Combustion Physics (Day 4 Lecture) Chung K. Law Robert H. Goddard Professor Princeton University
Princeton-CEFRC-Combustion Institute Summer School on Combustion June 20-24, 2016
1 Day 4: Laminar Premixed Flames
1. The standard premixed flame 1. Phenomenological and asymptotic analyses 2. Parametric dependence 3. Chemical structure 2. Limit phenomena 1. The S-curve concept 2. Extinction through volumetric heat loss 3. Aerodynamics of flames 1. Hydrodynamic stretch 2. Flame stretch 3. Flamefront instabilities 2 1. The Standard Premixed Flame
3 Structure of (Standard) Premixed Flame
1. Hydrodynamic, Rankine-Hugoniot level (strong discontinuity): o Flame sheet o Uniform upstream and downstream states 2. Reaction-sheet level (weak discontinuity): o Additional preheat, diffusion zone of
o thickness D : reaction frozen due to large Ta , diffusion-convection controlling 3. Complete structure
oo o Additional reaction zone: RD for large activation energy o Convection negligible relative to diffusion due 0 to small R , diffusion-reaction controlling • System is conservative
o o o c()T o T q Y f u u u u b u b , p b u c u 4 Flame Characteristics (1/5)
• Characteristic temperature change across reaction zone
(T ) [ TTxoo ()]~[/( wdwdT / )] ()/ T2 T R b fo b a T b w ~ exp( T / T ) a • Continuity of heat flux through preheat and reaction zones
oo()() TT2 R~ R b Ze 1 1 . (7.2.5) o o o DTTTTT b u() b u a
5 Flame Characteristics (2/5)
• Convection and diffusion balance in preheat zone:
ood d d d f~ [(/ cpp ) ], f ~(/ c ) d x d x d x d x
/ c p f o ~. o (7.2.6) D • Overall mass flux conservation:
Reactant mass flux entering flame ( o ) = Yfu Reaction flux through reaction zone ( oo ) Ywu b R
o o o (7.2.7) fw~ bR
6 Flame Characteristics (3/5)
o 표 • Solving for f and ℓ퐷 from (7.2.6) and (7.2.7), using (7.2.5) o (/) cwpb (/) c (f o )2 ~, o 2 p (D ) ~ Ze (7.2.8, 9) Ze w o b • Results show three fundamental quantities governing flame response
o /cp : diffusion o o w b : reaction o o Ze : activation (Ta) and exothermicity (Tb )
7 Flame Characteristics (4/5)
o o • f ~ ( / c p ) w b : propagation rate, which is a response of the flame, is the geometric average of the diffusion and reaction rates, which are the driving forces in forming the flame.
• Dependence on transport: o o Nonpremixed flame: f ~ / c p ; diffusion dominating o o Premixed flame: f ~ / c p ; “diluted” by reaction • (7.2.8) and (7.2.9) can be alternately expressed as oo o fc Dp ~/ : depends only on transport
f o w o o ~ b : depends only on reaction o D Ze 8 Flame Characteristics: Summary (5/5)
• Since there are only two controlling processes (diffusion w o and reaction: / c p , b ), flame characteristics are described by two independent relations, which can be expressed in three different ways to convey different messages o Balance of processes:
/ c p fo~ , f o ~ w o o o bR D o Explicit expressions for the responses:
o (/)(/)cp w b c p (fZoo e )22 ~ , ( ) ~ D o Z e w b o Explicit dependence on individual processes: f o w o fcoo~ / , ~ b Dpo D Ze 9 Specific Dependence on Pressure
on • w bp ~ pc ; / pressure insensitive (n 1 ) o o1 / 2 o 1 / 2 n / 2 o o o 2 o f ~[(/)]~() cp w b w b ~ p , suu~ f / ~ f / p ~ p
oo~ ( /c n / w )1 / 2 ~ p / 2 D p b • Implications:
o o For n = 2: s u f () p ; cancellation between density and reaction; this is not a fundamental result f o o o For 0 < n < 2: increases and s u decreases with increasing p o Dependence of o on p is through reaction, not D diffusion 10 Asymptotic Analysis
11 Governing Equations (Le =1) (1/2)
d T d2 T fo c q w , • Dimensional: pc d x d x 2
d Y d2 Y fDwo d x d x 2 d2 T d T o TTa / • Nondimensional: D aC Y e , d x2 d x (7.3.4)
d2 ()() T Y d T Y 0;Le 1 d x2 d x (7.3.5’) / c o o p ~ f Da C B C , x x o 2 f / c p
x : T T , Y 1, • B.C. : u o x : T Tb , Y 0 , d T d Y x : 0 . d x d x
12 Governing Equations (2/2)
• Integrating (7.3.5’) once; with b. c. at x ,
d() T Y (TYT ) ( 1). u (7.3.10’) dx x TTo 1 T ˆ o Evaluating (7.3.10’) at yields Tad : bu a d • Integrating (7.3.10’) again:
ˆ ()T Y To c e x T o bb 1 for boundedness at xˆ
ˆ o • Substituting YTT b into (7.3.4) yields the single equation that needs to be solved:
d2 T d T oo TTa / D aCb T T e (7.3.13) d x2 d x
o This is why TY ˆ ˆ is called a (de-)coupling function
13 The Cold Boundary Difficulty
d2 T d T o TTa / • Evaluating D aC Y e , (7.3.4) d x2 d x at the x freestream, where d2 T d T o 0 , 00 LHS d x2 d x o YTTRHS1,0 Thus the governing equation is unbalanced ill posed • Difficulty exists for many steady-state problems with premixture at ambience: reactive ambience has infinite time to react all reactants would be reacted before arrival of flame unphysical posing of problem • Recourse o Artificial suppression of reaction term at x o Asymptotic analysis: Rational freezing of reaction at due to large activation energy. 14 Distinguished Limit (1/2)
• Asymptotic analysis capitalizes on the largeness of activation energy which localizes reaction to a thin zone.
TT/ • Consider reaction rate: a w~ D aC Y e
• For T a 1 and Dac푌 fixed, w 0 • No reaction in the reaction zone, obviously wrong!
• Thus the limit T a must be taken rationally
15
o Thus the limit of must be taken rationally Distinguished Limit (2/2) • Distinguished limit:
T requires D aw such that is fixed . ac o TTab/ • Express D aC ~ D a e
Then 11 w~ D a e x pe x T p D a Z e To T ab TTo b • Thus for w to remain fixed, Ze → ∞ requires
TTo thin reaction zone b
16 Procedure for Asymptotic Analysis
• Separately obtain (partial) solutions for the three zones: o Broad upstream, convective-diffusive, preheat zone, subject to b. c. at x only. No downstream b. c. o Thin reactive-diffusive zone, without any b. c. o Broad downstream, equilibrium zone. • Partial solutions determined to leading order of reaction sheet ( Ze ) and next order of broaden reaction zone (Ze >> 1, but finite) • Asymptotically match these partial solutions to determine the various boundary conditions, hence completing the solutions
17 Structure Equation for Inner, Reaction Zone (1/2)
• Define inner “stretched” variable and inner solution for reaction zone as ~ 2 exO/(1) T in o e 1 O e
2 dd 2 o eZe • Then G.E. becomes 11 1 e eL eD a 1 e . dd2 • Observations: o Diffusion term: O(1); Convection term: O(e) o To retain exponential nonlinearity essential to chemical reaction: e Ze ~ O (1 ) e ~ Ze 1 identified o For reaction term to be O(1) in order to balance diffusion term:
e 2 Da o ~ O (1) Da 0 ~ e 2 o e2: one e from thin zone, one e from reduced concentration 18
• Xxx • xx Structure Equation for Inner, Reaction Zone (2/2)
2 o o d o 2 L e D a 1 e 1 . • Final structure equation: 1 , 2 (7.5.44,45) d 2 2 Ze 2
dd o 1 e 1 • Solution: 1 (7.5.16) dd1 Integrating with b. c.: 1 d 1 / d 0 at : 2
d oo 1 e11 d c 1. e c 1 1 in 1 in (7.5.47) d o 1 1. e 1 1 (7.5.48) Evaluating at : cx 1 1 lim 111 e lim 1 c e 0 (7.5.49)
• Final result: o 1 as eigenvalue • In physical terms: 2 ( / c ) B o o 2 p c T / T ( f ) e a b Ze 2 • Phenomenological derivation only missed the term 2 Le / Ze 19 Dependence on Tad
• Correlates well with Tad through heat of combustion o Equivalence ratio o C/H ratio
20 Dependence on Le
• More rigorous derivation shows (fo )2 ~ L e o Concentration effect
o Tf Tad is not affected for the standard flame, hence weak
effect; exaggerated for hydrogen (Lelean0.3, Lerich2.3) o Effects are more significant for stretched flames for which
Tf is affected
21 Dependence on Molecular Structure
• Flame speed increases with ethane (C2H6),
ethylene (C2H4), and acetylene (C2H2)
Air as oxidizer Modified air to match Tad
22 Dependence on Pressure
• Dependence of flame speed on pressure is through o Chemistry o Density • Observed decreasing trend of flame speed with pressure is density effect, not chemistry effect • f o is the proper parameter because it is only affected by chemistry • f o usually increases with increasing pressure
23 Dependence on Transport Properties
• Flame speed can be manipulated through inert substitution, while keeping oxygen mole fraction fixed
o N2 and Ar have similar molecular weights and hence diffusivities, but different cp, which affects the flame temperature.
o Ar and He have the same cp but different diffusivities and densities
24 Extraction of Global n and Ea
ln f o ln f o n 2. ER2.o a ln p (1 /T ) T ad ad p
• Results demonstrate the role of pressure on two-body branching (promoting with pressure) and three-body termination (retarding with pressure) reaction • Note: possible n < 0 25 Chemical Structure of Flames
26 Asymptotic versus Chemical Structure
• Asymptotic Structure o Broad, convective-diffusive, nonreactive zone followed by: o Narrow, diffusive-reactive zone at downstream end of flame o One-step overall reaction accounts for both activation and heat release o Chemical activation is thermal in nature
27 Asymptotic versus Chemical Structure
• Chemical Structure (with chain mechanism) o Termination reaction is temperature insensitive can occur in upstream diffusive zone reactions take place throughout entire flame structure o Termination reactions can be highly exothermic substantial heat release in preheat zone o Chemical activation through radicals produced at downstream, high-temperature end that back diffuse to the preheat zone o In homogeneous system initiating radicals are produced by original fuel-oxidizer species.
28 Premixed H2-Air Flame: Diffusive Structure
• H2 diffusion layer is thicker than those of O2 and heat (T) because of its high diffusivity.
• Rapid reduction in H2 concentration (due to diffusion, not reaction) causes a bump in mole fraction of O2. This is not physical, just definitional (on mole basis)
29 Chain Structure
• Active reaction zone: 0.04 cm to 0.1 cm; two zone structure • Trailing, H production zone
o O+H2→OH+H; OH+H2 →H2O+H o H back diffuses • Leading, H consumption zone
o Back-diffused H reacts with O2 at low temperature through H+O2+M →HO2+M. o HO2 subsequently forms H2O2,
o Contrasts with H2+O2 →HO2+H in homogeneous system.
• Maximum consumption rates of H2:O2 = 2:1; occurring at same location
• H2O generated through entire reaction zone. 30 Thermal Structure
• Major exothermic reactions o H consumption layer,
H+O2+M →HO2+M o H production layer
OH+H2 →H2O+H • Major endothermic reaction
o H+O2→OH+O which is the major branching step • Maximum heat release occurs at 800 K ! • 30% heat released in H consumption layer, at 1000 K • Chemical activation, indicated by maximum H production rate, occurs around 1400 K. 31 Summary of Contrasts with Asymptotic Structure
• Important reactions occur throughout flame structure • H radical needed for initiation at leading edge is produced in the downstream and back diffuses • Maximum heat release occur at front of the active reaction zone • Substantial heat evolved in the moderately low temperature region of the flame
32 2. Limit Phenomena
33 Concepts of Ignition & Extinction
• Thermal runaway: Feedback loop involving nonlinear Arrhenius heat generation and linear heat loss • Radical runaway: Radical proliferation through chain branching • Unsteady (ignition) analysis: Tracking the temporal evolution of a reacting mixture upon application of ignition stimulus • Steady, S-curve analysis: Identify states at which steady solution does not exist for a non-reacting situation, signaling ignition, or a strongly burning situation, signaling extinction • Ultimate (extinction) consideration: system adiabaticity
34 Principle of Well-Stirred Reactor
• In steady-state operation: • Using: ˆ T Ta d c F
TT/ TT/ V c() T T V Q B c e af T T D a(), T T e af o p f o c F f o Cad f B Characteristic flow time Da C (8.1.23) VV/ Characteristic collision time
• Solutions: 1: Weakly-reactive state 2: Strong-burning state I: Ignition state E: Extinction state 3: Triple solution nonmonotonicity and hysteresis
35 Concept of S-Curve
• Ignition/extinction turning points defined by
dln D a C 0 dT f cr • (8.1.23) yields the two roots
1 / 2 (TTTTTa do ){1 4( o a d / a )} T f , c r . (8.1.29) 2 (1 1 /T ) a • Folding possible when {·} > 0 in (8.1.29). Otherwise S-curve is stretched for: o Low activation energy reactions o High initial temperatures 36 Premixed Flame Extinction (through Heat Loss) (1/3)
• The standard flame, being adiabatic, does not exhibit any extinction o behavior, i.e. finite f for finite Yu. • Heat loss lowers flame temperature
from Tad, leading to abrupt extinction, at finite Yu. System becomes non-conservative • Radiation from flame is an inherent heat loss mechanism • Assume loss occurs only in the preheat zone, and with L being a loss coefficient, then amount of loss is / c D p q L d x D L L 0 f 37 Premixed Flame Extinction (2/3) • Overall energy conservation
/ c p o fc p()(). Tad T u fc p T f T u L (8.4.4) f / c 2 p 2 Tfa d Tad (1 L ) T (1 L '/ f ) (8.4.4’) fT2 ad 2 / c p o L ';/ L f f f o o 2 (A) ()fT ad • In analogy to standard flame result (/) cwo oo20p (f ) ; w e x p ( Ea / R T a d ) (B) Ze we can write
(/) cwp (f )20 ; w e x p ( E / R T ) af Ze (C) • Using (8.4.4’) in w
0 2 0 2 w~exp[( Eaa / R )/(1 L '/ f )] exp( E / R )exp( L / f ) 0 (D) LERL (/)'a 38 Premixed Flame Extinction (3/3) • C/B using D: o f 2 (/ f f oo ) 22 w / w exp( L / f ),
from which f22ln. fL (8.4.9) • (8.4.9) is the generalized equation governing flame propagation with loss. o For L0 , f 1, f f 0 o Extinction, turning point: dL ( )ex 0 df 2 1 1 / 2 o Solving: LEE e, f e
39 Other Limit Phenomena
• Flammability Limits: o For given mixture temperature and pressure, the leanest and richest concentrations beyond which flame propagation is absolutely not possible o Set the ultimate boundaries for extinction
• Blowoff and Flashback o Consequence of lack of dynamic balance between flame speed and flow speed o Has nothing to do with extinction and ignition
40 Flammability Limit
• Simulation for methane/air mixtures shows o Extinction f = 0.493 empirical: f = 0.48 o f/fo = e-1/2 ≈ 0.6
o (Tf)ext ≈ 1,450 K • Extinction temperature result corroborates with the concept of limit temperature for hydrocarbon fuels Stabilization Mechanism of Premixed Flame at Burner Rim Triple-Flame Stabilization Mechanism of Nonpremixed Flame at Burner Rim 3. Aerodynamics of Laminar Flames
44 Standard Flame vs. Real Flame (1/2)
Hydrodynamic Limit
Reaction-Sheet Limit
45 Standard Flame vs. Real Flame (2/2)
• Standard 1D Planar Flame foo f(,,) q L e w o c i, j k
oo o System is conservative, Tb T b() q c fo ~ L e o ij, • General Stretched Flame
o f f(,,;,) qc L e i, j w k K a L . Ka: Karlovitz number, representing aerodynamic effects of flow nonuniformity, flame curvature, flame/ flow unsteadiness . L: Generalized loss parameter o System could become locally or globally nonconservative
Tb T b(,,;,) q c L e i, j w k K a L o O(e) modification of flame temperature leads to O(1) change in flame speed locally intensified burning or extinction
46
The Stretch Rate
• Definition: Lagrangian time derivative of the logarithm of
area A of a surface 1 dA A d t
A(p,q,t) = (epdp) × (eqdq) = (dpdq)n
• Letting Vv f,, ts t v ()() V n n t s, t f vs, t n ( v s n ) • Sources of stretch:
Flow nonuniformity: vs Flame curvature: n
Flame oblique to flow vns 0.
Flame unsteadiness (Vf ≠0), with curvature n 0
47 Examples of Stretched Flames
a • Stagnation Flame: v x, a y , 0 , a 0 (k 1)
2 dR f • Expanding Spherical Flame: 0 Rf d t wasin 2 • Bunsen Flame: 0 2 R f • Each of above has its opposite analog
48 Effects of Stretch (1/2)
• Hydrodynamic stretch: Flame-sheet limit o Tangential velocity gradient: changes flame surface area and hence total burning rate, ∫ f dA o Normal velocity: Balances flame speed o Net effect: Distortion of flame geometry Modifies total burning rate (e.g. higher burning rate in turbulent flame through surface wrinkling)
49 Effects of Stretch (2/2)
• Flame stretch: reaction-sheet limit
o Tangential velocity affects normal mass flux f b entering reaction zone o For Le≠1, modifies temperature and concentration profiles differently modifies total enthalpy and flame temperature locally non-conservative • Hydrodynamic stretch and flame stretch strongly coupled • Stretch (i.e. convection) in thin reaction zone is unimportant
50 Example of Hydrodynamic Stretch: Corner Formation in Landau Propagation
o • Landau propagation: ssuu • Concave segment develops into a corner; convex segment flattens • Positive curvature and hence stretch dominate • Mathematically described by Burgers equation, similar to that for shock formation
51 Flame Stretch due to Flow Straining: The Stagnation Flame (1/2)
• Stretch is positive, >0; situation reversed for <0 • Consider total energy conservation in control volume o Diffusion: normal to reaction sheet o Convection: along (divergent, >0) streamline • Le>1: More heat loss than reactant mass gain system sub-adiabatic • With increasing : o Flame temperature decreases, until extinction o Flame at finite distance from stagnation surface at extinction o Complete reactant consumption at extinction
52 The Stagnation Flame (2/2) • Le<1: More reactant mass gain than heat loss system super-adiabatic • With increasing : o Flame temperature increases, extinction not possible as long as the flame is away from surface o Eventually flame is pushed to the stagnation surface, leading to incomplete reaction and eventually extinction
53 Flame Stretch due to Flame Curvature: The Bunsen Flame • For the concave flame curvature, <0; expect opposite response from the >0 stagnation flame • (Negative) flame curvature focuses heat and defocuses mass in the diffusion zone o Le>1: Super-adiabatic, burning at flame tip intensified relative to shoulder o Le<1: Sub-adiabatic, burning at tip weakens, lead to extinction (i.e. tip opening)
54 Flame Stretch due to Flame Motion: the Unsteady Spherical Flame
• For the expanding flame, >0 ; expect similar behavior as the stagnation flame.
• An increase in flame radius Rf by dRf leads to an 2 increased amount (4 d RR Tf ) of heat transferred to the 2 preheat zone, and (4 d RR Mf ) of mass transferred. o Le>1: Sub-adiabatic, more heat transferred away o Le<1: Super-adiabatic, less heat transferred away
• Stretch rate, = (2/Rf)(dRf/dt) continuously decreases
with increasing Rf ,
approaching the planar limit
55 Analysis (1/2) (Based on Stagnation Flame Analogy) • Energy loss/ gain in control volume TT fu Y u fu c p()()()()() T a d Tq fc D ss// T u T M u M
o () TT / f u T Normal heat flux; loss ()q / D Y o c u M Normal chemical energy flux; gain
o /(/) su Fraction of flux diverted out of control volume due to stretch • In nondimensional form
T T K a0( L e 1 1) / f 2 S 0 / f 2 (A) a d f o S0 K a 0( L e 1 1) 0 0 0 o Karlovitz number, K a /(/); s uT nondimensional stretch rate
56 Analysis (2/2)
• Following same analysis as that for premixed flame with heat loss, yields
22 o o o o 1 o ffln s , s Z e S Z e( L e 1) K a uu • -so has the same role as L in (8.4.9), showing extinction for s<0 (i.e. loss due to stretch) • Further define Markstein number as M aoo Z e L e 1 1 • Then: s o M a o K a o
57 Response of Stretched Flame
o • From (A): Tf (>,<) Tad for S (<, >) 0 Since So= (Le-1-1)Kao, influence is lumped for nonequidiffusion and stretch. o Tf > Tad for (Ka > 0, Le < 1) or (Ka < 0, Le > 1) o Tf < Tad for (Ka > 0, Le > 1) or (Ka < 0, Le < 1) o Tf Tad for either Ka = 0 (Le ≠ 1) or Le = 1 (Ka≠0) • Super-lumped parameter for flame speed: so= Zeo(Le-1-1)Kao:
Reactivity ╳ nonequidiffusivity ╳ stretch • Markstein number, Mao=Zeo(Le-1-1) :
Reactivity ╳ nonequidiffusivity; a property of the mixture.
58 Results on Stretched Equidiffusive Flame • Stretched flame ( ≠ 0) for equidiffusive mixture (Le = 1) is not affected by stretch: So=(Le-1-1) 0
59 Nonequidiffusive Mixtures
60 Results on Stretched Nonequidiffusive Flame
(a)Counterflow flame (b) Outwardly (c) Inwardly propagating flame propagating flame
61 Flame Images
> 0
< 0
62 Further Implications of Stretched Flame Phenomena
• Determination • Concentration • Flame of laminar and temperature stabilization flame speeds modifications in and blowoff flame chemistry
50
45
40
35
30
25 1920 1940 1960 1980 2000 BurningVelocity 298at (cm/s) K Year of Publication 63 Flame Front Instabilities (1/2)
Diffusional Thermal Instability Hydrodynamic Instability (Flame sheet, constant flame speed; density jump)
64 Flame Front Instabilities (2/2)
Diffusional-Thermal DL Diffusional-Thermal DL Le<1 Le>1 Le>1 Le>1 Le<1 Le>1 65 Closing Remarks of Day 4 Lecture (1/2) • The standard premixed flame o Concepts introduced: asymptotic analysis; cold boundary difficulty; distinguished limit; flame structure based on one-step and detailed chemistry; extraction of global kinetic parameters • Further studies: o Need chemical structure of hydrocarbon flames o Explain the lack of influence on laminar flame speed by: (a) low-temperature chemistry, and (b) molecular size for large n-alkanes o Is it possible to derive a semi-empirical expression for
the laminar flame speed based on Tad and extracted global kinetic parameters? 66 Closing Remarks of Day 4 Lecture (2/2) • Limit phenomena: Extinction is due to insufficient reaction time while flame blowoff/out is due to imbalance between flame and flow speeds • Extinction is in general caused by enthalpy loss in the preheat zone, leading to O(e) reduction in flame temperature and O(1) reduction in flame speed • Need to distinguish flame stabilization by auto-ignition or flame holding • What is the fate of a freely-propagating flame to increasing strength of imposed strain rate? • Need detailed study of the chemical structure of stretched flames 67 ! Daily Specials !
68 Day 4 Specials
1. Low-temperature, NTC flames 2. Pulsations in premixed and diffusion flames: a. Extinction b. Spiral patterns 3. Pulsation in self-propagating high- temperature synthesis 69
1. Low-temperature, NTC Flames
70 The Issue
5 10
4 10
• NTC behavior observed for 3
10
(ms) g
i 1 atm homogeneous mixtures in 2 3 atm 10 low- to intermediate- 5 atm 1 10
temperature range Ignition delay 0 10 atm 10 20 atm 50 atm -1 10 0.8 1 1.2 1.4 1.6 1.8 2 1000/T (1/K) • NTC behavior not observed o in extensive counterflow experiments & simulations o Finite residence time shifts ignition temperature to >1000K, hence moves ignition chemistry out of the NTC regime o Can NTC behavior be manifested for: . Low strain rate flows? . High pressures?
NTC Behavior Predicted at Low Strain Rates !
k = 200/s k = 100/s
Pressure:1 Atm. NTC Behavior Exaggerated with Increasing Pressure
2500 2500
2000 2000 1 atm 1500 1500 3 atm
1000 1000
Two-
Maximumtemperature (K) Maximumtemperature (K)
500 500 stage 0 500 Tair (K) 1000 1500 0 500 1000 1500 Tair (K) ignition k = 100/s 2500
• At lower pressures, ignition occurs in 2000 two stages, final ignition controlled by high-temperature chemistry 1500 10 atm • At higher pressures, ignition occurs 1000 in one stage, controlled by low- One-stage ignition Maximumtemperature (K) 500 temperature chemistry 0 500 Tair (K) 1000 1500 Experimental Observation of Nonpremixed LTC Flame
Infrared Images at Ignition Ignition Temperature (1 atm, strain rate: 60 /s)
A: Air vs DME
B: N2 vs DME C: Ignition (Air vs DME)
D: Ignition (N2 vs DME)
• Detailed chemical structure study shows that such flames are governed by LTC Experimental Observation of Premixed LTC Flame 0.8 1.0 1.2
1.4 1.6 1.8 • Different behavior from hot flames • Insensitive of flame location, implying small variation in flame speed with Φ variation
• Stronger chemiluminescence for richer cases 18 2a. Pulsating Extinction of Premixed and Diffusion Flames
76 The Issue: Heightened Sensitivity Near Extinction • S-curve analysis is based on steady-state considerations, showing on-off states • Near state of extinction, the burning intensity of flame is reduced, implying larger effective activation energies • Flamefront pulsating instabilities ae also promoted with increasing activation energy • Could extinction occurs in a pulsating manner, especially for Le > 1 flames?
77 The Issue: Intrinsic Chemical Influence Near Diffusion Flame Extinction
• Finite-rate chemistry plays no role in the reaction-sheet limit of diffusion flames • Flamefront (pulsating) instability intrinsically requires consideration of finite-rate chemistry • Since finite-rate chemistry is responsible for extinction, then could extinction occur in a pulsating manner even for diffusion flames, if Le > 1 for one of the reactants?
78 Oscillatory Extinction of Le > 1 Premixed Flame
Calculated extinction Calculated S- Experimental dynamics; extinction curve, showing observation of controlled by steady- earlier extinction oscillatory luminosity due to pulsation state extinction temperature
79 Oscillatory Extinction of Diffusion Flames: Computation
Calculated S-curve, Calculated extinction showing earlier dynamics; extinction extinction due to controlled by steady-state pulsation extinction temperature 80 Oscillatory Extinction of Diffusion Flames: Experiment
Steady and oscillatory extinction of diffusion flames experimentally observed
81 2b. Observations of Spirals over Flame Surfaces
82 Target Patterns (Le > 1)
Experimental conditions: • Lean butane-air, f = 0.59 • 30 atm. pressure • Consecutive frames at framing rate of 15000 fps • Frame dimension: 2.73cm × 5.46cm
Regular Spirals
70 Hydrogen, f=4.30, p=20 atm 60
50
40
Pixels 30
Starting Point 20 End Point
10 94 pixels/cm
0 0 20 40 60 80 100 120 Pixels • Rich hydrogen-air flame (f = 4.30) at 20 atm. • Spirals confined within hydrodynamic cells • Spiral can be either clockwise or counter- clockwise a b
Magnified View of the Spirals
c
Disordered Spirals
• Hydrogen-oxygen flame at 30 atm. and f = 6.00 • Disordered spirals
3. Pulsation in Self-propagating High-temperature Synthesis
87 Pulsation in Condensed-Phase Flames
• A curious result: o Propagation of solid flames exhibit temperature-sensitive Arrhenius behavior, e.g. pulsation o But reaction for individual particles is in diffusion flame- sheet limit no finite rate chemistry
Laminated product structure due to pulsation 88 Pulsation in Condensed-Phase Flame
• Explanation: Arrhenius behavior from temperature- sensitive solid-phase diffusivity
2 TTTTaa//a d a d • Gas-phase flame speed: fg a s ~ ( D )gg ~ ( D ) e ~ e • Condensed-phase flame speed: f2 ~ ( DD ) K ~ ( ) SHS s S H S s c
22 Since Kc (2 B D / A )ln(1 s Y B / s O )~ D s fSHS ~ D s
TTTTdd//ad But Ds ~ e , therefore: fSHS ~ e
89