Away from the recirculation zone, LMIN and LMODE:

1
Experiments on the Autoignition of
EthyleneInjected Concentrically into Confined
Annular Jets of Hot Air
Christos Nicolaos Markides and Epaminondas
MastorakosHopkinson Laboratory, Department of
EngineeringUniversity of Cambridge
Objectives
Motivation

• Interest in the effect of in-homogeneities and
  turbulence on autoignition is both fundamental
  and practical (critical in HCCI engines,
  important in diesel/CI engines and LPP gas
  turbines, to be avoided in SI engines, storage of
  flammables, et.c.)
• DNS of turbulent mixing layers marginal
  propensity for earlier autoignition as the
  turbulence intensity (uair') is increased
• Non-premixed counter-flow experiments higher
  air temperature necessary for autoignition as
  uair' is increased
• Non-premixed co-flow experiments autoignition
  delayed as uair' is increased
• Engine (e.g. HCCI) research earlier
  autoignition as mixing is enhanced
• So What is the effect of turbulence on the
  autoignition of in-homogeneous flows?

• Non-premixed co-flow experiments in this
  apparatus with H2 and C2H2 injected into pure
  confined co-flows (as in Fig. 1, but w/out the
  bluff body), showed that as Uair (and hence
  uair') and/or Uinj/Uair were increased
• The mean autoignition length increased
  non-linearly, so that,
• The mean residence time until autoignition was
  delayed
• Thus Investigate a case for which uair'
  increases independently of Uair and Uinj/Uair
• In a practically relevant mixing configuration
  (akin to LPP premix ducts)
• Provide well-characterized data in a turbulent
  reacting flow in which the chemical and
  fluid-mechanical processes interact on the same
  scales that can serve as a challenging test-bed
  for the validation of advanced turbulence
  combustion models

Experimental Methods

• Air was heated up to Tair of 1100K and flowed
  upwards through a 3.0mm grid, around a bluff-body
  with Uair up to 40m/s and into a well-insulated,
  fully transparent quartz tube
• The tube was open-ended, so experiments were
  done at atmospheric pressure
• Two tube/bluff-body sizes were used, but the
  blockage ratio, (DBL/DIN)2, was kept equal to
  0.17
• The grid ensured turbulent flow for all
  conditions the macroscale Reair, based on the
  annular hydraulic diameter (DIN-DBL) and Uair was
  1400 3600
• The fuel was N2-diluted C2H4, with mass
  fraction of C2H4 in C2H4/N2 equal to 0.74
• Fuel was injected continuously and
  concentrically into the annular air jet behind
  the bluff-body with Uinj 10 80m/s, Uinj/Uair
  1.1 4.4 and Tinj in the range 710 900K
• Autoignition occurred in the form of repeated
  spotty flashes accompanied by a popping sound

Fig. 1. Apparatus Schematic. Mixing patterns for
illustration. Fig. 2 (above). Instantaneous
(1ms exposure) OH (31010nm) chemiluminescence
of autoignition, from left-to-right 1st pair
Tair 1059K, Tinj 822K, Uair 17.8m/s,
Uinj/Uair 2.5. 2nd pair Tair 1051K, Tinj
848K, Uair 13.2m/s, Uinj/Uair 3.1.
Results

• Away from the recirculation zone, LMIN and
  LMODE
• Decreased with Tair, and, increased with Reair
• Were found not to be sensitive to Uinj/Uair
• Everything else being the same, LMIN increased
  relative to the pure (no bluff-body) co-flow
  experiments that were associated with lower uair'
• When LMIN reaches the re-circulation zone, the
  highly intermittent Spot-Wake Interactions
  behaviour replaces the Random Spots

• LIGN measured optically in the
  continuous-behaviour Random Spots and
  Spot-Wake Interactions regimes
• For each run, i.e. set of Tair, Tinj, Uair and
  Uinj conditions, 200 instantaneous images, like
  those in Fig. 2, were used to compile three
  processed images Average, RMS and PDF, as shown
  in Fig. 3
• The minimum length from the RMS (LMIN) and most
  probable from the PDF image (LMODE), were
  correlated with the conditions, as shown in Fig.4

Fig. 3 (left). Average, RMS and PDF
post-processed images, calculated from 200 images
taken during constant conditions Tair 1066K,
Tinj 745K, Uair 19.2m/s and Uinj/Uair 3.2.
Also, RMS and Abel transform of RMS for Tair
1091K, Tinj 832K, Uair 38.2m/s and Uinj/Uair
1.5. Fig. 4 (above). LMIN as a function of Tair
for various Reair and Uinj/Uair in Random Spots
and Spot-Wake Interactions regimes.
Showing re-circulation region extending from the
injector to 1 3 DBL downstream. Fig. 5 (right).
LMIN time series in Random Spots and Spot-Wake
Interactions. Note the high intermittency
occurring approximately every 5 10 s, during
which the autoignition location shifts abruptly
to very short LMIN.
Conclusions
Further Work

• Further evidence has been obtained, by
  comparison with homogeneous and more weakly
  turbulent flows, that turbulent mixing (through
  uair') inhibits autoignition
• Turbulent mixing, even in this simple flow, can
  lead to extreme, possibly dangerous autoignition
  behaviour (here termed Spot-Wake Interactions)
  if the turbulence is strong enough (as it is in
  HCCI, Diesel/CI, LPP, SI)

• The discrepancy with the DNS can only be
  clarified if a link can be made between the
  variables Reair, Uair, Uinj/Uair and uair', and,
  the mixture fraction (?) and scalar dissipation
  rate (?) in the flows in which these experiments
  were performed (Fig. 1)
• Preliminary results from acetone PLIF
  measurements of these variables suggest that the
  delaying effect can be explained in terms ? and
  ?, but on a problem-specific basis

() cnm24_at_cam.ac.uk, http//www2.eng.cam.ac.uk/c
nm24/ http//www.eng.cam.ac.uk/em257/