Accepted Manuscript

Effects of heat treatments to inner liner material, thermal barrier coating, and outer
shell material on lifetime of a combustion chamber

Toshiya Kimura, Shin-ich Moriya, Masaharu Takahashi

PII: S0094-5765(17)31923-9
DOI: 10.1016/j.actaastro.2018.05.019
Reference: AA 6875

To appear in: Acta Astronautica

Received Date: 31 December 2017

Revised Date: 9 April 2018
Accepted Date: 7 May 2018

Please cite this article as: T. Kimura, S.-i. Moriya, M. Takahashi, Effects of heat treatments to inner liner
material, thermal barrier coating, and outer shell material on lifetime of a combustion chamber, Acta
Astronautica (2018), doi: 10.1016/j.actaastro.2018.05.019.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Effects of Heat Treatments to Inner Liner Material, Thermal

Barrier Coating, and Outer Shell Material on Lifetime of a
Combustion Chamber

PT
Toshiya Kimura*$, Shin-ich Moriya*, and Masaharu Takahashi**
*Japan Aerospace Exploration Agency, Kimigaya, Kakuda, Miyagi, 981-1525, Japan
** Hitachi Solutions East Japan, Ltd. Honcho, Aoba, Sendai, Miyagi, 980-0014, Japan
$

RI
corresponding author [email protected]

SC
Abstract
Key factors were investigated to extend the lifetime of a combustion chamber of a large thrust rocket
engine. Two-dimensional finite element simulations were conducted to estimate creep damage and low
cycle fatigue damage to the inner liner of a combustion chamber wall. In these simulations, the effects
of difference of heat treatments to the inner liner material, heat shield by a thermal barrier coating, and

U
thermal expansion of the outer shell on lifetime were examined. It was found that heat shield by a
thermal barrier coating had the largest effect in extension of the lifetime though the effect depended on
AN
the physical characteristic of the inner liner material. Expansion or contraction of the outer liner
during combustion affects the restriction of the inner liner resulting in the extension of the lifetime. By
combination of these effects, the lifetime can be extended up to about a factor of 10 longer than that of
a combustion chamber without any effects.
M

Keywords: Combustion chamber; Lifetime prediction; Creep damage; Low cycle fatigue damages;
Thermal barrier coating; Copper alloys; Rocket engine
D

1. Introduction
TE

In the development of a reusable rocket engine, the combustion chamber is one of the most critical components
which defines the lifetime of the engine. Hot combustion gas under high pressure flows into the chamber, and a very
cold and high-pressure coolant flows inside cooling channels. Combustion gas and coolant are separated by a wall
made of a copper alloy with a thickness of around 1 mm. These conditions are drastically changed during start-up,
EP

shutdown, and throttling operations of a combustion cycle. Hence, the inner liner of the combustion chamber is
exposed to large and unsteady thermal stress change.
To extend the lifetime of a rocket engine combustion chamber, various concepts have been proposed [1, 2].
Quentmeyer discussed seven different design concepts to reduce the thermal strain and the material strength of the
C

inner liner of a combustion chamber and showed that the thermal barrier coating (TBC) and transpiration cooling had
the best life-enhancing capabilities. Popp and Schmit also presented five design concepts to extend the lifetime of a
AC

combustion chamber and concluded that TBC had the best potential to extend the lifetime by a factor 3 to 30.
Various approaches for prediction of lifetime have been made experimentally and analytically. Sun et al. [3, 4]
conducted experiments to produce failure due to creep damage and low cycle fatigue damages using a subscale
combustion chamber and conducted finite element analysis to predict the failure. Riccius et al. [5-7] attempted to
revise the analysis method. A flat panel (so-called TMF panel) was heated by a laser and the experimental results
were used to validate the analysis method. A multi-physics coupled simulation was developed and applied to a rocket
combustion chamber by Nishimoto [8] and Amakawa [9]. In this simulation, the thermo-fluid coupled dynamics
were solved between coolant flow and the chamber wall.
At the Kakuda Space Center of Japan Aerospace Exploration Agency (JAXA), a reusable rocket engine was
developed and hot firing tests were conducted from 2014 to 2015 [10, 11]. This engine was designed for 100 flights
of a reusable sounding rocket [12]. Due to small thrust of this engine, the combustion chamber of this engine could
be designed to be reusable for 100 flights by adjusting the operating condition, especially by supressing the chamber
pressure and the chamber wall temperature as low as possible. Through 42 hot firing tests conducted, the total
damage due to creep and low cycle fatigue was roughly estimated to exceed the expected damages of 100 flights,
proving that the combustion chamber could be reusable for 100 flights.

See nomenclature at the end of paper

 ACCEPTED MANUSCRIPT

However, the chamber pressure of rocket engine with a thrust of 1000 kN should be much higher than that of the
engine of the reusable sounding rocket. It will be very difficult to achieve such a high-pressure combustion chamber
by just adjusting the operating conditions. The key factors to realize a long-life combustion chamber with a thrust of
over 1000 kN were examined, especially focusing on copper alloy materials of an inner liner, material properties of
an outer shell, and a thermal barrier coating (TBC) on the inner surface of the chamber.
Two-dimensional finite element simulations were performed to estimate creep damage and low cycle fatigue
damage to the inner liner of the combustion chamber for the cross-section at the throat region for different copper
alloy materials of the inner liner and materials of the outer shell with or without TBC.
Simulation results showed that the lifetime of the inner liner of a combustion chamber was strongly affected by both
the characteristics of a copper alloy of an inner liner and the material property of the outer shell. TBC effectively

PT
shields the heat flux from combustion gas and reduces the chamber wall temperature. However, it was found that the
effect of TBC on the lifetime depended on the material property, especially the yield stress of the inner liner material.
This work will provide key information useful for the development of long-life combustion chambers of high power
reusable rocket engines with a thrust over 1000 kN.

RI
2. Numerical analysis

SC
2.1 Combustion chamber structure

Figure 1 shows a typical structure of a combustion chamber and the cross section of the chamber wall. One of the
major objectives of the present work was to examine the possibility of extending the lifetime of the combustion

U
chamber of a large thrust rocket engine by using FE analyses. For the present simulations, the combustion chamber
of the LE-X rocket engine was modelled, which has a thrust over 1400 kN. LE-X was studied intensively as the next
AN
generation of the Japanese rocket engine [13,14]. Therefore, the FE model was created based on the structure of the
inner shape of the chamber and that of the coolant channels.
M
D
TE
EP

Figure 1: Schematic of the combustion chamber and a cross-section of the chamber wall
C

2.2 Finite element modelling

AC

For the estimation of damages due to creep and low cycle fatigue, the cross-section of the chamber wall, where the
highest heat flux within the throat is expected, was modelled. The geometry of the coolant channel and the inner liner
are similar to that of the LE-X combustion chamber were used for the present model. Although the combustion
chamber of the LE-X has a thick throat support [13], we assumed that the outer shell of the present model has the
same thickness as the inner liner. With the simplified thickness of the outer shell, the chamber wall was estimated to
withstand hoop stress due to the chamber pressure.
Because the coolant channels are located at regular intervals in the circumferential direction and each coolant
channel has the same geometry, the finite element mesh was composed of only a half cooling channel due to
rotational symmetry reasons (see Fig. 1). The region between dashed lines in Fig.1 was modelled.
The layer of the TBC was assumed to have thicknesses of 50 µm and 100 µm in the present model. In Fig. 2, the FE
mesh without a TBC and the FE mesh with a TBC layer of 100 µm are shown. The spacing between nodes of the
mesh is typically 15 µm. For this mesh size, the total number of nodes and elements are around 62 000 and 20 000,
respectively.

2
 ACCEPTED MANUSCRIPT

The dependency of the present calculations on the mesh resolution was checked. One typical model was picked up
and calculated using four different mesh resolutions 8µm, 15 µm, 30 µm, and 60 µm. At four evaluation points,
which will be explained in the later section 2.4, the low cycle fatigue damage, the creep damage, the equivalent stress,
and the lifetime cycle were estimated and compared for these resolution models. From these comparison, we decided
that the mesh resolution with the spacing of 15 µm was the most appropriate accuracy of solutions and computational
cost.
In the present FE simulations, the type of 8-node biquadratic quadrilateral, reduced integration is used for FE
elements. As a boundary condition, a symmetry condition in a circumferential direction is applied to the symmetry
plane.

PT
RI
U SC
Figure 2: Finite element mesh of 2D cross-section of the combustion chamber wall
AN
For the surface on the combustion gas side of the FEM model shown in Fig.2, the condition of the combustion gas
at the throat was applied as a boundary condition. The temperature and pressure of the combustion gas were derived
using the one-dimensional equilibrium code ODE [15]. The heat transfer coefficient at the throat was estimated from
M

the Bartz equation [16].

For the surface of the coolant channel, the temperature, pressure, and heat transfer coefficient at the throat were
applied as the boundary conditions. These values were derived by solving flow equations along the coolant channel
from the inlet to the outlet using a quasi-one-dimensional coolant flow solver which was developed at JAXA.
D

A copper alloy is usually used for the inner liner of a combustion chamber due to its high thermal conductivity. For
TE

inner liners of rocket engines that have been developed in the US and Europe are made out of a copper alloy with
zirconium and silver, which is called NARloy-Z [17]. NASA has developed a copper alloy containing chromium and
niobium, called GRCop [18], for reusable rocket engines.
For Japanese rocket engines, copper alloys with chromium and zirconium (CuCrZr) have been mainly used. In the
EP

present work, three different copper alloys with chromium and zirconium, named Cu-A, Cu-B, and Cu-C, were
examined. These are all copper alloys with chromium and zirconium, and Cu-A and Cu-B have the same quantity of
constituents of Cr and Zr though they were taken from different batches. In the manufacturing process, Cu-B is
subjected to solution treatment and aging after hot forging. In processing Cu-A, a cold forging process is added
C

between solution treatment and aging to improve mechanical properties. Cu-C is subjected to heat treatment during
the brazing thermal cycle. Hence, Cu-A, Cu-B, and Cu-C are copper alloys with chromium and zirconium but have
different characteristics, especially mechanical properties, due to different manufacturing processes.
AC
 ACCEPTED MANUSCRIPT

Figure 3: Normalized thermal conductivity (left) and coefficient of thermal expansion (right)
for Cu-A, Cu-B, Cu-C, and the outer shell materials

Figure 3 shows the thermos-physical properties, thermal conductivity and coefficient of thermal expansion, for
materials used in the inner liner and outer shell. Cu-A and B have the same thermal conductivity and coefficient of
thermal expansion. On the left side of Fig. 3, the thermal conductivity is plotted normalized by that of Cu-A at room
temperature. Coefficients of thermal expansion are also normalized by that of Cu-A at room temperature and plotted
in the right of Fig. 3. As shown in Fig. 3, thermal conductivity and the coefficient of thermal expansion of Cu-C are
similar to those of Cu-A and Cu-B.

PT
RI
U SC
Figure 4: Normalized yield stress (left), creep curve (middle), and fatigue rupture curve (right) of Cu-A, B, and C

Figure 4 shows the mechanical characteristics for this alloy CuCrZr. Yield stresses are normalized by that of Cu-A
AN
at room temperature and plotted on the left-hand side of Fig. 4. Creep and fatigue strengths are shown in the middle
and right figures, respectively. In the graph of the creep strength, the time to rapture is normalized by that of Cu-A at
the stress of 300 MPa. For the fatigue rapture curve on the right-hand side of Fig. 4, the strain range is normalized by
that of Cu-A at the cycles to failure of 10. The mechanical characteristics of these copper alloys show large
M

differences. Especially the yield strength presents typical characteristic differences among these copper alloys. Cu-C
is an alloy heat-treated during the brazing thermal cycle and its yield strength becomes much smaller than those of
Cu-A and Cu-B. The creep and fatigue strengths of these alloys also show differences. The creep strength of Cu-A is
D

highest and that of Cu-C is much lower than the others. The fatigue strength of Cu-C, on the other hand, is the
highest among them. These differences of mechanical characteristics result in extension of lifetime as an inner liner
material.
TE

In the present paper, the chamber wall model, in which the material of the inner liner is made of Cu-A, is named
model A. The other models using the copper alloys Cu-B and Cu-C are called model B and model C, respectively.
Two types of the outer shell were assumed. One was a normal outer shell made of a high strength stainless steel
(A286). The other outer shell was characterized by low expansion (the so-called low-λ). To examine the effect of
EP

restriction suffered by the outer shell during a combustion cycle on the inner liner, an alloy with a very low
coefficient of thermal expansion, was used for the low-λ outer shell in the present simulation (the so-called Invar).
The normalized coefficients of thermal expansion of these materials are presented in Fig. 3.
Although the applicability of Invar to the outer shell of a rocket engine combustion chamber should be investigated,
C

the main objective in this study was to examine the effect of restriction by the outer shell on lifetime of the inner
liner.
AC

2.3 Inelastic model

The chamber wall of a reusable rocket engine will be subjected to cyclic thermal stress loading in a plastic
region. To simulate behaviour of inner liner materials of a combustion chamber wall, the nonlinear
isotropic/kinematic hardening model installed in ABAQUS/standard was used. This model consists of a
nonlinear kinematic hardening component and an isotropic hardening component [19].
The kinematic hardening component is given by

1
α& k = Ck (σ − α ) ε& pl − γ kα k ε& pl (1)
σ 0

4
 ACCEPTED MANUSCRIPT

σ 0 and ε& pl are the size of the yield surface and the equivalent plastic strain rate, respectively. C k and γ k are
the material parameters calibrated from cyclic test data.
The overall backstress is calculated by

N
α = ∑αk (2)
k =1
N is the number of backstresses.
The isotropic hardening component is given by

PT
(
σ 0 = σ |0 +Q00 1 − e − bε
pl

) (3)

RI
σ |0 is the yield stress at zero plastic strain. Q00 and b are material parameters.
For these copper alloys, uniaxial tensile fatigue tests were conducted at several temperatures. The parameters to
define the nonlinear kinematic hardening law in ABAQUS were defined from the 1st cycle data of fatigue tests,

SC
and those of the isotropic hardening properties were defined from several data sets between the 10th and 50th
cycles.
By using a nonlinear isotropic/kinematic hardening model of ABAQUS, the components of which were defined by

U
experimental data of fatigue tests, each test was simulated with a 2D axisymmetric FEM model with the equal
conditions related to the experiments. Figure 5 shows a comparison between calculated data and test data by a stress-
strain hysteresis curve for the material of Cu-B at 873 K. Figure 5 shows the 1st and 30th cycles.
AN
The calculations were conducted for the fatigue tests of Cu-A, Cu-B, and Cu-C, changing the temperature between
about 300 K and 900 K and the strain amplitude of 2 % and 4%. The deviation between the test and the simulation
depends on the model. The maximum deviations for strain and stress were approximately 20 %. The main objective
of the present work was not so much to predict the lifetime of a combustion chamber quantitatively as to investigate
M

possible keys to extend the lifetime of a combustion chamber with a large thrust. Hence, an inelastic model was not
further pursued in this work.
D
TE
C EP

Figure 5: Comparison of fatigue test and FEM simulation for one of the uniaxial fatigue tests of Cu-B. Stress-strain
AC

curves at 873 K for the 1st cycle (left) and the 30th cycle (right)

2.4 Estimation of low cycle fatigue and creep damages

In the present simulations, a typical combustion cycle was assumed to be start-up for 5 seconds, a steady state
combustion phase for 400 seconds, shut-down for 2 seconds, and a chill-down phase for 2000 seconds. This
combustion cycle was repeated three times. In the final cycle, the total range of equivalent strain, ∆ε, average
equivalent strain, σ, and average temperature, T, during a steady state combustion phase were estimated.
The number of cycles to rupture, Nr, was derived using ∆ε from the fatigue curve of each copper alloy. The time to
rupture due to creep, tr, was derived from the creep curve by using σ and T. The damages due to low cycle fatigue
and creep were derived using Eqs. 3 to 6. The total damage was the summation of df and dc, assuming a linear
cumulative damage rule. tc is the duration time of a steady state combustion phase. These equations were used to
evaluate damages as a post processing procedure and damages do not affect the finite element analysis.
 ACCEPTED MANUSCRIPT

Since equivalent stress showed a slight change with time even in the steady state combustion phase, the total creep
damage was derived by integrating creep damages estimated in a short time inverval.

1
df = (4)
N r ( ∆ε )
tc
dc = (5)
t r (σ , T )
D = df + dc

PT
(6)
1
Nf = (7)
D

RI
U SC
AN
Figure 6: Damage evaluation points

Figure 6 shows points to evaluate damages and lifetimes of the chamber wall. P1 is located at the center of a rib
facing combustion gas. P2 and P3 are located at the center of a ligament facing combustion gas and coolant,
respectively. P4 is located at the corner of the coolant channel. Damage was estimated for all points, the total damage
M

of P1 being the largest among all the present models. Hence, in this paper, effects on lifetime are discussed at the
point P1.
For all models, calculations were conducted for three cycles of the combustion sequence as stated above, and the
D

creep damage and low cycle fatigue damage were evaluated for the last cycle using Eqs. 4 to 7.

3. Thermal barrier coating

TE

3.1 Material and processing

EP

Although the effectiveness of the TBC for enhancing the life of a liquid rocket combustion chamber has long been
pointed out [1,2], examples of the practical use of the TBC have been very limited, such as Cr/Ni coating of the RD-
170 engine [20]. Superior bonding strength of the TBC layer onto the copper alloy substrate is quite necessary to
endure huge buckling stress caused by both the extremely high-heat flux from combustion gas and the restriction of
C

the outer shell. Ceramic TBC, such as yttria stabilized zirconium dioxide (YSZ), is often used in gas turbines.
However, such brittle ceramic TBC cannot be applied to a rocket engine combustion chamber because it delaminates.
In this study, therefore, NiCrAlY which is usually used as a bond coat between a ceramic top coat and a metallic
AC

substrate in a gas turbine of a TBC system, was investigated. NiCrAlY coating inside a combustion chamber was
formed by atmospheric plasma spraying (APS), and then the coating layer was re-melted by a laser beam and fused
to the copper alloy substrate to improve the adhesion strength considerably. Figure 7 shows a cross-sectional photo
of NiCrAlY coating layer on a copper alloy substrate before and after the laser re-melting treatment that was
observed by Scanning Electron Microscopy. Although some porosities in the coating layer and at the interface
between the coating and the substrate were observed before the re-melting process (indicated “as sprayed” in Fig.7),
almost no porosities could be seen after the laser re-melting treatment (indicated “re-melted by laser” in Fig.7). The
NiCrAlY coating layer was densified, and simultaneously the ingredients of both the coating and the substrate were
diffused into each other, resulting in improvement of the bonding strength.

6
 ACCEPTED MANUSCRIPT

Figure 7: Cross-sectional photo of NiCrAlY coating layer on copper alloy substrate before and after laser re-melting
treatment

PT
3.2 Evaluation of thermal conductivity

RI
The thermal conductivity of the laser re-melted NiCrAlY coating layer was evaluated [21]. A NiCrAlY coating layer
with a thickness of about 150 µm was formed on a CuCrZr specimen, and the thermal diffusivity was measured by
the laser flash method. Thermal diffusivity of the laser re-melted NiCrAlY coating layer (αTBC) on the specimen was

SC
determined by Eq. 8 as follows:

2
dTBC
αTBC = (8)

U
τ TBC
where
AN
6 ( CSb + CTBC ) ⋅ A − 6 ( CSb + 3CTBC ) ⋅ ASb
τ TBC = (9)
3CSb + CTBC
M

CSb = cSb ρ Sb d Sb (10)

D

CTBC = cTBC ρTBC dTBC (11)

TE

where dTBC is the thickness of the NiCrAlY coating layer, τTBC is the thermal diffusion time of the NiCrAlY coating
layer, cSb is the specific heat of the substrate, ρSb is the density of the substrate, cTBC is specific heat of NiCrAlY, A is
the area thermal diffusion time of the whole specimen, and ASb is the area thermal diffusion time of the substrate.
The thermal conductivity of the re-melted NiCrAlY coating layer was then determined by Eq. (9),
EP

κTBC = cTBC ρTBCαTBC (12)

where κTBC, cTBC, ρTBC, and αTBC are thermal conductivity, specific heat, density, and thermal diffusivity, respectively.
C

Because the laser re-melted NiCrAlY coating layer was so thin, it was difficult to measure its specific heat and
AC

density. Therefore, specific heat and density of the laser re-melted NiCrAlY coating layer were substituted by using a
sprayed thick material of NiCrAlY. Figure 8 shows the thermal conductivity of the re-melted NiCrAlY coating layer
on the copper alloy substrate. The thermal conductivity was normalized by that of Cu-A or B at room temperature
(see Fig. 3) As can be seen from Fig. 8, a decrease in thermal conductivity around 800 K is observed, but the
mechanism of such decrease is unclear at present and detailed analysis should be conducted to determine the thermal
conductivity. In this study, life time analysis of a combustion chamber was conducted by using the thermal
conductivity data indicated in Fig. 8.
 ACCEPTED MANUSCRIPT

PT
RI
Figure 8 : Thermal conductivity of re-melted NiCrAlY coating layer

SC
3.3 Cyclic combustion gas heating test of the TBC

Cyclic heating test of TBC was conducted by using a subscale calorimetric chamber whose inner wall of the throat
section was covered with TBC. Figure 9 shows an assembly drawing of the test specimen. The part from contraction

U
section to nozzle section was divided into four parts and the inner surface of the third throat section was covered with
TBC. The Inner diameter of the throat was 26 mm. A NiCrAlY TBC layer was formed on the inner surface of the test
AN
specimen by APS, and the NiCrAlY layer was re-melted by a laser. Finally, the coating surface was polished and
finished by a lapping method to a thickness of 150 µm. Cyclic combustion gas heating test conditions are listed in
Table 1.
M
D
TE

Figure 9 : Assembly drawing of a cyclic combustion gas heating test specimen (TBC was formed on the inner
surface of throat section part)
EP

Table 1 : Cyclic combustion gas heating test conditions

C

Propellant GOX/GH2

Coolant water
AC

Combustion pressure (Pc) 2 MPa or 2.5 MPa

Mixture ratio 6.5

Steady combustion time 10 sec

The cyclic combustion gas heating tests were terminated after 24 cycles because of the leakage of water from the
coolant channel of the contraction section. Figure 10 indicates heat flux and the maximum surface temperature of the
TBC calculated from chamber wall temperature measured by thermo-couples. Heat flux was about 21 MW/m2 and
estimated TBC surface temperature was about 1020 K when the chamber pressure Pc was 2 MPa, and heat flux was
about 28 MW/m2 and TBC surface temperature was about 1180 K when Pc was 2.5 Mpa.

8
 ACCEPTED MANUSCRIPT

Figure 11 shows an image of a TBC test specimen after cyclic heating tests. As can be seen from Fig.11, there was
no damage to the TBC coating layer. Because delamination of the TBC was observed after three cycles of the cyclic
combustion gas heating test without a laser re-melting process, the laser re-melting process is very effective. The
cyclic combustion gas heating tests will be continued to evaluate the durability of TBC for reusable rocket engine
combustion chamber application at the Kakuda Space Center of JAXA.

PT
RI
U SC
Figure 10 : Heat flux and maximum surface temperature of TBC
AN
M
D

Figure 11 : Appearance of TBC test specimen after cyclic heating test

TE

4. Results
EP

4.1 Effect of material properties of an inner liner

In this section, the effect of three material models of the copper alloy used in the inner liner on creep damage and
C

low cycle fatigue damage and lifetime were examined for the model A, model B, and model C. Furthermore, an outer
shell made of a normal material was considered and no TBC was assumed.
AC
 ACCEPTED MANUSCRIPT

Figure 12: Estimated fatigue damage and strain range (left) and creep damage and average stress (right) for three
copper alloys, Cu-A, B, and C of the inner liner

On the left-hand side of Fig. 12, low cycle fatigue damage and total strain range which was calculated from
equivalent strain are plotted by using a bar graph and a line with circles, respectively. The strength of the copper
alloy reflects the magnitude of the total strain range. The total strain range of Cu-A was the lowest because its yield
stress is largest. The total strain range of Cu-C, of which the yield stress is smallest, was the largest. However, since
the fatigue strength of Cu-C is the largest among the three copper alloys, the low cycle fatigue damage was much
smaller than those of Cu-A and B.
In the graph on the right of Fig. 12, the creep damage and average equivalent stress during the steady state

PT
combustion phase are presented. The average stresses are plotted by a solid line with symbols and become smaller in
the order of Cu-A. B, and C. As shown in Fig.4, the yield stresses of these copper alloys decrease in order of Cu-A,
B, and C. Hence, for the inner liner with the copper alloy of larger yield stress, the stress level during the steady
combustion phase becomes larger. Although the creep strength of Cu-C is smaller than those of Cu-A and B, the

RI
effect of low level stress resulted in the smallest creep damage of Cu-C.
In the left-had side of Fig.13, the total damage estimated at P1 in the inner liner for three copper alloys is plotted
together with low cycle fatigue damage and creep damage. The creep damage was much lower than the low cycle

SC
fatigue damages, and then the total damage was dominated by low cycle fatigue damage for these models. As a result,
the lifetime of the chamber wall of model C is longest, around five times longer than the others, as shown in the
graph on the right-had side of Fig.13.

U
AN
M
D
TE

Figure 13: The total, low cycle fatigue, and creep damages estimated at P1 of three models (left) and estimated
lifetime cycles (right) of the chamber wall for nominal outer shell without TBC
EP

4.2 Effect of a thermal barrier coating

A thermal barrier coating on the surface of the inner liner reduces heat flux from combustion gas and decreases wall
C

temperature. Hence, a TBC can improve the lifetime of the chamber wall. However, a decrease of wall temperature is
causing a rise in the yield stress of the inner liner material. This increased the stress level during the steady state
AC

combustion phase that results in an increase of creep damage.

Creep damage and average equivalent stress during a combustion phase for the inner liner material Cu-A and Cu-C
are presented in Fig. 14. For both copper alloys, by adding a TBC layer on the inner surface of the chamber wall, the
average equivalent stress increased. Since the yield stress of Cu-A is much higher than that of Cu-C, the stress
increase of Cu-A during steady state combustion is higher than for Cu-C.
On the other hand, the wall temperature decreased remarkably by the addition of a TBC layer on the surface of the
inner liner. The decreases of wall temperature from that of the model without TBC were about 90 K for 50 µm and
about 140 K for 100 µm, respectively.
Hence, the creep damage of Cu-A for the TBC thickness of 50 µm increased due to the increase of stress level,
despite the decrease of wall temperature. For the TBC thickness of 100 µm, the stress almost reached its yield stress
and was nearly the same as that of 50 µm. The wall temperature decreased and thus the creep damage decreased even
in model A.

10
 ACCEPTED MANUSCRIPT

PT
Figure 14; Effect of TBC on creep damage and average equivalent stress during the steady state combustion phase

RI
for model A and model C

Figure 15 presents estimated damages of the total, low cycle fatigue, and creep for models A, B, and C. The graph

SC
on the left shows damages of the models without TBC, which is the same as in the left of Fig.13. In the middle and
on the right-hand side show the influence on damage by using a TBC with a thickness of 50 µm and 100 µm. By
adding a TBC layer on the surface of the inner liner, the total damage decreases. Especially the decrease of creep
damage was remarkable.

U
As shown in Fig.14, in model A, the increase of stress due to the decrease of wall temperature resulted in an increase
of the creep damage, and the total damage of the model with a 50 µm TBC layer became almost the same as that of
the model without a TBC layer.
AN
M
D
TE

Figure 15: The total, fatigue, and creep damages estimated at P1 of three models. The left graph shows damages
without TBC, the middle for 50 µm, and the right for 100 µm
EP

The estimated lifetimes of model A, model B, and model C are plotted in Fig. 16 with or without TBC. The
improved effect of lifetime is clearly seen in models B and C. Especially, the lifetime of the inner liner of model C
showed remarkable improvement. Compared with the model without TBC, the lifetime of the model with 100 µm
TBC layer was four times of that of the inner liner without TBC.
C
AC

Figure 16: Comparison of lifetimes of the inner liner of three copper alloys with or without TBC
 ACCEPTED MANUSCRIPT

λ outer shell
4.3 Effect of a low-λ

The inner liner tends to expand while being heated from the combustion gas due to the temperature rise on the wall.
On the other hand, the outer shell is cooled by the coolant which flows in coolant channels and tends to contract. In
the present chamber wall model in which the outer shell is made of normal material as explained in the section 2.3,
the chamber wall showed contraction during the combustion phase. The radial displacement at position P1 of model
C in the last combustion cycle is plotted in Fig.17. In the start-up phase, the chamber wall rapidly contracted and
remained an almost constant radius during the steady state combustion phase. The magnitude of contraction during
the steady state combustion phase was about -0.3 mm. After the combustion finished, the chamber started to expand.

PT
In the present simulation, the effect of contraction on damages to the inner liner was examined. Therefore, a metal
material with very low coefficient of thermal expansion was applied to the outer shell. For the model C, Invar was
assumed as the material of the outer shell. The calculated radial displacement at position P1 during the last cycle is
plotted in Fig. 17. The chamber wall expanded during the start-up phase and showed a behaviour opposite to that of

RI
the normal case. The radial displacement during the steady state combustion phase was about +0.15 mm.
As shown in Fig.17, by using Invar for the outer shell, the magnitude of radial displacement became smaller than
that of the model using the normal outer shell, resulting in a decrease of the total strain range in the combustion cycle.
On the left-hand side of Fig.18, the fatigue damage and total strain range are presented for three models. For all

SC
models, the total strain ranges were smaller than those of the normal outer shell.

U
AN
M
D

Figure 17: Radial displacement during the 3rd combustion cycle of model C and of the outer shell made of stainless
TE

steel and Invar

The low cycle fatigue damage of the model of a low-λ outer shell showed smaller fatigue damages about 50% than
that of the model of a normal outer shell. On the right-hand side of Fig.18, the estimated lifetime of the chamber wall
EP

for all models are presented. The present simulation results suggested that the lifetime would be extended by using
Invar for the outer shell. Especially, the use of Invar for the model in which Cu-C was used as an inner liner material
was found to be most effective to extend lifetime.
C
AC

Figure 18: The low cycle fatigue damage and total strain range

12
 ACCEPTED MANUSCRIPT

5. Conclusions

Effects of material properties of the inner liner and outer shell and a thermal barrier coating on the lifetime of a
combustion chamber wall of large thrust rocket engines were investigated by calculating creep damage and low cycle
fatigue damage on the inner liner for repeated firing cycles. Major results of the present work are as follows,

Even for the same copper alloys with chromium and zirconium, the mechanical properties largely differ according to
their manufacturing process and thermal treatment. From the present simulations, their differences strongly affected
the lifetime cycle of the chamber wall. Especially, the yield stress of the inner liner material is one of the key factors

PT
affecting lifetime. For the combustion chamber of the inner liner material with lower yield stress, the low cycle
fatigue damage tended to be lower due to higher fatigue strength and the creep damage to be lower due to lower
stress level during steady state combustion, resulting in extending the lifetime.

RI
A thermal barrier coating on the surface of the inner liner reduces wall temperature and decreases the low cycle
fatigue damage and creep damage, resulting in the extension of the lifetime of the combustion chamber. However,
for the copper alloy with high yield stress, TBC may have the opposite effect of increasing creep damage.

SC
The outer shell with a low coefficient of thermal expansion reduces restriction of the inner liner and could extend the
lifetime of the chamber wall.

By combination of these effects, low yield stress material for the inner liner, a thermal barrier coating, and a low

U
thermal expansion outer shell, the lifetime of a combustion chamber wall could be extended about 10 times longer
than that of a combustion chamber without any effects.
AN
Nomenclature

A = area thermal diffusion time

M

c = specific heat
d = thickness
df = damage due to low cycle fatigue (LCF)
D

dc = damage due to creep

D = total damage
Nf = number of lifetime cycles of combustion chamber
TE

Nr = number of cycles to rupture

T = average temperature during steady-state combustion
tc = duration time of steady-state combustion
EP

tr = time to rupture due to creep

α = thermal diffusivity
∆ε = total strain range
C

k = thermal conductivity
λ = coefficient of thermal expansion
AC

ρ = density
σ = average equivalent stress during steady-state combustion
τ = thermal diffusion time
subscript
TBC = thermal barrier coating
Sb = substrate

References

[1] R.J. Quentmeyer, Rocket Combustion Chamber Life-Enhancing Design Concepts, 26th Joint Propulsion
Conference and Exhibit, (1990), AIAA Paper 1990-2116.
[2] M. Popp and G. Schmidt, Rocket Engine Combustion Chamber Design Concepts for Enhanced Life, 32nd Joint
Propulsion Conference and Exhibit, (1996), AIAA Paper 1896-3303.
 ACCEPTED MANUSCRIPT

[3] I.-K. Sung, B.G. Northcutt, K.S. Rubel, and W.E. Anderson, A Subscale-Based Life Prediction Methodology
For Rocket Engine Combustors, 39th Joint Propulsion Conference and Exhibit, (2003), AIAA Paper 2003-4900.
[4] I.-K. Sung and W.E. Anderson, A Subscale-Based Rocket Combustor Life Prediction Methodology, 41st Joint
Propulsion Conference and Exhibit, (2005), AIAA Paper 2005-3570.
[5] W. Riccius, B. Bouajila, and E.B. Zametaev, Comparison of Finite Element analysis and experimental results of
a combustion chamber type TMF panel test, 49th Joint Propulsion Conference, (2013), AIAA Paper 2013-3846.
[6] W. Riccius, E.B. Zametaev, and B. Bouajila, Inner liner temperature variation caused deformation localization
effects in a multichannel model of a generic LRE wall structure, 50th Joint Propulsion Conference, (2014),
AIAA Paper 2014-3988.
[7] R.G. Thiede, E.B. Zametaev, J.R. Riccius, M. Fassin, and S. Reese, Comparison of Damage Parameter Based

PT
Finite Element Fatigue Analysis Results to Combustion Chamber Type TMF Panel Test Results, 51th Joint
Propulsion Conference, (2015), AIAA Paper 2015-4070.
[8] M. Nishimoto, H. Negishi, S. Yoshimura, N. Kasahara, H. Akiba, H. Sunakawa, S. Hori, and N. Negoro,
Failure mechanism and life evaluation of liquid rocket engine combustion chamber under sever multi-physics

RI
condition, Trans. Japan Soc. Mech. Eng., 81 (2015), pp. 1-18 (in Japanese).
[9] H. Amakawa, N. Negishi, M. Nishimoto, and S. Hori, Numerical Investigation of Longer Life Combustion
Chambers of Liquid Rocket Engines Based on Coupled Thermal-Fluid-Structure Simulation, 58th Structures,

SC
Structural Dynamics, and Materials Conference, (2017), AIAA Paper 2017-1989.
[10] T. Kimura, T. Hashimoto, M.Sato, S. Takada, T. Yagishita, Y. Naruo, H. Ogawa, T. Ito, K. Obase, Y. Fukuda,
and T. Kai, Firing Tests for the Main Engine of the Reusable Sounding Rocket, Proceedings of 30th
International Symposium on Space Technology and Science, (2015).

U
[11] T. Kimura, T. Hashimoto, M. Sato, S. Takada, S. Moriya, T. Yagishita, Y. Naruo, H. Ogawa, T. Ito, K. Obase,
and H. Ohmura, Reusable Rocket Engine -Firing Tests and Lifetime Analysis of Combustion Chamber, J. Prop.
Power 32 (2016) , pp. 1087-1094.
AN
[12] T. Ito, S. Nonaka, and H. Ogawa, Progress of Technical Demonstrations for Reusable Sounding Rocket System,
Proceedings of 30th International Symposium on Space Technology and Science, (2015).
[13] A. Kurosu, N. Yamanishi, H. Sunakawa, M. Nishimoto, and K. Okita, LE-X -Japanese Next Liquid Booster
Engine, 44th Joint Propulsion Conference and Exhibition, (2008), AIAA Paper 2008-4665.
M

[14] H. Sunakawa, M. Nishimoto, N. Yamanishi, K. Noda, E. Sato, and M. Shiwa, Reliability and Life Time
Evaluation Approach of the LE-X Engine, 47th Joint Propulsion Conference and Exhibition, (2011), AIAA
Paper 2011-5933.
[15] S. Gordon and B.J. McBride, Computer Program for Calculation of Complex Chemical Equilibrium
D

Compositions, Rocket Performance, Incident and Reflected Shocks, and Chapman-Jouguet Detonations,
(1976), NASA SP-273.
TE

[16] D.R. Bartz, A Simple Equation for Rapid Estimation of Rocket Nozzle Convective Heat Transfer Coefficients,
J. Jet Propulsion, 27 (1957) 49-53.
[17] D. Ellis and M. Michal, Mechanical and Thermal Properties of Two Cu-Cr-Nb Alloys and NARloy-Z, (1996),
NASA CR-198529.
EP

[18] D. Ellis, GRCop-84 : A High-Temperature Copper Alloy for High-Heat Flux Applications, 2005, NASA TM-
2005-213566.
[19] 23.3.3 Models for metals subjected to cyclic loading, in :Abaqus Analysis User's Guide, (2016).
[20] V. Fedorov, V. Chvanov, F. Chelkis, A. Polyansky, N. Ivanov, I. Lozino-Lozinskaya, and A. Buryak, The
Chamber Cooling System of RD-170 Engine Family: Design, Parameters, and Hardware Investigation Data,
C

42nd Joint Propulsion Conference and Exhibit, (2006), AIAA Paper 2006-4363.
[21] Metallic and other inorganic coatings – Determination of thermal conductivity of thermal barrier coatings, ISO
AC

18555:2016. https://www.iso.org/standard/62875.html (accessed 5 April 2018).

14
 ACCEPTED MANUSCRIPT

The properties of the copper alloy of an inner liner strongly affect the lifetime of a chamber wall.

A thermal barrier coating on the surface of the inner liner effectively extends the lifetime.

The outer shell with a low coefficient of thermal expansion can extend the lifetime.

PT
By combination of these effects, the lifetime will be extended about 10 times longer than normal
one.

RI
U SC
AN
M
D
TE
C EP
AC