Laser Measurements for High-Pressure Combustion:

Challenges and Opportunities

Robert P. Lucht School of Mechanical Engineering

Purdue University West Lafayette, Indiana

Presentation at the Workshop on Techniques for High-

Pressure Combustion Argonne National Laboratories

August 30, 2011

Acknowledgments

•  For dual-pump CARS, funding from NASA Glenn under Cooperative Agreement Number NNX07AC90A , technical discussions with Drs. Yolanda Hicks, Clarence Chang, and Randy Locke. For fs CARS, funding from DOE Office of Basic Energy Sciences, AFOSR and NSF.

•  Graduate students: Mathew Thariyan, Ning Chai, Daniel Richardson, Aizaz Bhuiyan

•  Research staff: Scott Meyer, Yu Matsutomi, Sameer Naik

•  Colleagues: Profs. Hukam Mongia, Jay Gore (Purdue), Sukesh Roy and Waruna Kulatilaka (Spectral Energies), Jim Gord (AFRL)

Outline of the Presentation •  Optically Accessible Gas Turbine Combustor

Facilities

•  Dual-Pump CARS Measurements: Challenges and Optical System

•  Femtosecond CARS for Single-Shot Temperature at 5-10 kHz Data Rates

•  Laser System for 5-10 kHz PLIF and PIV, 5 kHz OH PLIF Measurements

•  Conclusions

Purdue Gas Turbine

Combustion Facility (GTCF)

Laser Laboratory10 Hz Nd:YAG/Dye5 kHz Nd:YAG/Dye

ControlRoom

FTIRandH

FID

Natural GasCompressor System

Non-VitiatedInlet Air Heater

OpticalTable

Aviation

GTTestRig

PowerGTTestRig

Laser Duct

CoolingWaterandFuelPumps

Purdue Gas Turbine Combustion Facility (GTCF)

High

Pressure Lab System

Maximum Flow

Capacity

Max Operating Condition

Natural Gas Heated High Pressure Air

9 lbm/sec 700 psi / 540 deg C 1000 deg F

Electric Heated Air or Nitrogen

1 lbm/sec 600 psi / 600 deg C

Nitrogen 2 to 5 lbm/sec 1,500 psi

Liquid Aviation Fuel (Kerosene)

1 lbm/sec/tank (2 tanks)

1,500 psi

Natural Gas 1 lbm/sec 3500 psi

NASA 9-Point LDI Assembly (Top-Hat) •  Nine simplex injectors arranged at throats of

nine converging-diverging venturis in a 3 x 3 arrangement.

•  Axial swirlers with helical vanes at 60° impart swirl to incoming heated air.

•  Only central injector used for testing.

•  Conventional “Single-Pump” CARS

•  Noninvasive •  Coherent Laser-Like

Signal •  Spatially and

Temporally Resolved •  Excellent Gas

Temperature Data (especially at higher temperatures)

ωpump

ωStokes

ωCARS

ωV

ωpump

ωpump

ωStokes

ωCARS

ωV

ωpump

ωStokes

ωPump1

ωPump1

ωCARS

Coherent Anti-Stokes Raman Scattering (CARS)

Dual-Pump CARS of N2/CO2

Overall Experimental System

Measurement Challenges in GTCF

•  Translation of probe volume inside the flame zone.

•  Installation of pin-hole for spatial overlap of CARS beams not possible, must use referemce leg alignment.

•  Measurement of non-resonant signal in the reference leg for spectral normalization of CARS signal.

•  Safety of thin window, CARS beams are focused tightly in the middle of the test section.

l/2 SPEX

l/2 Pol.

Pol.

l/2

Pol.

S

S

S

Main Leg

Reference Leg CL

ZL

Flow

Entrance Translation

Stages

Exit Translation

Stages

BD

BD BD

BD BD BD

A: Aperture BD: Beam Dump CL: Camera Lens IF: Interference Filter L: Lens l/2: Half-wave-plate Pol: Polarizer S: Shutter TS: Translation Stage ZL: Zoom Lens

L5

L2 L1

L3 L4

Probe Volume

TS

TS

Probe Volume

A

A

Window Assembly

IF

608 nm 532 nm 561 nm

496 nm

Optical System near GTCF

Raman Shift (cm-1)

1300 1320 1340 1360 1380 1400

(CA

RS

Inte

nsity

)1/2 (a

rb. u

nits

)

-15

-5

5

15

25

35

45

55

65

75

DataTheoryResidual

T = 1325 KCO2/N2 = 0.079 N2

CO2

N2

Comb. Pr. = 100 psia., Eq. Ratio = 0.8

Operating Conditions, Measurement Locations and Sample DP-CARS Spectra

Φ=0.4 Φ=0.59 Φ=0.80 Φ=1.0

100 psia (7.0 atm.)

■ ■ ■ ■

125 psia (8.5 atm.)

■

150 psia (10 atm.)

■

+ 3 mm

+ 6 mm

+ 9 mm+ 11 mm

- 3 mm

- 6 mm

- 9 mm- 11 mm

F = 0.6P = 100 psiDP/P = 4%

10 mm

Note: Distance between points along the centerline is 5 mm

•  Burner Inlet Temperature: 850 0F (725 K) •  Fuel: Jet-A •  Normalized injector pressure drop = 4%

Measurement Grid for DP-CARS

Flame Characteristics @ 100 psia

Φ =0.8 Φ =1.0

Φ =0.4 Φ =0.59

Temp PDFs Along Centerline Temperature PDF: Z = 15 mm, R = 0 mm

Cou

nts

Temperature (K)1000 1250 1500 1750 2000 2250 2500 2750

0

20

40

60

80

100

120

140

160Mean Temp. = 1820 K

Std. Dev. = 340 K(b)

Temperature PDF: Z = 10 mm, R = 0 mmC

ount

s

Temperature (K)1000 1500 2000 2500 30000

20

40

60

80

100Mean Temp. = 2020 K

Std. Dev. = 370 K(a)

Combustor Pressure: 104 psia, Equivalence Ratio: 0.4

Temperature PDF: Z = 25 mm, R = 0 mm

Cou

nts

Temperature (K)

1000 1200 1400 1600 1800 2000 2200 24000

20

40

60

80Mean Temp. = 1595 K

Std. Dev. = 267 K(a)

Temperature PDF: Z= 50 mm, R= 0 mm

Cou

nts

Temperature (K)

1000 1200 1400 1600 1800 20000

20

40

60

80

100

Mean Temp. = 1340 KStd. Dev. = 180 K

Temperature and CO2/N2 Correlation Plot Combustor Pressure: 150 psia., Equivalence Ratio: 0.48

Adiabatic Equilibrium

Accomplishments and Conclusions

  A new OPO/PDA system was used to generate the 560-nm pump beam in the dual-pump CARS system. Considerable care in laser system alignment was required to obtain good beam quality in the combustor test cell.

  The Zaber translation stages performed well, alignment was maintained over the entire spatial region of interest during the test.

  The reference leg was invaluable for alignment and for frequent recording of the nonresonant signal. Alignment was maintained before and after translation of the large 2-inch prisms.

Accomplishments and Conclusions

  Estimated uncertainty in temperature measurements :   Accuracy: 1-2%   Precision: 2-3%

  Uncertainty in CO2/N2 ratio measurements :   Very dependent on CO2 concentration and on the

temperature, approximately 10% relative standard deviation in the range of 5% CO2 concentration around 1500 K.

  Probe volume dimensions:   500-600 µm (FWHM) along the laser propagation

direction.   50 µm perpendicular to the laser direction.

Modified Combustor Window Assembly   Cross section increased from 3"x3“ to 4.2"x4.2". The

modified CWA is fabricated from Hastelloy-X instead of stainless steel. Brazing has been eliminated. Film cooling air passages are incorporated in the injector assembly rather than in the CWA. Thermal barrier coatings are being applied to the window assembly inner surfaces.

  Upstream spool section has been redesigned to accommodate the larger injectors and to ensure uniform flow into the injector.

  Downstream spool sections redesigned for larger flow cross section.

Test Rig with Modified Combustor Window Assembly (CWA) Installed

10-5

10-4

10-3

10-2

10-1

100

-1 0 1 2 3 4 5

300K900K1500K2100K

CA

RS

Sig

nal (

a.u.

)

Time (ps)

Chirped Probe Pulse

10-5

10-4

10-3

10-2

10-1

100

-1 0 1 2 3 4 5

300K900K1500K2100K

CA

RS

Sig

nal (

a.u.

)

Time (ps)

Chirped Probe Pulse

10-5

10-4

10-3

10-2

10-1

100

-1 0 1 2 3 4 5

300K900K1500K2100K

CA

RS

Sig

nal (

a.u.

)

Time (ps)

Chirped Probe Pulse

Single-Shot CPP fs CARS

+1 ps Probe Delay Nonresonant spike

20 Single-Shots with Chirped Probe Pulse

0

200

400

600

800

1000

17100 17200 17300 17400 17500

CARS

Sig

nal (

a.u.

)

Frequency (cm-1)

t0,pr

=1.0 ps

Φ = 0.5

Temperature Histograms from Single-Shot fs CARS in Flames

0

0.2

0.4

0.6

0.8

1

17100 17250 17400

Theory 1602 KExperiment

CARS

Sig

nal (

a.u.

)

Frequency (cm-1)

t0,pr

=0.5 ps

0

50

100

150

200

1510 1535 1560 1585 1610 1635N

umbe

r of L

aser

Pul

ses

Best Fit Temperature (K)

Tmean

= 1592 KT

adiab=1640 K

Std Dev= 10.5 K

• New laser system from Coherent delivered to Purdue (AFOSR DURIP Program).

• Laser rep rate of either 5 kHz with 2.6 mJ per pulse or 10 kHz with 1.0 mJ per pulse.

• Pulse durations of either 60 fs or 30 fs. • Pulse shaper integrated into the system

between the oscillator and amplifier.

Single-Shot Temperature Measurements at 5 kHz

Single-Shot Temperature Measurements at 5 kHz: Coherent Ultrafast Laser System

Silhouette 128-Pixel MIIPS Pulse

Shaper

Legend EliteTi:S Regenerative Amplifierand Single-Pass Amplifier

800 nm13 W at 5 kHz10.5 W at 10 kHz

Evolution-HENd:YLF Pump Laser

527 nm90 W 10 kHz

OPerA Solo Optical ParametricAmplifier + Nonlinear Crystals

200 nm - 10 µm

Mantis Mode-Locked Ti:SLaser with IntegratedOPSL Pump Laser

800 nm

• Demo camera systems loaned by Princeton Instruments and Andor.

• Princeton Instruments system: 512x512 chip, physical mask, spectral acquisition from 66 rows by 512 columns, 5 kHz or 10 kHz spectral acquisition with charge in columns vertically binned.

• Andor system: 1024x1024 chip, optical mask, 50 rows by 1024 columns, 5 kHz spectral acquisition with charge in columns vertically binned, also have 512x512, 10 kHz system.

Single-Shot Temperature Measurements at 5 kHz

Single-Shot Temperature Measurements at 5 kHz

Stokes Beam -ω2800 nm, 60 fs, 70 µJ

Pump Beam -ω1674 nm, 60 fs, 50 µJ

Probe Beam -ω3800 nm, 60 fs, 90 µJ

Delay Linefor Probe

CARS Signal Beam -ω4Turbulent Flameor Gas Cell

Chirped Probe Pulse2-3 psec

Raman Coherence

t

DispersiveRod

60 cm SF11

To Spectrometerand EMCCD

Single-Shot Temperature Measurements at 5 kHz

ω1 ω4

Hencken Burner Calibration Flames

0

50

100

150

200

250

700 800 900 1000 1100 1200 1300 1400

Cou

nts

Best Fit Temperature (K)

Tadiab

= 1210 KT

avg = 1112 K

σ = 95 K

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 1129 K

0

100

200

300

1600 1700 1800 1900 2000 2100 2200

Cou

nts

Best Fit Temperature (K)

Tadiab

= 1990 KT

avg = 1956 K

σ = 68 K

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 1945 K

Single-Shot Temperature Measurements at 5 kHz

Single-Shot Temperature Meas at 5 kHz: Nonresonant Background Suppressed

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 1788 K

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 1204 K

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 845 K

0

0.5

1

14700 14900 15100

CA

RS

Sig

nal (

a.u.

)

Frequency (cm-1)

Single-shot fitT

fit = 332 K

Single-Shot Temperature Meas at 5 kHz: Nonresonant Background Suppressed

500

1000

1500

2000

0 50 100 150 200 250 300 350 400

Bes

t Fit

Tem

p. (K

)

Time (ms)

Conclusions: Fs CARS

•  Temperature determined from single-shot fs CARS N2 spectra recorded at 1 and 5 kHz from laminar, unsteady, and turbulent flames.

•  Signals are very strong and reproducible from shot to shot. Precision comparable to or better than the best single-shot ns CARS systems.

•  Theoretical model developed to fit CPP fs CARS spectra for temperature, accuracy of measurements can be improved. Main issue is incomplete characterization of the laser beams.

•  CPP fs CARS concentration measurements in progress. Data analysis issues similar as for measurement of temperature.

Future Work: fs CARS

• Temperature measurements in turbulent flames. • Comparison of measurements with and without

background suppression. • Concentration measurements are still an open

issue. • More accurate characterization of the ultrafast

laser pulses. • Measurements in high pressure flames. • Electronic resonance enhancement.

High-Repetition-Rate Diagnostic Techniques Based on Diode-Pumped

Nd:YAG Lasers

  10 kHz PIV – dual-head Edgewave laser

  10 kHz OH PLIF – Credo dye laser pumped by the Edgewave laser

High-Repetition-Rate Laser System

Edgewave Diode-Pumped Solid State Nd:YAG Laser: 5 kHz Rep Rate, Dual-Head, 6 mJ/Pulse at 532 nm, 7 nsec Pulses

Sirah Credo Dye Laser 5 kHz Rep Rate, 500 mJ/Pulse at 283 nm (2.5 W average power in UV)

UV Camera Lens

10 kHz UV Intensifier

Phantom CMOS Camera

TOPVIEW

SIDEVIEW

283 nm Laser Sheet 500 µJ

•  Variety of gelled and liquid droplets have been investigated – methanol and JP8 fuels, HPC and aerosil gellants

•  Drop sizes on the order of a few mm

•  Combustion process is much more dynamic for the gelled droplets

•  Effects of pressure investigated in optical cell

5 kHz OH PLIF of Liquid and Gelled Droplet Combustion

5 kHz OH PLIF of Liquid and Gelled Droplet Combustion

5 kHz OH PLIF of Liquid and Gelled Droplet Combustion

5 kHz OH PLIF of Liquid and Gelled Droplet Combustion

High-Pressure Diagnostic Techniques •  Measurements in high-pressure flames complicated

because of the experimental apparatus (windows, etc.) and because of increasing collisional rates (increased quenching, absorption line widths). High-rep-rate PLIF and PIV will be extremely valuable for turbulent flames.

•  CARS has been demonstrated in high-pressure flames systems for temperature and concentration. Fs CARS offer the potential for high-rep-rate single-shot measurements in high-pressure systems.

•  Dual-resonance techniques like electronic-resonance-enhanced CARS or two-color polarization spectroscopy may help to overcome interferences due to overlapping absorption lines from different species.