ORIGINAL PAPER

Thermal fatigue analysis of small-satellite structure

Gasser Farouk Abdelal Nader Abuelfoutouh
Ahmed Hamdy Ayman Atef
Received: 11 November 2006 / Accepted: 12 January 2007 / Published online: 16 February 2007
Springer Science+Business Media B.V. 2007
Abstract The small-satellite thermal subsystem
main function is to control temperature ranges on
equipments, and payload for the orbit specied.
Structure subsystem has to ensure the satellite
structure integrity. Structure integrity should
meet two constraints; rst constraint is accepted
fatigue damage due to cyclic temperature, and
second one is tolerable mounting accuracy at
payload and Attitude Determination and Control
Subsystem (ADCS) equipments seats. First,
thermal analysis is executed by applying nite-
difference method (IDEAS) and temperature
prole for satellite components case is evaluated.
Then, thermal fatigue analysis is performed
applying nite-element analysis (ANSYS) to
calculate the resultant damage due to on-orbit
cyclic stresses, and structure deformations at the
payload and ADCS equipments seats.
Keywords Aerospace Satellite Structure
Finite difference Finite element Thermal
Mounting accuracy Fatigue
1 Introduction
Structure made of different materials can experi-
ence thermal stresses without external constraints
even under uniform temperature. Differences in
the materials coefcients of thermal expansion
produce incompatible strains and resulting ther-
mal stresses. These stresses balance when no
external constraints are present.
Thermally induced loads and stresses are
limited by deformation; once a material reaches
its proportional limit, or once the structure begins
to buckle, thermal stress no longer increases in
proportion to the change in temperature. Ductile
materials seldom rupture or buckle from a single
application of thermal stress, but they can fail in
fatigue from the many cycles of thermal loading
common to orbiting satellite.
Actual thermal-design problems for satellite
are complicated. The design problem typically
must combine multiple modes of heat transfer
with time varying boundary conditions that
require transient instead of steady-state solutions.
To predict satellites temperatures, the thermal
analysis problem combines two types of math
models. The rst one is a thermal-radiation model
G. F. Abdelal (&) A. Hamdy A. Atef
Egyptian Space Program, National Authority for
Remote Sensing and Space Science (NARSS), 23
Jozef Broz Teto Street, Elnozha Elgededah, Cairo
11769, Egypt
e-mail: [email protected]
N. Abuelfoutouh
Aerial Photograph Division Program, National
Authority for Remote Sensing and Space Science
(NARSS), 23 Jozef Broz Teto Street, Elnozha
Elgededah, Cairo 11769, Egypt
123
Int J Mech Mater Des (2006) 3:145159
DOI 10.1007/s10999-007-9019-1
to calculate external heating rates by simulating
the external geometry of the satellite including
surface properties. By subjecting this model to a
simulated orbit, the output from this mode
consisting of; the environmental heating rates
due to direct solar, albedo, and planetary emis-
sions; and external radiation between satellite
surfaces; becomes input for the second model.
The second thermal model uses a thermal
analyzer. The satellite is modeled much the same
as the structural analysis model with internal
details but the analysis is based on the nite
difference method. Then the heating sources are
dened. Finally, the model is solved to simulate
the heat transfer paths of conduction, convec-
tion, and radiation. The thermal analyzer calcu-
lates temperatures at all nodes for steady-state
or transient conditions by solving energy
equations.
A thermoelastic analysis for Small Sat struc-
ture due to on-orbit cyclic temperature is consid-
ered. It calculates stresses and distortions due to
temperatures cyclic loading. The results of ther-
moelastic analysis are used to calculate fatigue
damage due to on-orbit cyclic stresses. Moreover,
it is used to check mounting accuracy of the
precise equipments (MBIE and ADCS equip-
ments) after on-orbit thermal deformation. The
most critical structural module in Small Sat
regarding by on-orbit thermal deformation is the
basis unit block module. This module is very
sensitive to thermal deformation because it
carries the precise equipments of the satellite.
The rest of structural modules do not have
severe restrictions on their equipment mounting
accuracy.
2 Satellite overview
The satellite conguration is shown below in
Fig. 1. The structure is a 1.03 0.8 1.0255 m
cube with four solar panels attached to the
satellite by means of rotation mechanism. Four
heat shields are installed on the satellite structure
to prevent internal instruments from direct envi-
ronmental heat loads. The highest power con-
suming components will be placed away from hot
heat shields subjected to solar radiation. The
primary main elements of the satellite structure
are shown in Fig. 2.
3 Mission requirements and constraints
The temperature constraint is specied for each
component. Table 1 lists temperature ranges for
different satellite components. Component tem-
perature constraints contain an upper and a
4 Solar panels
4 Heat shields
Satellite case
Fig. 1 The satellite under study
Fig. 2 3-D model of satellite structure
Table 1 Satellite component temperature limits
Component Temperature constraints
Power subsystem 5 to 45C
Communication subsystem 5 to 45C
Attitude control subsystem 5 to 45C
Onboard computer 0 to 40C
Solar panel 100 to 85C
Structure 100 to 100C
146 Int J Mech Mater Des (2006) 3:145159
123
lower bound of each components preferred
operating temperature range. The thermal
design of the satellite must not allow any
component to exceed absolute operating
temperature ranges. Components operate most
efciently and effectively when preferred oper-
ating temperatures are maintained. In addition,
component life span will be degraded by oper-
ating the component outside of its preferred
temperature range.
The satellite orbits Earth at the following
orbital parameters:
Sun-synchronous orbit
Altitude: 576 km
Inclination angle: 97.721
Local time of descending node: 9 h: 45 min
b Angle is dened as the angle between the
solar vector and the normal to the orbital plane.
According to the orbital parameters listed above,
Beta angle (b) can be calculated mathematically
as (Gilmore 1994):
b sin
1
cos d
S
sin i sinXX
S
sin d
S
cos i
1
Where, d
S
= declination of the sun; i = orbit
inclination; (W W
S
) = local time of descending
node.
Figure 3 shows the b angle history during one
year. b Angle will reach its maximum value
(about 63) at summer solstice while the min-
imum value will be at winter solstice (about
55).
4 Thermal model
Thermal analysis involves constructing a geomet-
ric math model (GMM) and a thermal math
model (TMM) of the satellite and identifying
analysis cases to be run. The GMM and the TMM
serve different purposes. The GMM is a mathe-
matical representation of the physical surfaces of
the satellite and is used to calculate black body
radiation couplings between surfaces as well as
heating rates due to environmental uxes. The
TMM is most often a lumped parameter network
representation of the thermal mass and conduc-
tion and radiation couplings of the satellite, and is
used to predict satellite temperatures. The radi-
ation interchange couplings and environmental
heat uxes calculated by the GMM are used in
constructing the TMM. Both the GMM and
TMM are constructed and executed using I-
DEAS TMG. First, the GMM of the satellite
was constructed using I-DEAS TMG. The model
consists of a simple representation of the satellite
and the payload. It was constructed using rectan-
gular, circular, and cylindrical surfaces, and each
surface was assigned the appropriate solar
absorptivity and emissive. Dimensions and mass
of each component of the satellite can be found in
Table 2.
A transient-state analysis was conducted to get
the temperature variation of each component
with during one complete orbit.
The effects of the Beta angle are considered by
using different possible orbital parameters. This
angle plays an important roll in determining the
eclipse and sunlight time periods for the satellite
Fig. 3 b angle history
during 1 year
Int J Mech Mater Des (2006) 3:145159 147
123
and hence irradiation due to different sources
such as the Sun, Earth and albedo.
Black paint is used to coat the electronic boxes
and on the inside surfaces to improve energy
exchange with the other subsystems onboard the
satellite. Moreover, the external structure is
coated by white paint to minimize solar radiation
absorption. Finally, Multilayer insulation (MLI)
is used to control the temperature of the payload.
A catalog of the thermal properties of each
material used on the satellites is listed in Table 3.
The element chosen from the I-DEAS TMG
library is linear quadrilateral thin shell element.
Table 2 Number of nodes and elements for some selected
structural modules
Satellite structure
module
No. of elements No. of nodes
Basis plate 146 242
Basis walls 66 104
Star sensor bracket 32 61
Table 3 Mass, material properties, and thermo-optical properties of the different coatings applied
Subsystem and Component Mass (kg) Material Coating
Payload
Multi-band earth imager 45 Titanium alloys Chemglaze Black paint Z306 (inside)
Multi-layer Insulation (outside)
Payload CDHS unit 7.2 Aluminum 2024-Tx Chemglaze White paint A276
MEI signal processing unit 3.7 each Aluminum 2024-Tx Chemglaze White paint A276
ADCS
Star sensor 4 Aluminum 2024-Tx Chemglaze Black paint Z306
Angular velocity meter Gyro 1 each Aluminum 2024-Tx Chemglaze White paint A276
Interface unit for each gyro 0.92 each Aluminum 2024-Tx Chemglaze White paint A276
Magnetometer 1.5 Aluminum 2024-Tx Chemglaze White paint A276
Magnetorquer 0.38 each Aluminum 2024-Tx Chemglaze White paint A276
Reaction wheel 3.3 each Aluminum 2024-Tx Chemglaze White paint A276
Communications subsystem
X-band equipment
X-band electronic module 3.8 Aluminum 2024-Tx Chemglaze White paint A276
X-band antenna 1.6 Aluminum 2024-Tx Uncoated
S-band equipment
S-band electronic module 2.2 each Aluminum 2024-Tx Chemglaze White paint A276
S-band conical antenna 0.27 each Aluminum 2024-Tx Uncoated
S-band dipole antenna 0.13 Aluminum 2024-Tx Uncoated
GPS receiver
GPS electronic module 1.1 Aluminum 2024-Tx Chemglaze White paint A276
GPS antenna 0.15 each Aluminum 2024-Tx Uncoated
Platform CDHS
On-board digital computing complex 3.7 each Aluminum 2024-Tx Chemglaze White paint A276
Telemetry module 2.8 Aluminum 2024-Tx Chemglaze White paint A276
Power Subsystem
Battery cell module 16.5 Ni-Cd Chemglaze Black paint Z306
Power-conditioning unit (PCU) 3.6 Aluminum 2024-Tx Chemglaze White paint A276
Cells leveling unit (CLU) 1.9 Aluminum 2024-Tx Chemglaze White paint A276
Solar array panels 6.8 Gallium arsenide Cell side: a
s
= 0.75, e = 0.9 Backside:
Chemglaze Black paint Z306
Thermal subsystem
Heat shields 3.6 Honeycomb structure Chemglaze Black paint Z306
Multi-layer Insulation 0.2
Structure and mechanism subsystem
Satellite structural modules
Rotation mechanism 41 Aluminum 2024-Tx Chemglaze Black paint Z306
Locking and releasing mechanism 1.7 Aluminum 2024-Tx
Separation transducer 0.5 Aluminum 2024-Tx
0.1 Aluminum 2024-Tx
Satellite total mass 205
148 Int J Mech Mater Des (2006) 3:145159
123
This type of elements assumes a constant thick-
ness. Because of the small thickness of all the
elements, this thin shell was the most appropriate
element which could represent it without using an
excessive number of elements. The thermal
model consists of 8,574 shell elements with
different thicknesses and 1,0138 nodes. Figure 4
shows the structure of the satellite after complete
meshing and Table 2 dene the mesh character-
istics for some selected elements.
4.1 Modeling parameters
The I-DEAS TMG software uses a nite differ-
ence method to solve the heat balance equation
and to get the temperature distribution in the
model. The heat balance equation for a transient
run to be solved iteratively at iteration n + 1 and
time t + Dt for element i can be cast in the form
(Camack and Edwards 1960):
c
i;t
T
i;tDt;n1
T
i;t
_ _
Dt
Q
i;tDt
ta
s
A
s
q
s
i;tDt
t
aA
A
q
A
i;tDt
t
eA
E
q
E
i;tDt
t
r

N
j1
F
ij
A
ij
T
4
j;tDt;n1
T
4
i;tDt;n1
_ _

N
j1
K
ij
T
j;tDt;n1
T
i;tDT;n1
_ _
2
where, C
i,t
= the capacitance of the element;
Ti, t + dt, n + 1 = the temperature of element i at
time t + dt and iteration n + 1; n = the current
iteration value; Ti,t = the temperature of element
i at time t; Dt = the integration time step;
Qi,t + dt(t) = the heat generation rate of element
i as a function of time at time t + dt; Gij = the sum
of the conductive conductances between i and j; r
= Stefan Boltzmann constant; Fij = gray body
view factor between elements i and j; A =
radiating surface area of element i.
The Backward method was used to solve Eq.
(2). In particular, it is most effective for this
model where s min is small compared to the
solution interval. In addition, it is more accurate
than the ForwardBackward method under
conditions of rapid temperature change.
Listing of Transient Analysis Parameters,
1. Start Time = 0.0 s.
2. End Time = 116686.8 s.
3. Integration Method: Backward.
4. Time Step value = 58.8 s.
5. Maximum Number of Iterations = 500.
6. Relaxation Factor = 0.05.
7. Temperature Difference for Convergence =
0.1.
Listing of Solution Methods,
1. Solution Method: Conjugate Gradient
2. Iteration Limit = 300.
Fig. 4 Satellite structure after meshing on I-DEAS TMG
Fig. 5 Shown temperature distribution of base plate
Int J Mech Mater Des (2006) 3:145159 149
123
3. Convergence Criterion = 0.1.
4. Preconditioning Matrix Fill value = 10.
5 Thermal analysis results
Satellite heating and cooling occurs while it goes
through sunlight and eclipse respectively. Maxi-
mum and minimum temperatures for the satellite
will be calculated for the orbits sunlight and
eclipses zones. Temperature variations of basis
plate and basis walls during one orbit are shown
in Figs. 6 and 7. Table 4 shows maximum and
minimum temperatures for various satellite com-
ponents.
6 Thermo-elastic analysis
The most critical structural module in Small Sat
affect with on-orbit thermal deformation is the
basis unit block module. This module is very
sensitive for the thermal deformation because it
carries the precise equipments into the satellite,
Fig. 6 Temperature
variations of basis walls
during one orbit
Fig. 7 Temperature
variations of basis plate
during one orbit
150 Int J Mech Mater Des (2006) 3:145159
123
otherwise, the rest of structural modules have not
big restrictions on their equipments mounting
accuracy. The following points are taken into
consideration during building this model:
I. Basis unit block module consists of basis
plate, four basis walls, two diagonal struts,
and star sensor bracket.
II. The material is Aluminum alloy AMg6
(E = 7.2 10
3
MPa, m = 0.33, q = 2,630
kg/m
3
, r
y
= 150 MPa, r
u
= 310 MPa). For
thermal analysis, it has the following
properties:
III. Thermal conductivity (K) = 117 W/m.C
IV. Coefcient of thermal expansion (a) =
24.7 10
6
m/m/C
V. Assign the location of mounting the high
precise equipments located at basis unit
block module to calculate thermal defor-
mation at their mounting points and hence
their mounting accuracy.
VI. Equipment support is designed to elimi-
nate thermal loading as they allow the
sliding of this support under thermal
expansion. Therefore, in the simplied
thermal model all satellite equipment is
removed to avoid the appearance of any
articial thermal stresses that may result
from the improper representation of the
sliding support.
VII. Basis unit block module is meshed with
35,015 elements for the basis plate and
49,096 elements for the basis walls, while
the rest of structural modules are meshed
with coarse meshing. SOLID92 elements
are used during meshing process.
VIII. The nodes located at connection areas
between the different structural modules
are coupled at all degrees of freedom.
Figure 8 shows the entire nite element model
of Small Sat used during on-orbit thermal defor-
mation analysis.
Thermal deformations are evaluated for the
worst cases, which are dened by the on-orbit
thermal analysis for the satellite structure. Ther-
mal satellite engineer performs a complete on-
orbit thermal analysis of the satellite and supply
input data for surface temperatures gradient of
different satellite structure modules. These data
are considered as an input data to perform on-
orbit thermal deformation analysis. For Small Sat,
the satellite structure is solved thermally due to
on-orbit Input data for surface temperatures
gradient of different satellite structure modules
of Small Sat are listed in Table 2. It contains
maximum and minimum average temperatures
for each structural module.
Displacement boundary conditions are usually
specied at model boundaries to dene rigid
support points. During on-orbit operation, the
satellite is totally free without any xation points,
but to conduct a thermo-elastic analysis, the
displacement boundary condition must be dened
for the related model. Therefore, to simulate the
satellite thermal deformation due to on-orbit
cycling, the satellite model must be constrained
Table 4 Surface temperatures of different satellite struc-
ture modules
Satellite
structure
module
Corner
or Face
Maximum
temperature,
( C)
Minimum
temperature,
( C)
Base plate I 5.3 10.1
II 1.8 6.2
III 7.2 4.9
IV 5.4 1.6
Mounting plate I 22 11.9
II 4.5 7.3
III 21.2 9.4
IV 18.6 7.5
Basis plate I 17 9.5
II 19.8 4.7
III 17 10.6
IV 14.6 12
Basis walls I-II 32.8 31.3
II-III 21.9 20.2
III-IV 21.9 20.2
IV-I 21.9 20.2
Upper frame I-II 36.1 35
II-III 28.5 29.6
III-IV 32.5 31.7
IV-I 36.3 35.2
Lower frame I-II 18.5 16.3
II-III 12 5.9
III-IV 17.6 12.8
IV-I 13.5 7.2
II 14.7 10.3
III 15.2 8
Diagonal strut I 14.3 13.2
IV 16.6 15.5
Star sensor bracket 18.8 17.4
Int J Mech Mater Des (2006) 3:145159 151
123
by rigid support points. Selection of these points
locations and their xation manner should
provide minimum effect on satellite deformation.
For Small Sat model represented at Fig. 8, the
support points are selected at plan of star sensor
mounting with its bracket. This plan is selected
because the mounting accuracy of all precise
equipments installed at the basis unit block is
measured relative to the star sensor bracket.
Displacement boundary conditions are applied by
xing one of the star sensor four points of xation
(on bracket) in X direction. Then, x the next
point in the Y and one of the other in the Z.
By reviewing the input data listed at Table 2,
some of the satellite structural modules are
divided into more than one division according to
their position. Each division has its maximum and
minimum average surface temperature. Before
conduct a thermo-elastic analysis, a thermal anal-
ysis process is performed to expand temperatures
through all solid elements and calculate temper-
ature distribution for all nodes. This process is
performed twice for the worst cases, maximum
and minimum average surface temperatures.
A thermo-elastic analysis is conducted to the
entire satellite model based on the output results
from the thermal analysis. The analysis is per-
formed twice for the worst cases, maximum and
minimum average surface temperatures. Dis-
placement boundary conditions are applied as
shown in Fig. 8.
Thermal strains are given by a. (T-T
ref
),
where a is the coefcient of thermal expansion, T
is the current element temperature, and T
ref
is the
reference (ambient) temperature. In thermo-elas-
tic analysis, the reference temperature is the zero
thermal stresses temperature which is the assem-
bly room temperature (=20C) in Small Sat case.
7 Thermo-elastic analysis results
Temperature distribution for the entire satellite is
determined as a result of the thermal analysis.
Fig. 8 F.E model used during on-orbit thermal deforma-
tion analysis
Fig. 9 Temperature distribution for the entire satellite
during maximum average surface temperatures
Fig. 10 Temperature distribution for the entire satellite
during minimum average surface temperatures
152 Int J Mech Mater Des (2006) 3:145159
123
Figures 9 and 10 show the temperature distribu-
tion (C) for the entire satellite during maximum
and minimum average surface temperatures.
Stress-strain state of the basis unit block
structure module is determined as a result of the
on-orbit thermal deformation (thermo-elastic)
analysis for both worst cases, maximum and
minimum average surface temperatures. The
diagrams of equivalent stress distribution
(10
3
MPa) and displacements (mm) of this struc-
tural module are shown in Figs. 1114. The stress
values are determined according to Von-Misses
criterion. Displacements are relative to the points
of star sensor attachment to its bracket, which are
the xation points for the entire model of Small
Sat structure during thermo-elastic analysis as
presented at Fig. 6.
The maximum stresses values r
e
for basis unit
block module at each on-orbit thermally worst
load cases and there equivalent yield margin of
safety (MS
y
) are given in Table 5.
8 Mounting accuracy due to on-orbit thermal
deformation
From static point of view, the basis unit block
module is safe because the yield margin of safety
is a positive value in both of worst design cases.
But this is not enough to determine that the
design is applicable and can withstand to on-orbit
thermal loads. Satisfactory performance of the
satellite requires accurate predictions of thermal
Fig. 11 Basis unit block stress distribution due to on-orbit
thermal deformation at maximum average surface tem-
peratures
Fig. 14 Basis unit block displacement due to on-orbit
thermal deformation at minimum average surface temper-
atures
Fig. 12 Basis unit block displacement due to on-orbit
thermal deformation at maximum average surface tem-
peratures
Fig. 13 Basis unit block stress distribution due to on-orbit
thermal deformation at minimum average surface temper-
atures
Int J Mech Mater Des (2006) 3:145159 153
123
deformations to verify pointing and alignment
requirements for Mounting Accuracy. Therefore,
it is important to calculate the angular positioning
deviations for all high precise equipments (pay-
load and ADCS devices) relative to the star
sensor due to on-orbit thermal deformation.
These values must not exceed the relative limiting
deviations specied for mounting the precise
equipments. Table 6 lists the limiting angular
positioning deviations for the most precise equip-
ments relative to the star sensor which are
derived from the structure requirements.
The angular positioning deviation between any
two equipments is calculated by measuring the
deviation angle between normal vectors for their
mounting plans. The deviation angle is deter-
mined by subtracting the measured angle after
on-orbit thermal deformation from the initial
angle before deformation. The criterion used to
calculate the angular positioning deviation
between the precise equipments and the star
sensor is explained below. Figure 15 shows
graphical sketch to explain the criterion used to
calculate the angular positioning deviation be-
tween any two equipments.
All precise equipments mounted on basis
unit block are xed through three or four
xation points. The mounting plan can be
dened by only two vectors connecting at least
three xation points. In case of equipment A,
V
*
1A
, and V
*
2A
identify its mounting plan before
on-orbit thermal deformation. While, V
*
1B
and
V
*
2B
identify the mounting plan of equipment B.
The normal vectors, V
*
and V
*
nB
for mounting
plan of equipment A and B respectively before
on-orbit thermal deformation, can be calculated
by applying vector cross product as following:
V
*
nA
V
*
1A
V
*
2A
V
*
nB
V
*
1B
V
*
2B
The angle between both of these normal vectors
is calculated by the next formula:
h cos
1
_
V
*
1A
V
*
2A
j V
*
1A
j j V
*
2A
j
_
After on-orbit thermal deformation, the equip-
ment mounting plan is usually deformed. And so
the normal vectors, V
*
nA
and V
*
nB
, are modied to
V
*
0
nA
and V
*
0
nB
respectively. They are calculated
for the deformed mounting plan as following:
V
*
0
nA
V
*
0
1A
V
*
0
2A
V
*
0
nB
V
*
0
1B
V
*
0
2B
Where: V
*
0
1A
& V
*
0
2A
and V
*
0
1B
& V
*
0
2B
identify
the mounting plan of equipment A and B
respectively after on-orbit thermal deformation.
These vectors are calculated from the displace-
ment deformation results of on-orbit thermal
deformation analysis in X, Y, and Z direction.
The modied angle between normal vectors for
the deformed mounting plan is calculated as
following:
h
0
cos
1
V
*
0
1A
V
*
0
2A
j V
*
0
1A
j j V
*
0
2A
j
_
_
_
_
Table 5 Maximum Von-Misses equivalent stresses and
yield margins of safety for basis unit block module at each
on-orbit thermally worst load cases
Design case r
e
(MPa) MS
y
Case 1 94.158 0.59
Case 2 125.937 0.19
s Case 1 represents the thermal deformation due to
on-orbit maximum temperatures.
s Case 2 represents the thermal deformation due to
on-orbit minimum temperatures.
Table 6 Limiting angular positioning deviations for the
most precise equipments relative to the star sensor
Equipment Limiting angular positioning
deviations, (arcmin)
MBEI (optical-
mechanical unit)
30
Angular velocity
meters (gyro)
60
Reaction wheels 60
Magnetometer 60
Magnetorquers 60
154 Int J Mech Mater Des (2006) 3:145159
123
The angular positioning deviation angle between
equipment A and B is calculated by the next
formula:
h
dev
h
0
h
For Small Sat, the angular positioning devia-
tion angle h
dev
(arc-min) is calculated for the
precise equipments mounted on basis unit block
relative to the star sensor. Table 7 lists the values
of angular positioning deviation angle due to on-
orbit thermal deformation relative to the star
sensor. It shows the results for both maximum
and minimum on-orbit temperatures. By compar-
ing these results with the limiting values listed at
Table 6, the performance of the satellite is not
affected with on-orbit thermal deformation under
maximum or minimum temperatures.
9 Fatigue damage due to on-orbit thermal cycling
One of the most essential aspects during perform-
ing on-orbit thermal deformation analysis is to
evaluate the fatigue damage due to on-orbit cyclic
thermal stresses. Ductile material (AMg6 alumi-
num alloy), used to manufacture the satellite
structure modules, does not rupture or buckle
froma single application of thermal stress, which is
clear from the analysis until now. But the material
can fail in fatigue from the many cycles of on-orbit
thermal deformation loading. The results data of
thermo-elastic analysis is used to evaluate the
thermal fatigue damage for the basis unit block.
By reviewing the results data of thermal
deformation analysis and the thermal stresses
distribution for the basis unit block module under
maximum and minimum temperatures (Figs. 11
14), the maximum thermal stress is occurred at
the same locations under both of the worst cases.
These locations are illustrated in Fig. 16. The
entire satellite structure is affected by a cyclic
thermal stresses along the operation life time of
the satellite. However, these points are the most
locations that are affected by fatigue damage due
to on-orbit thermal cycling.
Thermal fatigue damage must be calculated for
the critical points located at basis plate module
and basis walls module. For thermal fatigue
damage calculation, the maximum and minimum
cyclic thermal stresses for the critical points are
taken from the thermal deformation analysis
results. The time life cycles (N
th
) corresponding
to each given stress ratio is calculated (Incropera
and David 1990),
log N
f
_ _
a
1
a
2
log S
max
1 R
n
a
3
2
R S
min
=S
max
; a
1
20:68; a
2
9:84; a
3
0
S
max
= the highest algebraic value of stress in the
stress cycle. S
min
= the lowest algebraic value of
stress in the stress cycle.
The number of cycles (n
th
) corresponding to
the operation life time of the satellite is calculated
as following:
The satellite duration period (T
dr
), which is the
time interval for completing one orbital cycle
(revolution), is dened in the following way:
Equipment A Equipment B
Mounting plan after
deformation
Mounting plan
before deformation
1A
v
h
nA
v
h
2A
v
h
1A
'
v
h
2A
'
v
h
nA
'
v
h
1B
v
h
2B
v
h
nB
v
h
nB
'
v
h
1B
'
v
h
2B
'
v
h
Fig. 15 The criterion of
angular positioning
deviation calculation
between any two
equipments
Int J Mech Mater Des (2006) 3:145159 155
123
T
dr

2p

l
e
p a
3
2
_
1
e
e
l
e
a
2
1 e
2

2
_
2
5
2
sin
2
i
_ _
1 e
2
_ _3
2
1 e cosu x
2

1 e cosu x
3
1 e
2

1 3 sin
2
i sin
2
u
_ _
__
3
Where, a = the major semi axis, (= R
e
+ h); R
e
=
the mean earth radius, (= 6378.14 km); h = the
satellite altitude; l
e
= the gravitational constant of
the earth, (= 3.986005 10
5
km
3
/s
2
); e
e
= the
earth oblateness parameter, (= 2.6333 10
10
km
5
/
s
2
); i = inclination; e = eccentricity; u = the
argument of latitude; x = the argument of perigee
For Small Sat, the following parameter values
are taken during calculation, h = 668 km; i =
98.085; e = 0.001; u = 0; x = 0. By substituting
with these values into the previous equation, the
draconian period for Small Sat equals to:
T
dr
5881:9 s
The number of cycles (n) corresponding to the
operation life time is dened as:
n
th

T
opr
T
dr
Fig. 16 Location of the worst thermal case
Table 7 Angular positioning deviation angle for the precise equipments relative to the star sensor
Equipment Angle before deformation,
h (deg)
Modied angle after
deformation, h (deg)
Angular positioning deviations
angle, h
dev
(arcmin)
Case 1 Case 2 Case 1 Case 2
MBEI 136 135.9871 135.9955 0.777 0.269
AVM, gyro M
x
90 89.9982 90.0063 0.105 0.379
AVM, gyro M
y
46 46.0042 46.0117 0.252 0.7002
AVM, gyro M
z
44 43.9885 43.9908 0.691 0.5512
AVM, skewed gyro 44 44.0233 44.0238 1.3996 1.428
Reaction wheels M
x
90 90.0127 90.0068 0.7614 0.40734
Reaction wheels M
y
-1 134 134.0076 133.9993 0.4549 0.0395
Reaction wheels M
y
-2 134 134.0035 133.9959 0.2078 0.2446
Reaction wheels M
z
44 44.0221 44.0202 1.3242 1.2117
Magnetometer 136 135.9947 135.9979 0.3168 0.1239s
Magnetorquers 44 43.9828 43.9892 1.0341 0.6458
Note:
s Case 1 represents the thermal deformation due to on-orbit maximum temperatures.
s Case 2 represents the thermal deformation due to on-orbit minimum temperatures.
156 Int J Mech Mater Des (2006) 3:145159
123
Where: T
opr
is the operation life time of the
satellite, which equals to 5 years.
T
opr
5 365 24 60 60 157:86 10
6
sec
Hence,
n
th
26; 808 cycles
The thermal damage at the critical points is
calculated by the following relation and listed in
Table 8:
D
thermal

n
th
N
th
Thermal fatigue damage calculations show
that the satellite structure can withstand all
cyclic stresses due to on-orbit cyclic thermal
loading during its life span, because the total
cumulative fatigue damage does not exceed one
for all critical points located in the basis unit
block modules.
10 Conclusion
Satellite structure thermal analysis is done to
verify functional temperature prole, mounting
accuracy of equipments and validate acceptable
thermal fatigue damage. By referring to the
thermal results gained above applying IDEAS,
the structure will not experience high tempera-
tures during the mission. Inspection of the above
results appears that the maximum temperatures
occur at the onboard computer frame modules
(25.2C). As one can notice from Table 6, there is
a difference of 6C between two sides of the rst
onboard computer frame module. This tempera-
ture difference is due to the temperatures of the
heat shields. The heat shields have a direct impact
on the temperature of the frame modules as from
facing the heat shield facing the sun or deep
space. As it has been showed in Fig. 7, the
temperature gradient between the plates did not
exceed 5C.
Temperature drop in height of reinforcing ribs
was taken equal to 0.2C. For plates this temper-
ature increases in the direction of an mounting
plate, and for a case of a basis unitinwards of a
basic unit. In calculation of thermal strains of the
plates and basic unit case, a temperature drop in
height of reinforcing ribs was taken into account.
In calculation of cases of frame modules this drop
was not taken into account, since it effects slightly
on angular deviations of instruments requiring an
accurate installation. These temperature distribu-
tions are applied in a nite-element model as
boundary conditions and thermal analysis is
executed but using FE technique this time. This
step is done to account for the mesh difference
between nite-difference model and the nite-
element one.
During on-orbit operation, the satellite is
totally free without any xation points. To con-
duct a thermoelastic analysis, displacement
boundary conditions must be dened for the
related model. Therefore, the satellite model
must be constrained by rigid support points.
Selection of these points locations and their
xation manner should provide minimum effect
on satellite thermal deformation. For Small Sat,
the support points are selected at the plane of star
sensor mounting with its bracket. This plane is
selected because the mounting accuracy of all
precise equipment installed on the basis unit
Table 8 Thermal and overall fatigue damage calculation for the critical points
Critical points Thermo-elastic analysis Thermal fatigue analysis
r
max
r
min
N (cycles) n (cycles) Thermal Damage
MPa Ksi MPa Ksi
12 23.575 3.422 27.956 4.058 1.00E+12 26808 2.68E-08
13 29.530 4.286 28.697 4.165 1.00E+12 26808 2.68E-08
14 1.855 0.269 4.066 0.590 1.00E+16 26808 2.68E-12
24 94.158 13.666 125.937 18.278 5.00E+06 26808 5.36E-03
Int J Mech Mater Des (2006) 3:145159 157
123
block is measured relative to the star sensor
bracket. Displacement boundary conditions are
applied by xing one of the star sensor four points
of xation (on bracket) in the X direction, xing
the next point in the Y and one of the other points
in the Z direction.
The stress-strain state of the basis unit block
structure module is determined as a result of the
on-orbit thermoelastic analysis for both worst
cases, in which the maximum and minimum
average on-orbit surface temperatures take place.
Displacements are relative to the points of star
sensor attachment to its bracket. From static
point of view, the basis unit block module is safe
because the yield margin of safety is a positive
value in both design cases. But this is not
sufcient to decide whether the design is satisfac-
tory. Satisfactory performance of the satellite
requires accurate prediction of thermal deforma-
tions to verify pointing and alignment accuracy
requirements for sensors. The angular positioning
deviation angle h
dev
(arcmin) is calculated for the
precise equipments mounted on the basis unit
block relative to the star sensor due to on-orbit
thermal deformation. These results are compared
with the specied limiting angular positioning
deviations for the most precise equipment rela-
tive to the star sensor. Results show that the
performance of the satellite is not affected with
on-orbit thermal deformation under maximum or
minimum temperatures.
Fatigue damage due to on-orbit cyclic thermal
stresses is evaluated after performing on-orbit
thermal deformation analysis. The maximum
thermal stress occurs at the same location under
both worst cases. This location is specied as
Point 24 as shown in Fig. 15 is most severely
affected by fatigue damage due to on-orbit
thermal cyclic loading. The total fatigue damage
for any point located in the basis unit block
module is the dynamic fatigue damage due to
mechanical vibration acting on the satellite during
transportation and launch plus the thermal
fatigue damage due to on-orbit cyclic thermal
stresses. Therefore, to calculate the overall
fatigue damage for the basis unit block, dynamic
fatigue damage must be calculated for Point 30
based on the fatigue analysis of dynamic vibra-
tions. Moreover, thermal fatigue damage must be
calculated for the critical points located in the
basis plate and basis walls modules. Thermal
fatigue damage calculations show that the satel-
lite structure can withstand all cyclic stresses due
to on-orbit cyclic thermal loading during its life
span, because the total cumulative fatigue dam-
age does not exceed one for all critical points
located in the basis unit block modules. Fatigue
damage due to thermal loading is evaluated and
the satellite structure is safe for the functional
period of life time (5 years).
References
Periodicals
Grooms, H.R., DeBarro, C.F., Paydarfar, S.: What is an
optimal spacecraft structure? J Spacecraft Rockets.
29, 480483, No. 4, July-August (1992)
Tsai, J.R.: Overview of satellite thermal analytical model. J
Spacecraft Rockets 41, 120125 (2004)
Books
Bruhn, E.F.: Analysis and design of ight vehicle struc-
tures. Tri-State Offset Company, USA (1973)
Camack, W.G., Edwards, D.K.: Effect of surface thermal
radiation characteristics on the temperature control
problem in satellites. In: Clauss, F.J. (ed.) Surface
effect on spacecraft material. John Wiley & Sons, Inc.
New York (1960)
Crandall, S.H.: Random vibration. Technology Press,
Wiley, NY (1958)
Dnepr SLS users guide, Issue 2, Nov. 2001. http://
www.kosmotras.ru
Dornheim, M.A.: Planetary ight surge faces budget
realities. Aviation week and space technology 145,
4446, No. 24, 9 Dec. (1996)
Gilmore, D.: Satellite thermal control handbook. The
Aerospace Corp. Press, El Segundo, CA (1994)
Hatch, M.: Vibration simulation using MATLAB and
ANSYS. Chapman & Hall/CRC, Florida (2001)
Hibbeler Russell, C.: Structural Analysis, 3rd. PRENTICE
HALL, Inc (1995)
I-DEAS TMG Reference Manual, MAYA Heat Transfer
Technologies Ltd., January 2003
Incropera, F.P., David, P.: Introduction to Heat Transfer.
John Wiley & Sons, Inc. (1990)
Karam, R.D.: Satellite thermal control for system engi-
neer. McGraw-Hill, New York (1998)
Prediction of satellite structure life on-orbit under thermal
fatigue effect, National Authority of Remote Sensing
and Space Science, NARSS, Egypt (2006)
158 Int J Mech Mater Des (2006) 3:145159
123
Saran Thomas, P., Larson Wiley, J. (eds.): Spacecraft
structures and mechanisms - From concept to launch.
Microcosm Press and Kluwer Academic Publishers,
Torrance, CA (1995)
Schiff Daniel.: Dynamic analysis and failure modes of
simple structures. John Wiley & Sons, Inc. NY (1990)
Shegly, J.: Mechanical engineering design, 6th edn.
McGraw-Hill, Boston, MA (2004)
Tedesco, J., McDougal, G., Ross, C.: Structural dynam-
ics theory and applications. Addison Wesley Long-
mann, CA (1999)
Terster, W., NASA Considers Switch to Delta 2, Space
News, Vol. 8, No. 2, 13-19 Jan. 1997, pp., 1, 18
Wertz, J.R., Larson, W.J.: Space mission analysis and
design. Space Microcosm Press and Kluwer Academic
Publishers, Torrance, CA (1992)
Proceedings
Moftt B.A.: Predictive thermal analysis of the combat
sentinel satellite. Proceedings, 16th Annual AIAA/
USU Conference on small satellites, Logan, Utah
State University (2002)
Computer Software
ANSYS Professional, Software Package, Ver. 10, ANSYS,
Inc., Southpointe, 275 Technology Drive Canonsburg,
PA 15317
I-DEAS TMG reference manual, MAYA Heat Transfer
Technologies Ltd., January 2003.
Int J Mech Mater Des (2006) 3:145159 159
123