Singleshot Raman measurements in diesel spray systems as a tool to differentiate twostage ignition f

1
Single-shot Raman measurements in diesel spray
systems as a tool to differentiate two-stage
ignition from single stage ignition

• Terry Parker, Chris Dryer, Manfred Geier,
  Jennifer Labs
• Engineering Division
• Colorado School of Mines
• Presented at
• American Chemical Society
• 233rd National Meeting Exposition
• March 26, 2007

2
Acknowledgements

• NSF, Army ARO, NASA
• Graduate student support, GANN award, Dept. of
  Ed.
• Custom drilling of injector nozzle, Raycon
  Corporation
• New injection system, Sturman Industries
• CSM contributors to the project
• Dr. Tom Grover
• Dr. Tony Dean

3
Diesel engines are a critical part of the
transportation sector of our economy

• Heavy duty hauling relies almost exclusively on
  the diesel
• The diesel provides long engine life and superior
  fuel economy
• Fuel tolerance for the cycle is typically greater
  than for spark ignition engines
• Maintenance for diesels is typically lower than
  for spark ignition engines
• The diesels advantages are due to
• Higher compression ratios
• The basic heterogeneity of the cycle which
  provides variable power with no throttling
  penalty
• As implemented today, the diesel is a much
  dirtier cycle than the spark ignition engine
• The Diesel, as implemented, is more efficient
  (less carbon emissions)

4
Overall emission levels for the diesel are THE
driving factor in diesel engine research

• The 2010 emissions targets are exceptionally
  aggressive
• Emissions for diesels are harder to control than
  those for the spark ignition engine
• Exhaust after treatment is difficult
• Soot is difficult to burn out
• Catalytic treatment of exhaust is ineffective
• Equivalence ratio range is inappropriate for NO
  reduction
• Catalysts are susceptible to plugging and
  deactivation
• The diesel cycle is plagued by the soot-NOx
  tradeoff

5
Significant improvement in emissions requires a
detailed understanding of diesel combustion

• CSM Research focus
• Experimental investigations of the fundamentals
  of diesel combustion
• Diesel spray measurements
• Post combustion exhaust measurements
• Line-of-sight temperature (time resolved)
• All have led to a unique view of NO destruction
• Critical question is ignition behavior in
  experiments
• Todays discussion
• Experimental capability apparatus and ignition
  dilema
• Sprays measurement results
• Nitric oxide measurement techniques and results
• Temperature measurement results
• Raman measurements resolution to dilema
• Conclusions

6
A unique facility is used to mimic the diesel
combustion environment

• Simulator is a cold-wall pressure vessel with a
  heated air core
• Hard wall interaction simulated with a plate on
  top of packed bed
• 1000 K and 50 atm capability
• System includes central air flow and side arm
  nitrogen flows
• limited oxidant used to control pressure rise in
  system
• central air flow used to control temperature
  gradients ( 15oC)

7
The simulator provides a controlled environment
for diesel engine research

• Temperature uniformity was ensured using constant
  temperature walls and the packed-bed heater
• Stainless steel plate acts as a hard wall for
  vapor jet to interact with
• Standard operation was at 873 K and 12.5 atm
• Maximum 1173 K, 16.5 atm
• Air and nitrogen flows used to maintain
  temperature and avoid window fouling
• Significant differences between an engine and the
  simulator
• Simulator is isobaric
• Total pressure for simulator is low
• 12.5 atm versus 30 to 100 atm.

8
Ignition Occurs During Injection Event
Combusting Spray

• Injection Start (0.0 ms)
• Ignition (2.2 ms)

• Injection End (3.05 ms)

Why the Delay
9
Droplet Size and Volume Fraction Measurements Use
a Pair of Infrared Lasers

• Lasers
• NdYAG (1.06mm) and CO2 (9.27mm) lasers are
  co-aligned and focused to a 150 mm waist
• Waist size is experimentally verified
• Scattering measurements
• 1.06mm at 90
• 9.27mm at 10
• Extinction measurements at both wavelengths
• Beam power monitored to compensate for power
  fluctuations

10
Scattering Measurements Provide Spatial
Resolution Within the Spray

• The spray event is exceptionally repeatable
• Direct comparison of events is possible
• Data Acquired
• Axially every 5 mm from 10mm to 50 mm
• Radially every 0.3 mm from centerline to spray
  edge
• All measurements acquired at 500 kHz for an event
• Cold spray and reacting spray data

Center Line
Axial Position
Radial Position
11
Spray Measurements show a jet with a high
velocity core and low momentum exterior

• Liquid Penetration Length
• Predicted 32 mm
• Measured 35 mm
• Spray Half Angle
• Predicted 2.8-4.3
• Measured 5.0

Higgins, B.S., C.J. Mueller, and D.L. Siebers,
SAE Paper No. 1999-01-0519. Wu, K.-J, C.-C.
Su, R. L. Steinberger, D. A. Santavicca, and F.
V. Bracco, Journal of Fluids Engineering
105406-413 (1983).
12
Combusting Volume Fraction Movie (t 0 ? 4.3 ms)

• Time averaged over 0.1 ms per frame ? 43 frames
  
• Look for
• Spray development
• High volume fraction area visible during steady
  state
• Loss of any structure after injection shut-off

13
Current View of Diesel Combustion

• Put forth by Flynn, et. al.

• Two Phases
• Premixed
• Diffusion
• Soot considered a product of fuel-rich central
  spray zones
• NO is formed in the
• Diffusion flame front
• Post-combustion hot gases

Flynn, P.F, R.P. Durrett, G.L. Hunter, A.O. zur
Loye, O.C. Akinyemi, J.E. Dec, C.K. Westbrook,
SAE Paper No. 199-01-0509.
14
Post Combustion Measurements Expected NO/CO2
Trend

• NO formation rate hypothesized to be constant for
  steady state diffusion combustion
• Low NO/CO2 levels for small injection events
• Larger fraction of fuel consumed in fuel-rich
  premixed combustion
• NO/CO2 level rising asymptotically to a constant
  NO production rate
• Majority of fuel burned in the diffusion phase

15
NO/CO2 for Simulator and Test Engine is Maximized
at an Intermediate Load
McCormick, R.L., M.S. Graboski, T.L. Alleman,
A.M. Herring, and K.S. Tyson, Environmental
Science Technology 35(9)1742 (2001)
16
Line averaged temperatures can be used to
monitor the temperature evolution of a system

• Temperature measurements rely on
  emission/extinction of soot

17
Temperature measurements indicate a very similar
temperature trend for very different NO levels

• Temperatures are very similar
• Crudely, this would imply similar NO levels

18
NO Production Does Not Scale with Temperature

• Results indicate that the production of NO in
  combusting diesel plumes is more complex than
  simply Zeldovich

19
Global equilibrium indicates that Diesels should
make high levels of NO

• Overall diesel combustion is fuel lean but
  majority of burn is at equivalence ratio 1
• Simple equilibrium indicates that current
  emission levels should be hard to meet

20
Injection and Combustion Time Scales Overlap IF
we understand ignition

• Ignition delay and start of secondary burn
  relatively constant
• Negligible NO/CO2 for injection events shorter
  than ignition delay
• Peak NO/CO2 production occurs when injection
  event begins to interact with major heat release
  event
• Steady state reached at long injection events

21
The ignition controversy two stage ignition and
the NTC region

• Classic shock tube and RCM work has shown a first
  and second stage of ignition for lower pressures
  in the 700 to 900 K range
• Two stage ignition Initial reactions produce
  radicals and then heat release as intermediate
  products are formed. The increase in temperature
  acts to shift the reaction pathways so that
  radical consumption is enhanced which slows the
  overall reaction down. Slower reactions rebuild
  the radical pool and after a time, vigorous
  reaction occurs.

R-H O2 gt R HO2 R O2 ltgt ROO

• Second reaction, net rate falls
• as T increases, shuts off overall reaction
• Observed ignition behavior
• Two stage ignition
• Ignition of all premixed fuel, mixing delay,
  diffusion burn

22
Ignition results as a function of pressure are
inconclusive

• At low pressures, ignition results indicate
  two-stage ignition
• At 12.5 atm., results are inconclusive

23
RAMAN signals are wavelength shifted from the
pump beam and are weak

• System setup
• 100 mJ _at_ 355 nm, imaging spectrometer with gated
  intensified camera
• Wavelength shifts

24
Signals from pressurized cold airand CO2 provide
the basis for interpretation

• Vignetting correction is applied to spatial
  dimension
• Parabolic smoothing is applied across 39 pixels
  in in space and 5 pixels in wavelength
• Results are as expected and show good quality in
  a simple well understood field

Distance along Beam
25
Continued investigation illustrates expected
trends at elevated temperatures

• At high temperatures (873 K) signal falls as
  number density scales with inverse T
• Nitrogen normalized signals
• subtle spectral changes
• identifying temperature via spectral shape
  probably not viable

26
To investigate Ignition, signals must be
acquired at appropriate time and spatial position
post-burn Mid-burn Pre-burn

• Times are pre, mid, and post burn
• Spatial location is 5 mm from tip and at spray
  edge
• Avoid plasma formation due to laser interaction
  with spray

27
Preburn results show a broadband fluoresence
that complicates signal processing

• Nitrogen and C-H signals clearly identifiable
• Broadband fluorescence in regions that we expect
  to see fuel

28
Post Burn Results indicate spatial variation,
C-H, and CO2

• Signals at edge are dominated by nitrogen and
  oxygen
• Central regions also include CO2 and CH (unburned
  HC)
• Carbon Dioxide signals are weak

C-H
Nitrogen
29
Mid burn results indicate full combustion

• Signals are elastic, CO2, O2, CO, N2, C-H
• Central region shows evidence of burning
• CO signal reasonably strong
• CO and CO2 indicate combustion with final
  products
• First pressure pulse is true ignition

center edge
center edge
30
Ignition results indicate simulator provides
true ignition during the spray

• NO/CO2 expected to be constant at steady state
• NO/CO2 levels begin to fall for both simulator
  and test engine experiments

• There is no kinetic pathway to destroy NO in air
• From examining the data, it is postulated that
• Products of combustion re-entrained into
  hydrocarbon rich plume
• Nitric oxide destroyed via NO re-burn
• Reactions between NO and unburned hydrocarbon
  radicals
• Radicals present in the interior of jet
• Initially, formation/destruction is transient

31
Conclusions

• Raman signals confirm the observed ignition
  behavior is not two stage ignition
• NO in the simulator does not follow expected
  trends