Weapons and Materials Research Burning-Rate Models and Their Successors Martin S. Miller MURI Kickoff Meeting 17 OCT 02.

Weapons and Materials Research

Burning-Rate Modelsand Their Successors

Martin S. Miller

MURI Kickoff Meeting

17 OCT 02


Goals of Briefing

• Convey complexity of phenomena

• Concepts - continuum-mechanics paradigm

• Recent modeling approaches

• Frozen ozone

• RDX

• Propellants

• The MURI challenges


PHENOMENA


98% NC (13.16%N)

2.8 mm/s @ 1 MPa


28.7% NC(12.68%N)22% Nitroglycerine

47.3% Nitroguanidine

2.3 mm/s @ 1 MPa


76% RDX (5 micron)12% CAB

4% NC(12.6%N)8% Energetic Plasticizer

0.8 mm/s @ 1 MPa


80% HMX (200/20 micron)20% Polydiethylene Glycol Adipate

0.5 mm/s @ 1 MPa


CONCEPTS


Propellant Combustion & Modeling Abstraction

RDX CompositePropellant M43

@ 15.5 atmD

ark

Zon

eV

isib

le f

lam

e

Caricature of Molecular Nature of 3-Phase Processes

GAS-PHASE FLAME elementary reactions thermal conduction convection molecular diffusion multi-component transport thermal diffusion

LIQUID/FOAM c-phase reactions thermal density changes mixture properties thermal conduction convection molecular diffusion bubble formation

UNREACTED SOLID thermal conduction convection thermal density changes

MULTI-COMPONENTEVAPORATION &GAS/SURFACE REACTIONS

{

MELTING{


Conservation Equations for 1-D, Steady-State Combustion at Constant Pressure

Chemical-Species Conservation in Each Phase:

)(,..,1 specieschemicalKk d

dxY V m

dY

dxWk k

kk k a f 0

Energy Conservation in Each Phase:

d

dx

dT

dxmc

dT

dxY V c h Wp k k p

k

k

K

k k kk

K FH IK

1 10

Mass Conservation:

0mdx

d

conduction convection diffusion reaction

diffusion convection reaction

constm

SOLID LIQUID GASx

x = 0x = -xliq

T(x)


0dx

dTliq

liqK

kkk hYm

1

00

liqK

kkkk hVY

1

000

0dx

dTgas

gasK

kkk hYm

1

00

gasK

kkkk hVY

1

000

SOLID LIQUID GASx

x = 0x = -xliq

solliqliq

K

k

xk

xk hYm

1

liq

liqliq

K

k

xk

xk hYm

1

liq

liqliqliq

K

k

xk

xk

xk hVY

1

liqx

liq dx

dT

liqx

sol dx

dT

Energy Fluxes at the Phase Boundaries

convection

convection

diffusion

diffusion

conduction

conduction

T(x)


An Example of a Surface Regression Mechanism:Single-Component Evaporation

EQUILIBRIUM

Equilibrium vapor pressure pLiquid

M W nvcondensing 14e j

M Mescaping condensing

npRT

E

s

v 8RTW

s

12

FH IKE

NON-EQUILIBRIUM

Liquid

M W X pRT vtotal

scondensing FH IK

1

40

M WpRT

RTW

E

s

sescaping FH IK

14

812e j

Surface Regression escaping condensing m M M= -

E

Vapor pressurereduced from p by reactions

totals pXTm ,, 0


The 3-Phase Mathematical Problem Posed

• Integration over solid phase for heat flux:

• Integration over liquid phase for heat flux :

• Integration over gas phase for heat flux :

• Surface regression mechanism: (p, Ts)m

Solution eigenvalues: xliq , Ts, } m

)()( liqliqliqsol xx

)0()0( gasliq

),( sTpmm

),,,( 0 mTxT sliqsol

),,,( 0 mTxT sliqliq

),,( mTp sgas


Iteration Scheme for 3-Phase Problem

Solve Conservation Eqns. in Liquid Layer

Liq./Solid BC Satisfied?

Solve Conservation Eqns. in Gas Phase

Surface Regression Mechanism Satisfied?


Gas/Liq. BC Satisfied?

Converged Eigenvalues

xliq, Ts, and m

Choose trial values of xliq, Ts, and m

totals pXTm ,, 0m

)()( liqliqliqsol xx

)0()0( gasliq


MODELS


Solid-Propellant Combustion-Modeling Timeline

1971Guirao & Williams

AP(14s,10r)

1986Hatch

NG(19s,60r)

1950Parr & Crawford

Rice & GinellDB

1990MeliusRDX

(38s,158r)

1996Miller

O3

(3s,3r)

RDX(45s,232r)

Liau & Yang1995

RDX(45s,232r)Davidson & BecksteadPrasad, Yetter, & Smooke1997

2000Miller & Anderson

O3, RDX, NG

Double Base(59s,365r)Miller & Anderson2002

2003Miller & AndersonRDX/BAMO-AMMOCL20/BAMO-AMMO(80s,550r)

MURIContribution?

1950's1940's 1960's 1970's 1980's 1990's 2000's 2010's

Single Overall Irreversible Rxn, Analytic Approx.Elementary Rxns, Decoupled Transport & Kinetics, Numerical Soln.Two Overall Rxns, Coupled Transport & Kinetics, Numerical Soln.Elementary Rxns, Coupled Transport & Kinetics, Numerical Soln.

Gas-Phase Treatment:

RDXBenReuven & Caveny

1977


FROZEN OZONE


• 3 Reversible Gas-Phase Reactions

• Heterogeneous Reaction Considered

Ozone Chemistry

I: O M O O M H 25.65 kcal / mole3 2 298.15K0

II: O + O O + O H = - 93.41 kcal / mole3 2 2 298.15 K0

III O O M O M H kcal moleK: . /. 2 298 150 119 06

IV O l O g O g O g H kcal moleK: ( ) ( ) ( ) ( ) . /.3 2 2 298 15 92 67


Frozen Ozone:Simplest Case of 3-Phase Deflagration

Distance from Surface ( microns )

-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70

Te

mp

era

ture

(

K )

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

Mo

le F

ractio

n

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

GA

S

SO

LID

r (cm/s) = 0.2526 xliq (micron) = 24.07 Ts (K) = 158.06

FROZEN OZONE COMBUSTION AT 1 ATM & 40 K

T

O2

O

O3

LIQ

UID

Distance from Surface (microns)

0.1 1 10 100 1000

Spe

cies

Pro

duct

ion

Rat

e (m

oles

/cc-

s)-4

-3

-2

-1

0

1

2

3

4

5

R(O3)

R(O)

R(O2)

ZONE BZONE A ZONE C ZONE D

P = 1 atm T0 = 40 KGas/Surface Reaction Probability = 0

Gas-Phase Species-Production Rates in Frozen-Ozone Combustion

RxnsIb & IIIf

Rxn IIf Rxn If Rxn IIIf


Comparison of Ozone Model to Experiment

• 10% O2 / 90% O3 liquid at 90 K: rexptl ~ 0.4 cm/s (Streng 1960)

• Single-component evaporation model with mixture-corrected liquid density, thermal conductivity, and enthalpy:

rcalc = 0.30 cm/s

What can explain the discrepancy?

• Multi-component Evaporation

• Liquid-Phase Diffusion

• O3/O2 Phase Separation in Liquid


Multi-component Evaporation in Ozone Model

• O2 at surface evaporates faster than O3, enriching the surface concentration of O3 from feedstock value

• O3 surface concentration becomes new eigenvalue; necessitates 4th iteration loop

• Necessitates consideration of molecular diffusion in the liquid phase

O3-O3 Strong interaction

O2-O2 Weak interaction

O2-O3 Moderate interaction


Liquid-Layer Molecular Diffusion

O3

O2

LOW RATE OF ESCAPE

HIGH RATE OF ESCAPE

DiffusionalFlow of O3

DiffusionalFlow of O2

LIQUID GAS

mg

cm sY

g

cmr

cm

sO O O3 2 30

3 3FH IK FH IK FHIK

mg

cm sY

g

cmr

cm

sO O O2 2 20

3 2FH IK FH IK FHIK

Condition for steady-state surface regression: rO3 = rO2

m

Y

m

YO

O

O

O

3

30

2

20

YO30=> New Eigenvalue


Calculational Price of Including Multi-Component Evaporation in Continuum Model







Converged Eigenvalues

xliq, Ts, and m








Converged Eigenvalues xliq, Ts, and



Surface Mole Fraction Relation Satisfied?

xliq

Ts

xliq

Ts

m

m

0O3

Y

Multi-ComponentEvaporation – Ozone

Single-ComponentEvaporation – Ozone

3 Eigenvalues3 Nested Loops

4 Eigenvalues4 Nested Loops

Multi-ComponentEvaporation – 6 species

6+2 Eigenvalues6+2 Nested Loops

Solve

Liq./Solid

Solve

Surface

Solve

Gas/Liq.

Choose

Solve

Surface

Solve

Liq./Solid

Solve

Surface

Solve

Gas/Liq.

Conv

Solve

Surface

(Assuming Oatoms not in liq.)


RDX


RDX Burning-Rate Model Results Compared

Davidson & BecksteadLiau & YangPrasad, Yetter, & Smooke

MODEL VAPORPRESSURE

SENSITIVITY OFBURN RATE TO

VAPOR PRESSURE

LIQUID RDXDECOMPOSITION

P, Y, & Snone - input exptl. Ts

instead none 40% at 1 atm

L & YPRDX(600K)=0.69 atm

(vapor-liquid data) high <1% at 1 atm

D & BPRDX(600K)=0.17 atm

(vapor-solid data) low 25% at 1 atm


C-Phase Decomposition Mechanisms Used by Different Models for RDX

3 CH2O + 3 N2O H570K = - 47 kcal/mole (D&B, PY&S, L&Y)

3 H2CN + 3 NO2 H570K = + 180 kcal/mole (D&B, PY&S)

3 HCN + 3 HONO H570K = + 19 kcal/mole

3 HCN + 3 NO2 + 3 H H570K = + 256 kcal/mole

3 HCN + (3/2) NO +(3/2) NO2 + (3/2) H2O H570K = + 34 kcal/mole (L&Y)

RDX(liq.)

{NO2 + CH2O NO + CO + H2O H570K = - 42 kcal/mole (D&B, PY&S, L&Y)

k1

k2

k3


RDX Liquid-Phase Reactions Assumed by Different Models

MODEL RDX-> 3CH2O+3N2O

k1 = A exp(-E/RT)

A E

RDX-> 3NO2+3H2CN

k2 = A exp(-E/RT)

A E

CH2O+NO2-> CO+NO+H2O

k3 = A T2.77 exp(-E/RT)

A E

P, Y, & S 6.0E13 36.0 1.6E17 45.0 802.0 13.73

L & Y 6.0E13 36.0 3.2E17 ** 45.0 802.0 13.73

D & B 6.0E13 36.0 0.8E17 45.0 802.0 13.73

** RDX => (3/2)NO + (3/2)NO2 + (3/2)H2O


GUN PROPELLANTS


ARL Burn-Rate Predictor: A New Approach

• Assumption 1: Universality and availability of an empirical pyrolysis law for the given class of propellants

• Assumption 2: Condensed-phase decomposition products can be estimated for each ingredient, e.g.,

• Assumption 3: Decomposition of the propellant into gas-phase reactants can be approximated as the non-interactive decomposition of each of its ingredients

r = As exp(-Es/RTs)

NG 2 H2CO + 2 NO2 + HONO + CO

[ 2 H2CO + 2 NO2 + HONO + CO ] x 0.14

[ 2 H2CO + (CHO)2 + 2 NO2 + NO + CO + HCO ] x 0.59

[ 3 H2CO + 2 NO2 + CH2 ] x 0.27 }2.3 H2CO + 0.6 (CHO)2 + 2.0 NO2

+ 0.1 HONO + 0.6 NO + 0.7 CO

+ 0.6 HCO + 0.3 CH2

Gas-Phase Reactants

GASSOLID


“Pyrolysis” Laws from Zenin Microthermocouple Data


CYCLOPS v1.0: Burning-Rate Predictorfor Multi-Ingredient Propellants with NC

Mononitrate: 6 w%

Trinitrate: 75 w%

DEGDN: 25 w%

NC (13.1 %N): 60 w%

CODEINPUT

MONTE-CARLOCALC. OF NITRATE-STATE

DISTRIBUTION FOR GIVEN %N

NC2 2 NO2 + 0.5 CHOCHO + 1.5 CH2O + 1.5 HCO + CH + CO

NC3 3 NO2 + CHOCHO + 2 CH2O + CH + CO

NC1 NO2 + 3 CH2O + 2 HCO + CH

NG 3 HONO + 2 HCO + CO

DEGDN NO2 + 3 CH2O + CH2

0.288 NO2 0.070 HONO 0.061 CHOCHO

0.285 CH2O

0.083 HCO

0.045 CH2

0.076 CH

0.092 CO

NASCENTGAS-PHASE REACTANTS(mole fractions)

Condensed Phase Gas Flame

DECOMPOSITIONPRODUCTSDATABASE

NG: 15 w%

Dinitrate: 19 w%

59 S

peci

es, 3

65 R

eact

ions

,C

ondu

ctio

n, C

onve

ctio

n,

Mol

ecul

ar D

iffu

sion


Nitrate-Ester Linear Burning Rates

Pressure ( MPa )

0.1 1 10 100 1000

Lin

ear

Bu

rnin

g R

ate

( c

m/s

)

0.1

1

10

100Experiment, M10: Miller (1985)Experiment, M10: AFATL (1982)Experiment, M10: Atwood, et al. (1988)CYCLOPS Code

M10

Pressure (MPa)

0.001 0.01 0.1 1 10 100

Lin

ear

Bu

rnin

g R

ate

(cm

/s)

0.001

0.01

0.1

1

10

100

Data of Andreev (1957) Data of Andreev (1940) 3% gelData of Andreev (1959) 3% gelCYCLOPS Code (NG4 Product Set)

NG


Flame Structure

Distance from Surface in Gas (cm)0.01 0.1 1

Gas

Tem

per

atu

re (

K)

500

1000

1500

2000

2500

3000

CYCLOPS Code PredictionVanderhoff, et al. (1992) by NO AbsorptionVanderhoff, et al. (1992) by OH Absorption

JA2 @ 16 atm

Distance from Surface in Gas (cm)0.01 0.1 1

NO

Mol

e F

ract

ion

0.00

0.05

0.10

0.15

0.20

0.25

CYCLOPS Code PredictionVanderhoff, et al. (1992)

JA2 @ 16 atm


Species Mole Fractions in the Dark-Zoneof Double-Base Propellant (~ M9)

P a ra m e te rs H e lle r &G o rd o n (1 9 55 )

L e ng e lle , e t a l.(19 8 4 )

V a nde r ho ff, e ta l. (1 9 9 1)

C Y C L O P S(p re se nt ca lc .)

P (a tm ) 1 6 9 1 7 1 7

T D Z (K ) 16 0 0 15 0 0 15 0 0 15 4 3

N O 24 % 21 % 24 % 25 %

C O 3 3 3 8 3 2

H 2 8 8 8

N 2 4 2 0.4

H 2 O 2 0 2 0 1 9

C O 2 1 0 9 1 0

H C N 0.4 0.4

C H 4 0.8 2 .6 0.9

C 2 H 4 0.8 0 .8 0.1

EXPERIMENTAL


Nitramine-Propellant Burning Rates& Flame Structure

RDX/TPE Propellant @ 293K:80% RDX, 20% BAMO/AMMO (50/50 Mix)

Pressure (MPa)

0.1 1 10 100

Lin

ear

Bu

rnin

g R

ate

(cm

/s)

0.01

0.1

1

10

Data of Zenin (2000)CYCLOPS: NTRB pyrolysis lawCYCLOPS: RDX/BA pyrolysis law


CHALLENGES&

OPPORTUNITIES


Barriers to Development of a Predictive Model

LIQUIDSOLID GASM

EL

T I

NT

ER

FA

CE

GA

S /

SU

RF

AC

E

• Chemical kinetics

• High-density transport

• Evaporation of mixtures

• Critical phenomena of mixtures

• Heterogeneous reactions

• Non-planar surface phenomena

• Reactions

• Bubble formation, dynamics

• Mixture equations of state

• Mixture molecular diffusion

• Mixture thermal conductivity

• Reactions

• Mixture equations of state

• Mixture thermal conductivity

• Mixture melting

• Polymer softening


Issues in Developing a Molecular-Dynamics Description of EM Combustion

• Gas phase:

– Most easily and accurately done with continuum-mechanics formulation (>80 species, 550 rxns)

• Condensed phases:

– no reliable reaction mechanisms, and those that exist have only a few reactions with uncertain rates

– MD would likely have no competitor for the foreseeable future

– How to couple a MD description with a continuum description of the gas-phase processes?

• Surface-regression mechanism:

– MD coupled with quantum-structure calculations might be able to rationalize pyrolysis law data and provide predictions

– How to couple a MD surface-regression mechanism to the continuum description of the gas phase, as in multi-component evaporation

• MD-calibrated continuum models the answer?


rz

d

L IQ U ID

G A S

a 0

z = 0

dV adadz2=a

n dV n dz r r drLJ2 ( )

612

4)(rr

rLJ

where

( ) ( ( )) ( ( ))d n f x d g x d 8 3 x dd

a( )

0

f xa

x

x x x xx( )

( ) ( ) ( ) ( )arctan ( )

1

10 9

9

8 1

63

48 1

315

192 1

945

384 1

105

384

105

7680

9

2 4 2 3 2 2 2

g xa

x

xx( )

( )arctan ( )

1

4 2 1

1

2 40

3

2

6

0

3

0

32

32

71

aanH vap

5 0 50.3

0.2

0.1

0

V( )d

d

(d)

Continuum Model of theMolecule/Liquid-Interface Potential


Heat-of-Vaporization Estimation Theory for Pure LJ Fluids(Gas-Phase LJ Parameters Used)

6

0

3

0

32

32

71

aanH vap


Vapor-Pressure Estimation Theory for Pure LJ Fluids

5 0 50.3

0.2

0.1

0

V( )d

d

out liq c e cn P E E v E E 1

4( ) ( )

E f H

f c c H

c

c

c vap

vap

1 2

1

26

0 725

9 8 10

.

.

pRT

RTM

vapout

4

8


THE END

Weapons and Materials Research Burning-Rate Models and Their Successors Martin S. Miller MURI Kickoff Meeting 17 OCT 02.

Documents

materials research concepts

materials research models

materials research phenomena

liquid slide

liquid phase slide

mpa slide

phase problem slide

atm slide