboratorsPaul Danehy (ASOMB), Craig Johansen (NIA)

NASA Langley Research Center

June 16, 2011

1



2

Background

• CFD methods employing semi-empirical models used in analysis of hypersonic airbreathing engine flow paths

– RANS (or Favre averaged) codes have models for turbulent stresses, mass and energy transport, turbulence chemistry interactions

– LES methods have models for subgrid scale turbulence

• Models depend upon experimental data for validation– Information on mean flow and statistics of the turbulent fluctuation in flow

properties

• Data requirements– Simple well defined supersonic combustion flows with well-known boundary

conditions– Time and spatially resolved– Good instrument precision– Converged statistics

• Approach– Dual-pump CARS

• T, mole fraction species

– Planar laser-induced fluorescence imaging of OH radical (PLIF)• Flow-visualization


3

Flow Fields

Scramjet combustor

University of Virginia’s Dual-Mode Scramjet

Supersonic free jet flames (hot center jet with H2 co-flow)

Laboratory supersonic flame with 10 mm center jet


4

Dual-Pump CARS

• Four-color mixing process at beam focus / intersection– Two narrowband pump (Green+Yellow) + broadband Stokes (Red) laser– Probe molecular Raman transitions– Two simultaneous processes

• Coherent signal beam (Blue)– Analyze with spectrometer– Fit spectra to theory to obtain T, mole fraction species

• Measurement time ~10 ns, volume ~50 µm x 1.5 mm

Rot.-vib. states

Virtual states

P1

S

P2

AS

P2

S

P1

AS

Process 1 (N2, H2) Process 2 (O2, H2)


5

“Mobile” CARS System

• Laser cart– Nd:YAG (532 nm,1.2 J, 20 Hz, seeded) =

Green– YAG-pumped broad-band dye laser (~9.6

nm or 263 cm-1 at 605 nm) = Red– YAG-pump narrow-band dye laser (~ 0.07

cm-1 at 553 nm) = Yellow– Remote beam steering (picomotor mounts)– Control of beam size (telescopes) and

energy (polarizers/waveplates)

• Detection system– Focusing lens, polarizer, remote beam

steering– 1 m monochromator– Cooled CCD

Laser cartTransmission Collection Detection

Beam Relay System (BRS)

Narrowband and broad-band dye laser beams on top of each other

Flame


6

Beam Relay System at UVa• Moves beam crossing within flow field; 3 components, typically 2

motorized

Detection

Beams from laser cart

Mot

ion

Motion

Mot

ion

Linear bearings/rails*

*Horizontal motions are motor driven

5’x8’ optical table

Dual mode combustor

Measurement point

Transmission

CollectionCARS window


7

Transmission and Collection Optics

• Laser beams focused and combined with separate mirrors and lenses• Focal plane imaging system

– Beam focusing– Beam crossing

From BRS and Cart

Focal plane imaging system: microscope objective and CCD

camera

To BRS and Detection

Measurement volume

Multi-pass dichroic splitter

Signal beam

Lenses

ND filter

Uncoated glass splitter

Focal plane images


8

Optical Implementation at UVa

Transmission optics platform

Monochrometer


9

Method for Fitting Spectra

• Sandia CARSFT code– Computes theoretical spectra

• New algorithm* fits spectra by interpolating from library of pre-computed spectra

• Novel feature is structure of library (sparsely packed) and method for interpolation of spectra from library– Smaller libraries, faster library generation, faster fitting

• Allows fitting of more chemical species and faster turn around– Enables “WIDECARS”

*Cutler, A.D., Magnotti, G., J. Raman Spectroscopy, 2011.

no. chem

species

sparse library size

regular library size

size ratio

1 10 10 1.002 60 100 1.673 270 1000 3.704 1080 10000 9.265 4050 100000 24.696 14580 1000000 68.59


10

Measurements in Hencken Burner

• Hencken flat flame burner– H2-air flame– Flame temperature and composition known from theory

• Computed from gas flow rates assuming adiabatic reaction to equilibrium

Average DP CARS spectra at 3 equivalence ratios

Hencken burner10

100

1,000

10,000

100,000

0 200 400 600 800 1,000 1,200

Pixel

Inte

nsity

(cou

nts)

H2 S(9)

H2 S(5)O2

N2

H2 S(6)

=2.34 =1.17 =0.23


11

10

100

1000

10000

100000

2050 2100 2150 2200 2250 2300 2350

Wavenumber/cm-1

CA

RS

In

te

ns

ity

/Arb

itr. U

nit

s

Theory

Experiment

10

100

1000

10000

2050 2100 2150 2200 2250 2300 2350

Wavenumber/cm-1

CA

RS

In

te

ns

ity

/Arb

itr. U

nit

s

Theory

Experiment

10

100

1000

10000

100000

2050 2100 2150 2200 2250 2300 2350

Wavenumber/cm-1

CA

RS

In

te

ns

ity

/Arb

itr. U

nit

s

Theory

Experiment

10

100

1000

10000

2050 2100 2150 2200 2250 2300 2350

Wavenumber/cm-1

CA

RS

In

ten

sit

y/A

rb

itr. U

nit

s

Theory

Experiment

(a)

(b)

(c)

(d)

Typical spectra normalized and fitted

(a)=0.23 (single shot)

(b) =1.17 (single shot)

(c) =0.23 (mean)

(d) =1.17 (mean)

O2

N2

H2


12

Fitted Temperature in Hencken Flame

• Means agree well with theory• Actual standard deviation (SD)

in flame due to unsteadiness is unknown but believed small

• Fitted parameters have noise– Mode noise in broad band dye

laser, camera noise, photon shot noise

• SD of fitted parameters depends on selection of residual minimized by fitter

– R2 = noise-weighted least squares fit to signal intensity (Snelling et al, 1987)

– (R1, R3 commonly used in the literature)

• We found R2 < R3 < R1 • Similar results for fitted N2, O2,

H2 mole fractions

0

500

1000

1500

2000

2500

0 0.5 1 1.5 2 2.5 3 3.5

Equivalence ratio

Te

mp

era

ture

(K

)

Standard deviation of fit x10

Fitted mean

Theory

R1

R3

R2

SD in T reduced x2 by proper selection of residual! Important

for turbulence studies


13

Saturation Effects

• Signal Ip1 Ip2 IS d2 L2

– Want to minimize L– Must maximize Ip1 Ip2 IS to maintain

signal – Saturation effects limit Ip1 Ip2 IS

• Saturation effects studied in Hencken flame– Mixture of Stark shift and

stimulated Raman pumping

• Fitted mole fraction more sensitive than temperature– Error in O2 up to 30%– H2 most sensitive to saturation– Saturation thresholds determined– We now know how to avoid

saturation

O2

N2

0.1

0.11

0.12

0.13

0.14

0.15

0.16

1 10 100 1000 10000 100000 1000000

O2

mo

le f

ract

ion

Ip2*Is

Ip1=40 GW/cm2

Ip1=140 Gw/cm2

850 mm f.l 500 mm f.l 300 mm f.l

(GW2/cm4)

0.1

0.11

0.12

0.13

0.14

0.15

0.16

1 10 100 1000 10000 100000 1000000

O2

mo

le f

ract

ion

Ip2*Is

Ip1=40 GW/cm2

Ip1=140 Gw/cm2

850 mm f.l 500 mm f.l 300 mm f.l

(GW2/cm4)

0

2

4

6

8

10

2150 2200 2250 2300

Sq

rt o

f C

AR

S i

nte

ns

ity

(arb

itra

ry u

nit

s)

Raman Shift (cm-1)

Ip2=100 GW/cm2

Ip2=400 GW/cm2

Saturation effects in Hencken burner flame = 0.3

Mean spectra

Fitted O2 mole fraction vs. irradiance


14

Status of CARS

• CARS data base in laboratory supersonic flame– Data acquisition completed– Data analysis in progress

• CARS system currently installed at UVa


15

Supersonic Flame*• Center jet 10 mm, hydrogen vitiated air,

variable Mach (0.75-2.0) and temperature• Unheated, low-speed coflow • 10 test cases; variables

– Nozzle exit Mach number (Me)– Center jet temperature (equivalent flight Mach

Mh)– Coflow gas

• C = H2 coflow, combustion• M = N2 coflow, mixing but no combustion

• Nominally 100,000 spectra for each test case, fitted for T, and mole fractions N2, O2, H2

– Sampled at 5 axial locations, 23-30 radial locations at each

• Will be used by NCHCCP modelers

Me/Mh 5 6 70.75 C+M

1 C+M1.6 C C+M C+M2 M

Exit Mach number (Me)

Equivalent flight Mach number (Mh)

*This work mostly supported by NASA Hypersonics NRA grant (NNX08AB31A)


16

Typical Results in Supersonic Flame

0

400

800

1200

1600

2000

2400

-20 -10 0 10 20

1

15

35

65

100

0

400

800

1200

1600

2000

2400

-40 -30 -20 -10 0 10 20 30 40

2

16.2

36.6

63.4

98.9

Mixing (coflow = N2)Me=1.6Mh=7

Combustion(coflow = H2)Me=1.6Mh=6

Plots of mean T vs. radial distance at several axial locations

Center jet diameter = 10 mm, surveys are at ~1, 15, 35, 65, and 100 mm


17

“WIDECARS”*

• Enable DP CARS in C2H4 + H2 – air flames

• New broadband Stokes laser– 2x increase in spectral width allows

additional chemical species to be simultaneously probed

• Temperature, mole fraction N2, O2, CO2, C2H4, H2 and CO measured at 300 K.– Most species ever simultaneously

measured with CARS!– Fit with new software

• Future development required– Validate at flame temperature– Modeling/calibration for C2H4

Stokes laser Spectrum

N2

CO2

CO2

CO

C2H4

H2 S(3)

H2 S(4)

3% CO2, 5% CO, 5% C2H4, 25% H2, 62% N2

Demonstration in a room T gas cell

*Collaboration with Tedder and Danehy (LaRC)Tedder et al., Appl. Optics, 2010


18

OH PLIF Preliminary Setup

• Modified CARS laser cart to produce UV light needed for OH PLIF– Installed doubling crystal– Optimized laser for power

• Performed OH PLIF measurements in laboratory– Set up sheet forming optics– Set up laminar OH combustion with water welder– Investigated camera settings– Determined optimal PLIF transitions

• Flow visualization only– Signal roughly ~ OH density (not calibrated)


19

Tunable Laser

Planar Laser-Induced Fluorescence (PLIF)

CCD camera detects

Laser sheet excites molecules

Excited molecules fluoresce

Gas flowGround state

Excited state

LIF ~ nOH

H2 O2

Water welder provides reactants for combustion

λ =

281

.135

nm

A2∑+←X2ΠΩ(1,0)


20

Ongoing Work and Future• Summer 2011

– Acquire CARS data bases in UVa Dual Mode Scramjet Configuration B (no isolator) and C (with isolator)

– OH PLIF flow visualization (same cases)

• 2011-2013– Develop and validate “WIDECARS” for C2H4 (+ H2) in Hencken

flame – Acquire CARS data base supersonic flame with C2H4 + H2 coflow– Acquire CARS/PLIF data base in UVa scramjet with ethylene

flame/cavity flame holder

Combustor ExtenderTDLATCombustor ExtenderTDLAT

ExtenderCombustorIsolator TDLAT ExtenderCombustorIsolator TDLAT

Config. C

Config. B


21

Summary• Motivation• Dual pump CARS• “Mobile” system and implementation at UVa• Method of fitting spectra• Measurements in a Hencken burner (H2-air) and

supersonic jet• Saturation effects• Measurements in a supersonic flame• WIDECARS• OH PLIF setup and preliminary data• Ongoing and future work

dual-pump cars and oh plif pi: andrew cutler

Documents