Modeling Pyrolyzing Ablative Materials

with COMSOL Multiphysics®

Tim Risch
Deputy Branch Chief
Aerostructures Branch
NASA Armstrong Flight Research Center
COMSOL Day Los Angeles
March 21, 2019

1
 Outline

• Definition and Applications

• General Pyrolyzing Ablator Problem
• Solution Example Using Finite Element Model
• Summary and Conclusions

2
 Definitions

• Ablation is removal of material from the surface

of an object by melting, vaporization, chipping, or
other erosive processes.
• Pyrolysis is the decomposition of a material
brought about by high temperatures
• Pyrolyzing Ablator is a material that undergoes
sub-surface decomposition when heated and
when becomes sufficiently hot, loses surface
mass loss through melting or vaporization
3
 Reentry Heat Shields

Applications of
Pyrolyzing Ablator
Laser Machining
Modeling Wood Combustion

4
Rocket Nozzles Plasma Heating
 Heated Pyrolyzing Ablator

Virgin Material
Char Material

From Başkaya, A. O. “CFD Analysis of The Cork-Phenolic Heat Shield Of A Reentry Cubesat In Arc-Jet
Conditions Including Ablation And Pyrolysis”, 15th International Planetary Probe Workshop

5
 General Problem Illustration
External Flow

Radiation Ablation
Radiation
In Products
Out
Chemical
Species Convection
Diffusion In
Frontface
Char or Residue
s

y In-Depth
Pyrolysis Gas Pyrolysis Zone
Conduction

Virgin Material

Backface
6
 Modeling Requirements for Pyrolyzing Ablators
• Non-linear heat conduction in solids
• Non-linear, thermal boundary conditions
• Moving boundaries
• Non-linear, time-dependent quasi-solid in-depth reactions
• Transport and thermal properties as a function of material
state as well as temperature
• Inclusion of the thermal effects of gas flow within the solid
material
• In-depth pore pressure due to pyrolysis gas transport (not
always employed)

7
Material Pyrolysis

10 K/min

8
 Decomposition Model

• Material consists of three constituents (although the number

could be increased)
𝜌 = Γ 𝜌𝐴 + 𝜌𝐵 + 1 − Γ 𝜌𝐶
• Components A and B decompose according to:
𝜓𝑖
𝜕𝜌𝑖 𝐸𝑖 𝜌𝑖 − 𝜌𝑟,𝑖
= −𝐴𝑖 exp − 𝜌
𝜕𝑡 𝑦
𝑅𝑇 𝑜,𝑖 𝜌𝑜,𝑖
• Material properties are a function not only of temperature,
but also material state

9
 In-Depth Temperature History
• In-depth temperature time history can come from:
– Thermogravimetric Analysis (TGA)
𝑇 = 𝛽𝑡 + 𝑇0
– Steady-State energy balance (1-D transformed coordinate)
𝜕 𝜕𝑇 𝜕𝑚ሶ 𝑔 ℎ𝑔 𝜕𝜌ℎ𝑠
𝑘 + + 𝑠ሶ =0
𝜕𝑦 𝜕𝑦 𝜕𝑦 𝜕𝑦

– Transient energy balance (1-D transformed coordinate)

𝜕𝑇 1 𝜕 𝜕𝑇 𝜕𝜌 𝜕𝑇 1 𝜕𝑚ሶ 𝑔 ℎ𝑔 𝐴
𝜌𝐶𝑝 = 𝑘𝐴 − ℎത 𝑇 + 𝑠𝜌𝐶
ሶ 𝑝 +
𝜕𝑡 𝑦
𝐴 𝜕𝑦 𝜕𝑦 𝑡
𝜕𝑡 𝑦
𝜕𝑦 𝑡
𝐴 𝜕𝑦 𝑡

– Transient Energy Balance (1- and 2-D fixed coordinate)

𝜕𝑇 1 𝜕𝜌 1
𝜌𝐶𝑝 ത
= 𝛻 𝑘𝐴𝛻𝑇 − ℎ(𝑇) + 𝛻 ∙ 𝒎ሶ 𝒈 ℎ𝑔 𝐴
𝜕𝑡 𝐴 𝜕𝑡 𝐴

10
 Surface Energy Balance and Pyrolysis Gas Flow
𝛼𝐼 𝜌𝑒 𝑢𝑒 𝐶𝐻 ℎ𝑟 − ℎ𝑤 𝐹𝜎𝜀 𝑇𝑤4 − 𝑇∞4 𝑚ሶ 𝑤 ℎ𝑤

𝜕𝑇
−𝑘 𝑚ሶ 𝑐 ℎ𝑐 𝑚ሶ 𝑐 ℎ𝑐
𝜕𝑦
𝜕𝑇
𝛼𝐼 +𝜌𝑒 𝑢𝑒 𝐶𝐻 ℎ𝑟 − ℎ𝑤 +𝑘 − 𝐹𝜎𝜀 𝑇𝑤4 − 𝑇∞4 − 𝑚ሶ 𝑤 ℎ𝑤 − 𝑚ሶ 𝑐 ℎ𝑐 − 𝑚ሶ 𝑔 ℎ𝑔 = 0
𝜕𝑦

𝜕𝜌
𝛻 2Φ =
𝜕𝑡

𝑚ሶ 𝑔 𝒎ሶ 𝒈 = 𝛻Φ

11
 Surface Thermochemistry – Normalized Mass Loss
100 B'g = 10
Surface P = 1 atm B'g = 7.5
B'g = 5.5
thermochemistry B'g = 4

conditions 10
B'g = 3
B'g = 2.4
computed from B'g = 1.9
B'g = 1.5
equilibrium B'g = 1.2

thermochemistry in B'g = 1
B'c

1 B'g = 0.9

terms of normalized B'g = 0.8

Increasing B'g B'g = 0.7
mass fluxes. B'g = 0.6
B'g = 0.5

𝐵𝑐′ = 𝑚ሶ 𝑐 /𝜌𝑒 𝑢𝑒 𝐶𝑀 0.1 B'g = 0.4

B'g = 0.32
𝐵𝑔′ = 𝑚ሶ 𝑔 /𝜌𝑒 𝑢𝑒 𝐶𝑀 B'g = 0.25
B'g = 0.2
B'g = 0.15
B'g = 0.1
0.01
𝐵𝑐′ = 𝐵𝑐′ (𝑝, 𝐵𝑔′ , 𝑇𝑠 ) 0 1000 2000 3000 4000
B'g = 0.07
B'g = 0.04
B'g = 0.02
Temperature (K) B'g = 0

12
Implementation

In-Depth Conduction and Surface Energy Balance

Decomposition Reactions

In-Depth Pyrolysis Gas Flow

Surface Thermochemistry

Moving Boundary

13
 Two-Dimensional Transient Example
2
𝐼 = 𝐼𝑜 ∙ exp −𝐶 𝑟/𝑟0
• Problem is for a two- 𝐼𝑜 = 1 × 107 W/m2 : 𝐶 = 5
dimensional, axisymmetric 𝐼𝑜
puck
• Top of puck heated with
Gaussian flux profile
• Pyrolysis gas flow calculated
from potential flow
• Full surface thermochemistry 𝑟
with recession
• 2-D COMSOL Multiphysics
results compared to a series
of 1-D results

14
2-D Problem Animation

Animation is twice actual speed

15
Original and Deformed Mesh

16
 Summary
• COMSOL is a suitable tool for modeling pyrolyzing ablative
materials
• General capabilities of COMSOL Multiphysics allow for a
wide variety of geometries and problems to modeled
• COMSOL allows for modifications to model to be made quickly
and easily
• Solution algorithms are efficient and stable
• Integrated environment provides a very user friendly and
powerful system for modeling
• Multiphysical modeling capability allows for structural and
external flow to be incorporated into analysis (in progress)

17
 For Additional Information

Risch, T., “Verification of a Finite-Element Model for Pyrolyzing

Ablative Materials”, presented at the AIAA 47th AIAA
Thermophysics Conference, Denver CO, June 5-9, 2017.

18
QUESTIONS?

19
 Final Recession Profile at 30 s
0.010
2-D
0.009 1-D
0.008 Quasi 1-D

0.007
0.006
Height, m

0.005
0.004
0.003
0.002
0.001
0.000
0 0.002 0.004 0.006 0.008 0.01
Radius, m 20
 Example Problems

• Look at four examples solved with COMSOL

– Thermogravimetric Analysis (TGA)
– Steady-state one-dimensional thermal and density
profile
– One-dimensional transient temperature and
recession history
– Two-dimensional transient temperature and
recession history

21
Thermogravimetric Analysis (TGA) Example

22
 Thermogravimetric Analysis (TGA) Example

• Three component TACOT model

• Linear ramp increase in temperature at 10 K/min
• First-order time integration, not a spatial problem
• Results provide density and reaction rate for three
components as a function of time
• COMSOL Multiphysics results compared to independent
fourth-order Runge-Kutta calculation

23
 TGA Results - I
290 0.16%
 = 10 K/min
280 0.14%

270 COMSOL 0.12%

Runge-Kutta
Difference
260 0.10%
Density, kg/m3

Difference
250 0.08%

240 0.06%

230 0.04%

220 0.02%

210 0.00%
300 400 500 600 700 800 900 1000 1100 1200
Temperature, K 24
 TGA Results - II
0.040 0.05%
COMSOL  = 10 K/min
0.035 Runge-Kutta
Difference
0.03%
0.030
Decomposition Rate, kg/m3-s

0.025
0.01%

Difference
0.020

-0.01%
0.015

0.010
-0.03%
0.005

0.000 -0.05%
300 400 500 600 700 800 900 1000 1100 1200
Temperature, K 25
Steady-State Profile Example

26
 Steady-State Profile Example

• After long times in an infinite sample with a fixed surface

temperature and recession, temperature and density profile
will reach a steady state
• Problem solution becomes independent of time
• Specified surface temperature (3000 K) and steady recession
rate (110-4 m/s)
• COMSOL Multiphysics results compared to independent
second order finite difference calculation and results from the
Fully Implicit Ablation and Thermal Analysis Program (FIAT)

27
Finite Difference Temperature Profile Comparison
3500 0.05%
Finite Difference
3000 COMSOL
0.00%
Solution Difference

2500
-0.05%

Relative Difference
Temperature, K

2000
-0.10%
1500

-0.15%
1000

-0.20%
500

110-4 m/s
0 -0.25%
0 0.02 0.04 0.06 0.08 0.1
Distance, m 28
 Finite Difference Density Profile Comparison

290 0.04%
110-4 m/s
0.02%
280
0.00%
270

Relative Difference
-0.02%
Density, kg/m3

260 -0.04%

250 -0.06%

-0.08%
240
Finite Difference -0.10%
230 COMSOL
-0.12%
Solution Difference
220 -0.14%
0 0.02 0.04 0.06 0.08 0.1
Distance, m 29
 FIAT Temperature Profile Comparison
3500 0.0%
FIAT
3000 COMSOL SS -0.5%

Difference
-1.0%
2500

Relative Difference
-1.5%
Temperature, K

2000
-2.0%
1500
-2.5%

1000
-3.0%

500 -3.5%
110-4 m/s
0 -4.0%
0 0.02 0.04 0.06 0.08 0.1
Distance, m 30
 FIAT Density Profile Comparison
290 2.5%
110-4 m/s
280

2.0%
270

260

Relative Difference
Density, kg/m3

1.5%
250

240
1.0%
230

220
FIAT 0.5%

210 COMSOL SS
Difference
200 0.0%
0 0.02 0.04 0.06 0.08 0.1
Distance, m 31
One-Dimensional Transient Example

32
 One-Dimensional Transient Example

• Problem is for a planar, finite width slab heated on one

surface
• Frontface free stream enthalpy of 40 MJ/kg, a heat transfer
coefficient of 0.1 kg/m2-s, and reradiation
• Backface is adiabatic
• Full surface thermochemistry
• Thermocouples located at 0.001, 0.002, 0.004, 0.008, 0.016,
0.024, and 0.050 m
• COMSOL Multiphysics results compared to FIAT results

33
 FIAT Surface Temperature Comparison
3,000 2.0%

COMSOL
2,500
FIAT
Difference 1.5%
2,000

Relative Difference
Temperature, K

1,500 1.0%

1,000
0.5%
500

0 0.0%
0 10 20 30 40 50 60
Time, s 34
 FIAT Recession Comparison
0.0045 2.0%

0.0040 COMSOL
FIAT
Difference
0.0035
1.5%
0.0030

Relaive Difference
Recession, m

0.0025
1.0%
0.0020

0.0015
0.5%
0.0010

0.0005

0.0000 0.0%
0 10 20 30 40 50 60
Time, s 35
 Char and Pyrolysis Surface Mass Loss Rates
0.10 5%
COMSOL mc
0.09 COMSOL mg 4%
FIAT mc
0.08 FIAT mg 3%
0.07 Difference mc
Mass Loss Rate, kg/m2-s

Difference mg 2%

Relative Difference
0.06
1%
0.05
0%
0.04
-1%
0.03
-2%
0.02

0.01 -3%

0.00 -4%
0 10 20 30 40 50 60
Time, s
36
 FIAT In-Depth Temperature Comparison
3000
COMSOL Surface
COMSOL TC1 - 0.001 m
2500
COMSOL TC2 - 0.002 m
COMSOL TC3 - 0.004 m
COMSOL TC4 - 0.008 m
2000
Temperature, K

COMSOL TC5 - 0.016 m

COMSOL TC6 - 0.024 m
COMSOL TC7 - 0.050 m
1500 FIAT Surface
FIAT TC1 - 0.001 m
FIAT TC2 - 0.002 m
1000 FIAT TC3 - 0.004 m
FIAT TC4 - 0.008 m
FIAT TC5 - 0.016 m
500 FIAT TC6 - 0.024 m
FIAT TC7 - 0.050 m

0
0 20 40 60
Time, s
37
 FIAT Temperature Profile Comparison after 60 s
3000 2.0%
COMSOL
FIAT
2500 Difference
1.5%

2000

Relative Difference
Temperature, K

1.0%

1500
0.5%
1000

0.0%
500

0 -0.5%
0 0.01 0.02 0.03 0.04 0.05
Distance, m 38
 FIAT Density Profile Comparison after 60 s
290 0.0%

-0.1%
280
-0.2%
COMSOL Density
270
FIAT Density -0.3%
Difference

Relative Difference
260
Density, kg/m3

-0.4%

250 -0.5%

-0.6%
240
-0.7%
230
-0.8%
220
-0.9%

210 -1.0%
0 0.01 0.02 0.03 0.04 0.05
Distance, m 39
Two-Dimensional Transient Example

40
BACKUP

41
 Density Comparison 1-D vs 2-D

1-D 2-D

42
Pyrolysis Gas Flowrate

43
 Thermophysical Properties
2,500 3.0
Thermophysical
properties defined 2.5
2,000
separately for virgin and

Thermal Conductivity, W/m-K

char constituents.

Specific Heat, J/g-K

2.0
Composite properties 1,500
determined by mixing
1.5
rule based on mass.
1,000
𝑘 = 𝑥𝑘𝑣 + (1 − 𝑥)𝑘𝑐
1.0
Virgin Specific Heat
𝐶𝑝 = 𝑥𝐶𝑝,𝑣 + (1 − 𝑥)𝐶𝑝,𝑐 500 Char Specific Heat
0.5
Virgin Thermal Conductivity
𝜌𝑣 𝜌𝑐
𝑥= 1− Char Thermal Conductivity
𝜌𝑣 − 𝜌𝑐 𝜌 0 0.0
0 1,000 2,000 3,000 4,000
Temperature, K

44
 Material Enthalpy

Virgin and char 7,000

enthalpies computed Virgin
6,000 Char
from integration of
specific heats. 5,000

Enthalpy, kJ/kg 4,000

𝑇
ℎ = න 𝐶𝑝 𝑑𝑇 + ℎ0 3,000
𝑇0
2,000
ℎ = 𝑥ℎ𝑣 + (1 − 𝑥)ℎ𝑐 1,000

-1,000
0 1,000 2,000 3,000 4,000
Temperature, K

45
 Pyrolysis Gas Enthalpy

Pyrolysis gas enthalpy 60,000

computed from 0.01 atm
50,000 0.1 atm
equilibrium 1 atm
thermochemistry as a 40,000
function of
30,000
temperature and Enthalpy, kJ/kg
pressure. 20,000

10,000
ℎ𝑝𝑔 = ℎ𝑝𝑔 𝑝, 𝑇
0

-10,000

-20,000
0 1,000 2,000 3,000 4,000
Temperature, K

46
 Surface Thermochemistry – Normalized Mass Loss
100 B'g = 10
Surface P = 1 atm B'g = 7.5
B'g = 5.5
thermochemistry B'g = 4

conditions 10
B'g = 3
B'g = 2.4
computed from B'g = 1.9
B'g = 1.5
equilibrium B'g = 1.2

thermochemistry in B'g = 1
B'c

1 B'g = 0.9

terms of normalized B'g = 0.8

Increasing B'g B'g = 0.7
mass fluxes. B'g = 0.6
B'g = 0.5

𝐵𝑐′ = 𝑚ሶ 𝑐 /𝜌𝑒 𝑢𝑒 𝐶𝑀 0.1 B'g = 0.4

B'g = 0.32
𝐵𝑔′ = 𝑚ሶ 𝑔 /𝜌𝑒 𝑢𝑒 𝐶𝑀 B'g = 0.25
B'g = 0.2
B'g = 0.15
B'g = 0.1
0.01
𝐵𝑐′ = 𝐵𝑐′ (𝑝, 𝐵𝑔′ , 𝑇𝑠 ) 0 1000 2000 3000 4000
B'g = 0.07
B'g = 0.04
B'g = 0.02
Temperature (K) B'g = 0

47
 Surface Thermochemistry –Gas Phase Enthalpy
35000
B'g = 10
Enthalpy of gases at the 30000
P = 1 atm B'g = 7.5

wall computed similarly 25000

B'g = 5.5
B'g = 4

from equilibrium B'g = 3

thermochemistry. 20000 B'g = 2.4

B'g = 1.9

ℎ𝑤 = ℎ𝑤 (𝑝, 𝐵𝑔′ , 𝑇𝑠 )
B'g = 1.5
Enthalpy (J/g) 15000 B'g = 1.2
B'g = 1

10000 B'g = 0.9

B'g = 0.8
B'g = 0.7
5000 B'g = 0.6
B'g = 0.5
0 Increasing B'g B'g = 0.4
B'g = 0.32

-5000 0 1000 2000 3000 4000 B'g = 0.25

B'g = 0.2
B'g = 0.15
-10000 B'g = 0.1
B'g = 0.07

-15000 B'g = 0.04

Temperature (K) B'g = 0.02
B'g = 0

48
COMSOL Multiphysics User Interface

49
Example Uses of Pyrolyzing Ablator

50
 Objective

• NASA primarily relies on custom written codes to analyze

ablation and design TPS systems
• The basic modeling methodology was developed
50 years ago
• Through the years, CFD, thermal, and structural
mechanics calculations have migrated from custom, user-
written programs to commercial software packages
• Objective is to determine that a commercial finite
element code can accurately and efficiently solve
pyrolyzing ablation problems

51
 Advantages of Commercial Codes

• Usability (e.g. GUI)

• Built–in pre- and post-processing
• Built-in grid generation
• Efficient solution algorithms
• Multi-dimensional capability (planar, cylindrical, 1-D,
2-D, & 3-D)
• Built in function capability (predefined, analytic, and tabular)
• Validated by a wide user base
• Reduced life cycle cost
• Regular upgrades and maintenance
• Modeling flexibility
• Better documentation

52
 Material Selection

• For comparisons, utilize Theoretical Ablative

Composite for Open Testing (TACOT) Material
Properties
• Open, simulated pyrolyzing ablator that has been used
a baseline test case for modeling ablation and
comparing various predictive models
• Properties Required
– Solid virgin and char specific heat, enthalpy, thermal
conductivity, absorptivity and emissivity
– Pyrolysis gas enthalpy
– Surface thermochemistry mass loss and gas phase
enthalpy

53