darling_nathan_2013_web.pdf
TRANSCRIPT
![Page 1: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/1.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 1/152
BOSTON UNIVERSITY
COLLEGE OF ENGINEERING
Thesis
DESIGN ND
TESTING o F A
NOVEL MODUL R
N NOS TELLITE STRUCTUR L
BUS
by
N TH N
THOM S
D RLING
B.A. , Saint Michael s College, Colchester , Vermont, 2004
Submitted in partial fulfillment of the
requirements for
the
degree of
aster of Science
2013
![Page 2: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/2.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 2/152
First
Reader
Second Reader
Third Reader
Approved by
Raymond Na gem
Ph
Professor of Mechanical Engineering
Theodore A F r i t z Y n
~ · -
Professor of Astronomy
![Page 3: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/3.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 3/152
©
opyright by
Nathan
Thomas arling
2 13
![Page 4: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/4.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 4/152
Es verdad; pues reprimamos
esta fiera condici6n
esta furia esta ambici6n
por si alguna vez soiiamos;
s{ haremos pues estamos
en mundo tan singular
que el vivir solo
es
sonar;
la experiencia
m
enseiia
que el hombre que vive sueiia
lo que es hasta despertar.
Pedro Calderon de la arca
![Page 5: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/5.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 5/152
cknowledgments
hanks to Professor Theodore A. Fritz, who offered me enough rope on the first
day
of the job, as he has done
with
great success for a great many students. Also to
the many
undergraduate and graduate
students
(past and present) who have worked
harder
than most anyone I know
to
juggle a huge course load nd long hours on
hardware development, design review posters, and whatever else it takes
to
get the
project out the
door on time. Finally, I would like to thank Krista, who doesn't like
satellites very much, for supporting me when I really needed it.
Figure ·1: Sipping from the fire hose.
![Page 6: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/6.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 6/152
DESIGN ND TESTING OF
NOVEL
MODUL R
N NOS TELLITE STRUCTUR L BUS
N TH N THOM S D RLING
ABSTRACT
Launch cost per unit mass is the most consistent and largest cost driver for any
modern
spaceflight mission,
with
quoted per-kilogram launch f s in
the
thousands
or
tens
of thousands of dollars. This is followed
by the
costs associated with long
development schedules, shifting mission requirements, radiation-hardened electron
ics
and
other similar systems engineering constraints common to satellite develop
ment
efforts. The Boston University
Student
satellite for Applications and Training
(BUSAT) responds directly to these issues by providing a modular satellite
bus
archi
tecture addressing
the
early hardware lifecycle from concept development
to launch
.
BUSAT s novel approach to
the
logistics of satellite structural integration supports
a novel plug-and-play software and physical-layer architecture. Unique pressure-lock
fastening techniques not only allow very fast integration but also display excellent
frequency response characteristics when subjected
to
stringent vibration tests. The
unique cubic geometry of BUSAT s modular structure allows mission
planners
better
control over cost and risk before launch, including the ability to predict features of
the structure s frequency response over a variety of payload configurations.
vi
![Page 7: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/7.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 7/152
Contents
Introduction
1.1 Small Satellites and Scientific Space . 1
1.2 B
oston
University and Scientific Space 3
1.2.1 Background:
The
University
Nanosat
Pro
gram
UNP)
4
2 Requirements
Development
5
2.1 Requirements Identifi
cation
and Flow-down . . . . . . . . . . . . . . 5
2.1.1
The
BUSAT
Primary
Science Mission: Ionosphere-
Magnetosphere
Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Space Plug-
and-Pla
y
Proto
col
Operationally
Responsive Space
ORS)
and the
Getting to
Space . . . . . . . . . . . .
2.3 Mission Overview
and Concept
of
Operations
2.4 High-level Mechanical Bus R
eq
uir
ements
2.5 Low-level Design R
eq
uirements
2.6 Solar
Panel
Deployment . .
2.7 Nanosat-5
Inherited
Design
Featur
es
3
Structural Design
3.1
Components
3.1.1
Hub
3.1.2
Cube
Stack
3.1.3 Endoskeleton
3.1.4 Exoskeleton .
vii
7
1
11
13
13
14
7
18
18
23
24
25
![Page 8: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/8.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 8/152
301 5 Deployables
302
~ s s e n a b l y
0 0
303
Mass Budget
Vibration
Testing
401
Background 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
401 1 Vibration Testing for Snaall Satellites
401.2 Mechanical Vibration Testing
5 Data
Analysis
50003
Observations
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
500 4 Sunanaary of
ata
Gathered and Conclusion
6
Analytical Model
601
Introduction 0 0 0 0 0 0 0 0 0 0 0
602 Danaped Mechanical Oscillators
603 Multiple Degrees of Freedona 0 0
6.4 Danaped Frequency Response for Multiple Degrees of Freedona
605 Spring Coefficients
and
Forcing
Input o t i o n ~
606 Three Degree of Freedona Model 0 0 0 0 0 0
6 7 Nine Degree of Freedona Model Coupled)
608 Twenty-seven Degree of Freedona Model Coupled)
609 Franaework of ~ n a l y s i s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
609
01 Building Mass, Spring and Danaping Matrices
6010
Initial Results
6011 Discussion 0 0
7
Results
and Conclusion
701
Design 0 0 0 0 0 0 0 0 0
Vlll
26
29
30
4
34
34
36
44
44
59
6
60
61
64
64
66
69
74
77
78
85
86
92
95
95
![Page 9: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/9.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 9/152
7 2 Testing . .
. . .
. . . . . . . . . . 95
7 3 Thned
27
Degree of Free
dom
Model and Comparison to Vibration T
est
Dat a .
.
.
.
96
A
MATLAB Script
for 1 DOF
Model
98
B
MATLAB Script
for
3 DOF Model
102
MATLAB Script
for
9 DOF Model
107
D
MATLAB Script
for 27 DOF
Model
115
References
130
E Vita
132
lX
![Page 10: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/10.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 10/152
List of igures
0·1
Sipping from the fire hose. . . .
1·1
The
Pumpkin CubeSatKit 1U picosatellite structure, commercially
available and used widely accross academia, industry and even the
v
military. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1·2
The
ANDESITE mission, a Boston University 3U CubeSat currently
approved for funding from the Air Force Office of Scientific Research.
Andesite will explore fine-scale filamentation in the Earth s magneto-
sphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2·1 Requirements originate from the Principal Investigator via high-level
mission definition, the University Nanosat Program via industry stan-
dards for safety
and
best practices. Additionally, BUSAT has identified
a niche opportunity in the pico- and nanosatellite indust ry from which
requirements are sourced. . . . . . . . . . . . . . . . . . . . . . . . . . 5
2·2 The Space Experiments Review Board (SERB) 2012 rankings include a
total CubeSat volume of over 40U in experimentation awaiting launch
opportunities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2·3 BUSAT seeks
to
optimize
the
pre-launch lifecycle,
with particular
em
phasis on tolerance
to
late, game-changing requirements shifts and the
lengthy integration process. . . . . . . . . . . . . . . . . . 9
2·4 The BUSAT Low-Earth orbit (LEO) concept of operations.
![Page 11: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/11.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 11/152
2·5 BUSAT's 1U solar panel design costs
about
150 in
parts and takes
less
than three
hours
to
build using refiow solder techniques but do
not
perform as well as more expensive cells.
PCB substrate
is shown at
left, finished component is shown at right. . . . . . . . . . . . . . . . 14
2·6 Nanosat-5 model
structural
spacers (not
to
scale). From left: exterior
spacers arranged in
top
panel configuration, interior spacer (original
model, unmodified), hexagonal spacer. 15
2·7 1-U, 2-U
and
3-U Cube Structures. . . 16
3·1 Early concept sketch of
the
BUSAT mechanical architecture. 17
3·2
The
BUSAT satellite shown before solar panels are deployed. 18
3·3 Exploded views of
the
BUSAT Hub (upper assembly) and Endoskele-
ton
showing
the
hub
PC\
104 printed circuit
board
stack, connectors,
and four routing boards. . . . . . . . . . . . . . . . . . . . . . . . . . 19
3·4 Detail from
the PC\1 4
embedded systems
standard
showing fastener
placement , header (right) and keep-out areas.
3·5 Detail view of
the
hub stack assembly showing
three
midplane stand-
offs Each standoff fastens
to the
external
hub
walls in a slotted groove
that
allows variations in
hub
stack height
due to
non-standard
PC\1 4
20
overhead height in
COTS
components. . . . . . . . . . . . . . . . . . 21
3·6 Detail view of
bottom cap
and
three hub
walls integrated.
Aper
t ures
for 25-pin and SMA connectors are shown, as well as a tongue-and
groove fastening technique
that
minimizes
the
need for fasteners along
the
entire length of the hub, and the labeled
PC\
104
bolt
pattern. . . 21
Xl
![Page 12: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/12.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 12/152
3· 7 Top left : schematic
top
view of
the
BUSAT wiring harness. Composed
of several standardized
short
cable harnesses, this wiring method alle
viates the need for costly and time-consuming one-off harness design
and
allows the hub to be removed from the satellite at any time during
integration. Top right: side vew of the harnessing scheme from the
payload (left)
to
endoskeleton to hub (right) . . . . 22
3·8 The
hub
. Inset: A 45° filet fixes
the hub
in place.
3
3·9 The Birtcher WedgeLok is
the
principal fastener used in BUSAT s eas-
ily integrable design. A torque applied to the tension bar results
in
a
lateral
pressure applied to the
cube
stack.
3·10 Twelve
WedgeLok™s are distributed around
the cubestack and pro
vide sufficient clamping force to restrain BUSAT s payload bay during
launch . . . .
3·11 The BUSAT exoskeleton.
3·12 The BUSAT-NMT Langmuir Probe shown
protruding
through the ex
oskeleton
top
panel. . . . . . . . . . . . . . . . . . . . . . . . .
3·13 The BUSAT satellite showing modular solar panels deployed
3·14 The TINI Aerospace Frangibolt (shown
in
red)
with
associat
ed
housing
3
24
5
5
6
and custom
titanium
flange for heat isolation. . . . . . . . . . . . . 27
3·15 Cross-sectional view of the cup-cone kinematic mounting structure. 27
3·16
Th
e BUSAT deployable
panel
stowed and completely
constrained
(at
left) by the TINI Aerospace 4 Frangibolt and
(at
right) by a titanium
groove (to avoid cold-welding under extreme vibration. . . 28
3·17 The tongue-and-groove kinematic mount after deployment. 28
Xll
![Page 13: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/13.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 13/152
3·18
The
BUSAT assembly procedure (top left
to
bottom right) begin
ning with
the
exoskeleton bottom plate, endoskeleteon, cube stack,
WedgeLok™s and hub. . . . . . . . . . . . . . . . 29
3·19 BUSAT s mass breakdown arranged by subsystem. 30
4·1 Nanosat-5 integration and test flow (from NS-7 User s Guide) 34
4·2 Mechanical Test Flow (from NS-7 User s Guide). . . . . . . . . 35
4·3 Concept schematic of Thermotron vibration
table
control loop. (5) . 36
4·4 Schematic view of the structural test as performed at the Draper ETF
showing control accelerometer
mounted
on the
shake table
armature.
37
4·5 Conceptual representation of SRS plot generation
and
vibration table
control showing (in red) control accelerometer input
to
a virtual mass
spring-oscillator array implemented within the control software. The
model simulates a system response based on
the
control accelerometer
input , and this signal is plotted (red). Control software commands
based on
this
simulation are
then
passed to a power amplifier which
sends current to the
table
s inductive coils, imparting an acceleration
to
the table and again causing the control accelerometer to feed back to
the
software oscillator model. The number of oscillators in the software
model (n) determines resolution (5) . . . . . . . . . . . . . . . . . . . 38
4·6 Typical plot of sine sweep acceleration vs. frequency. This is an in-axis
response. . . . . . . . . . . . . . . . . . . . ·. . . . . . . . . . . . . . . 39
4· 7 Typical plot of sine burst acceleration as a function of
time
. This is an
in-axis response. . . . . . . . . . . . . . . . . . . . . . . . . .
4
4·8 Random vibration response raw
data
from 20Hz
to
2000Hz 42
xiii
![Page 14: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/14.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 14/152
4·9 Typical
time
history of shock test.
This
is an in-axis response. Ramp
up and ramp down motion can be seen from 0.00 ms
to
0.04 ms
and
from 0.06 ms
and
0.09 ms. . . . . . . . . . . . . . . . . . . . . . . 43
5·1
C-channel assembly. . . . . .
46
5·2 First Diagnostic Sine Sweep. Run 1: Z-Axis. 48
5·3 Channel 4 only. Diagnostic Sine Sweep Runs 1, 3 5 and 9.: Z-Axis. 49
5·4 Diagnostic Sine Sweep, Run
14:
X-Axis. All Channels.
5·5 Diagnostic Sine Sweep,
Run 19:
Y-Axis. All Channels.
5·6 Sine Burs t,
Run
02:
Z-Axis. All Channels.
5·7 Sine Burst , Run 11: X-Axis. All Channels.
5·8 While still within specifications, the BUSAT structure does display
some shift in lowest
natural
frequency. Generally, the peak shifted
lower by
around
10 to 20 z after each subsequent test (frequency
51
51
53
53
and
magnitude are labeled for each of the four response curves). 56
5·9 Elastic modulus as a function of
temperature.
. . . . 56
5·10 Evidence of wear on
the
anodized surfaces of a cube. 59
6·1 The model developed in this section seeks to determine the BUSAT
satellite's frequency response based on the characteristics of
the
pay-
load bay. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
6·2 A basic mass, spring,
and
damper system where the motion of mass
min
response
to
an excitation .;(t) is described by the
state
variable x
measured from
the
origin in
the
upper
left
hand
corner of
the
diagram
. 61
6·3 A single degree of freedom mass, spring and
damper
system with two
points of coupling to the oscillating platform. . . . . . . . . . . . . . . 63
6·4 Single degree of freedom mass-spring-damper system displacement re-
sponse plotted over frequency.
68
xiv
![Page 15: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/15.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 15/152
6·5 Single degree of freedom mass-spring-damper system acceleratation fre-
quency response
plotted
over frequency. . . . . . . . . . . . . . . . . . 69
6·6 A three degree of freedom mass, spring and
damper
system
with
four
points of coupling to the oscillating platform. . . . . . . . . . . . . . . 70
6·7 Generalized free-body diagram for the three degree of freedom system. 71
6·8 Three degree of freedom forced mass-spring-damper acceleration re
sponse
plotted
over frequency.
Undamped
natural frequencies
are
marked by vertical lines. .
73
6·9 Schematic representat ion of a general planar cube array element sur-
rounded
by
four masses, four springs
and
four dampers. . . . . . . . . 74
6·10 Masses 1-3 of a nine degree of freedom forced mass-spring-damper fre
quency response plotted over frequency. Undamped natural frequencies
for each mass are marked by vertical lines. . . . . . . . . . . . . . . .
75
6·11 Masses 4-6 of a nine degree of freedom forced mass-spring-damper fre
quency response plotted over frequency. Undamped natural frequencies
for each mass are marked
by
vertical lines. . . . . . . . . . . . . . . . 76
6·12 Masses 7 9 of a nine degree of freedom forced mass-spring-damper fre
quency response
plotted
over frequency. Undamped natural frequencies
for each mass are marked by vertical lines. . . . . . . . . . . . . . . . 77
6·13 The BUSAT
standard
satellite-fixed reference frame. The unit vectors
i
3
and are assigned
to
the
x
y and
z
axes, respectively. . . . . . . 79
6·14 A one-to-one transformation between payload cube position and state
variable subscript, as shown in Table 6.2. . . . . . . . . 80
6·15 Two-dimensional representation of the satellite cell array. 82
6·16 SolidWorks FE displacement analysis of Top Plate. . . 87
X
![Page 16: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/16.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 16/152
6·17
Acceleration response for masses XI x
2
and x
3
.
Subscripted mass
notation can be converted using Table 6.2
• • 0 • •
88
6·18
Acceleration response for masses x
4
x
5
and x
6
.
Subscripted mass
notation can be converted using Table 6.2 • • • •
88
6·19
Acceleration response for masses x
7
x
8
and x
9
.
Subscripted mass
notation can be converted using Tab le 6.2 • • • • •
89
6·20
Acceleration response for masses
x
10
x
11
and xi
2
.
Subscripted mass
notation
can
be converted using Table 6.2
89
6·21 Acceleration response for masses
XI
3
XI
4
and
5
. Subscripted mass
notation can be converted using Table 6.2
90
6·22 Acceleration response for masses
6
XI
7
and XIs· Subscripted mass
notation
can
be converted using Tab le 6.2
90
6·23 Acceleration response for masses x
19
x
20
and x
2
·
Subscripted mass
notation can
be
converted using Table 6.2
91
6·24 Acceleration response for masses x
22
x
3
and x
24
. Subscripted mass
notation can
be
converted using Table 6.2 91
6·25 Acceleration response for masses x
5
x
26
and x
27
. Subscripted mass
notation can be converted using Table 6.2
92
6·26 Acceleration response for mass XI3 located in the center of the top
layer of the satellite cube stack. . . . . . . . . . . . . . . . . . . . . . 93
6·27 Acceleration response for mass x
1
9
located
at
a corner of the top layer
of the satellite cube stack.
6·28
xvi
93
94
![Page 17: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/17.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 17/152
List of bbreviations
AIAA . .
AFOSR
.
AFRL/RV
BU NSP . .
BUARC .
COG .
COTS .
EPS .
ETF
. . .
FCR .
IES .
MGSE .
PCB
.
PSD
.
RMS .
RPI
.
SIF . .
SIP . . .
.
SRS . . . . . .
TID . . .
UNP
. . .
VLF .
American Institute of Aeronautics and Astronautics
Air Force Office of Scientific Research
Air Force Research Laboratory s Space Vehicles
Directorate
Boston University Near Space
Program
Boston University Amateur Radio Club
Center of Gravity
Commercial Off-The-Shelf
Electronic Power Supply
Environmental Test Facility
Final Competition Review
Imaging Electron Spectrometer
Mechanical Ground Support Equipment
Printed Circuit Board
Power Spectral Density
Root Mean Square
Remote Polar Imager
Scientific Instruments Facility
Satellite Interface Plane
Shock Response Spectrum
Total Ionizing Dose
University Nanosat Program
Very Low Frequency
xvn
![Page 18: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/18.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 18/152
hapter
Introduction
1 1 Small
Satellites and
Scientific
Space
Since 1957
and
the launch of the Russian
Sputnik
1, the first Earth-orbiting artificial
satellite, the concept of spaceflight has
taken
a superlative position in the politics,
technology, education
and
cultural psyche of
the
United States of America. Despite
budget
cuts
to
spaceflight programs within NASA, the Air Force and elsewhere, names
like Yuri Gagarin, Neil Armstrong and now Elon Musk resonate even with people who
are unfamilar with engineering and space physics.
Scientific exploration was integral
to
the first successful US attempts
to
fly a satel
lite-
James Van Allen s inclusion of a Geiger counter on
the
Explorer I mission (1958)
enabled discovery of the Earth s radiation belts
that
now
bear
his name. As would be
expected from early, ground-breaking missions developed under the political pressure
of the early Cold War, these early
US
satellite missions were
o
ne-off engineering
concepts.
More recent trends in spaceflight hardware (and especially in the funding avail
able for space missions) indicate an industry shift towards cheaper spaceflight plat
forms. Perhaps
the
single biggest game-changer in
the
small satellite
industry
has
been
the CubeSat, a containerized satellite concept originated
at
California Polytech
nic
State
University. The core concepts
that
have enabled the CubeSat
standard to
tally roughly 75 successful launches since 2003 are miniaturization, containerization,
interface standardization, and open-source development.
![Page 19: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/19.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 19/152
2
igure
· :
The Pumpkin CubeSatKit
1U picosatellite
structure,
com
mercially available and used widely accross academia, industry and even
the military.
Weighing
under
2 kilograms at a volume of 1 liter, the basic single-unit 1U)
CubeSat contains all modifications to the satellite structure within an external stan-
dard,
allowing any
CubeSat
rideshare on a launch vehicle
via
the
standardized
Pol
y-Picosatellite Orbital Deployer (P-POD). This gives
greater
freedom to the
pay
load developer by relieving the paperwork, cost, risk and
time spent
interfacing
to
the
l
aunch
vehicle.
igure ·2: The ANDESITE
mission, a Boston University 3U
CubeSat
currently
approved for funding from
the
Air Force Office of Scientific
Research. Andesite will explore fine-scale filamentation
in
the Earth s
magnetosphere.
![Page 20: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/20.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 20/152
3
1 2
Boston University and
Scientific Space
The Boston University Student satellite for Applications and Training (BUSAT)
program is a grassroots spaceflight hardware development program entirely staffed
and
directed by students of engineering, physics, astronomy, and computer science.
BUSAT was initially formed in 2007 in response to the University Nanosat Pro-
gram (UNP) Nanosat-5 competition for the purpose of delivering a completed space
we
ather
nanosatellite. Since
that
time, the BUSAT program has grown to include
the
Boston University Near Space
Program
(BU NSP, a scientific high-altitude bal
looning program for
the
Boston University community
and the
local MATCH
charter
public high school) , the Boston University Amateur Radio Club (BUARC) ,
and
ul
timately
a new space weather satellite design for the
UNP
Nanosat-7 competition.
While professional and academic mentorship is essential to the program, BUSAT is
entirely staffed and directed by students of Boston University and Boston Univer
sity Academy. Cooperat ive efforts by
students
at Georgia
Institute
of Technology
and New Mexico Institute of Mining and Technology also contribute considerably
to
the
BUSAT effort . BUSAT s programmatic goals include
the
laying
the
groundwork
for future spaceflight hardware and general sensor development programs at Boston
University.
BUSAT s current scientific space weather satellite effort seeks
to
characterize mag
netosphere ionosphere interactions from low-Earth orbit by simultaneously measuring
energetic electron flux and subsequent light emissions in the polar
latitude
s. Comple
mentary information about the local magnetic field, plasma environment, and internal
system
health
will also be collected
and
telemetered to ground, along with structural
health monitoring and total radiation dose
data
![Page 21: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/21.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 21/152
4
1.2.1
Background:
The University Nanosat Program UNP)
The
University Nanosat
Program
is a joint program between
the
Air Force Research
Laboratory s Space Vehicles Directorate AFRL/RV), the Air Force Office of Scien
tific Research (AFOSR)
and
the American
Institute
of Aeronautics and Astronautics
(AIAA).
The
objectives of the program are to educate and train the future workforce
through
a national
student
satellite design and fabrication competition
and
to enable
small satellite research and development, payload development , integration, and flight
testing.
The
N anosat
Program
has two distinct stages.
The
first stage is a design
and
proto-flight build phase, which lasts approximately two years
and
culminates
in the
AIAA Student Satellite Flight Competition Review (FCR). During this first phase the
universities are partially funded by the AFOSR and construct a proto-flight satellite
while participating in various design reviews and program-sponsored hands-on activ
ities and workshops over a two-year period. Teams are evaluated based on criteria
including program
maturity
, hardware
maturity
and flight readiness. FCR judges are
a distinguished panel of government and industry professionals.
The UNP is currently in its seventh competition cycle. Boston University
and
BUSAT participated in Nanosat-5; currently the BUSAT program is a
participant
in
Nanosat-7.
![Page 22: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/22.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 22/152
5
Chapter 2
Requirements Development
2.1
Requirements
Identification and
Flow down
In
a
project
as complex
and
long-lived
as
the
development of a satellite
it
is essen
tial to ground truth development at regular intervals to keep the
project
progressing
efficiently towards the originally stated goal. The process of requirements identifica
tion
is complicated
by
the fact
that there are at
least
three
customers:
the BUSAT
Principal
Investigator,
the
satellite
industry
represented
by
the
Air Force),
and the
UNP
whose
dual
mission is
education
and technology development)
Figure
2· : Requirements
originate from
the Principal Investigator
via
high-level mission definition,
the
University
Nanosat Program via
industry standards
for safety
and best
practices. Additionally,
BUSAT
has identified a niche opportunity in
the
pico- and nanosatellite industry
from which
requirements are
sourced.
![Page 23: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/23.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 23/152
6
2.1.1 The
BUSAT Primary
Science
Mission: Ionosphere-Magnetosphere
Coupling
BUSAT s primary scientific instrumentation includes the Remote Polar Imager (RPI),
a pushbroom-style geographically resolved spectrometer sensitive to visible light at
391.4 nm, 427.8
nm
, 557.7 nm,
and
630.0 nm. These wavelengths correspond
to
auro
ral
phenomena produced by energetic particle interactions with oxygen
and nitrogen
in
the ionosphere. Imaging is accomplished using several off-the-shelf components
mounted to an optical bench, including the Lippert Cool Space Runner single-board
computer , an ATIK 314L+ astronomical camera, and a sophisticated arrangement of
customized commercially available mirrors, lenses and a diffraction
grating
. Together
these components provide the ability
to
bend
the
required focal length into a 3U form
factor , the sensitivity required for low-light measurements,
and
the processing power
needed
to
compress, buffer
and translate data to
BUSATs native SPA-1 I
2
C
data
protocol.
Working in tandem with the RPI is the Compact Half-Unit IES for Cubesat
Operations
(CHICO) . Complementing
the RPI s
data
, CHICO allows simultaneous
in-situ measurement of energetic electron flux.
CHICO
is based
on
a line of legacy
hardware
flown on
the
Cluster (launched 2000) and
Polar
(launched 1996) missions, as
well as the Fixed Sensor Head (FSH) , currently awaiting launch on
the Demonstration
and Science Experiments (DSX) mission (Gunda, 2008). Filling a total CubeSat
volume of less that 0.5U, CHICO represents a powerful new adaptation of this line of
legacy hardware, capable of detecting particles
with
energies between 30keV
and
500
keV and resolving pitch angle to within 10 degrees in a plug and play s nsor package
Originally developed for Boston University s
Twin
Imaging of the Moving Ele
ctron
trapping boundary (TIME) mission, CHICOs uti lity is demonstrated by its ability to
be
integrat ed into
the
BUSAT
bus
with a minimum of redesign.
![Page 24: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/24.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 24/152
7
The RPI s requirement for pointing and stability is the primary scientific con
straint on the mechanical satellite bus design. These requirements are satisfied by
the
BUSAT bus three-axis torque coil control and
an
on-orbit
attitude
control scheme
that matches
the
satellites rotational period to its orbital period (described as a
Thomson Spinner ). In this way, the auroral imager points towards the earth and
CHICO points towards zenith
at
all times.
It
is interesting
to
note that future spec
trographic imaging missions with higher resolution
and
much finer pointing would
be
possible with minimal recurring engineering cost if several reaction wheels were
included in Mission Group A A further requirement levied on the current BUSAT
mission by the space weather instrumentation is the need for a low-Earth orbit in
sertion. Additionally,
both
RPI
and
CHICO require direct access
to
open space for
measurement.
2.2 Space Plug-and-Play
Protocol,
Operationally Responsive
Space ORS) and the Getting
to
Space
The
Space Test
Program (STP)
has been
the
primary
provider of
Department
of
Defense (DoD) integration and launch services for almost 50 years since it was char
tered
in 1965,
and
had over 200 missions
to
its name as of 2010 (Galliand). Serving
the entire DoD, the STP coordinates the Space Experiments Review Board (SERB),
which meets on
an
annual basis with the Air Force, AFRL, the Navy,
the
Army,
DARPA,
the
NRO and
the
MDA. The panel hears a series of 15-minute briefs from
spaceflight mission representatives and a ranking is then assigned based on percieved
military relevance and DoD interest in providing access
to
space via sounding rocket,
the
International Space Station or a launch vehicle to orbit. While this process
may
seem challenging at best, current plans for the U.S. fiscal year 2013 budget have in
dicated that both
STP
and the Operationally Responsive Space (ORS)
program
may
![Page 25: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/25.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 25/152
8
be closed altogether.
Against this background, small scientific spaceflight systems that actively seek to
op timize t he massive cost of
sate
llite development and
laun
ch are abso
lutel
y essen
tial
for t he U.S. to
maintain its current
level of industrial, educational, scientific and
defense activity
beyond
t he lower atmosphere.
Figur
e (2·2) shows a portion of the
2012 SERB list, showing rankings for programs currently seeking launch opportuni-
t ies. Twelve CubeSat missions are included on the
ranked
list,
with
a total volume
of 42.5U.
With
appropriate
funding for development, BUS T could el
iminate
more
than 20
of
the CubeSat missions on the SERB list in just two deployments, with
room to spare. This co
uld
very eas
il
y be accomplished wit h
ju
st one launch.
:
-
I I
-
.
-
-
.
- I - I ·-··
-
-
~ ;
-
--
_
.
· .. dl.AN Product
o..
- _::_
-
-
-
-·
· n v : r ~ 1 i ;
-·
-
- -·
-
-
- ·/
NPS Solar (@II Arrav
T@ster
-
-
•'
-
·-
Operationally Responsive
1\
.
;
-
-
~
Space Enabler Satellite
- -
..
Peregrine
'
- ·-
ParkingsonSat Remote Data
-
7
.
Relav
----
Pica Satellite Solar Cell
Testbed
-
-·
v
-
-
....
,
Rapidprototyped
..
.
\
Mlcroelectromechanlcalsyste
·-
' ~ , ;
7
·--
-
\
m Propulsion and
Rad
iation
:::: ::=:::
· -
..
Test Cubeflow Satellite i
-
-
-
-
..
SMOC n o s t e l l i t e ~
-
--
_ ;:--...
--
-·
.
l>ro«ram
/f-:: :
-
·-
·
- I
- I ·
-
·-
-·
-
-
-
--
-
-·
. .
.
.
.
-
igure
2·2: Th
e Space Exp
er
iments Review Board (SERB) 2012 rank-
ings include a
total
CubeSat volume of over 40U in experimentation
awaiting launch opportunities.
The cost of launch is not t he only
problem in
need of engineering in the cur
rently challenging fiscal environment.
Many
satelli
te
development programs follow
![Page 26: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/26.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 26/152
9
a one-of-a-kind design
path
in which a unique satellite is developed specifically to
support the needs of a
distinct
set
of instrumentation. This method
is both expen
sive and time-consuming [4] . Because of the complexity and variability of
satellite
missions, a universal satellite has yet to develop fully,
although
at least one modular
demonstration
mission has received
substantial
support [2] BUSATs novel
satellite
nanosatellite
bus
provides
both
the general utility of a multi-mission structure while
at the same
time
retaining mission-specific adaptability to a wide range of in
strument
requirements.
igure 2·3: BUSAT seeks to optimize
the
pre-launch lifecycle, with
particular emphasis on tolerance to late, game-changing requirements
shifts and the lengthy integration process.
While BUSAT strives
to
optimize
the
design cycle
by
reducing
recurring
engineer
ing costs, the superlative and elegant universal satellite
bus
is
not
possible, especially
on
a university budget.
This
is supported
by
the fact
that programs such
as Space
Plug and Play
SPA) , developed
at
AFRL and in the
ORS
program
, have
met with
limited success operating under much more favorable budgetary constraints. How
ever, while the concept of universality is unwieldy in its most general form, it can be
effectively applied to a slightly narrowed subset of spacecraft missions,
namely
those
requ
iring only basic
system
support functions.
The broad
satellite
bus
functionalities
![Page 27: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/27.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 27/152
10
required by
this
subset include
rough
pointing, low
data
rate
telemetry an
d
tolerance
of
duty
cycling,
and
are achievable on a university budget.
In other
words,
the
elegant
solution s practical if properly constrained.
BUSATs
target
role in the spaceflight
industry
is
that
of a mission broker for small
scientific
instrumentation.
By adhering
to standardized
interfaces
in
a
star
network,
taking
advantage of symmetries
that
exist
in
software, hardware, mission design and
even documentation, and
by
arranging
the
misssion around a centralized bus,
BUSAT
facilitates mission sharing
and
drives down cost for payload developers
and
overall
risk for launch providers.
2 3 Mission Overview and
Concept
of Operations
The
dual
design
intent that
drives BUSAT s requi rements addresses both the geo
physical environment of the
upper
atmosphere
as much as
it
addresses
the
economic
and
technical environments
in
which a small sate llite develops. Initially,
BUSAT
was
planned
as a scientific
platform
for space weather research. As
the
design
ma-
tured budget and
personnel
constraints stimulated
the
development of a parallel
engineering mission
in
order to optimize the pre-launch lifecycle. The
two
missions
are
co-dependent and fully
supported by
the satellite s mechanical
architecture.
![Page 28: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/28.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 28/152
11
OIII NTIITION
Separation from ESPA rln1vi•
PSC U1htbond Adapter.
Detumble (3 weeks max.)
Fine
Polnlinc(lwoek
.)
lleain hea lth beacon
lleain Hub Initio- on Mode
- GIOCI ID OfiERAJIOfVS
S IIIINtll around stlltlon
at BU
Cno rles Rver campus
func liOM
oullonomously to own I
nk
dato .
f ll 0 O L I ~
sate lli
te or
ien
ted for max
i
mum drae:
durlna:de-orb ft phase .
Al l systems poM
red
d
own
.
Pav load health
assessed
poss ib
le
new sc iencemisslonsstarted.
Amateur repeater fun ction enabled.
Soler panels depl
oy
Bec ln
pa
yl
oad
boy
lnitiolizotlon
1
week
)
I
·0
NOIIMAI.
Of fiiATIONS
SCifflCE MODE)
Ill
.
CH
I
CO
.
Ma1nt tommr.
SHM takl
nedateo
v r
po l
es
{
La
titude > 50 ' .
Opti
on
to
duty cy
cl
e
Instrumentation at anv
location in orb it
OUTY CYCLING
Low-power
or
sleep mode
dependi
n a
on stlllte of
charp. SHM and
Ma10etometer tokin1
data
Figure
2·4: The
BUSAT
Low-Earth orbit
(LEO) concept
of opera
tions.
2.4
High level Mechanical Bus Requirements
Modularity
, scalability,
limited
resources
and
a robust science mission
ar
e
the
four
mam themes
to
BUSAT s
mission. Science
requirements
have
been
introduced
and
have
little
direct effect
on the
structure of the satellite
(other than
what was
mentioned
in Section 2.1.1).
The
following list of high-level
requirments
pertaining to the satellite
mechanical
bus
fall
under
the remaining three headings.
![Page 29: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/29.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 29/152
12
Modularity
1. Any 1 U BUSAT payload shall be capable of integrating with the spacecraft in
at
least two different angular orientations.
n. Any 1U BUSAT payload shall have at least one to three open look-directions into
space.
m. Any BUSAT payload shall interface to
the
mechanical bus via a single
standard-
ized connector and harness.
iv. Integration shall take less
than
6 hours nominal) for any given payload bay
configuration.
v. Integration shall be repeatable i.e., several build up/ take-down sequences
must
be possible without new machining, fabrication, or undue effort.
Scalability
i
The
final flight configuration shall
be
scalable from a single central module
to
a
larger, more capable aggregate spacecraft.
ii. The avionics module shall be capable of removal and use in
other
compatible
satellites.
iii. BUSAT compatible payloads shall be capable of flying in multiple configurations
on multiple missions.
Low
Cost
i Hardware should be radiation tolerant wherever possible, but BUSAT shall pro
vide enough shielding to
protect
off-the-shelf
radiation
sensitive electronics.
![Page 30: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/30.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 30/152
13
n. BUSAT shall be compatible with Commercial Off-The-Shelf (COTS) components
commonly available to the satellite developer, thereby eliminating the cost of in
house development.
iii. Payloads shall require no more than 30 seconds between plug and play (i.e.,
physical layer integration completed and functionl testing commenced
within
that
time).
1v
Payload structural raw materials shall
be
sourced from commonly available and
affordable stock.
v. Machining tolerances specified in engineering drawings shall be 0.005 inches or
greater wherever possible.
v1
Locking helicoils shall only be used in final integration phases. Alternative meth-
ods shall be used during early build/ rebuild sessions.
2.5 Low level
Design
Requirements
The
process of requirements identification is iterative and new design specifications
generate new dependencies
and
new design requirements arise.
The
following design
constraints are more evolved than
the
requirements listed in Section (2.3) and begin to
suggest
the
actual design specifications and choices described later
in
this document.
2.6 Solar Panel Deployment
While a single side panel covered with conventional satellite solar cells (such as the
Spectrolab Ultra Triple Junction UTJ) Solar Cell) would provide power for the en
tire satellite, a cheaper modular solar panel based on Triangular Advanced Solar
Cells (TASCs, also from Spectrolab) was chosen instead.
The
TASCs are custom
manufactured at Boston University using a solder refl ow oven
and
are therefore very
![Page 31: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/31.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 31/152
4
cheap
to
manufacture, install
and
replace. Because they do not produce sufficient
power for a comfortable margin of error (one side of BUSAT s exoskeleton can pro
duce
approximately
20 W
with
TASCs
- which leaves a power
budget margin
of only
8 W without considering
the
power
draw
from
the payload
bay.
The
power budget,
therefore, drives a requirement for BUSAT to deploy
additional
solar
panels
once in
orbit.
Figure
·5:
BUSAT's 1U solar
panel
design costs
about
150
in
parts
and takes less
than three
hours
to build
using refiow solder techniques
but do not perform as well as more expensive cells.
PCB substrate
is
shown at left, finished component is shown at right.
2.
7 N
anosat 5
Inherited
Design Features
Even given
the
geometric
and
conceptual simplicity of BUSAT's cube stack, it has
proven very difficult for both the Nanosat-5 and Nanosat-7 teams to provide an
elegant solution to more advanced design issues
such
as cabling, integration
and
access to key electrical components
after
assembly. As more
information
about the
original development of the Nanosat-5 structure
became
available, it was clear that
requirements for easy assembly of the structure were compromised late in the design
![Page 32: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/32.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 32/152
15
process because of a break down in modularity. The spacers became specific
to
the
geometry of adjacent cubes, removing
the
ability to reorient or reposition payloads
as
originally intended. As
the
design evolved incrementally,
with
new modifications
to both machine drawings and fabricated parts , each spacer was
edited
in such a way
that many distinct modifications to the original part were necessary.
Due to these
non standard spacers, as well as the sheer number of unique
parts
and attachments,
the
assembly process
became
very difficult.
igure
2·6: Nanosat-5 model
structural
spacers not to scale). From
left: exterior spacers arranged in top panel configuration, interior spacer
original model, unmodified), hexagonal spacer.
Interior, exterior and· hexagonal standoff) spacers
ultimately made
up 15 dis
tinctly modified manufactured parts. Since
the
N anosat -5 part naming scheme lacked
enumeration and because there was no intuitive way to quickly assemble components,
build up
was
belabored
and slow.
Each
assembly step was not ·necessarily
intuitive
and
would
not
have been possible to rebuild without a working CAD model. The
CAD
model was
not without
its own challenges, however.
Due
to
revision control
issues
during
Nanosat-5, even the SolidWorks structural model lacked coherence and
did not reflect
the
fabricated
structure
, leaving ample room for interpretation as
the
NS-7 team assembled the structure.
![Page 33: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/33.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 33/152
6
igure ·7:
1-U, 2-U and 3-U Cube Structures.
Once a UNP satellite passes the Final Competition Review (wins the competi
tion), the satellite and the team are supported for a further two years of integration
and testing in preparation for launch. Integration,
the
process of forming a single co
hesive unit (electrically, mechanically
and
in software) from
the
satellite s substems, is
undertaken by professional technicians at AFRL s in-house facility. One ofthe largest
challenges to
the
Nanosat-5 structure would have been
in the integration
phase. Even
with
a very good assembly manual to explain the complex assembly procedure, tech
nicians working
to integrate
the
satellite
with
a
launch
vehicle would
be
challenged
by cabling and by the inevitable need to assemble and disassemble the structure re
peatedly
for troubleshooting.
This
is especially important for centralized subsystems
like the data processing unit or the communications system. Since the Nanosat-5
structure
provides no re-usable means of restraining cabling, some semi-p
ermanent
solution such as epoxy staking would have to be adopted, effectively locking sub
systems
into
a complicated mechanical puzzle. Though the intent and
end
goal of
the design was novel and impressive, the actual prototype did
not
satisfy original
requirements for ease of use and modularity.
![Page 34: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/34.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 34/152
7
hapter 3
Structural Design
Figure
3· :
Early concept sketch of
the
BUSAT mechanical architec
ture.
With
a total payload bay of 27U, BUSAT reserves
the
core 3U volume for its cen
tral avionics unit, leaving a total of 24U open for payloads seeking a mission-of
opportunity launch into space. One of the best advertising points for future BUSAT
missions is the centralized electrical
bus
architecture. While conventional 1U, 2U
and 3U satellites must necessarily use substantial portions of their meager mass and
payload allotments on support subsystems like radios and batteries, BUSAT payloads
optimize the ''rideshare launch opportunity
and
can use entire mission resource bud-
gets (mass, volume, power, time, and fiscal) on the payload and the mission itself.
![Page 35: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/35.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 35/152
8
Figure
3· :
The BUSAT satellite shown before solar
pan
els are de
ployed.
The BUSAT bus is organized into five specific groups of hardware: Hub ,
En
doskeleton, Exoskeleton, Deployables
and
Cubestack. An explanation of each is nec
essary to fully understand BUSATs
intended
function and mechanical design.
3 1 Components
3 1 1 Hub
The Hub Figure 3·3) is literally and figuratively the core of the satellite
bu
s, hold-
ing all of
the
subsystems essential for supporting a satellite mission. This includes
command and data handling, communications, power generation and
regulation
at -
titude determination
and
control,
and
a single magnetometer implemented
on
fifteen
standard printed circuit boards using a
standard
set of stack-through connectors.
The Hub
is completely enclosed in 6061-T6 aluminium
to
shield
the
satellite
bus
subsystems.
![Page 36: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/36.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 36/152
19
Figure 3·3:
Exploded
views of the BUSAT Hub
(upper
assembly)
and Endoskeleton showing the hub PC\104 printed circuit board stack,
connectors, and four
routing
boards.
The PC\104
embedded
systems standard defines printed circuit board dimensions
as
seen in Figure 3·4). Following
suit with
much of
the industry
that
has
developed
surrounding
the
CubeSat
standard
such as hardware manufacturers AstroDev
and
Clyde Space) , BUSAT
s hub
includes
5 printed
circuit
boards that
conform
to the
PC\ 104 standard. Also in line with common practice in
the
embedded systems in
dustry, the BUSAT design rejects some features of the PC\104 standard, including
the
use of several specified electrical headers. While the
modularity
of
the standard
s beneficial in lowering non-recurring engineering costs i.e. every
student
designer
begins from the same
template), mating and de-mating
PC\104 headers is
very
dif
ficult
without damaging
components
and there must
certainly
be
a
better
solution
for
modular
circuit design
in
small satellites. Additionally, standard PC\104 circuit
board dimensions as shown in Figure 3·4) are within press-fit tolerances of the inner
diameter
of BUSAT s 4-inch
extruded aluminum cube
design,
meaning that either
the standard
must be
compromised or that other
materials must be
used for
payload
![Page 37: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/37.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 37/152
20
cube design
(t
he hub
ca
n a
ll
ow standard P C
\1
04 dimensions because it is machi
ned
from four pieces of a
lu
m
inum
).
-1:>0 (6-'J ) } OIA
1-.AI)
1 2 ~
( J . II ) OIAHOl£
l"P P
..A
C£5
.3:.:1 (
e.
ll l )
D i m ~ n ~ i o n ~
a r ~ in
i n c h ~ m i l l i n t ~ t ~ r s
_ , : , ; . ~ ; : tney _oett eft; m
l4 · e o ~ , ~ ~ . ,
. , ; . ~
t. '
, , , ,
(1(-.< Udft
.. 9 (Of' tetor) ... , , , ,.
A
},
,
, ;
_,
' , , ... ,
. ' " , " .-
'
, ' ,
">, . ' ",, ...., .", .' .',""
,
.,"..", ' .., ' .." "; )
~
¥o@
a o
0 0
0 •
0 0
n o
0 0
0 .
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 .
. CoO
0 0
( 2 . ~
".
t
u 0
-
f-
0.
0 0
0 0
:;
0 •
0 0
0 0
0 0
0 0
0 0
/
I I
~
I
- - : - : - x
- > ; '
/ / / / / /
/ /
/ /
/ X ' , ' / /
<-Xl
J
; " : ' , . ' ,
'
; ~ f / 0
c ; . , . ~
tnO)'
~ . n q
In
t\ te
~ . . , ; ,
.. , ,...
,
A
[, , $ ; - ~ , ~ < ; o ~ t ~ ~
, . ,
/ ' ~
,
j
~ I ~
; ; ;
c
. -
.... . . c . , ~ . . / . . , . ,
~ ~ . . J
. - ( , . ~ ~ . ~ , . . ,
. . , . .
- . , . . _
_
. t . ~ J
I
i ~ ~
g g
§
..
..
?)
...
, , / i
,",", ' ,/
, ..
,
.. .
)
,.t I ,
J
"'""
... , . ' ,
, .. " ,,"I'
,
l
, , ·
, . ·,
, £ - / ,
/
,r.
)
, ;
,
' rf ..,"
/ A
/ c3 ...
' fJ
,;
/ / "
,,.
. 1 ~ § · ~ ' - "
)
/ , ' / I
/ /
l t - ~ 0
' :::. ,.•,,/J
, , , , J
', , ..
;
..
, ; _ , ~
,
.
'/,".,'', '
.;.
I/ / / , I
I
,. , ,
/ .I I I'
P'' J /
J_ J
r----t-
I
3 ~ 5 0
Q ; )
' . l l >
{e
1
liZ)
.J-2:1
2 6 )
:l:Kl
( ;
J }
2
00
. :::00
-
. lOO
(-1 2 .70)
Figure 3·4: Detail from
the
P C\104
embedded
systems standard
showing fastener placement, h
eader
(right) and keep-out areas.
The PC\104 circuit
board
stack wit hin the hub is supported by t hree sets of four
mid-plane standoffs
that
fasten directly
to
t he four wa
ll
s of t he hub
(Figure
3·5) ,
which allow each of t he avionics subsystems to be fully integrated (with standoffs)
before enclosing them in the
structure.
This allows for easy trouble
shooting
should a
![Page 38: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/38.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 38/152
2
component
within
the
hub
stack
fail as often happens
during
satellite
environmental
testing and integration.
igure
·5:
Detail view of
the hub stack
assembly showing
three
mid
plane standoffs.
Each
standoff fastens to
the external
hub walls in a
slotted
groove that allows variations in hub
stack
height due to non
standard PC\104 overhead height
in COTS
components.
igure ·6:
Detail view of
bottom cap and
three hub walls inte
grated . Apertures for 25-pin and .SMA connectors are shown as well
as a tongue-and-groove fastening technique that minimizes
the need
for
fasteners along
the
entire length of
the
hub and
the
labeled
PC\104
bolt pattern
![Page 39: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/39.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 39/152
22
The hub interfaces electrically to the payload bay via four axially symmetric walls
of the Endoskeleton (Figure 3·3) . Both data and power are routed
through
four 25-
line wire harnesses
that
connect at the
hub
's
top
cap
and run
to one of the four
Endoskeleton walls (Figure 3·7). As shown in Figure (3·6), the hub bottom cap
allows
routing
to
the
external components of
the
BUSAT bus, including solar
arrays
,
external
sensors and radio
antennas
.
igure
·7:
Top left: schematic
top
view
of the
BUSAT wiring
har-
ness. Composed of several standardized short cable harnesses,
this
wiring method alleviates the need for costly
and
time-consuming one
off harness design and allows
the hub to be
removed from
the satellite
at
any time during integration. Top right: side vew of the harnessing
scheme from
the
payload (left) to endoskeleton to
hub
(right).
In order to satisfy low-cost requirements for the satellite
that in
a structural
design, translate
to
judicious use
of
machining tolerances, a 45° surface has
been
included in
the
hub
bottom
cap
to
provide regis tration
and
a
lateral restraint. When
the hub
is pressed down
into the
satellite structure
by the
exoskeleton
top plate
this surface alleviates the need to access the bottom of the hub for fastening (Figure
3·8).
![Page 40: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/40.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 40/152
3
Figure
3·8:
The hub. Inset: A 45° filet fixes the hub in place.
3 1 2 Cube
Stack
BUSAT s containerized structure is based on a broad interpretation of the CubeSat
standard Due to
the
availability
and
price of extruded aluminum
tubing
(4-inch
outer
diameter, 0.125-inch wall) as well as
the
relative ease with which it is machined, the
BUSAT standard 1U cube is 4 inches on edge and contains a volume of 64 in
2
. The
Birtcher WedgeLok™provides a novel
and
effective means of fixing the cube
stack
in
place (Figure 3·9).
Compressive
movement
Lateral movement
Figure 3·9: The Birtcher WedgeLok is the principal fastener
used
in
BUSAT s easily integrable design. A
torque
applied
to
the
tension bar
results in a lateral pressure applied to
the
cube stack.
![Page 41: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/41.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 41/152
![Page 42: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/42.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 42/152
5
3 1 4 Exoskeleton
Figure 3· : The
BUSAT exoskeleton.
The
Exoskeleton fastens directly
to
the Endoskeleton
s
four identical side walls
s
the
satellite is assembled. Deceptively simple
in
concept,
the
Exoskeletons primary
purpose
is
that
of a space-born shipping container. The external
stru
ctures two
key features are its ability
to
allow standard
(and
customized
apertures
for
payload
instrumentation and the
ability
to support the
forces associated
with launch and later
solar
panel
deployment .
Figure 3· 2: The
BUSAT-NMT
Langmuir Probe
shown
protruding
through
the exoskeleton top panel.
![Page 43: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/43.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 43/152
6
3 1 5 Deployables
Since BUSAT will
be
sun-facing under normal operating conditions
but
will begin
its
operational
life in a
tumbling
state, solar panels that begin the mission arranged
around
the satellite and are moved to face
the sun
are optimal.
This deployment
configuration is easily realized in
the
form of a simple door, as seen in Figure 3·13).
Figure 3· 3: The BUSAT satellite showing
modular
solar
panels
de
ployed.
While the spring-loaded hinge and
latch
concept is
not
challenging to design, cer
tain features of the deployment mechanism requi red sufficient design effort and
testing
due
to
the
risky
nature
of deployable
structures in the
eyes
of
past UNP competition
reviewers. Because the deployment hinge was
predicted to
be the weakest part of
the
design all analysis was completed
s i
the hinge pin were not
in
place Actuation
is accomplished using the TINI Aerospace Frangibolt Figure 3·14), a non-explosive
actuator NEA)
that
functions
by heating
a
shape-memory
alloy
to expand and break
a pre-scored 4 bolt. While the Frangibolt has flight heritage on many successful
spaceflight missions,
any
moving parts designed by
Boston
University students no
matter how good) will not be as well trusted. t is therefore necessary
to
find a
method of allowing the
Frangibolt to
completely
constrain the
deployment
without
the use of additional moving
parts
in this case, hinges).
![Page 44: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/44.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 44/152
27
Figure
3· 4: The TINI
Aerospace
Frangibolt
(shown in red) with
associated housing and
custom
titanium flange for
heat
isolation.
The
Frangibolt itself required
that no
shear
force
be transmitted to the
breakable
bolt at
any time
before deployment
at the
risk of early
actuation
at the test facility
not
dangerous)
or on
launch (very dangerous) . A cup-cone kinematic
mount
was
employed (Figure 3·15)
with
five
mated
conic surfaces of slightly
interfer
i
ng dimen-
sions (i.e. the male cone was very slightly more oblique than the female cup and
guarunteed
zero shear at
the
interface). Titanium was chosen for
the
female fixture
in order
to
isolate
the actuator during its
15 second
heating
phase.
Figure
3· 5:
Cross-sectional view
of
the
cup-cone
kinematic
mounting
structure.
This heat
isolation was shown
to be particularly important
for successful de
ployments during microgravity
testing, and the addition
of a titanium
washer
was
necessary
to
ensure
that
the titanium
pad did not act
as a
heat
sink
and
delay
or
![Page 45: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/45.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 45/152
8
even deny actuation.
igure 3· 6: The BUSAT deployable
panel
stowed and completely
constrained (at left) by the TINI Aerospace 4 Frangibolt and (at
right)
by
a titanium groove
(to
avoid cold-welding
under extreme
vi
bration.
While the cup-cone assemblly is able to produce
contact
forces in any direction
radially from or along the axis of the Frangibolt,
there
is only friction at
the
cup-cone
interfaces to keep it from
rotating.
For
this
reason a tongue-and-groove
feature
was
added to hinged end of the deployment
panel
to provide further support
(Figure
3·16).
Because of an offset hinge, the deployment
naturally de-mates
the tongue-and-groove
before allowing
the panel
to
swing free Figure 3·17).
igure 3· 7: The tongue-and-groove kinematic mount
after
deploy
ment.
It
is
important to note that
the tongue-and-groove feature allows all structural
analysis of the assembly
to
exclude
the
hinge. In other words, the only functional
![Page 46: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/46.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 46/152
9
reason for
the
hinge is
to
guide
the
solar panel into
the
correct position after deploy
ment.
3.2 Assembly
.
~
Figure 3·18:
The BUSAT assembly procedure (top left to bottom
right) beginning with
the
exoskeleton bottom plate, endoskeleteon,
cube stack, WedgeLok™s,
and
hub.
BUSATs assembly procedure is shown in Figure (3·18) , and demonstrates the de-
sign s minimization of non-recurring engineering by exploiting symmetry about
the
vertical axis. Four vertical Exoskeleton walls (identical
with
small modifications
post-
fabrication) are assembled
with
a bottom panel
to
form
an
aluminium (6061-
T6
box.
The
central component of the satellite is the Endoskeleton, a routing node that pro
vides power and data transfer to each of up to 24 payload modules. The Birtcher
WedgeLok™is
used
to
compress
the internal structure
(center). Once
the
Exoskele
ton panels have been fully fastened,
the
WedgeLok™assemblies are tightened
and
torque
coils are installed. Solar panels
and
antennas
are then
installed
and
connected
to the power system and
the
deployable mechanism is ready for testing, as shown
from a top view in Figures (3·16)
and
(3·17).
![Page 47: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/47.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 47/152
3
3 3
Mass Budget
The
cost of launch is directly related
to
a system s mass as it
is
acceler
ated
into
orbit,
with
per-kilogram cost of launch to low-Earth orbit ranging from
thousands
of dollars
to
tens of
thousands
of dollars, depending on
the
launch vehicle. Like
any
other
satellite, BUSAT s structural mass is closely tracked. The tables in
this
section
summarize BUSAT s mass budget and are organized by major assemblies.
Margin
Exoskeleton
Central Routing Cube
ommunications
- -
Magnetometer
CHICO
Auroral Imager
C DH
Figure 3·19:
BUSAT s mass breakdown arranged
by
subsystem.
![Page 48: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/48.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 48/152
31
Name
Mass (g)
Qty.
Total (g)
Hu
Outer Structure
Hu
b B
ottom Cap
405.83 1 405.83
Hub Top Cap
263.65
1
263.65
Side Hub Wall
33
0.38
2
660.76
Front/Back Hub Wall 194.26
2
388.52
P C s and Components
Fl
ex
UEP
S
3
1.
31
1
3
1.
31
30 Whr CubeSat St andalone Batt ery 1
03
.51
1
103 .51
Helium 100
Ra
dio
33
.00
1
33.00
I
nter
face,
Vo
ltage Regul
ato
r Stack
45
.
36
1
45 .36
RF Sw itching PCB
22 .68
1
22.68
Magnetometer Small Pl
ate
3.79 1 3.79
Magnetometer Large
Pl
ate
5.80
1
5.80
Magnetometer PCB
22.68
1
22.68
AD&C PCB
22.68
2
45
.36
C&DH
22 .68
1
22
.68
Na
noCDH
6.
02
1
6.02
Expa
nsion PCB
22 .68
4
90
.
72
Rout ing Node PCB
22 .68
1
22 .68
Right Angle Cable Plug
0.87 2
1. 75
Right Angle Jack
0.61
2
1.21
Stando s
0.25in Stando
ff
0.
13
24 3
.1
2
1524 Standoff
0.29
48
13.92
2200 Standoff
1.11 4
4.44
Midplane
Sta
nd
o
ff
1
0.41
3 1.23
Midplane
Sta
ndoff 2
0.66
3
1. 98
Midplane Standoff 3
0.59
3
1.
77
Midplane Standoff 4
0.41
3 1.23
Other Components
AirBorn Connector
1. 29
4
5.16
Fasteners
Washer
0.01
20
0.2
40
-40 1/
4
Bu
tto
n Head
Ca
p Screw 0.06 8 0.50
4
-40
3/
16
Bu
tton Head Screw 0.05 4
0. 18
6-32 1
/4
Bu
tton Head Cap Screw 0. 1 12
1.13
4-
40
Machine Hex Nut
0.07 4
0.28
Total ub
Mass
(g)
2212.38
![Page 49: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/49.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 49/152
32
continued)
Mass g) Qty.
Total g)
Endoskeleton
Endo Wall 512.88 4 2051.52
Endo Routing
PCB
44.31
4
177.24
Endo AirBorn Connector
17.0109
24
408.26
Endoskeleton Spacers
Bottom
Spacer
55.53
8 444.24
Interior Spacer
47.72
4
190.88
Half Interior Spacer
56.4
1
56.4
Total Endoskeleton Mass
3328.54
xoskeleton
Bottom
Plate
2532.
1
1
2532.
1
Top
Plate
2611.81
1
2611.81
Back
Plate
1776.07
1
1776.07
Side Plate 1
1755.73
1
1755.73
Side Plate 2
1754.01
1
1754.01
Front Plate
1755.12
1
1755.12
Total Exoskeleton Mass
12184.84
Empty Payload Cubes
Single Cube Assembly 295.10 1 295.10
Double Cube Assembly
260.75
3 782.25
Triple Cube Assembly
654.64
5
3273.2
Total
Empty
Payload Cube Mass
4350.55
Payloads
CHICO
295.10
1
295.10
. LPP-SHM Assembly
755.47
1
755.47
Total Payload
Mass g)
1050.57
We eLoks
WedgeLok Corner
Pad
196.03
4
784.12
WedgeLok Side Pad
207.2
10
2072
Total
WedgeLoks
Mass
2856.12
![Page 50: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/50.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 50/152
33
continued)
Mass g) Qty.
Total g)
FrangiBolt ssembly
X-Wing EXO
197.45
2
394.9
Cup
Attachment for X-Wing EXO 21.87 2 43.74
MiniActuator Housing
2.67
2
5.34
MiniActuator Bolt
1.59
2
3.18
MiniActuator FD04
10.62
2
21.24
X-Wing SOL
74.35
2
148.7
Total
FrangiBolt Mass
617.1
Torque Coil ssembly
TCH Bottom Side
72.19
3 216.57
TCH
Left Side
72.24
3 216.72
TCH Right Side
72.24
3 216.72
TCH
Top Side
72.24
3 216.72
TCH Standoff Special
4.62
24
110.88
TCH
Standoff Corner
5.92
12
71.04
Total Torque
Coil Assembly
Mass
1048.65
Solar Panel ssembly
Solar Panel
Plate
295.91
2
591.82
Support Flange
433.95
2
867.9
Spring Hinge Mounting Extension
39.15
2
78.3
Dowel
Pin
1.74
4
6.
96
Setscrew
0.21
4
0.84
Spring
0.03
4
0.12
Solar Panel PCB
17.46
18
314.28
Total Solar Panel Assembly
Mass
1860.22
MGS
ssembly
MGSE Vertical Support
1354.5
2
2709
MGSE Corner Plate
188.7
4
754.8
MSGE Corner Fixture
93.01
4
372. 4
Total
MSGE Assembly
Mass
3835.84
Total Mass g)
33344.80
Total
Mass
kg)
33.34
![Page 51: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/51.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 51/152
4
Chapter
4
Vibration Testing
4 1
Background
4 1 1
Vibration Testing for Small Satellites
Integrated
environmental
testing
makes
up
one part of the overall lifecycle of a small
satellite and occurs
after
fabrication and before launch. For BUSAT, environmental
test procedures including cleaning, functional testing, mass and volume verification,
and
final review will likely occur at
the
AFRL High Bay facility
at Kirtland
Air Force
Base in Albuquerque, NM (see Figure 4·1).
Integrated
environmental testing is one
part of this process and includes several mechanical vibration tests (usually performed
on a shake
table and
shock
testing apparatus
interspersed
with
electronics testing.
Figure 4· : N anosat-5 integration
and test
flow (from NS-7 User s
Guide)
![Page 52: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/52.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 52/152
35
This
is followed
by
a bake
out test,
during which
time
the satellite is
subjected to
temperatures
high enough
to
vaporize many of the volatiles that would
othe
rwise later
contaminate the vacuum chamber the volatiles will also boil
out
at low pressure).
Thermal and vacuum testing then ensues, followed by electromagnetic interference
and
electromagnetic compatibility
testing
(see Figure 4·2). These
last tests
ensure
that
the
satellite can function within the ambient electromagnetic environment
of
low-Earth orbit, as well as perform all electronic tasks without adversely affecting its
own circuitry.
Repea
te
d
ro
r X Y and Z
Axes
Pre-Test Sine
Si
ne Bur
st
Intra- Test Si
ne
1
sweep
Sweep Al iveness
Te
st
n
L
Random
Intra-Test Si
ne
1
VIOratlan
Test
Sweep Al tvene
ss T
es
t
n
L
Bal<eO
U
Sha
ck
Test
Past-Te
st Si
ne
Fu
Fun
ctional
Therma l vacuum
sweep
Te st
1
Therma l Balance
EM
VE
MC
igure 4· : Mechanical Test Flow (from NS-7 User s Guide).
![Page 53: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/53.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 53/152
36
4.1.2
Mechanical Vibration
Testing
r { R
~
~
Cl - CONTliOL 1:5POtiSC
D -
DRM:
0 - VTPUT
Figure 4·3: Concept schematic of
Thermotron
vibration table control
loop. 5)
Mechanical testing (the main focus of the testing described in this document in
cludes four categories of vibration testing. All four can be administered on a shake
table
, although other apparatus are also used such as centrifuges, drop
test
machines
and Hopkinson bar or hammer blow test rigs. A shake table can
be
explained as
a very large audio speaker or solenoid connected to a large power
supply
and con
trolled by a software system with closed-loop feedback from a control accelerometer
(see Figure 4·3, Figure 4·5). The software package models
an
incremental series of
mass-spring-damper oscillators. Control software causes an initial vibration
in
the
shake
table
by commanding the power supply to deliver an electrical oscillation to
the
coils in
the
vibration table, which in
turn
induces a mechanical oscillation in
the
armature, test object
and
control accelerometer.
The
control accelerometer provides
confirmation of the
actual
vibration induced
in
the armature to the control software,
which is then able
to
change the properties of the actual mechanical oscillation de-
livered
to
the
armature
.
In
this way, the control accelerometer becomes an active
![Page 54: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/54.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 54/152
37
part of the test equipment and at the same time provides a means of estimating the
forcing function
that
excites
the
test object
Control
ccelerometer
Channell)
Fasteners
Shake Table
Figure
4·4: Schematic view of
the structural
test as performed at
the
Draper
TF
showing control accelerometer mounted on
the
shake table
armature
![Page 55: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/55.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 55/152
8
F t)
mplifier
F t)
Figure 4·5:
onceptual
representation
of SRS
plot
generation
and
vi
bration table
control showing (in red) control accelerometer input
to
a
virtual
mass-spring-oscillator
array
implemented
within the
control
software.
The
model simulates a
system
response
based
on
the
control
accelerometer
input
,
and this
signal is
plotted
(red). Control software
commands
based
on this
simulation
are then passed to
a power
am-
plifier which sends
current
to the table
s
inductive
coils,
imparting an
acceleration
to the table and again
causing
the
control accelerometer
to
feed back
to the
software oscillator model.
The
number
of oscillators
in
the
software model (n)
determines
resolution (5)
Sine Sweep
The sine sweep
test
is
performed on
a shake
table and subjects the test object to
sinusoidal
motion
from 20 Hz
to
2 kHz, keeping
maximum
acceleration values very
low.
This
is
meant
to
identify
any resonant
frequencies
that the satellite may
have
without subjecting the
satellite to
potentially destructive
forces. When a sine sweep
is administered, several accelerometers around the structure
monitor
the vibration
response of various mechanical elements
of the
satellite.
When plotted against
fre-
![Page 56: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/56.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 56/152
39
quency, the resulting acceleration data clearly identifies resonant responses at each
accelerometer location.
In
the normal test flow, the sine sweep is
performed
first
to
establish a baseline response, and then again after each random vibration, sine burst
and shock test It can
be
assumed that any
structural
change such as a loosening
bolt
or plastic deformation) will result in some shift in
the
baseline sine sweep acceleration
responses.
-
r
0
·;:::;
ro
_
(IJ
(IJ
u
u
4:
ro
_
J
ro
z
Channel2
4
3
2
1
0.6
0.4
0.3
0.2
0.1
0.06
0.04
0.03
0.02
0 01
0.006
0.004
0.003
0.002
0.001
20
30 40 50 60 70809 ]00 200 300 400 500600 800100 0 1500 2000
Frequency (Hz)
Figure
4·6:
Typical plot of sine sweep acceleration vs. frequency. T his
is an in-axis response.
Sine Burst
The sine burst test is meant to expose
the
structure to inertial loading similar to
that experienced
on
a launch vehicle.
The
number
of actual
strokes
of the
table
as
![Page 57: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/57.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 57/152
40
it delivers the sine burst is very low (around 50 cycles), and the test is over very
quickly (around 5 seconds). Typically, sine burst
tests
are performed at a frequency
of
at
least one third the lowest identified natural frequency of the structure. In
the
case of the BUSAT structure, this lowest resonance occurred
at
relatively high
frequencies (rv300 Hz and rv500 Hz). Based on recommendations from UNP staff,
the BUSAT sine
burst test
was performed
at
50 Hz to more accurately recreate
the
launch environment, a lthough future sine burst testing would likely be performed
at
levels as low as 20Hz. Acceleration levels were set
at
24 G, as specified by the UNP
User s Guide.
Because the main purpose of
the
sine burst
test
is to simulate inertial loading
not
necesarily oscillatory loading), similar testing can also
be
performed in a centrifuge.
![Page 58: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/58.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 58/152
41
Burst Run 2
25
l
channe12l
20
15
10
c:
c
5
0
h
< J
Q
Q
u
u
<
-5
ro
::J
z
10
-15
-20
-25
0
0.1 0.2 0.3
0.4 0.5
0.6
0.7
Frequency (Hz)
Figure
4-7:
Ty
pi
ca
l
plot
of
sine
bur
st
acceler
at
ion
as
a fun
ct
ion of
t ime. This is an in
-ax
is response.
Random Vibration
Rather t han
ta
king its
inpu
t cont rol from an oscillatory s
in
e wave, t he random vib ra
t ion test is gen
erated
by a whi
te
noise source (4) and provides a
better
a
pp
roximation
of t he launch environment . Random vibrat ion data is plotted as a root mean square
(RMS)
va
lue
a
lso referred
to
as
Power
Sp
ectra
l Density
or
P SD) over fr
equen
cy.
This
prov
ides a quali
tat
ive com
pari
son between t he n
at
ural fre
qu
encies (re
pr
esented
by
p
ea
ks) in bo t h t he di
ag
nostic s
in
e sweep
test
and the ran
do
m v
ib
ra t ion
test,
al
though t he va
lu
es
are
shown in Gs in the former and G
2
z in the latter (see
Fig
ure
4·8 .
![Page 59: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/59.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 59/152
42
I
10.0000
· ·-·- • -----
--
-
c : : : _ ' ~ ~ ~ = · o c ~ : : = ± c F . = : ~ ¥ · ~ ~ ~ ~ ~ ~ ~
~ = = = · = = , ' = = = = -
1.0000
0.1000
0.0100
0.0010
0.00 1
20.00
100.00 1000.
2000.00
Frequen y (Hz
)
igure
4·8: Random
vibration
response raw data from 20Hz
to
2000Hz.
Shock
The
effect of
transient
events
that
are not
emphasized
in the random vibration and
sine burst
tests
are
measured
directly in
the
shock
test.
Shock,
or
a
sudden
forcing over
a very short
time period,
often affects stiff
or brittle
satellite
components (components
with higher natural frequencies) like circuit boards and solder joints. Shock tests
are
delivered very quickly and therefore require
testing
apparatus capable of rapid and
large accelerations. Consequently, shock
tests
for large objects are
not often
delivered
on shake tables. Instead, hammer blow tests, Hopkinson bar
tests, or
drop tests
apparatus are
employed.
Because cost and time were test
constraints
on
BUSAT s
three days at the ETF,
shock tests were delivered
at
the maximum possible acceleration values that the vi-
bration
table
and power
supply
could deliver. This
amounted to
around 60 of
the
given UNP test specification for shock tests
in
the Z-axis, and 40 of the
same
for
![Page 60: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/60.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 60/152
4
those delivered in the X andY axes (the difference is accounted for by the use of a
large slip plate
and
air bearing in
the latter
two tests, adding to
the
effective mass of
the
objects driven
by the
shake table) . Additionally, shock tests delivered on a shake
table include some amount of ramp up and ramp down motion before and after
the delivered shock (see Figure 4·9).
Shock Run 25
100
80
60
40
20
( )
0
-20
-40
-60
-60
0
0. 1 0.02
0.03 0.04 0.05 0.06 0.07 0.08 0.09
Time Milliseconds)
igure
·9:
Typical
time
history of shock test. This is an in-axis
response.
Ramp
up
and ramp
down motion
can
be seen from 0.00
ms to 0.04 ms
and
from 0.06 ms
and
0.09 ms.
![Page 61: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/61.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 61/152
44
Chapter
5
Data Analysis
5 0 3
Observations
Meeting the
100
Hz Stiffness
Requirement
The first
natural
frequency was higher
than 100Hz
consistently in all diagnostic sine
sweep
and random vibration tests, thus passing the minimum design
criteria
from
the UNP. Not only did the natural frequency remain higher
than
100 Hz
during
all
sine sweep and
random
vibration tests, but
it
occurred along a very clear and
distinct
curve toward the first peak above the 60 Hz noise, which can be seen in Figure 5·2.
All comparative graphs can be found in Appendix C
The
dominant first natural frequencies were found as an average of peak response
frequency values over each successive diagnostic sine sweep. Using this
method
a
good approximation of the exact lowest resonance frequency in each axis
can
be seen
in Table
5
.
1)
Random vibration data plots are qualitatively very similar to the diagnostic sine
sweep plots. This similarity could
be
indicative of a structurally resilient design - even
under the
more stressful
random
vibration test the N anosat-5 structure responds to
. z
233.61 Hz
221.16 Hz
Table
5 1:
BUSAT s lowest natural frequencies when subjected to a
sine sweep
test at
a fixed base in each of three orthogonal directions
described in Figure6·13
![Page 62: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/62.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 62/152
45
vibration inputs in the same way it did under the low-stress diagnostic sine sweep.
The
interplay between WedgeLoks ™ normal and friction forces in all three axes could
be one cause of
this
trend
and
a more thorough treatment of the
test data and
ana-
lytical models would likely yield proof of this.
omain
Peak
Shifts
Data analysis of
the
graphs from before and after each diagnostic sine sweep shows
a progressive shift toward a lower frequency with each successive
run
. Though the
shift signifies a decrease in stiffness, the changes are relatively small. Because
the
shifts are no more than 5 for each comparative diagnostic sine sweep, the use of
the WedgeLok™ proves to be promising for a flight-ready structure. It
maintains
a
level of stiffness
that
would not have been possible without strict tolerancing
and an
increase in manufacturing costs of the cubes.
Table 5 2: Percent change between Runs 3 and 5 Before
and
after
the random vibration tests in the Z- Axis)
and Run 12
and 14 before
and after
the random
vibration
tests in
theY-Axis .
Range Peak
Shifts
The
amplitude of the acceleration associated with the first natural frequency varies
greatly between runs and channels. As seen in Table 5.3, the greatest percent change
![Page 63: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/63.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 63/152
46
between any two runs and among all channels occurs between Runs 3 and 5 and
Runs 12 and 14. Each pair is separated by a random vibration test in t he
Z-
Axis and
X-Axis, respectively.
T he outliers in t he percent shift are Channel 4 along the Z-Axis (which is discussed
in a later section) and
Channe
l 6 along theY-
Ax
is. However t he Ch
anne
l 6 response
amplitude is so small along
the
Y-Axi
s,
the large percent (197.69 ) be lies the sig
nificance of the shift in acceleration. In Run 19, Channel 6 has a 0.024
gn
response,
wh ile in Run 22 , it has a 0.073
gn
response.
O lam el
Rill
2
3
4
5
6
7
z
Axis
(gn)
X Ax is (gn) Y Ax is ( )
1
3
5
•
1 2
•
•
22 3
9.13249 9.385m 15.10611
14.16861
1.237602
1.194257 1.412469 1.383674 0.225371
0.248292 0.248292
14.26768 14.70858
21.57141 18.62549 0.138114 0.
11
9972 0.041706 0.07677 0.710027
0.803721 0.803721
0.335918
0.432391 1.167563 1.207651 0.743424 0.711831 0.721749 0.770507 10.24388 10.59418
10 .59418
0.209738 0.198381
0.314879
0.106336
7.306512 7.06328 8.580392 a 194292
0.608061
0.657268
0.657268
11 .65426
11 .87445 18.46641 15.95839 0.663216 0.654204 0.923633 0.833619 0.024359
0.072515 0.072515
0.680448
0.676037 0.77299 0.425868 11 .06-493
10.71398
13.14629 12.52688 0.941785 1.010123
1.010123
Ta
bl
e 5 3 : Response amplitude measured in uni
ts
of Normal Accel
eration gn= 9.80665 m/s
2
) . Run 24 is a
dup
li
cate
of Run 22
due
to
human error.
Data Outlier: Th
e
C Channel
--
Figu
re 5·
:
C-channel assembly.
3
0.216734
0.738183
10.17026
0.782798
0.063465
1.18017
![Page 64: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/64.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 64/152
7
Ol
a
nn
el
z
x is( )
X
- Axis ( ) Y
- Ax is ( )
RW I 1 3
3 5
5 9
10 12
12 14 14 19 22 22 a4
24 26
2
277
-
60
.96
6.21 3.
50
-18.
27
2.02
10.17 0.00
3
-
295
-46.66
13.66 13.14
65.24 -84.
08 11
.93 0.00
4 -28.72 -
17
4.65 -1.69 4.25 -1.39 -6.76 3.42 0.00
5
5.41
-58.72 66.23 3.33
-21.48 4.50
8.
09
0.00
6
-1.89
-55.51
13.
58
1.
36
-41.18 9.75 197 .69 0.00
7
0.
65
-14.
34
44.91
3.17
-22.70
4.71 7.26 0.00
a.nge
30 10
-12.44
-.3.08
CN•axis
Table 5.4:
Percent change in acceleration
amplitude
for each succes
sive diagnostic sine sweep test according to axes. Run 24 is a duplicate
of
Run
22 due to
an
unrecorded
test
run, hence the
0
change.
Change
over axis is a comparison of averaged
peak
shift for
the
first
and last
diagnostic sweep in each axis.
Table
5.5:
Percent change in acceleration
amplitude
between Runs 3
and
5 Before
and
after the random vibration
tests
in
the
Z- Axis)
and
Run
12 and 14 Before and after the
random
vibration tests in the
Y-Axis).
12.71
8.15
4.00
19.10
12.48
16.83
Channel
4 was attached to a c-channel WedgeLok™ retainer see Figure 5·1)
and
showed the most discrepancy in relation
to
other
channels, with peaks well below
the
first natural frequency of the rest of the structure Figure 5·2) . However, since the
c-channel was
not
well constrained
it had
more degrees of freedom
than
most other
structural
elements.
This
allowed different acceleration
and
lower resonance peaks
than
in data from other channels.
![Page 65: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/65.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 65/152
48
n
39 .8107 ------------;------------------
10000
0.1000
0.0010
~ ~
f t ~
~
~
~
W J \ n
l l . \ inpu16 Q
inpu17 Q
0.0006 ------------------ ---- --------- -------- ------- -
20.
00
100.00 1000.00 2000.00
Fre uency l-Iz)
igure
5·2:
First Diagnostic Sine Sweep. Run 1: Z-Axis.
Off-axis accelerometers reflect more
variation in
the
plotted data than
on axis
channels for the
same
sine sweep,
as
seen in
Figure
5·2.
Note
the curves on
channels
4, 5 and
7.
Channels 5
and
7 have noticeable
peaks
before the on-axis
channels reach
the
first
natural
frequency
at approximately
590 Hz.
Channel
4
has
two
noticeable
peaks,
although neither of those peaks can be
definitively established
as
first
natural
frequencies for this channel. It is also important to
note that
most
of the
activity
in
the
off-axis channels is occurring below
the
0.25
gn
control accelerometer
read-
ing
Channel
plotted in
blue
as a relatively straight horizontal line from 20 Hz to
2000 Hz .
This
suggests
that
the channels
in
question are experiencing a lower level
of
acceleration than
the
forcing that is applied
by
the shake
table
,
indicating damping.
![Page 66: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/66.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 66/152
49
Channel4
25
18
12
8
6
4
3
2
run1
i
'
l
- · - · -
- - run3
i
i
i
'
i I
i
runS
·l
r- i
.
--------
---
--·
--- -----
· : i -i-
-· -
run9
i
I
i
i
'
i
i
I
i
I I
i
I
I
'
I I I
I
I
'
I
'
i
i
i
i
i
i
I I
i
I I I
0.6
0
.4
0.3
0.2
( )
0.1
0
.0
6
0.04
0.03
0.
02
0.01
0.006
0.004
'
'
i
i
'
i i
i
i
i
i I i
I '
i
'
-
·
r---- i
1·
--· -
i-
i - '- - - --
-- - i
1
i
---
· · ·
I
I I
I
~ ~
·--
---
--
i--
·
---:
·-
~ l .
1-P
~
--
--
··_
:.....:
-
r---
,-
- ---
I
--
I
'
I
I , / . . /
i
r
·'
: .. ;j
-V-
R
----·
-- · 1
--.----l
i .
i
I
'
l"" .t/
rr
-- --
---- -
: i
I I
-i-
I I I
i
_;_
: - -
r-·-·---
------
--- --
--·-- -
-----.
------
r -
0.003
0.002
i i i i i
i
:
i
i
: : l i
0 0 ~ 0
30 40 50 60 70 8090 00 200 300 400 500600 8001000 1500 2000
Hz
igure 5·3: Channel 4 only. Diagnostic Sine Sweep Runs
1
3 5
and
9.: Z-Axis.
While mostly below the 0.25 Gn control level, Channel 4 does exhibit some res
onance amplitude values
that
rise above this level, indicating a lower first natural
frequency than the rest of
the
satellite
structure
.
An
isol
ated
view of
Channel
4
during all runs along the Z-Axis shows that there are multiple peaks before
the
en
tire system's overall first
natural
frequency
at
approximately 590 Hz. Aside from
this channel's disparity with the first natural frequency of all other channels, the
WedgeLok
™
maintained a level of stiffness in
the
structure
that
was
not
catas
trophic , but rather consistent with the design goals.
In
spite of all
the
questionable
and numerous
structural
elements that were continually edited up until
test
date , the
structure held up well and maintained a high natural frequency.
![Page 67: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/67.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 67/152
50
On Axis
Versus Off Axis
Data
urther analysis of data from X and Y axis vibration tests show that the anomalies
can
be greatly attributed to
the
axis of vibration.
The
on-axis sensors have higher
acceleration for all diagnostic sine sweep and
random
vibration tests, while
the off-
axis sensors reflect fluctuant responses, relative to the on-axis curves.
In Figure 5·4, the two X-axis channels, 5 and 7 have
the
highest acceleration
and
the
most congruous curves, while all
the other
channels are off-axis
and
have lower
acceleration,
but
more erratic curvature. All
the
channels except 5
and
7 have
at
least
one, but frequently more,
maxima
and minima prior to
the
system natural frequency
at 237 Hz. For example, Channel 3 in Run 14 has four maxima
and
fo
ur minima
between the 200 and 300 Hz range. This behavior is analogous to the behavior of
Channel 4 in the Z-Axis Figure 5·3).
Similarly, Figure 5·5 shows
that
the only Y-axis channel, Channel 4 now
has the
highest acceleration response and a smooth curve as it approaches the first natural
frequency
at
222 Hz. Whereas Channel 4 was the outlier in previous runs here it is
the only curve
with
a distinct resonant frequency.
![Page 68: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/68.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 68/152
51
gn
40.0000
------------------------------------
10.0000
1.0000
0.1000
0.0010
1.00E-04
20.00
20111020 BUSATronH x-oxls busat01 dillgnostlc-
si l
e
sw
eep· -
pr
ofile( )
. o - = o : · , : c c : ~ · = - · - - · - . : : : c ;
- c ' ~ ' : : : ' _ . . : . - . _ _ , 0 : . . . : : _ : ; - . : - c . ~ - = : . . ~ - : - ~ - : : : = - ~ - = - ~ ; , _ . . . : . - : . : : : : . .
- = . . - · n
high-abort( )
- 1
- - - - .. - 1 t- - low-abort( )
input2
1)
inp
ut3 1)
input4 1)
100.00
1000.00
2000.00
Frequency (Hz)
igure 5· : Diagnostic Sine Sweep, Run 14: X-Axis. All Channels.
gn
40.0000-.----------------------------------
-20111020 BUSAT run19 y-IXis-busatOi·dlllgnos ic
si e
sweep
10.0000
1.0000
0.0010
1
OOE 04
L---- ----- -- --- -------------- -----------__J
profile( )
high-abort( )
low-abort( )
inp
ut2 1)
inp
ut3 1)
input4 1)
input5 1)
input6 1)
lnp
ut7 1)
control( )
20.00 100.00 1000.00
2000.00
igure 5· : Diagnostic Sine Sweep, Run 19: Y-Axis. All Cha
nn
els.
![Page 69: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/69.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 69/152
52
rumheading and
Separation
Sine
urst
graphs show
out
of phase acceleration in all axes. In
the
Z-Axis,
Channel
3
is out of phase
with
Channel 2 and 6 Because accelerometers are located on different
surfaces, all with
the
potential of drumheading,
it
is likely that Channel 3, attached
to
the End Cap of the center 3U Auroral Imager cube, was drumheading out of phase
with the
Top Panel of
the
Exoskeleton (Channel 6). Since Channel 2 and
Channel
6 were very close,
though
slightly
out
of line in
the
Z-Axis, the in-phase sinusoidal
waves indicate
that
they were accelerating in the same direction at all times during
the
Sine Burst
test
.
There is no evidence of separation within the system because the on-axis sensors
reflect a clean sinusoidal curve that does not have any aberrations throughout the
time domain. f separation had been an issue, the resultant chatter or knocking
would have shown up in
the
sine burst
data
as transient , shock-like waves . The sine
burst test displayed only smooth sinusoidal curves indicating that separation was
probably not an issue.
![Page 70: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/70.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 70/152
53
on
32.0000
28.0000
control( )
24.0000
20.0000
16.0000
12.0000
8.0000
4.0000
0
-4.0000
-8.0000
-12.0000
-16.0000
-20.0000
-24.0000
-- 1
inpu12 1
: _ _ _ ~ ---
l_ftMl r·:- inpU1
4
(t)
J
- inputS( )
'iJVIvv --:-- inpU16 t)
i\f N
" input7(1
tv re-
- ·
-
I
i
npU13
(1)
~
l
j
J
__
_ ____
-28.0000
-32.0000
-34.0000
-41
0
100 200 300
400
500 600 700 758.594
n me (Milliseconds
igure 5·6: Sine Burst
Run
02 : Z-Axis. All Channels.
gn
34.0000
. . . · · ·
- - -
control( )
28.0000
..
24.0000
· ·· ·
input2
1
)
20.0000
-
input3
1
)
16.0000
12
.00
00
input4
1
)
8.0000
input5(1)
4.0000
~
input6(t)
0
F
input7(1)
-4.00 00
-8.0000
-12.0000
-16.0000
-20.0000
-24.0000
-28.0000
-32.0000
-34.0000
-41 0 1
00
200 300 400 500 600 700 758 .594
Time
igure ·7: Sine Burst Run 11: X-Axis. All Channels.
![Page 71: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/71.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 71/152
54
WedgeLok
™
Loosening
Any significant decrease in n
at
ural fre
qu
ency is indica
tiv
e of a decrease in
st
iffness,
system components loosening, or
st ru
c
tur
al elemen
ts
reaching a plas tic limit. An
in
cremental decrease in the lowest na tural frequency of every channel, alt hough small,
l
ea
ds to a h
ypot
hesis that there was loose
nin
g in t he WedgeLoks™over t he t hree
day
test pe
riod . The data in Table 5.6 is a list of the p
ea
ks along t he curve of all
diagnostic sine sweep test s. A comp
ar
ison of the frequencies shows a progressive
decrease in magnitude with cont inued vibration testing. However Tab le 5. 7 reflects
an overall
stab
ili
za t
ion toward the l
ast
few tests.
Th
e decrease in
st
iffness for the
Y-Axis is much smaller in relation to
both
t he X-Axis and Z-Axis.
Th
e apparent
st iffening between runs 14 and 18 is an anomaly and any conjecture
to
the cause or
effect
wo
uld be unfounded.
Channel
A..
2
3
4
5
6
7
n Ja
Torque
Z - Axis (Hz) xis z) Y Axis (Hz)
(
111-32
)
1
3
•
1D
12 4 11
•
22
a
16: : 1 589.03
582.45 556.63 541 .39 237.42
236
.35 229.60
230
.32 221 .93 220.43 220.
43
220.68
16: :1 569.03 582.
45
556.63 541 .39 237.68 236.35 234.76 230.32 221 .68 220.43
220.43
220
.68
6:t
569.03
561 .79 556.83 541 .39 236.88 235.82 229.54 229 .60 221 .68 220.43 220.43
220.68
6:t
592.35
589.03 565.66 568.85 237.42 236.35
229.60 230 .32 22243
221 .43 221 .43 221.43
6 :t
589.03 582.45 556.63 541 .39 237.42 236.35 229.60
230
.32 224.« 219.
94
219.94 220.19
6:t
569.03
582.45
5.. 0
.63 541 .39 237.42 236.35 229.60
230 .32 222.43
221 .43
22
1.43
22
1.43
Table 5.6: Na
tural frequencies for each channel determined by
th
e
first peak along the curves of t he diagnostic sine sweep tests. Run 24
is a duplicate of Run
22
due to human e
rr
or .
Frequency
Shift:
Analysis
Jna
Torque
0-32)
in-lb)
7: : 1
7:t
7: :1
7: :1
7: :1
7: :1
While not outside of the specifica tions given
by
t he UNP for st ru
ctur
al
test
ing, some
shifting was obse
rv
ed
in
th
e lowest na
tur
al
fr
e
qu
ency of
th
e BUSAT
st
ruc
tu r
e over
t he t hree-day
spa
n of testing. Shown in Figure
5-8
are t he four diagnostic s
in
e sweeps
th
at
were interspersed between shock, r
an
dom vi
brat
ion and sine bur
st
tests. There
is a clear
and
somewhat regul
ar
pe
ak
shift from Run 1 (blue, at right) to Run 9 teal,
at left ). Addit ionally, t he quality of the peaks appear to indicate decreasing damping
![Page 72: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/72.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 72/152
55
Channel
Z -Alcis ( ) X -Alds (%)
y - Alcis (% )
Torque
(%)
im 1 3
sas
sa•
1t1 1Z
ZA 4
,..,.
••:zz
:ZZA:M
....
o Il
2 1.12
4.40 2.77 0.45 2.77 -0.23 0.
67
0.00
0.
11
56.25
3 1.12
4.40 2.77 0 .56 0.67 1.89 0.56
0.
00
0.11 56.25
4 1.23 4.29 2.77 0 .45 2.66
-0
.
11
0.56 0.00 0.
11
56.25
5 0.56 3.97
-0.56 0 .45
2.77
-0.23 0.45
0.00 0.00
56.25
6 1.12
4.40 2.77 0 .45 2.77 -0.23 2.00 0.00
0.
11
56.25
7 1.12 4.40 2.77
0 .45
2.77
-0.23 0.45
0.00 0.00
56.25
::.::.
7
U 1 1 ?1
• .25
Table 5. 7: Percent change of natural frequencies for each succes
sive diagnostic sine sweep t est acco
rdin
g to axes. Change in nat ural
fre
qu
encies are c
ompar
ed to percent change in t orque r
ea
dings from
Wedge Lok
s™
.
Run 24
is a duplicat e of
Run 22 du
e to errors in record
keepin
g,
hence t he 0% change.
Ch
ange over axis (last row) is a com
par ison of t he average p
ea
k of all cha
nn
els for t he fir
st
and last run
along each
ax
i
s.
with each subsequent test .
One possible explana
tion
of t his phenomenon is that
th
e WedgeLoks™ co
uld
be
loosening over t ime
du
e to relaxation of shear st resses in the
we
dges t hemselves, due
to t hermal cyclin
g,
or due to t hreaded tension rods vibrating loose. This would in
t urn cause an overall loose
nin
g of the structure a
nd
therefore a lower fundamental
frequency.
![Page 73: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/73.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 73/152
0.1
0.06
0.04
0. 3
56
Channel
0.02
------ -------- ----- ----- --
- ----------- ------ ----
520 530
540 550 560
Hz
590
600
igure
5·8: While still within specifications, the BUSAT structure
does display some shift in lowest natural frequency. Generally, the
peak shifted lower by around
1 to
2 z after each subsequent test
frequency and magnitude are labeled for each of the four response
curves).
![Page 74: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/74.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 74/152
3
;
:
25
i
20
i
15
0
;
'Z
10
Iii
w
5
0
57
-
--eng ineeringtoolbox.com
1
Carbon Steel, C < 0.3'4
Nickel Steels, Ni 2 4- 9'4
Cr
-Mo Steels,
Cr
2'4.3'4
Copper
leaded Ni-Bronze
Nickel Alloys - Monel
400
Titanium
Aluminium
-500
.300
-100 100 300 500 700 900 1100 1300 1500
Temperature (deg F)
Figure 5·9: Elastic modulus as a function of temperature
(www. engineeringtoolbox.com) .
The regularity of the peaks, however, suggests that
at
least one
other
hypothesis
is warranted . While
temperature
was not controlled or measured during testing, it
can be assumed that the climate inside the Draper
test
facility was relatively
stable
with
fluctuations of only a few degrees
and
that
the test
facility
heating system
was
sufficient to offset the chilly New England fall weather.
For a two-dimensional solid in which both transverse and longitudinal waves are
possible, the lowest
natural
frequency can
be
~ r i t t n as a function of
the
speed of
sound in the solid (aluminum,
in this
case):
5 .
1)
where n and are mode numbers, N and M represent characteristic lengths of
the
solid,
r
is the speed of transverse waves in aluminum and
£
is the speed of
longitudinal waves in aluminum. If n = and N = M
![Page 75: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/75.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 75/152
58
Inserting the Lame constants, (5.2) becomes
f
_ _': -J :_ 11
2N p p
·
Simplifying and using the definitions 1-" = l ~ v ) and A
Poisson s ratio,
5
.2)
(5.3)
v
where
v
is
( l+v)( l-2v) '
(5.4)
Given that p
~
2 7 ~ and
v ~
0.35 for Aluminum, natural frequency as a function
of the elastic modulus
E
is given as
fn = (.021) V (5.5)
Since the elastic modulus E varies inversely with temperature (see Figure 5·9) , a rough
relationship between temperature change and elastic modulus can
be
established.
From Figure (5·9), this corresponds
to
a reduction of
about
2.75
GPa
for an increase
of 35°C. Assuming an inverse linear relationship, this corresponds to about 0.078 GPa
per degree. Substitution into (5.5) reveals
that
even a small change in temperature
can produce a large shift in natural frequency:
n n n
:lfn
=
2
N.021yi :lE(T)
=
2N .021
X
V0
.
78
X
10
9
~
2
N186H
z) (5.6)
Therefore, for values of
n
and N near unity, response frequency is very sensitive
to
changes in temperature. Additionally, since the elastic modulus decreases with in-
creasing
temperature
, the lowest natural frequency also displays
an
inverse relation-
ship with temperature , suggesting that future testing efforts should include either
![Page 76: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/76.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 76/152
9
precise temperature monitoring or fine temperature control.
Wear
Figure 5·10 shows an example of surface wear that was found after testing and disas
sembly. The pressure exerted by the Wedgeloks and the frictional forces
acting
along
the surface was large enough to remove
portions
of the anodized material.
Results
showed less evidence of wear
than
expected.
Figure
5·1 :
Evidence of wear on the anodized surfaces of a cube.
5 0 4 Summary of Data Gathered and Conclusion
The
vibration
tests supplied proof that the
WedgeLok™
concept
upholds
UNP vi
brational specifications and suggests that a redesign of the Exoskeleton Cubes
and
Spacers while necessary
may
need
to
include more features of the original structural
design than previously thought Considering the positive results of the vibration t ests
the
WedgeLok™ may be
used
with
the
right design
constraints
for
the
cube
arrange
ments in
the
next engineering design unit . Structurally the NS-5 prototype exceeded
the requirements from
the
UNP thus
the
primary objectives of the new design should
be modifications geared
toward
modularity ease of assembly
and
accommodations for
integrating new ideas in
the
other subsystems.
![Page 77: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/77.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 77/152
6
Chapter
6
Analytical Model
6 1 Introduction
Because
vibration testing presents
a
potential point of
failure for
many aspects of
a
satellite mission,
it
is desirable for
the systems
engineer
to
have tools
at
hand
that
will accurately
predict whether
a given satellite structural will fail when subjected
to the sine sweep and other tests described in Section 4.1) . Additionally,
prediction
of the vibration environment within subsystem
component housings is valuable in
identifying critical
points
of failure
within
payloads. In other words an excellent sys
tems engineering model would
produce
frequency response charts from
characteristic
information
about
a variable configuration
of
mission payloads.
Figure
6· :
The
model developed
in this
section seeks
to determine
the BUSAT satellite s frequency response based on
the
characteristics
of the payload bay.
Due to the cubic and
modular
nature of the BUSAT mechanical bus , an iterated
mass-spring-oscillator model can be developed to assess the frequency
response
of
![Page 78: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/78.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 78/152
6
the
system
for a given
payload
configuration. This is accomplished by recognizing
that each of
the
satellite payloads
can be
modeled as an assembly of masses, springs
and
dampers.
The
model developed in this section houses 27 1U point masses in a
cubic matrix of springs and dampers constrained by a vibrating, rigid housing. As a
mesh of relatively simple oscillatory systems, the primary challenge is organizing the
hundreds of variables so that control of
the
model is retained
throughou
t . or
the
remainder of this section,
it
is noted that
the terms
mass , cube
and
payload
are analagous and interchangeable.
6 2
Damped
Mechanical Oscillators
Development of
the
equations of motion for a coupled 27 degree of freedom mass,
spring, and damper system starts with
an
elementary model.
The
most basic model
of a vibrating sy
stem
has three basic elements arranged as in Figure (6·2) .
Figure
6· :
A basic mass, spring,
and damper
system where
the
mo
tion of mass m in response to
an
excitation ~ t ) is described
by
the
state variable measured from the origin in the
upper
left hand corner
of
the
diagram.
The
equations of motion are found
by
considering
the
free-body
diagram
for m ,
![Page 79: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/79.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 79/152
62
using Newton s second law
x
-
: ;, = ~ ) , where is the sum of forces applied
to the mass m and f
)
= f )) is the periodic forcing delivered
to the
mass via
damper and
spring and caused by
the
periodic motion of the platform:
mx
ex kx =
~ t ) +
~ ( t ) (6.1)
and
~ t ) is the motion of the rigid platform. Written in
standard
notation for a
second-order system
x
+
2 wnX
+
~ x
=
. : _ ~ ( t )
+
} 5 _ ~ t )
m m
(6.2)
with
forcing terms from the platform motion placed in the right hand side of the
equation. The
undamped
natural frequency of
the
system is
(6.3)
which can be solved immediately upon inspecting the single spring
and
mass.
The
damped natural
frequency is given
by
6.4)
where
the
damping ratio, ( can
be
solved from quation (6.2)
and
is
c
=
2Vkm
(6.5)
f
the
mass is placed inside a vibrating box
with
more
than
one coupling
to
the
input motion
x t)
as in Figure (6·3), the equations do not differ greatly, but deserve
mention as the next
step
in evolving the model to a larger system.
![Page 80: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/80.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 80/152
63
igure
6· : A single degree of freedom mass, spring
and damper
sys
tem
with two points of coupling
to the
oscillating platform.
The equations of
motion
for m
3
are
(6.6)
Several comments can
be
made about
the
utility of the enclosed single-mass system
shown in Figure (6·3).
irst
,
it
is recognized as
the
same
system as
the
one
represented
in
igure
(6·2) with c
=
c
2
c
4
and
=
k
2
k
4
. Second, while redundant , this
system will prove
to
be a useful simplification in the larger model.
It
is also
noted
that
the
subscript convention used to describe masses, springs and
dampe
rs captures
an element s position
with
respect
to
the origin. Finally, the vertical motion of the
platform is a very good model for the shaking of
the
vibration table, which can only
impart vibrations in one axis at a time.
![Page 81: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/81.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 81/152
64
6 3 Multiple Degrees of
Freedom
In
general,
the
equations of motion for a vibrating system with more
than
one degree
of freedom can be
written
in matrix form as
Mx +Cx + Kx =
f t)
(6.7)
where M and K are symmetric matrices representing the mass, damping
and
stiffness of
the
system,
and f
) some forcing function affecting
the motion
of the
system. In
the
case of
Equation
(6.6),
the
forcing
term
is dependent on
the motion
of
the rigid supporting platform as well as
the
connecting dampers and springs, therefore
j t) =
c ~ : 4 ~ t )
+ k 4 ~ t ) .
Undamped natural frequencies are found by solving the eigenvalue problem
Kx=
.Ax
(6.8)
where
the
system s natural frequencies Wn and mode shapes are recovered from the
eigenvalues
and
eigenvectors of Equation (6.8) .
6 4 Damped
Frequency
Response
or Multiple
Degrees
of
Freedom
In the case of the BUSAT
structure
undergoing a sine sweep over a range of frequencies
with
the magnitude of acceleration held constant, the forcing function is sinusoidal
by
definition and
can be
represented
with
complex exponentials:
(6.9)
Taking the real
part
of
f t)
can be expressed as a function of
f t) and
its complex
![Page 82: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/82.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 82/152
65
conjugate yielding:
6.10)
Similarly, one
can
assume a steady-state displacement
Xp
t) of
the
matching form:
6.11)
Taking the real part of Xp t) yields:
6.12)
Substituting f t) and xp t) into quation
6. 7)
and reorganizing terms yields:
-w
2
M + iwC + K) Xpeiwt
-w
2
M + iwC +
K)
~ X = ~ F e i w t F - i w t
2
p
2 2
This can be separated into two equations - one real and one imaginary:
-w
2
M + iwC + K) ~ X
iwt
= ~ F e i w t
2 p 2
-w
2
M iwC + K)
~ X e - i
=
~ F e i w t
2 p 2
6.13)
6.14)
6.15)
Because
8)
is the complex conjugate of 7), they are redundant equations.
The
term
-w
M iwC K) is a matrix, and can
be
inverted
to
find Xp w):
Xp w) = -w
2
M+ iwC
K -
1
F
Xp w) = G w))-
1
F,
6.16)
6.17)
![Page 83: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/83.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 83/152
66
where Equation (6.17) is the complex displacement frequency response,
and
is
valid for oscillating systems of any size and configuration for which
G w)
can be
inverted, including a single degree of freedom damped oscillator in Section (6.2).
Since acceleration frequency response plots are desired (based on the
data
collected
during shake testing
and
based on industry standards , Equation (6.17) is differenti
ated twice and the magnitude is taken. Since
taking
the second time derivative
in
the
time domain is equivalent
to
multiplying by a factor of w
2
in the complex frequency
domain
, the final
quantity
to be
plotted
is
(6.18)
6 5
Spring Coefficients
and Forcing Input
o t i o n ~
Initial values of mass are then set
at
1
k
and spring constants k are calculated using
Hooke s law
(6.19)
where the cross-sectional area is based on
the
dimensions of 4-inch
extruded
alu-
minum
tube with a 0.0032 m (0.125-inch) wall, modulus of elasticity
E
is taken to
be 69 x
1
9
Pa (Aluminum 6061-T6), and Lis O.lm.
Looking again at Equations (6.9) and (6.5),
the
forcing term is written as
(6.20)
Since the motion of the vibration table discussed in Section (
4
, is the sinusoidal
displacement
that
will drive the network of masses,
![Page 84: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/84.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 84/152
67
~ t )
=
iwAeiwt
~ t ) = -w2Aeiwt,
6.21)
6.22)
6.23)
where A is the magnitude of displacement and w is the frequency sweep over which
the
frequency response is plotted.
By definition of
the
sine sweep test , the forcing magnitude is a constant 0.25G
0
,
where G
0
is 9.8 From Equation 6.23) ,
6.24)
and
therefore, A= A w) = Now Equation 6.20) is written
=
t w - - e
+
e
(
. 0.25G
iwt C2 +
C4) 0.25G
iwt k2 + k4
w
2
m
3
w
2
m3
6.25)
and the
magnitude
can be
recovered as
JSRJ +
<;Sj or
6.26)
Next,
the
magnitude of the complex displacement response
JXP
is pl
otted
over
frequency for a range of damping constants using
quation
6.17) , as seen
in
Figure
6·4). Qualitatively, the plot is similar to the test
data
in Section 4), but displays
far higher resonance frequencies because of
the
low mass
to
spring coefficient ratio.
![Page 85: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/85.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 85/152
68
10°
Single DOF model
10-
2
10 .
10 '
110 '
0
3
1 ·1
;
$
§
1 ·12
c.
i l
0::
10· 4
10
1 ·18
10-
20
10°
10 '
10
2
10' 10
10
5
10°
10
7
10'
Frequency
(Hz)
igure 6·4: Single degree of freedom mass-spring-damper system dis
placement response
plotted
over frequency.
I t
is also noted that this is a plot of displ cement over frequency that , while not
directly applicable
to
the test
data
discussed in Section
4),
does illustrate
the
effect
of
Equation 6.24
which constrains
the
sine sweep
to
a constant
cceler tion input.
Finally, the system's acceleration frequency response is
plotted
, as shown in Figure
(6·5).
As in Figure
(6·4), the
high resonance frequency response reflects the relatively
low mass of the system when compared
to the
spring constants applied.
![Page 86: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/86.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 86/152
10
1
3
10 '
10°
10
1
10
2
69
Single DOF model
\
0
3
10
4
10
5
10°
10
8
Frequency (Hz)
Figure 6·5: Single degree of freedom mass-spring-damper system ac
celeratation frequency response plotted over frequency.
6 6 Three
Degree
of Freedom
Model
Extrapolating the vibrating single-mass system into a three-mass system, position
vectors x
1
to x
3
are assigned to each of
the three
masses and the system is driven
by the
motion
x t),
as in Figure 6.6). It is also
noted
that in Figure 6·6), the
simple numbering scheme
adopted
in Figure 6.6) has been extended to the larger
system. Proceeding vertically down from
the
origin
at
the top left (at k
=
0 springs
and
dampers appear
at
even positions beginning at k
=
2 while masses only occupy
positions k = 1, k = 2 and k =
3.
Since
the
model does not
extend
in
the
i or
direction, there is no subscript applied.
![Page 87: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/87.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 87/152
70
igure · :
A three degree of freedom mass spring
and damper system
with
four
points
of coupling to
the
oscillating platform.
The equations of
motion
are again
obtained
from free-body diagrams of each mass
however a general method of analyzing
the diagram can be
used.
y noting that
any mass m is connected to other mass elements by springs or dampers in positions
1 a general formula for the
kth
mass is found.
![Page 88: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/88.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 88/152
7
•
l
igure
·7:
Generalized free-body
diagram
for the
three
degree of
freedom system.
and finally, rearranging quation ??) according to state variable,
6.27)
quation
6.27) represents the generalized
equation
of motion for the three
mass
![Page 89: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/89.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 89/152
72
system, where subscripts = 3 5 7 call appropriate values for each of
the
masses,
springs
and
dampers
that
are relevant
to the kth
equation of motion for
the
system.
Since the three-mass system is bounded above and below the three masses m
3
, m
5
and m
7
,
the general equation for state variables
±
2
will result in
state
variables
x
1
and
x
9
entering
the
equation. Looking again
at
Figure 6·6),
the
position values
k = 1 and k = 9 indicate the rigid structure surrounding the system and are therefore
variables of motion associated
with the
forcing term. Therefore, the full equations of
motion for th is system are
c
2
c
4
• C4 .
X - [ 0 ] + X - X5
m3
m3
k2 + k k c2 · k2
- [ 0]
+ X3
x5
= - ~ t ) +
- ~ t )
m
3
m
3
m
3
m
3
.. C4 • C4 C5 . C5 .
X5 -
X3
+ X5 - X7
m
5
m
5
m
5
k k + k6 k6
- X3 X5 - X7 =
0
m
5
m5 m5
.. C5 • C5 + Cg . [ O
X7
X5
+
X7
m7 m7
k6 k6 ks
cs ·
ks
- x5 + X7 [ 0] = - ~ t ) ~ t )
m
7
m
7
m7 m7
6.28)
Where
the
terms on
the
right hand side of Equations 6.28) are the motion x
1
=
x
9
= and ±
1
= ±
9
= of
the
rigid enclosure. he equations of motion
are
therefore
shown in Equations 6.30)
and the
normalized) frequency response is shown
in
Figure
6·8).
![Page 90: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/90.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 90/152
73
Where
[
1
0
~ ] ,
=
m2
0
m3
[k
+
4
K = -k4
k +
k6
0 k6
10
2
s
~ ~
-
C4
0
C =
-
c4
C4
+
C5
~
0
- C6
C5 + Cs
0
{ ~ t )
+
}
~ ]
j t)
= 0
k6 + ks
c s ~ t )
+
s ~ t )
3DOF System in Rigid Enclosure
1) 1
10
4
Frequency Hz)
3DOF System in Rigid Enclosure
1)2
\
'
Cll
'
:
0
a.
'
ll
::
10°
10
2
10
3
1) 1
10
4
Frequency Hz)
3DOF System in Rigid Enclosure
10
4
Frequency Hz)
y
}_[\
P
-.....:::::::::
Figure
6 ·8: Three degree of freedom forced mass-spring-damper accel-
eration
response plotted over frequency. Undamped
natural
frequencies
ar
e marked by vertical lines.
(6.29)
(6.30)
![Page 91: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/91.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 91/152
7
6.
7
Nine Degree of Freedom
Model
Coupled)
Figure ·9: Schematic
representation
of a general planar
cube
array
element
surrounded by
four masses, four springs and four dampers .
As
the
image in
Figure
6·14) shows, a nine degree of freedom model is also
simply
constructed
using a slightly more complex version of the
three
degree of freedom
method, as
shown
in Figure
6·9) , and the general
equation
of mot ion that describes
any
of the nine state variables is
C i+1,k) C i-1,k)
C i,k+l)
C i,k-1) .
X i,k) X i,k)
m i,k)
C i+1,k)
. C i-1,k) .
X i+2,k)
- X i-2 ,k)
m i,k,k) m i,k,k)
C i,k)
.
C i,k)
.
X i,k) - X i,k)
m i,k,k)
m i,k,k)
C i,k+1) . C i,k-1) .
X i,k+2) - X i,k-2)
m i,k,k) m i,k,k)
k i
,k+1)
k i
,k-1)
X i,k)
m i,k)
k i,k+l) k i,k-1)
m
X i,k+2)
- m X i,k-2) = 0
i,k,k)
i,k,k)
6.31)
Using
Equations
6.31) and
this time
varying
mass
values for the
system, Figures
![Page 92: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/92.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 92/152
75
(6
·1
0) through 6· 1
2)
we re generat ed to show qualitatively t hat t he model develop
ment model reflects a real system. Simul
at
ing a system wit h one very heavy (around
50 kg) node, the lowered first n
at
ural frequency verifies
that
the system behaves
s
it should . Before cont inuing to t he full model, it is also useful
to
note that
the
sig
nificant asymmetry of a large mass in
on
ly one node of t he system generates very
different high-frequency e
ff
ects in each of the 9 degrees of freedom (as expected).
10
5
I
~
i
i
)
10°
/)
<
0
0..
/)
Q)
0:::
10-
5
10
3
10
5
I
/)
:§_
Q)
10°
---------
----
/)
<
0
0..
/)
Q)
0:::
10-5
10
3
~
10
5
N
Q)
10°
/)
I
0
0..
/)
Q)
0:::
10-
5
10
3
- -- -- - ---1--------
1 o 9DOF System in Rigid Enclosure
i [ I ' l I '
i I :
j
1
i
i 1
2 ~ 1
i
~
~ ~ ~
~
t
:
10
4
Frequency (Hz)
2 o 9DOF System in Rigid Enclosure
10
4
Frequency (Hz)
3 of 9DOF System in Rigid Enclosure
' j '.
)
: I
i
i
I I
1
I : '
I I ' : I I
I
---f -
-+-
- :T.: .
---
:
-
l
10
4
Frequency (Hz)
Figure
6·1 :
Masses
1-
3 of a nine degree of freedom forced mass
spring-
damper
frequency response pl
otte
d over frequency. Un
da
m
pe
d
natural frequencies for each mass are marked by vert i
ca
l lines.
![Page 93: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/93.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 93/152
i
l
l
i
I
T
;
76
4 o 9DOF System in Rigid Enclosure
i
i
I
_.1.
I
- -----
I
10
4
Frequency (Hz)
[
5 o 9DOF System in Rigid Enclosure
I
I
i
1
I
T
'
I
~
i
i
10
4
Frequency (Hz)
)
I
'
.
A
i
Y
'
'
i
i
)
i
'
I
i
I
I
I
i
i
-
A
I
I
1
--·
f--
~
i
i
I
6 of 9DOF System in Rigid Enclosure
. ~ . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . ~ ~
s
~
f = = = = = = = i = = =
~
= = = f = = f = i =
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
~ ~ . l k d
i
0-5 L_____________j___
----'-------l_____J_______L_j_
j l_____
____
______
__ __
_____1_-- --------i.__
i
10
3
10
4
Frequency (Hz)
Figure
6· :
Masses 4-6 of a nine degree of freedom forced
mass-
spring damper frequency response plotted over frequency. Undamped
natural frequencies for each mass
are
marked by vertical lines.
![Page 94: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/94.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 94/152
77
7 of 9DOF System n Rigid Enclosure
- - - - - - - - - - - - - - ~ - - - - - - - - - - - . - - - - - - - - - - ~
f
'
=
= t = = r = = =
~
~
~
y
r
F = 4 ±
±=b
iiJ
I l y .
I
o:: 1
o·s - - - - - - - - - - - ~ - - ~ __ _ _ _ _ _ - - - - - - - ~ - - - - - - - - - - - - - - - - ~ - - L . J _ ~ _ , _ ~ _ . _ _ _ _ . _ j
10
3
10
4
l
· ---·-
Frequency (Hz)
8 of 9DOF System in Rigid Enclosure
I
i
i
I
:
i
I
I
I
I
-----
- -
:
10
4
Frequency (Hz)
J
iY
i
i
I
I
i
i
I
i}
I
'
i
"i"" '
I
I
i
1
Figure
6 12:
Masses 7-9
of
a nine degree of freedom forced mass
spring-damper frequency response plotted over frequency. Undamped
natural
frequencies for each
mass
are
marked by
vertical lines.
6.8 Twenty-seven Degree
of
Freedom Model Coupled)
This model being developed is essentially a matrix that organizes all masses, springs,
and
dampers
within
a framework that reflects the orthogonal modularity of the
BUSAT
structure
itself. Because the models components are
distributed
regularly
about a 10x10x10
matrix,
they can be manipulated
in
a regular fashion using itera
tive loops in
MATLAB, and a generalized
method
for dealing
with the
rather large
(27
-DOF
system becomes more manageable.
In
the full model
the
subscripts i, j know run from 1 to 10
in order
to accomodate
![Page 95: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/95.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 95/152
8
all masses, dampers and springs within the model.
It
is notable
that
positions
i, j
k
0, 10 are empty, that masses occupy the
odd
position values i, j k 3, 5, 7 only, and
that
dampers
and
springs occupy even position values
i,
j
k
2, 4, 6, 8 only.
Thus
the model contains 27 masses, 108 dampers and
36
springs.
Only
motion
in
the
vertical z) direction will be considered in this model
6 9
Framework of Analysis
In order to predict frequency response, an equation of motion must first
be
developed.
The standard
notation
for this system is
Mx + C±+
Kx
f t)
(6.32)
where M, C and K are diagonal matrices representing the mass, damping
and
stiffness
of the system, and
j t)
some forcing function affecting
the
motion of t he system.
Since the BUSAT satellite has up
to
27 mass elements of interest (excluding the
exoskeleton, which is considered a rigid external
structure to
which forcing motion is
applied), there must necessarily be 27 degrees of freedom (one for each mass) and M,
C and K are therefore square
27
x 27 matrices.
Due to the high number of discrete elements (high for an analytical model, at
least) in the system, a regular and easily accessed data structure is needed for storage
of the satellite's spring, mass
and
damping characteristics. These characteristics will
also be the knobs
to
be adjusted in order to tune the model
and
calibrate it to
reflect
the data
described in Section (4.
1
, so
they must
also
be stored
in
a logical
manner for ease of access. The most intuitive structure reflects
the
cubic nature of
the satellite itself: a
10
x
10
x
10
MATLAB cell array holds values for the system's
dampers, springs and masses, allowing intuitive access by the systems engineer as well
as indexed access by
the
frequency response algorithm implemented in MATLAB.
![Page 96: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/96.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 96/152
79
Subscripts for
the
discrete modelling elements placed within
the
cell array will be
governed
by
the following rules:
• Masses will occupy odd subscript combinations of 3 5 and 7 only.
• Springs and dampers will occupy indexes within the cell array that have at least
one even value.
• Any mass subscripted with a single value of 1 or 9 will indicate the position of
the
rigid
external
frame.
Further
simplifications
are brought about by setting
all spring
constants
to
zero
except those that are subscripted with
odd
integer values of 3 5 or 7 Physically
this reflects the fact that damping will likely occur only between cubes or
between
the cubes
and the
structure
but
that spring forces are likely
to
be important only
in
the
vertical direction. In other words because
the
system is clamped primarily
in the lateral direction because of the placement of WedgeLoks vertically around the
perimeter of the cube stack it is assumed that
the
vertically mated surfaces of the
wedgeloks exoskeleton
and
endoskeleton will
be
better
by
use of friction forces
than
spring forces.
igure
6· 3: The BUSAT standard satellite-fixed reference frame.
The
unit vectors
i J and
k are assigned to the x y
and
z axes respec
tively.
![Page 97: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/97.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 97/152
8
A
method
of
naming
each variable for
tracking throughout the
model is also essen
tial. While
the
numbering
system is
arbitrary,
strict adherence to the identifications
given in Table 6.2) is absolutely necessary. Additionally, a frame of reference for the
structure to
be
reflected in
the subscripted
indices
of
the MATLAB cell
array
will
ensure
fidelity
in
the model. The frame of reference is shown in
Figure
6·13) .
igure 6· 4: A one-to-one transformation
between
payload cube po-
sition and state variable subscript, as shown
in
Table 6.2.
Because each
point mass
is
arranged in
a
regular
network,
the motion of
any given
mass
is coupled
with
the
motion
of
as many as
six different masses
surrounding it
four
lateral
adjacent,
two
vertical
adjacent .
Since
the
forces applied by dampers
and
springs to a given
mass are
a function
of
both
the
motion of that mass and
the
surrounding
ones,
it
is important
to establish
the positional relationship
between
mass
values in the
MATLAB
cell
array
and their corresponding values within
the
M , C and K matrix equivalents.
In
other words, a
regular
relationship
between
the
position
of a
mass
in
the
MATLAB cell
array and
the
subscript of
its
state
variable in
the
equations of
motion must be established. This relationship is given by the linear
function
f i , j, k =
4.5i
1.5j 0.5k-
18.5, 6.33)
![Page 98: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/98.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 98/152
81
ube Mass
--+
~ t t e Variable
X333
xl
X335
- t
X2
X337
- t
X3
X353
t
X4
X355
_-r.
xs
X357
- t
x6
X373
t
X7
X375
t Xg
X377
Xg
X533
- t XlQ
X535
- t
X n
X537
t
X12
X553
- t
X13
X555
- t
X14
X557
t
X15
Xsn X15
X575
- t
X17
X577
t
X18
X733
t
X19
Xns
- t
X2o
Xn7 t X21
X753
t
X22
X755
- t
X23
X757
- t
X24
X773
t
X25
X775
- t
X26
X777
t
X27
Table
6.1:
State
variable conversion from
the
satellite matrix
to
sub
scripts used to describe
the
27 equations of motion.
where j k) are the positions of each mass value within
the
MATLAB cell array
and
f i
j
k) gives
the
index of
the
mass within
the
mass
matrix
, also
the
subscript
value in Table 6 .2) . f a relationship is discussed between a mass mi j k and the mass
mi, j ,k
2
directly below it,
the
index of the lower mass is
f i j k + 2) = 4.5i + 1 5j +
0.5 k
+ 2) -
18
.5
= 4.5i
+ 1 5j + 0.5k
- 17.5
=
f i
j k
+
1
6.34)
6 .35)
6.36)
indicating a net motion of + in the mass matrix index value for degree of freedom
,
j
k
+
2)
in the
downward or
k
direction within
the
MATLAB cell array model.
In
the following sections this convention will be used to developed a series of analyses
![Page 99: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/99.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 99/152
8
starting with
simplified one and
three
degree of freedom models.
n illustration of
the
satellite
matrix
is shown in
Figure
(6·15) .
0
igure
6· 5:
Two-dimensional
representation
of the satellite cell array.
Intuitively,
this
is
the
equivalent of beginning
at
the
cube
3
,
3
,
3
)
and progressing
first down
the
column of cubes in
the
positive k-direction,
then
shifting one
column
in the
positive i-direction and progressing down
that
column. Once
the fi
rst row
of
columns is complete, the process starts over
with
the
adjacent
row of columns in the
positive j-direction.
![Page 100: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/100.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 100/152
83
Th
e generalized
equation
of motion for the mass mi,j,k is
.. + C(i+1,j,k) +
C i
,j,k)
+ C i,j+ l ,k) + C(i,j
,k)
+ C i,j,k+l) +
C i,j ,k-1) .
X i,j,k)
X i, j
,k)
m i , j
,k)
C i+l, j ,k) . C(i-1
,j ,k) .
X i
+2
,j,k)
-
X i
- 2,j,k)
m i,j ,k)
m i,j,k)
C i,j
+ l ,k) . C(i,j-1,k)
.
X i,j+2,k)
-
X i
,
j -2
,k)
ffi(i
,
j,k) ffi(i,j,k)
C i,j,k+ l ) .
C(i,
j,k-1)
.
X i,j,k+2)
-
X i,j ,k-2)
ffi(i,j,k) ffi(i,j,k)
+ k i, j,k+l) + k i, j,k-1)
X i,j,k)
m i ,j,k)
k i , j
,
k+l)
X . . _
k i
,
j ,k-1)
X . .
=
Q
(t,J,k
+2
) t, ]
,
k-2)
m i,j ,k)
m i,j,k)
where
i , j ,
k = 3,
5, 7.
As stated
in
Rule (6.9), any state variables xi, j ,k where
i = 1 o r 9, j = 1 o r 9 or k = 1 or 9 then = 1
Xi,j,k
= ~ t and it is
expressed as a forcing term on
the
right side of
the
equation. Additional simplifica-
tions
ca
n be made for Ci,j,k where (i,
j
k) is equal to (2, j k), (i, 2, k
,
(i, j 2), (8, j k) ,
(i,
8,
k , (i,j, 8)-
these values
can be
simplified to a group of
outer
dampin
g values
and
separated from the inner damping values between cubes. Additionally,
the
three
dampers
installed at the
extreme
faces of each corner cube all depend on the
motion
between
the
rigid
outer structure and that
corner cube,
and can be
combined
into one value.
The
above equation can
be
re-written as
![Page 101: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/101.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 101/152
84
X(·
k)
+Ax
·
k)
,J, t,J,
+Ex ·
·k)
J
where the coefficients
are
matched with indexed with state variables according
to
Table
(6.2).
C:oethcient
---
::;tate Vanable
--- X i,i,k)
lh
--- X i+2
,j ,k)
H2
--- X i-2
,j ,
k)
G\
---
X i,i+2 k)
G 2
---
X i,.i-2,k)
l
---
X i,j,k+2)
l
---
X i,j,k-2)
.b
---
X i,.i,k)
1 1
---
X i,j,k+2)
1 2
---
X i,i,k-2)
able
6.2: Coefficients of the
i,
j, kth
equation
of motion. Since the
acceleration term always
has
a coefficient of 1,
there
is no letter as-
signed.
It
is recognized
that
while
the
coefficients
of the
velocity
and position
terms will
vary
with values of i, j,
k,
the shifting subscripts also indicate an
interdependence on
the adjacent cube
motions, i.e.,
the motion
of
cube m i,j,k)
is coupled
to
the
motion
of the two cubes above and below m i,j ,k±2) and the four laterally adjacent
cubes
m i±2,j,k)
and m
,j±2,k)·
In
addition to the
linear
Equation
(6.33) that governs
transformations
from the
![Page 102: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/102.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 102/152
85
MATLAB cell array
to the
equations of motion,
it
is also useful
to
note that
f i
± 2 j k f i
,j,
k ± 9
f i
, j ± 2 k
f i
, j, k ±
3
f i,j,
k ±
2) f i,j, k
±
1
6 9 1 Building
Mass Spring
and
Damping
Matrices
f
is the newly indexed position of variable ±i,j,k in the 27-DOF velocity vector
X.
then PA
= g ± i,j,
k
=
f i
,j,
k
and
PA 4.5i + 1.5j +
0.5k-
18.5
6.37)
This means that when X. is multiplied with the damping matrix
C
the PAth
term
of X. will multiply the PAth term of the rows of
C and
PA therefore dictates the
position of the damping coefficients within
the
rows of C:
c1 1 c1 2
c
PA
c1 21
Xl
C2,1 C2,2
C2 PA
c2 21
X2
X
6.38)
XpA
C27,27
X27
In this way the coefficient A which, as above, is multiplied by velocity ±i,j,k) is
appropriately indexed to the PAth position in the rows of
C.
Similarly, the positions
of the remaining damping coefficients are
![Page 103: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/103.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 103/152
86
Ps
1 ,2
=
g i: i±2,j
,k))
=
f i,j,
k
±
9
=
4.5i 1.5i
0.
k
18.5
±
9
Pc
1
,
2
=
g i:
(i ,
j±2
,k)) = f i
, j k ± 3
=
4.5i + 1.5i + 0. k 18.5 ± 3
l 2 = g i: i ,j,k±2)) = f i , j , k
±
1 = 4.5i 1.5i 0.5k -18.5
±
1
and the
position of each of the stiffness terms are
P
= g X
i
j k))
=
f i
,
j
k
=
4.5i 1.5i
0.5k-
18.5
1
,
2
= g x i ,j ,k±2)) = f i, j
, k ±
1 = 4.5i
+
1.5i
+ 0. k 18
.5
±
1
Therefore, one may build both
the
stiffness and damping matrices K
and
C by
placing the coefficients A through in the
Pth
positions along each row.
6 10 Initial Results
Created using the MATLAB scripts included in Appendix ??), Figures (6·17) (6·25)
show
the
frequency response for each of 27 payloads. The plots shown here
are based
on a model that assumes the following characteristics:
1.
A 3U cube is placed
at
the center of the satellite matrix . This is accomplished
by
fusing the central column of cubes by increasing the spring constants between
them
by
a factor of 100 (line 145) .
2. Drumheading of the otherwise rigid exoskeleton top plate is assumed
to
play
a role, so spring constant values at the top and bottom plate are lowered by
just over an order of magnitude (lines 163, 170).
The
value used is based on a
SolidWorks Finite Element Analysis of
the
top
plate
under 1N loading
at
the
![Page 104: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/104.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 104/152
87
center. Hooke's law was used to solve for
the
equivalent spring const
an
t using
the displacement from
the FEA
plot (Figure 6·16).
t.co4
Figure
6·16:
SolidWorks
FEA
displacement analysis of Top
Plate
3. 10 random noise is added to each ofthe values
in
the mass, spring and damping
matrices. Because the model is still very
symmetric
,
this
helps to
simulate
a
more realistic response.
4. Light damping from ( = 0 to ( = 0.3 is considered (line 316).
5.
Redundant
corner
dampers
are eliminated. For example,
the cube
at
position
(3,3,3) had three faces with dampers connecting to the cube, but is now reduced
to
just
one (lines 344, 354).
![Page 105: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/105.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 105/152
88
x1
:
g)
'
10
2
10
3
10
10
5
10.
Frequency (Hz)
x,
10
:
'
)
'
10
2
10
10
10
5
10.
Frequency
(Hz)
x3
10°
:
Q)
'
10
10
3
10
10
5
10.
Frequency (Hz)
igure
6 · 7:
Acceleration response for masses x
1
, x
2
and x
3
. Sub-
scripted mass notation can be converted using Table 6.2
10
2
10
3
10
Frequency (Hz)
10
3
10
Frequency (Hz)
x
10
5
10.
F i
gure
6 · 8: Accel
erat
ion response for masses x
4
,
and
6
•
Sub-
scripted mass notation can be converted using Table 6.2
![Page 106: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/106.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 106/152
89
x7
rf
10
°
:
<=
8
0::
10
2
10
3
10
10
10
8
Frequency Hz}
xe
Q c
10°
:
<=
8
0::
10
2
10
Frequency Hz}
Xg
rf
100
:
<=
8
l
:
:
0::
10
2
10
3
10
10
5
10
8
Frequency Hz}
i ure 6·19: Acceleration response for masses
x
7
x
8
and
x
9
•
Sub-
scripted mass notation
can be
converted using T
ab
le 6 2
10
2
10
3
10
Frequency Hz}
x
10
Frequency Hz}
x12
10
3
10
Frequency Hz}
10
5
10
8
10
10
8
i ur
e 6 ·2 : Acceleration response for masses x
10
x and x
12
Sub-
scripted mass
notation
can be converted using Table 6 2
![Page 107: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/107.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 107/152
10
2
10
2
9
10
3
10
4
Frequency (Hz)
x
Frequency (Hz)
x s
10
5
10°
10
5
10°
J
·rt
rllll \·
lr I
, 1
tft
11
10
2
10
3
10
4
10
5
10
6
Frequency
(Hz)
Figure
6·2 :
Acceleration response for masses
xi
3
XI
4
and xi
5
. Sub-
scripted mass notation
can be
converted using Table 6.2
10
2
10
2
10
3
10
4
Frequency (Hz)
x17
10
3
10
4
Frequency (Hz)
x s
1
3
1
4
Frequency (Hz)
10
5
10°
10°
10°
Figure 6·22: Acceleration response for masses xi
6
X 7
and
XIS · Sub-
scripted mass notation can
be
converted using Table 6.2
![Page 108: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/108.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 108/152
91
x g
r
;
i
j
~ i
0
2
10
10
4
10
5
10
6
Frequency (Hz)
x2o
10
8.
ill
0::
10
2
10
10
5
10
6
Frequency (Hz)
21
10°
8.
ill
0::
10
2
10
3
10
4
10
5
10°
Frequency (Hz)
Figure 6· 3: Acceleration response for masses x
19
, x
20
and x
21
. Sub-
scripted mass notation
can
be converted using Table 6.2
10°
::
10
2
10
10
4
10
5
10
6
Frequency (Hz)
23
10
ll
.
ill
0::
10
4
10°
Frequency (Hz)
x24
r
10
2
10
10
4
10
10
Frequency (Hz)
Figure
6· 4:
Acceleration response for masses
x
22
x
23
and
x
24
. Sub-
scrip
ted
mass notation can
be
conver
ted
using Tab le 6.2
![Page 109: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/109.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 109/152
9
~
i • i l
'' ' '
: : : :
: : : : : : : ::
: : : : :
:
: : :
: : : : : : : :
:
J
l
p:;
r
:··
;,rrw
·:
:··¥ .·p,:'
vm
U L W ~
10
2
10
2
10
3
10'
Frequency (Hz)
:ze
10'
Frequency (Hz)
x27
10
3
10'
Frequency (Hz)
10
5
10°
Figure 6·25: Acceleration response for masses x
25
, x
26
and x
27
. Sub
scripted mass
notation
can be converted using Table 6.2
6 11 Discussion
t is noted that of the 27 acceleration response plots in Section (6.10), only x
13
, x
14
and
x
15
resemble what is ordinarily
thought
of as a mass-spring-damper oscillation,
with
the quality of the response smoothing
out
as damping is increased. The re
maining 4 masses also display a smoothing characteristic,
but
instead of smoothing
down from the sharp,
undamped
peak, they
smooth
up and
the
magnitude of the
damped response approaches the magnitude of the undamped response
at the
reso
nance frequency as damping is increased. This seemingly
odd
behaviour is
illustrated
by Figures (6·26) and (6·27).
![Page 110: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/110.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 110/152
93
10 ' 10
4
Frequency Hz)
igure
6· 6:
Acceleration response for mass x
1
3, located
in
the center
of the top layer of the satellite cube stack.
9
rl
10
2
10
3
10
4
10.
1
Frequency (Hz)
igure 6· 7:
Acceleration response for mass x
1
9, located
at
a corner
of
the
top layer of
the
satellite cube stack.
While this effect is
not
intuitive, it is entirely the result of the system s depen-
dence on dampers as part of the forcing term (see Section 6.5) .
In the
analytical
model presented here, mass x
13
has fewer dampers
attached
to
it
than mass x
19
and
is therefore acts more like a normal oscillatory system. This can
be
verified by
increasing the value of the forcing damper constant in the model shown
in
Figure
(6·4) by two orders of magnitude. The resulting plot is shown n Figure (6·28).
![Page 111: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/111.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 111/152
9
Single
DOF
model
~ ~ ~ ~ r ~
4
- : - - - - - - - ~ - = - - - - - - - - _ _ _ _ . . _ - = - - _ . _ _ ____ -- -::-----------
Frequency
Hz)
igure
6 28:
Single OF system experiencing variable damped forc
ing.
![Page 112: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/112.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 112/152
9
Chapter
Results
and Conclusion
7 1 Design
Developed under a requirements-based design method, the BUSAT structure is able
to respond
to
almost all of
the
structural, scientific
and
logistical ideas
that
were
introduced in
the
first BUSAT University Nanosat
Program
competition (Nanosat-5,
2007
to
2009). This is remarkable, considering
the project s
100% personnel turnover
rate and low budget ( 55,000
per
year in 2007, 2008, 2011 and 2012). In addition, if
BUSAT continues
to be
successful in
the
University Nanosat
Program
and beyond,
the concept of
an
aggregate
and
containerized nanosatellite will carry a huge impact
on
the
current scientific spaceflight industry.
The structural
system's
potential
cost
savings alone will ensure this. However,
the
development documented here is only
an
early portion of
the
potential life cycle of
the
BUSAT modular bus as an
industry
game-changer;
there
is much work to be done
in
qualifying with a first successful
launch.
7 2 Testing
Not only did the vibration testing performed
at
Draper Laboratories in October,
2011 prove that BUSAT's novel fastening technique works well, it also served as a
good
starting
point for development of a standardized , modular
testing
protocols for
potential payloads. Because strict controls must be placed at the interfaces between
![Page 113: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/113.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 113/152
9
bus and subsystems
in
the modular BUSAT system, particular care must be taken
with the mechanical interfaces to see that they will satisfy the
minimum
needs for
both
sections of the satellite.
This
includes predicting the inner
cube vibration,
radiation
and
magnetic enviroments so that potential payload providers will know
what to expect on launch
and
on orbit .
Future
testing
will necessarily include electronic
hardware
, or at least
mass
models,
as well as thermal
and electromagnetic environmental procedures
to
verify structural
performance. Due to the sensitivity of
the
system to changes in temperature, im
proved monitoring, control and modeling is necesary to ensure that any temperature-
dependent frequency shift can
be
predicted
and
removed from results.
Th
is will
ensure that marginal
test
results do not result in the need for a complete redesign of
the structure.
7 3 Tuned 27
Degree
of Freedom Model and
Comparison
to
Vibration Test Data
The
first significant result of
the
27-mass model is
that
the method
of forming equa
tions
of motion is verified. Based on the results of both
the
simplified and the final
models , the systems behave qualitatively as they should
in
the frequency response
plots
and the remaining work to be done is
in
tuning and refining
the mo
del itself.
Second, while initial simple models suggested
that
the spring coefficients were far
more important
than
damping values
in
tuning the
system
, the 27-mass model high
lights an interesting feature of the dampers. Probably because of the sheer number
of
dampers
surrounding
the
cube stack (54),
the
low-frequency peak seen
in
Figure
(6·17) ris s to
meet
the lightly damped resonance,
smoothing
it. From this it can
be
hypothesized that BUSAT s dependence on a friction force oriented in t he vertical
direction for fastening
the
WedgeLok ™) could provide an
inherent tolerance to
![Page 114: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/114.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 114/152
97
high-energy environments and may be more appropriate for aerospace applications
than
standard threaded fasteners .
The fact that the model s frequency response begins
to
approach actual data
when
the
top layer of springs is weakened suggests
that
as-built ,
the
model may need
improvement. n its current configuration,
the
model
can
only approximate the effect
of the exoskeleton top
plate
resonance feeding back into the cube stack, and a more
thorough
look
at plate
bending under vibration is
warranted
.
![Page 115: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/115.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 115/152
Appendix A
MATLAB Script for 1-DOF
Model
clc
clear
close
al l
Specify
di rec tory
to which M-f
i l
e and
subsequent
p l ots wi l l
save
dir = 'C: \Users \Nate\Documents \newschool \1000 docs \1200
c l a s s e s \
;
subdir = Thes i s \ Thesis_Finals \10
_ModeLBuilding\
' ;
subsubdir
= '100F\ ' ;
cd
( s t r ca t (
dir
,
subd
i r , subsubdi r
) ) ;
Vector of colors for p lo t t ing
co l o r s= [
'b
' , ' g ' , r ' , 'c ' , 'm' , 'y ' , 'k ' , 'b: ' ' g: ' r : ' , 'c : ' , 'rn: ' , y: ' , k : '
, g ,
r ' , ' c ' , m ,
'y
' ,
k
) ' b :
)
,
g
' , ' r ' , ' c ' , m , y
' ,
'k ) )
'b :
)
' b ' , '
g
, ,
, m ,
' y) '
'k ) )
'b :
)
s
=
1 ·
for
stepping
thro
u gh
Vo
Spring
Constant Calcula t ion
cubewidth = . 1 ;
c u
be
lengt h = . 1 ;
wal l th ickness
= . 1 25*.0254 ;
A
L
E
4*cubewidth*wal l th ickness ;
c u
bewidth;
69*10 A9 ;
Set
mass and
spring
charac te r i s t i c s
m3
k2
k4
;
A*E /L;
A*E /L;
98
'g :
)
,
r :
'g :
)
, r :
'
g:
)
, r :
colors
o/cm
n
)
)
)
) )
c:
)
, m:
)
'y :
)
c:
)
, 'm:
)
' y :
)c:
)
, 'm: 'y :
inches
* .0 2 54
m/ inch
)
k:
)
)
'
k :
)
)
k:
) ;
![Page 116: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/116.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 116/152
K =
k2
+ k4);
M= m3;
%
Find
natural
frequency
of system
omega_n = sqr t K/m3);
l ightly_damped
=
0 . 001;
cri t ical ly _damped
=
1;
%Set
ve ctor of frequencies
for
plot
domain
omega = l inspace
1
,2000000 ,1000);
%In i t ia l ize
Magnitude
Response Vector
X_p
=
zeros
1,1000);
Wo KNOBS AND BUTIDNS
%
Scal ing factor and
s
witch for second-order forcing A omega)
scal ing
=
1;
%Pr in t
Button
ppr int =
t rue;
ggrid =
fa lse ;
box = fa lse ;
ax
=
fa lse ;
omegalines t rue;
%fo r - lo o p t o
i t e ra te
plots over in teres t ing
damping
values
for zeta = l ightly _damped:
.1 :
cr i t ical ly
_damped
%Set d amping va lues
c2
=
0.5* 2* zeta*omega_n);
c4
= 0.5* 2* zeta*omega_n);
C
= (
c2
+
c4);
o fo
r - loop
to co
n
t ruct
t
he vecto
r
of ma g nitude response for the
range
%
of fr
e
qu
e
ncies
s
pecif ied
.
Note tha t
t
he forcing function
is
dep
endent
fr e quency.
for l = l length omega)
![Page 117: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/117.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 117/152
1
X_p(l) =
1/(-omega(l)h2*M+
1i*omega(l)*C +K)* · · ·
scal ing*
. . .
end
omega(l) h
2*(0.25/rrili)*
sqr t
(((k2+k4)h2/omega(l)h4
+
1000*(c2
+ c4)
2/
omega
( l ) 2) ) ;
end
Plot
log log (omega, abs (X_p) , spr in tf (colors ( s ) ) )
hold
on
s
=
s+1;
Finish Plot
i f ax
axis
([20,
2000,
.001, 25])
end
y_lim
x_lim
get
(
gca
, 'YLim ' );
get (
gca
, 'XLim ' );
if
omegalines
l ine (omega_n*[1 1J,y_lim, ' Color ' , ' k ' )
l ine (x_lim,10h0*[1
1],
' Color ' , '
k
)
end
t i t le ( ' Single OOF
model
' )
ylabel
( ' R e
sponse Magnitude
(m/ s h2) ' )
x l a bel ( ' Frequency (Hz) ' )
Block out area where we want the plot to appear
if box
end
l
ine (20*[1
1] ,y _lim , '
Color
, ' k ' )
l ine
(x_lim,10h-3*[1
1], Color ' , ' k ' )
n e ( 2 0 0 0 * [ 1 1 J , y _lim , ' Color ' , ' k ' )
l ine (x_lim,25*[1
1],
C
olor ' , ' k ' )
if ggrid
grid on
end
![Page 118: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/118.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 118/152
1 1
hold
off
if
pprint
Change
to directory for saving plots
foldername = s t rcat
( ' plo t
s\
, datest r now,
30))
mkdir (s t rcat ( dir , subdir ,subsubdir , foldername)) ;
cd
(
s t r ca t
(
dir
, subdir , subsubdir ,
foldername));
fi lename
s t rca t ( ' fi g _l ' )
pr in t (
- pa
i
nt
e
rs
' , dpng ' , ' -
r600
' , filename)
end
![Page 119: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/119.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 119/152
Appendix
B
MATLAB Script for 3 DOF Model
cl
ea
r
close al
l
clc
Specify directory
to which
M-file and subsequent plots will
save
dir =
C:
\Users \Nate\Documents\newschool \1000 docs \1200 classes
;
subdir
= '
Thesis\
Thesis_Finals
\10
_ModeLBuilding\ ' ;
subsubdir = ' 300F\ ' ;
cd ( s t rca t ( dir , subd ir , subsubdir)) ;
Vector of
co lors for plot t ing
and
color increment counter
colors=[
b ,
'
g
'
r
' '
m ,
y ' ' k ' '
'
b:
g:
' r :
c:
'm:
y:
b '
g '
r
c '
m ,
y ' ' k ' '
' b:
g:
' r:
c:
'm:
y:
k:
b '
g '
r ' '
c
'
m
,
y ' k '
'
b:
g : r :
c:
'm:
y :
k:
k
b , ' g , r ' , c ' , m , y ,
k
, b: ' g: ' r : '
c:
' , 'm: ' y: '
k:
' ] ;
s= O
;
S p r i n g Constant
Calculat ion
cubewidth
= .
1;
cu belength = . 1 ;
wa ll
th ickn e
ss
= .
125*.0254;
A =
4*cubewidth*wa ll
t
hi
c k
ness
;
L
E
cubewidth;
1 2
o/<1 11
n
i
nch
es
* .0 2
54
m/
inch
o/111
o c N m
2
![Page 120: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/120.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 120/152
1 3
%Set
in i t i a l values for
mass,
springs
m = ;
k
=
A*E/1 ;
m3 =
5
j
m5 = 3*m;
m7
=
m;
M=
[
m3,
0 ,
0
k2
k·
k4
k·
k6
k·
k8
k·
K=
[
k2+k4,
-k4
0,
0 )
0
m5, 0
0 ,
m7];
-k4 , O·
k4+k6,
-k6;
-k6
k6+k8
%kilogram
<J{Njm
];
%Find
eigenva
lues and
eigenvectors
for system
[Phi , Omega] = eig (K, M);
omega_n
=
sqrt
( d i
ag (Omega));
%Define
size
of
system (OOF)
and
number of
data
points
OOF =
length
( diag
M)) ;
X_p = zeros
(OOF,5000)
;
counter = 1;
%
Begin
for -
loop for plot t ing over
a range of
damping values
for z e t a=
0 .
001: .1 :1;
%
Construct damping matrix
based
on
range
of damping ra t ios in for
- loop
c2 =
2*zeta*
sqr t (k*m);
c4 2*zeta* sqr t (k*m) ;
c6
2*zeta* sqr t (k
*m) ;
c8
2*zeta* sqr t
(k*m);
![Page 121: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/121.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 121/152
104
C=
[ c2+c4,
- c4 ,
O·
-c4
,
c4+c6,
-c6;
0,
-c6 , c6+c8
l
%
In
c r e m en t
c olor counter for th i s
plot
s
= s +
1 ·
%Set up fr e
qu
ency range of i n t e res t for plo t t ing
omega = l inspace (1
,200000
,5000);
% fo r -lo o p to cons truct the
vector
of magnitude response for the rang e
%s p eci fied. Note
tha t
the forcing funct ion
is
dependent on fr e quency
for r = 1: length (omega)
X_p(: , r) = inv ((-omega(r) 2 ) * M +
1i*omega(r)*C
+ K) * · · ·
omega(r)
2 * [ c 2 / o m e g a r )
+ k2/omega(r)
(forcing function)
0;
end
c6/omega(r) + k 6 / o m e g a r ) ~ 2 ]*9.8*0.25; %( forcing
funct ion
%
i f
sta t ement to allow figure set up only once in this loop
i f counter = 1
scrsz
=
get
(O,
'ScreenSize
' ) ;
f igure
( ' Pos i t i o n '
,[5
5
scrsz (3) *.995 scrsz (4) *.9])
hold
on
counter =
counter
+ 1;
end
%P 1o t f i r s t OOF
sub p1o t (OOF, 1 , 1 )
z=1;
loglog
(omega,
ab s
(X_p(z
, : ) ) ,
spr in t f
(co lors (s ) ) )
t i t l e
s t r c a t s p r i n t f
OOF), 'OOF System
in
Rigid Enclosure ' ))
ylabel
( '
Response
m_3 (m/ s ' )
xlabel ( 'Freq uency (Hz) ' )
1 n e ( [ omega_n (
1)
, omega_n (
1)
] , [ 1 0 - 15 , 1 0 5])
text ( .9*omega_n(1), , ' \
omega
_{nl} ' , 'Color ' ,
'k ' ) ;
l ine ( [
omega_n (2)
,omega_n (2)], [10 - 15 ,10 5])
text ( .9*omega_n(2), 1 0 ~ 6 , ' \omega_{n2} ' , ' Color ' , ' k ' ) ;
l ine ( [
omega_n
(3) ,omega_n (3)], -15 ,10 5])
![Page 122: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/122.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 122/152
105
text
( . 9*omega_n(3), 10 ' 6 , om ega_{n3} ' ,
axis
([10
' 3 10 ' 5 10 ' -2 10 ' 2])
hold
on
C olor ' ' k ' ) ·
Plot
second OOF
subplot (OOF, 1 , 2)
z=2;
end
loglog (omega , ab s (X_p(z , : ) ) , spr in t f (co lors (s ) ) )
t i t l e
(
st rcat
( spr in t f ( 'o/;< ' ,
OOF),
'OOF System in Rigid E
ncl
osure ) )
ylabel
( ' Response m_5 (m/s A2) ' )
xlabel ( '
Fr
equenc y (Hz)
' )
1
n
e ( [omega_n ( 1) , omega_n ( 1) ] , [ 1 0 ' -15 ,1 0 A ])
text
(.9*omega_n(1), 10 ' 6 , om ega _{n1} ' , ' Color ' , ' k ' ) ;
l ine ( [omega_n
(2)
,omega_n
(2)],
[10 ' -15 ,10 ' 5])
text
(.9*omega_n(2), 10 ' 6 , '
\omega
_{n2} ' , '
Color
' ,
k ;
l ine ( [omega_n (3)
,omega_n
(3)] , [10 ' -15 ,10 ' 5])
tex t (.9*omega_n(3), 10 ' 6 ,
om
e
ga_{n3}
' , ' Color ' , k ) ;
axis ([10 A3 10 ' 5 10 ' -2 10 ' 2])
hold on
Plot
third
OOF
subplot
(OOF, 1 ,3)
z=3;
1o g 1o g (omega '· a bs ( X_p ( z , : ) ) , s p r i n t f ( co 1or s ( s ) ) )
t i t l e
( s t rca t ( spr in t f
(
o
;<
' ,
OOF),
'OOF Sy
st
em
in Rigid
Enclosure
'
))
ylabel
( ' Re sponse m_7 (m/ s A2) ' )
xlabel
( '
Frequ
ency (Hz) ' )
1 n
e ( [omega_n (
1)
, omega_n (1) ] , [ 1 0 • -
15
, 1 0 '
5])
text
(.9*omega_n(1),
10 ' 6 ,
\ om
ega-{n1} ' , C olor ' ,
' k '
);
l ine ( [omega_n
(2)
, omega_n (2)] , [10 ' -15 ,10 ' 5])
text
( . 9 * omega_n(2) , 10 ' 6 , ' \o m ga_ {n2 } ' , Co lor , ' k ' );
1
n e ( [omega_n ( 3) , omega_n ( 3) ] , [10 ' -15 ,10 A5])
tex t
(.9*omega_n(3), 10 ' 6 , ' \ om ga _{n3} ,
C
lor ' , k ) ;
a xis
([10
' 3 10 ' 5 10 '
-2
10 A
2])
hold
on
hold
off
pngname=
s p r in t f ( ' 300FBOXED ' ) ;
![Page 123: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/123.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 123/152
1 6
pr n t
(
-dpng
, -
r500
,
pngname)
![Page 124: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/124.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 124/152
p p ~ n d i x
MATLAB
Script
for 9 DOF Model
clea
r
close
a ll
c l c
<7<
LOGISTICS
Specify directory to which M-file
and
subsequent plots will save
d i r = 'C:\ Users\Nate\Documents\newschool\1000 docs\1200
classes
\ ' ;
subdir =
'
Th es is
\ Thesis_Fina l s \10
_ModeLBuilding\
' ;
subsubdir =
'9DOF
Coupled\ '
;
cd
( s t r ca t (
dir
, subdir , subsubdir ) ) ;
Vector
of
colors
for
plot t ing
and
co
l
or increment
co
u
nter
colors = [
' b ' , ' g ' , ' r ' , ' c ' ,
'm'
, ' y ' , ' k ' , ' b : ' , ' g : ' , ' r : ' , ' c : ' 'm: ' , 'y : ' , ' k
.
' b ,
g
, r ' ,
'c
' , 'm ' ,
y
, 'k ,
'b:
' 'g: ' ' r : ' , c: ' , 'm: ' y : ' ' k : '
b , g , ' r ' , 'c ' , 'm ' , y
,
' k ' , ' b: ' , 'g: ' r : ' , c : ' , 'm : ' y
: '
' k : '
' b ' ,
g ,
' r ' , c , 'm ' , y
,
' k ' , ' b: '
g: '
'
r:
' 'c : ' , 'm: ' ,
y: '
'
k:
' ];
s=O;
< < Spring Constant Calculat ion
cubewidth
= .
1 ;
cu belength = . 1;
wal l th ickness =
.125*.0254 ;
A =
4*cubewidth*wall thickness;
L
E
cubewidth;
69*10
A9;
107
11
o m
inches*
.0254
m/ inch
nA2
n
<Y<N m
A
![Page 125: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/125.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 125/152
108
<71 UNDAMP ED SYSTEM
Set
v alues
for mass, springs
m=
1·
k = A*E/L;
m33 = m;
m35 =
hm;
m37 =
hm;
m53 =
hm;
m55 =
hm;
m57 = m;
m73 = l*m;
m75 = l*m;
m77 = m;
M3=
[
m33,
0, 0
0 ,
m35 ,
0
0 0,
m37];
M5=
[
m53 ,
0,
0
0,
m55,
0
0 0 ,
m57] ;
M7=
[
m73 ,
0 ,
0
0,
m75,
0
0 0,
m77];
M = [M3, zeros (3
,3) ,
zeros (3
,3) ;
zeros
3 ,3) ,
M5
, zeros (3 ,3) ;
zeros
3 ,3) ,
zeros
3 ,3) ,
M7];
k32
k·
'
k34
k·
k36
=
k·
'
k38
k·
k52
k·
k54
k·
'
k56
= k ;
kilogram
Njm
![Page 126: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/126.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 126/152
1 9
k58 = k ;
k72 =
k;
k74 = k;
k76
=
k;
k78
=
k;
K3 =
[
k32+k34 '
-k34
-k34
k34+k36,
0,
-k36,
K5 =
[
k52+k54 '
-k54
-k54
,
k54+k56,
0,
-k56 ,
K7 =
[
k72+k74
'
-k74
,
-k74
k74+k76,
0,
-k76,
K = [K3 , zeros 3 ,
3)
, zeros 3
,3);
zero
s ( 3 , 3 ) , K5 , zeros ( 3 ,3 ) ;
zeros 3 , 3) , zeros 3 , 3) , K7] ;
0·
'
-k36 ;
k36+k38
O·
'
-k56;
k56+k58
O·
-k76;
k76+k78
%
Find eigenvalues and eigenvectors for
system
[Phi ,
Omega] =
eig
K, M);
omega_n
=
sqrt
(
diag
Omega)
) ;
DOF =
length
(
omega_n)
;
X_p = zeros (OOF,5000) ;
scrncounter
= 1 ;
%Vector
and
constants to change plot locat ions
vee
= [ 1 , 2 , 3 , 1 , 2 , 3 , 1 , 2 , 3] ;
plot1 1;
plot2
=
3;
DAMPED SYSTEM
l
l ;
l
%
Begin for
-
loop for
plot t ing
over
a range of
damping values
for z e t a =
0.001: .1:1;
![Page 127: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/127.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 127/152
110
%Construct damping matr ix based
on
range of damp i
ng
ra t ios in for - loop
c32
2*Zeta
*S
qr t
(hm) ;
c34
2 * z e t a *
s q r t k ~ ) ;
c36
2*zeta*
sqr t k*m);
c38 2 * z e t a * s q r t k ~ ) ;
c52
2*zeta* sqr t k*m)
;
c54
2*zeta*
sqr t
k*m) ;
c56 2*zeta*
sqr t k*m);
c58 2*zeta*
sqr t k*m);
c72 2*Zeta*
sqr t
(hm) ;
c74
2*zeta*
sqr t k*m);
c76
2*Zeta*
sqr t k*m);
c78
2*
zeta*
sqr t
k*m)
;
C3
=
[ c32+c34 ,
-
c34
,
-
c34,
c34
+
c36,
0 ,
-c36,
C5
[ c52+c54
,
-
c54
,
-
c54,
c54
+
c56,
0 ,
-c56
,
C7
=
[
c72
+
c74,
-c74
,
-c74 ,
c74+c76 ,
0 ,
- c76 ,
C = [C3 , zeros
3
,3) ,
zeros 3
, 3) ;
zeros 3
,
3) ,
C5 , zeros
3
,3);
zeros
( 3 ,3) ,
zeros
( 3 ,3) , C7] ;
O·
-c36;
c36
+
c38
O·
- c5 6 ;
c56+c58
O·
-
c7
6 ;
c76+c78
%
Increment co
lor counter for th is
plot
s
=
s + ;
l
j
l
l j
%
Set up frequency range of in te res t for plot t ing
omega
= l inspace ( 1 , 200000
,5
000);
%
for
-
loop to
const ruc t
the
vector
of magnitude response for the range
%speci f ied . Note that
the
forcing funct ion
is
dependent
on f requency
![Page 128: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/128.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 128/152
111
for r =
1:
length (omega)
end
X_p(: , r ) = inv ((-omega(r) A2)*M+ l i*omega( r )*C+K)*· · ·
[ c32 /
omega(
r) + k32;
(forcing function)
0 ;
c36
/ omega(r) +
k36
;
c52 / omega(r) + k52;
(forcing
funct ion)
0 ;
c56 / omega(r) +
k56
;
c72/omega(
r)
+
k72; (forc
i ng function)
0;
c76/omega(r) + k76 ]*9.8*0.25; (forcing function
%Create f i rs t figure
only
if
s crncounter == 1
end
s c rsz
=
get (0 , ' ScreenSize ' ) ;
f igure
( '
Posi t ion
' , [5 5 scrsz ( 3)*.995 scrsz(4)* .9])
hold on
scrncounter
=
4 ;
%Create three ver t i ca l ly oriented plots
within
each figure
for dl = 1 : 3
end
f igure
(1 ) ;
hold
on
subplot
(3 , l , d l )
loglog
(omega , abs (X_p(dl ,: ) ) ,
spr in t f
( color s ( s ) ) )
t i t
e ( s t
rca
( spr in t f ( o/crl ' , d 1 ) , ' { } ' , ' of ' , ' { } ' , . . .
sprint£ ( 'o/crl' , DOF), 'DOF System
in
Rigid Enclo su re ' ))
ylabel
( '
Response (m/sA2)
' )
xlabel
( F
requency
(Hz)
' )
axis ([10 A3 , 10 A5 , 10 -5 , 10 ' 5])
l ine
( [
omega_n(
dl)
,
omega_n
(
dl )]
,
[lO
A
-15
,
lO
A
5])
t ex t (homega_n(dl) , 10 ' -16 , s t rca t ( ' \omega' , spr in t f ( o/crl ' ,d l ) ) )
grid on
hold
on
hold off
![Page 129: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/129.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 129/152
112
Create
second
figure
only
if scrncounter 4
end
scrsz = get (0
, '
Screen Size
' ) ;
f igure
( '
Position
' , [5 5
scrsz (3) *.995 scrsz
(
4) *. 9])
hold on
scrncounter =
7;
Cre
ate
three
ver t i ca l ly
oriented plots
within each
figure
for
d2
= 4 :6
end
f igure
(2) ;
hold on
subplot
(3 ,1
,d2-3)
loglog
(omega
,
ab s
(X_p(d2,:)) , spr in t f (co lors (s ) ) )
t i t
e ( s t
rca
t (
spr in t
f ( 'o/
rl
' ,
d2)
, ' { } ' , '
of
' , ' { } ' ,
. . .
spr in t f ( 'o/
rl
' ,
OOF),
'OOF System in Rigid Enclosure
))
ylabel
(
Response
(m/s ' 2) ' )
xlabel
( '
Fr
equen cy
(Hz )
)
axis ([10 ' 3 , 10 A5, lO A-5 , 10 A
5])
n
e ( [
omega_n
(
d2)
,
omega_n
(
d2)
] , [ 1 0 A-
15
,1 0 A
5])
text
(homega_n(d2) ,
10 A
-16
,
s t r ca t ( omega
' , spr in t f ( 'o/
rl
'
,d2)))
grid on
hold on
hold off
Cre
at
e third figure only
if
scrncounter
7
end
scrsz
=
get (0,
' S c r
ee
nS i
ze
'
) ;
f igure
( ' Po s i t ion ' ,[5 5
scrsz(3)*
.
995
scrsz(4)*.9])
hold on
scrncounter =
8;
Create three
ver t i ca l ly or iented
plots within each f igure
![Page 130: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/130.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 130/152
e
nd
113
for d3 7:9
end
figure
(3);
hold on
subplot (3 ,1 ,d3-6)
loglog
(omega,
ab s (X_p(d3,:)) , spr in t f (co lors (s ) ) )
t i t l e (s t rca t ( spr in t f ( 'o/crl' ,
d3),
' { } ' , ' of ' , '
{}
,
. . .
spr in t f ( 'o/
crl
' ,
OOF),
'OOF System in
Rigid
Enclosure ))
ylabel ( Respons e (m/ s ' 2) ' )
xlabel
( ' Frequency
(Hz)
' )
axis ([10
'
3 , 10
'
5, 10 -5 , 10 5))
line ( [omega_n (
d3)
, omega_n (
d3)
] , [ 1 0 ' -15 , 1
0 5])
t ext
(homega_n(d3) ,
10 -16 ,
s t rca t ( \
omega ' , spr in t f ( 'o/
crl
' ,d3)))
g rid
on
hold on
hold
off
hold
off
o/c
all f igures to t ime-stamped folder called PLOTS
%
Change to
directory for
saving plots
foldername
= s t rca t
( '
plots
\ ,
dates t r (now, 30))
mkdir ( s t rca t ( dir , subdir , subsubdir ,
foldername)) ;
cd ( s t rca t ( dir , subdir ,
subsubdir
,
foldername));
for f 1:3
fi g
ur
e
(f)
fi lename s t rca t ( ' fig_ ' , spr in t£ ( 'o
crl
' ,
f ) ) ;
(£,
'- p a in t e r
s
, ' -
dp n
g ' , '- r
600
' ,
fi lename)
end
pngname=
s print f ( '
DOFCOUPLED__I
- 3 ' ) ;
pr int ( '-£1 ' , '-dpng , r500 ' ,
pngname)
![Page 131: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/131.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 131/152
114
pngname=
s pr in t
f
' 900FCOUPLEDA-6 ' ) ;
pr int ( ' -
f2
,
dpng
' ,
r500
' ,
pngname
pngname=
s
pr in t f
' 900FCOUPLED_7- 9 ' ) ;
pr in t ( ' - f3 ' , ' - dpng ' , ' _:_r500' , pngname
![Page 132: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/132.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 132/152
Appendix D
MATLAB Script
for
27-DOF Model
clear
close
all
clc
L O G I S T I S
Specify
directory
to wh ich vl
fil
e
and
subsequent p l ots
will
save
dir = C : \Users
\Nate
\Documents\newschool \1000 docs \1200
classes
;
subdir = ' Thesis \ Thesis_Finals
\10
_ModeLBuilding\ ' ;
su
bsu
bdir = '
2700F
Coupled\ ;
cd ( s trcat ( dir , subdir , subsubdir)) ;
colors=[
c , ' k ,
' b ,
c
,
b
,
' b ' , c , b ,
b ,
c
, ' b ' ,
b ,
c
, , c , b , c , , c , ,
c
, , c
c
,
b c
b , ' c , b ,
c
, b , ' c ' , ' b , ' c '·
.
c ' b c ,
' b, ,
c , ,
'
c
'
b
' , c , b , c
.
c ,
b , c
,
'
b,
,
c
,
b
,
c b ' c ' ' b ' c ]·
s=O;
Recover
screen size
scrsz = get
(0,
Scr
nSiz '
);
Cr
e
ate
27
f igures
with
which
to
plot
27 DOFs
for
figz = 1:10
f i g ur e (
Pos i t ion
, [5 5 scrsz (3) * .995 scrsz (4) *.
9])
end
Cr eat e switch
for
figure plot t ing
f igswitch
=
1;
115
...
![Page 133: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/133.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 133/152
%
Set
log-log plot axes
xmin
=
20;
xmax
=
1000000
;
ymin
= .0 0 0 0 1 ;
ymax = 25;
%
Set
log-log plot
axes
% xmin
=
20;
%
xmax =
2000;
% ymin = 10· 2 ;
% ymax = 4 0;
116
%Create
Satel l i te
Structura l Matrix
L = ce ll
(10
,10 ,10);
Vo
INITIALIZE
ALL
MATRICES USED IN SCRIPT
M= zeros
(27);
K = zeros (27);
C = zeros (27);
F_l
zeros (27, 1)
; %
Forcing through
dampers
F_2
=
zeros (27,
1);
%
Forcing through
springs
Vo
SPRING
CONST NT
CALCULATION BASED ON AREA
AND
LENGTH
OF
CUBES
%
cubewidth =
.1;
o
on
%
c ubelength
=
.1;
o
on
%
wa ll
t hickness
=
. 125* .
0254·
%inches* .0254
m/i n ch
%
%
AA = 4*Cubewidth*wall thickness;
%n · 2
%
1
=
cubewiclth ·
%u
%
E =
69*10
. 9; m ;nt 2
%
%
Set
values
for mass sp
rings
% m =
1·
%kilogram
%
k_s
= AA*E/1 ;
m ;
m
![Page 134: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/134.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 134/152
117
10 SPRING OONSI'ANr CALCULATION BASED ON AREA
AND
LENGIH OF SPACERS
spacerwidth
= .01;
spacerlength = .01;
spacer thick = .
01;
AA
= 4*Bpacerwidth*Spacerlength;
l = spacer thick;
E = 6 9 1 0 ~ 9 ;
Set
values
for mass, springs
m 1 2;
k_s
= AA*E/ l ;
kilogram
~ / m
10 SATELLITE STRUCI'URAL MATRIX: MASSES AND SPRINGS
Create sa te l l i te
s t ruc tura l
matrix as a
10x10x10
cell
array
with mass
values in each
of
the [3 ,5 ,7]x[3 ,5
,7]
pos i t ions and 2x1 column vectors
(
<damping coeff ic ient>; <st i f fness coeff ic ient>] in o f f - d i a g o ~ a l
pos i t ions
within the matrix
i . e .
,
any odd-even index such
as
323 or
454
w i
11
cal l a
damper
or
spring
co
ef f i c i en t )
.
Springs
for x =
(2,4,6,8]
for
y = (3
,5,
7]
fo r
z= [ 3 , 5 , 7 ]
L{x ,y, z }(2) k_s;
end
end
end
for X=
(3 ,5, 7]
for
y =
(2 ,4 ,6 ,8]
for z =
(3 ,5 ,
7]
L{x
,y,
z }(2)
k_s;
end
end
end
for X=
(3
,5,
7]
for
y =
(3
,5, 7]
for z =
[2 ,4 ,6 ,8]
![Page 135: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/135.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 135/152
L{x , y , z } 2)
k s
;
end
end
end
Masses
for x = [3 ,
5,7]
end
for y = [3 , 5 , 7]
end
for z = [3,5,7]
L{x,y
,
z}
= m;
end
o dro TEST
MODUlE
118
ooFor i n
se
r t in g a
dif ferent mass size at any point in
the
sa te l l i t e
s t
ru ct ur
a l
matrix 0
ii 5 ;
JJ
= 5 ;
kk 3 ;
L{ i i , j j , kk} = 50 0 ;
teste d =
4
5* i i
+ 1.
5 *
j j + 0
5 * kk -
1 8 0
5 ;
s t r
ca
t ( '
The
OOF
being
tes ted
is :
,
spr int f ( '
d , tes ted))
o dro INSERT CUSIDM PROPERTIES:
3U CUBE
IN CENTER OF SATELLITE
Sim
u l
at
e s r ig id i t y in
center
3U
s t ructure by se t t ing
all ver t i
ca
l ly
o ri en t e d in t e rnal
values
of spring
constant
to very large numbe
rs
a ppr oac hing in f in i ty 0
for v = [4,6]
L{3
, 3,v}(2) =
lOO*k-s
;
end
Notes: no si g nific a n t shif t tow a rds lower frequenci e s
0
o dro
INS
ERT
CUSIDM PROPERTIES: ALL 3U CUBES
fo r x = [3 , 5 , 7]
fo r y = [3 ,5 , 7 ]
f
or
v = [4 , 6 ]
L{x,y , v} ( 2 ) lOOOOO * k -s ;
end
end
![Page 136: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/136.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 136/152
119
% end
%Not es : no s i gn i f i ca nt s hif t to w ar d s lower
fr e
qu e n c ies .
o/1
INS
ERT
CUSTOM
PROPERTIES : SIMULATE
DRUMHEADING
BY LOWERING S
PRIN
GS IN
o/1 CENIERMCBT CUBES ('IDP AND OOTIDM)
beta = . 11 ;
L{5 , 5 , 2} 2)
L{5,5,8} 8)
5 1
b e t a h 1
o c:NOTES:
Success
full y low e r s lowest nat u ra l fr e qu ency
o/1 INSERT CUSTOM PROPERTIES : SIMULATE DRUMHEADING BY LOWERING SP RINGS IN
o/1 PERIP
HERAL
TOP
CUBES
for
X [3 , 7]
for
y [5]
for
v = [2 , 8]
L{x , y , v} 2)
end
end
end
for
x
[5]
for
y
[3 ' 7]
for v = [2 , 8]
L{x
,y
,
v} 2)
end
end
end
for
X
[3 ' 7]
for
y
[3 ' 7]
for
v = [2 ,8]
L{x,y,v} 2)
end
end
end
o/c:N
OTES:
Succ ess
full
y
lo w
er s
lo w
e
st
natural
fr
e
qu
e n
cy
a
nd
oth
e
rs follow
o/1 INSERT CUSTOM PROPERTIES : TEST TO
SEE IF
LOvVERING SPRING UNDER OOF 1
o iDOES
WHAT IT SHOULD
%
L{3
, 3 ,4 } (2 ) = .
05
*
L s
;
% L{3 , 3 ,2} (2 ) = L{ 3 ,3 ,4 } 2) ;
![Page 137: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/137.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 137/152
120
Yo
INSERT CUSIOM PROPERTIES:
WW R
SPRING EXTERIOR
LATERAL
SPRING
CONSTANTS
% for x = [2
,8]
% for y = [2
,8]
%
for
v = [3,5 ,7]
% L{x ,y,v
} 2)
= k_s ;
% end
%
end
%
end
Yo
INSERT STRUCTURE PROPERTIES: 3U CUBES
AS
IN DRAPER TEST CONF1G
fuse
=
% for v = [
4,6]
% L{3,3,v} =
fuse*L{3,3,v};
% L{5,5,v} = fuse*L{5,5,v};
% L{7,7,v} = fuse*L{7,7,v};
% end
Yo INSERT STRUCTURE PROPERTIES:
2U
CUBES AS IN DRAPER TEST CONF1G
%
L{3
,5 ,6}
% L{4
,
7,7}
% L{3 ,7,4}
L{5
,3
,4}
%
L{6
, 3 ,7}
=
fuse*L{3
,5
,
6};
fuse*L{4 , 7 ,7};
= fuse*L{3 ,7,4};
fuse*L{5 ,3 ,4};
=
fuse*L{6,3
, 7} ;
Yo BUILD
lv IASS
AND SPRING MATRICES
for
I
=
[3, 5 ,
7]
for J = [3 , 5, 7]
for k = [3, 5, 7]
o/
<Row
is
giv
e n by
r
=
4 .
5*I + 1.5*J +
0 .5*k - 18 . 5 ;
%Populate mass
matrix:
M r , r ) =
L{I ,J ,k};
![Page 138: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/138.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 138/152
Yo
Yo
end
end
2
Coefficient va l ues :
E = (L{I ,J
,k+1}(2)+L{I
,J , k-1}(2)) ;
F1
=-L{I
, J
,k+1}(2);
F2
= -L{I ,J ,k-1}(2);
Coefficient indices i n
r th
row:
%The co e
ff
i c i en t PE
PE
= 4.5*I + 1.5*J +
0 . 5 k - 18.5;
K(r , PE) = K(r , PE) + E; o oE
will never
be out
of
bounds
PF1 =
4.5*
I + 1.5*J + 0. 5*
(k+2)
- 18.5;
if k = 7
F _2 ( r , 1) = F _2 ( r , 1) -
F1;
else K(r,
PF1) = K(r ,
PF1)+ F1;
end
PF2 = 4 .5 * I + 1. 5 * J + 0 . 5 * (k - 2) - 1 8 .5 ;
i k = 3
F _2 ( r , 1) = F _2 ( r , 1) - F2 ;
e s e K ( r ,
PF2)
= K ( r ,
PF2)
+ F2 ;
end
end
Yo KNOBS
AND
BUTIONS
force_scale = 1; %scales magnitude of plot
signal
=
false
; false
turns
off constant
accelerat ion
damping
epsilon_c . 1;
epsilon_k . 1;
epsilon_m .1;
Yo
Yo
Yo ADD NOISE TO THE SPRJNG
:NIATRIX
for
u =
1:27
for v = 1:27
![Page 139: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/139.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 139/152
122
K u,v) K(u,v)*(1 + epsilon_k*(
rand
(1)
- .5 ) ) ;
end
end
11'( ADD NOISE
ID
'lliE MASS MATRIX
for u
1:27
for v =
1:27
M(u,v)
M(u,v)*(1 +
epsilon_m*( rand
(1) - .5 ) ) ;
end
end
11'(
UNDAMPED SYSTEM
%Find eigenvalues and eigenvectors
for
system
[Phi, Omega] eig
(K,
M);
Recover all system natural frequencies
omega_n sqrt ( diag (Omega));
omega_n_sorted
·
sor t (omega_n) ;
%Set
damping
values ranging
from
undamped
to
l igh t ly damped
zeta_i
0;
zeta_f
=
.3;
D e fine region of
i n t e res t
at lowest natura l
frequency
- i .e
. ,
the
lowest
%nat
ural
fr equency of the sa te l l i t e displays a very sharp
peak
a t around
590 Hz , and ind ica tes tha t
damping
values
should be
set to
re f l ec t
l ight
o/odamping
around
the lowest natural frequency (min(omega_n))
c_i
2*zeta_i*min
(omega_n)*
min
(
diag
(M));
c_f
2*zeta_f*
min (omega_n)* min ( diag (M));
res 11;
in te rva l
plot
c _ f -
c_ i ) / r e s+ l ;
+1 i s a t r i c k
to make sure the
l
as
t
c
omes
out
dark
blue .
. .
based
on
color
,
s
for c
c_i:
i n t e rva l :
c_f; Beg in damping fo r - loop
for x
=
[2 ,4 ,6 ,8]
for y
[3,5,7]
![Page 140: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/140.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 140/152
end
end
for
z
= [3,5,7]
L{x,y,z} 1)
end
for
x
= [3,5,7]
end
for y = [2,4,6,8]
for
z
=
[3, 5, 7]
L{x,y,z} 1)
end
end
for x
=
[3,5,7]
for y
= [3
,5, 7]
c·
c·
for z = [2,4,6,8]
L{x,y,z} 1)
= c;
end
end
end
<}'( INSERT
CUSIDM
PROPERTIES :
for x =
[3,
7]
end
for
y
[3,7]
end
for
v
=
[2,8]
L{X y v
} 1
end
·
123
TURN OFF TOP AND
:oor:roM
OORNER DAMPING
<}'( INSERT CUSIDM PROPERTIES :
TURN
OFF ONE OF IW ) RENIAINING DAMPERS THAT
<}'( ATTACH TO
OORNERMCBI CUBES
for x = [2, 8]
end
for
y
[3,7]
end
for
v
=
[3,
7]
L{
X
y
v } 1
end
·
<}'( INS
ERT CUSIDM
PROPERTIES: INTERIOR DM.n:>ING ~ M E PROPORTION OF
![Page 141: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/141.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 141/152
EXTERIOR
ro
DAMPING
Scaling
factor :
alpha
=
1;
for x =
[4,6]
end
for y =
[4,6]
e
nd
for
v = [3 , 5 , 7]
L{x,y,v}(1)
end
124
otes: May
help
to i solate cubestack
as sing l e mass
ro DRIVE CUBE 5 ,5 , 8 WITH LARGE DAMPING
Scaling factor :
alpha = 1 ;
for
X
[5]
for y =
[5]
for
V =
[8]
L{x,y ,v}( l ) alpha*c;
end
end
end
Notes: May
help
to
isola te
cubestack
as single mass
ro
BUILD
DAMPING MATRJX
for
I = [3 , 5 , 7]
for J = [3, 5, 7]
for k =
[3,
5 ,
7]
o oRow
is
given
by
r = 4 .
5*1
+
1.5*J
+ 0
.5*
18.5;
Coefficient values:
A = (L{I+1,J ,k}( 1 ) +
L{I-1
,J ,k}(1)
. .
.
+L{I ,J+1,k}(1)
+ L{I , J -1 ,k} 1) . . .
+L{I , J ,k+l} l ) + L{I ,J , k-1}(1)) ;
![Page 142: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/142.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 142/152
25
B
=
-L{I+1,J
,k} 1);
B2
= -L{I-1,J
, k} 1);
C =
-L{I
,
J+1,k} 1)
;
C2
=
-L{I
,
J-1
,
k} 1)
;
D = -L{I
,J+1,k} 1);
D2 =
-L{I , J -1,k} 1) ;
Co e fficient i
ndices
in rth row:
PA
=
4.5*I
+
1.5*J
+
0.5
*
18.5
;
C r , PA) = C r , PA)
+A;
PB1
= 4.5* I+2)
+ 1.5*J + 0 . 5 * k - 18.5 ;
if I=
F _ ( r ,
1) =
F _ ( r ,
1)
-
B
;
e l se
C r , PB1) = C r, PB1) + B1;
end
PB2 = 4 . 5 * ( I -2) +
1.
5* J + 0. 5* k - 1 8 . 5 ;
i f
I
F
_
( r , 1)
=
F _ ( r ,
1)
- B2 ;
else
C r,
PB2) =
C r ,
PB2) +
B2;
end
PC1
= 4.
5 * I +
1.
5 * (
J
+ 2) + 0 . 5 * k - 18 . 5 ;
if J = 7
F
_
( r ,
1)
=
F
_
( r ,
1)
- C1;
else
C ( r , PC1)
=
C ( r , PC1) + C ;
end
PC2 = 4. 5 * I +
1.
5 * ( J - 2) + 0 .5 * k - 18 . 5 ;
i f J
=
3
F _ ( r , 1)
=
F _ ( r , 1) - C2;
else
C r,
PC2)
= C r
, PC2) +
C2;
end
PD1
=
4.5*I + 1.5*J + 0 .5* k+2) - 18.5;
i
k
=
7
F
_
( r ,
1) =
F
_
( r ,
1)
- D1;
![Page 143: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/143.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 143/152
end
end
end
126
else
C(r , PD1) =
C(r
, PD1) + D1;
end
PD2 = 4.5*1 + 1.5*J + 0.5*(k-2) -
18.5;
if k
=
3
F
_1
( r ,
1)
= F
_1
( r ,
1)
- D2 ;
else
C(r ,
PD2) =
C(r ,
PD2) + D2;
end
1o
ADD
NOISE ID THE DAMPING MATRIX
for u = 1:27
for v = 1:27
C(u,v) = C(u,v)*(1 + epsilon_c*(
rand (1)
- .5 ;
end
end
1o DAMPED SYSTEM
Set
up frequency range of i n t e res t for plot t ing
omega= linspace
(1,10'6,9000);
%Ini t ia l ize complex response vector
X_p =
zeros (27 ,5000)
;
A for
- loop to cons truct the vector
of magnitude
response
for
the
range
%
of
fr
e
qu
e
nci
e s
spec if ied.
Note
tha t
the
forcing function
is dependent
o
o0n
fr equ e ncy .
if
signal
for l = 1 :
length
(omega)
X_p( : , l ) = in v
(-omega(l)
' 2*M+ 1i*omega( l )*C+K)* · · ·
0 . 25* . . .
![Page 144: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/144.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 144/152
else
end
27
(F _homega(l) '
-1 + F_2*omega(l) ' -2)*force_scale;
end
for
l
= 1:
le
ngt
h
(omega)
end
X_p (: , l) = in v (-omega ( l ) • 2*M + 1 i *Omega ( l ) *C + K) * . . .
0. 25* . . .
omega( l ) ' 2*(F
_homega(
l ) A-1
+
F
_2*omega(
l ) A 2)dorce_sca le ;
In cre m ent the color counter s for th is plot
(corresponding
to current
lo o p e d v a
lue of damping
ra t io zeta
s
=
s
+
;
V a r ia b l e d1 for incr e menting OOF after each plot
d1
=
0;
for
fig2
= 1:9
for
plot2 = 1:3
end
f igure ( fig2) B egin f igure for
each degree of
·fr e
edom
Increment to next OOF
and plot
d1
=
d1 +
1;
subplot 3 ,1 ,p lo t2)
loglog (omega, ab s (X_p(d1
, :)
) ,
spr in t f color s s ) ) )
if
f igswitch
Block
of
code
to
f i r s t
gr
a
ph
on
fig10
end
figure (10)
subplot (2 ,1 ,1)
loglog (omega, ab s (X_p(d1 , : ) ) , sp r in t f
color s s ) )
f igswitch = 0; fig10
only
axis (
[xmin,
xmax, ymin,
ymax])
y label ( ' Response (G_n) ' )
xlabel (
'Frequency
(Hz) ' )
grid
on
t i t l e (
spr int f
( 'x-{%d}' ,
d1))
hold
on
![Page 145: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/145.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 145/152
128
end
Plot
natural frequ
e
ncies
as ver t ica l l ines on each figure
e1
=
0 ;
o cDOF
Counter.
ten
=
4;
for fig1
= 1:9
end
for
plot1 = 1:3
end
f igure
(
fig1)
subplot (3 ,1 ,p lo t1)
Plot
natural
frequency
for
lowest
5 DOF
for
e1
=
1 :5
end
y
_lim =
get ( gca , 'YLim ) ;
l ine
( real
(omega_n_sorted(e1))*[1 1],y_lim
,
C ol
o r ' , ' r ' )
hold on
tex t ( l .02*omega _
n_sorted(e1) ,
ten*10 ' 3,
s t r c a t ( ' \ omega ' , s p r i n t f ( ' d ' , e 1 ) ) ) ;
hold on
ten = ten
A
1 ;
Plot 0 .25G forcing magnitude for all omega
lin e ( [
xmin
, xmax ]
,[ .25
* 9 . 8 , .
25*9
. 8])
Plot 10 ' 1
peak
measure
l ine
([xmin
, xmax] ,
[10
,
10]
, '
Color
,
k
' )
end End damping for - loop
o/c FIG 10 : UNDAMPED
SYSTEM
RESPONSE WITH AT. ~
AND
PHASE
ANGIE
f igure (10)
t i t l e
( '
Un
damped R
es
po n
se
with Natur
a l
Fr
e
qu
e n
cies
a
nd
Ph
ase
Angl
e
)
subplot (2
,1
,1)
grid off
ylabel ( ' R espon se (m/ s A2) ' )
xlabel
( ' Frequen
cy
(Hz ) )
for e1
=
1:
length (omega_n)
y
_lim
=
get
(
gc a
, 'YLim ' ) ;
![Page 146: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/146.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 146/152
![Page 147: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/147.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 147/152
eferences
Avery K., et al., Total Dose Test Results for CubeSat Electronics, 2011 IEEE Radia
tion Effects Workshop {REDW),
Sevilla, Spain, 2011
Bruhn F., et al.,
QuadSatjPnP: A Space-Plug-and-play Architecture SPA) Compliant
Nanosatellite,
lnfotech@Aerospace Conference, Unleashing Unmanned Systems St.
Louis, Missouri, 29-31 March 2011
Clavier
0
and Cutler
J.,
Enabling low-cost, high-accuracy magnetic field measure
ments on Small Sats for space weather missions, Proceedings of the 25th Annual
AIAA/USU Conference on Small Satellites, Logan, Utah 2011.
Darling, Nathan T.; Harris, David. Boston University Student Proposal for Deploy
able Solar and Antenna Array Testing On Board
NASA
Parabolic Flights December
16, 2011. Available upon request from author.
Darling, Nathan T. ; Mendez, Joshua. Student Payload Integrated First Flight SPIFF)
Proposal for NASA / LSU High
Altitude
Student Payload December 16, 2011.
Available
upon
request from author.
Finlay
t
al.,
HOW DRS-
Modular Space Vehicle on the
T2E
Mission,
Proceedings
of
the
25th Annual
AIAA/USU
Conference on Small Satellites, Logan, Utah 2011.
Giurgiutiu V.,et al., Space Applications
of
Piezoelectric Wafer Active Sensors for
Structural Health Monitoring,
Journal of Intelligent Material Systems
and
Struc
tures vol.
22
no. 2 1359-1370, August 2011.
Lindstrom
C. D., et al., Characterization
of
Teledyne Microdosimeters for Space
Weather Applications, Proceedings of Solar Physics and Space Weather Instru-
mentation IV, Torino, Italy, August 2011.
Mendez J. , et al., Compact Half- Unit Imaging Electron Spectrometer for CubeSat
Operations {CHICO) ,
Proceedings of the 25th Annual AIAA/USU Conference on
Small Satellites, Logan,
Utah
2011.
Thermotron Industries 2011). Vibration Test Systems Retrieved December 18, 2011
from http:/ j www thermotron. com.
130
![Page 148: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/148.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 148/152
![Page 149: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/149.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 149/152
ppendix
Vita
32
![Page 150: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/150.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 150/152
133
Nathan Thomas arling
Center
for Space Physics
725 Commonwealth A venue
Room
506
Boston,
MA
02215
Personal
Born
1981
in
New York,
NY
Education
Phone:
Fax:
Email:
Homepage:
B.A. Spanish Language and
Literature,
2004.
M.S . Mechanical Engineering, 2013.
Graduate Coursework
61 7
353-0285
617) 353-6463
www.busat.org
ME 515- Vibration
of
Complex Mechanical
Structures
ME 560 - Machine Design and Instrumentation
ME 900 - Research:
Preliminary
Vibration Testing and Analysis
of
a Novel Nanosatellite Structure
Licensure
and Certification
USCG
500 Ton
Ocean Master
AB Unlimited, STCW 95
Medical person
in
charge
MPIC)
NAUI
Scuba
Diver course
completed
December, 2009)
Summary
of
Experience
NASA
project management,
proposals
rocurement and
budget
management
Personnel
management and
leadership
NASA
hardware
design cycle
and
reviews
![Page 151: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/151.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 151/152
34
Manual
and
CNC fabrication
CAD / Modeling: SolidWorks, MATLAB
Microgravity parabolic flight
testing
GIS: ArcMap 10, Agisoft PhotoScan
Rigging and heavy lifting equipment
STEM
education
and
outreach
Selected
Work
Experience:
Boston
University Center
or Space Physics (Jan. 11
to
present)
http :/ / www.bu.
edu/csp
Program Manager
Currently overseeing
the
Boston University
Student
satellite for Applications
and
Training BUSAT). Duties include personnel
and
logistics management for a 25
member
student engineering and physics team, budget oversight, as well as environ
mental testing of a scientific spaceflight structure. Familiar NASA project develop
ment
and review process, including documentation procedures and export control.
Duties include regular formal technical presentations , public relations and K-12
out-
reach. Direct experience with satellite mechanical and electronic hardware, vibration
testing, high altitude balloon flights,
and
microgravity testing on
board
parabolic
aircraft.
Boston
University Center
or
Space Physics (Jan. 11 to
present)
www.busat.org
Structures Team Lead
In charge of concept development , fabrication
and
testing of a novelmodular nanosatel
lite structure. Modeling
and
analysis performed in SolidWorks
and
MATLAB. Ex
perience with shock
and
vibration testing procedures
and
data analysis.
In
charge of
organizing a
rotating
student
team
of designers, also responsible for final documen
tation
and
design review presentations.
Scientific Instruments Facility (Oct.
10 to
Jan. 11)
sif.bu.edu Research Assistant
Assisted with general shop maintenance
and
production. Familiar with fabrication
process associated with precision scientific instrumentation. Experience
with
manual
![Page 152: Darling_Nathan_2013_web.pdf](https://reader030.vdocument.in/reader030/viewer/2022021216/577c7c981a28abe0549b3876/html5/thumbnails/152.jpg)
8/16/2019 Darling_Nathan_2013_web.pdf
http://slidepdf.com/reader/full/darlingnathan2013webpdf 152/152
135
lathes and mills.
National
Undersea Research
Center
(Jan. 10
to
May
10)
www.nurc. uconn.
edu
Intern
Overhauled the hydraulic manipulator valve pack for the scientific remotely operated
vehicle K2
at the North
Atlantic
and Great
Lakes National Undersea Research
Center
NURC) facility in
Groton
, CT.
The
Boat Company
(Apr.
to Sept.
06, 07, 08, 09,
Aug. 10)
www.
theboatcompany.org
Chief Mate
Stood
navigational watch, worked as wilderness guide
and
answered
to the captain
in
planning logistics of a small wooden cruise vessel. Oversaw safety
and
maintenance,
including weekly drills.
Operated
hydraulic crane
in
deployment
and
recovery of four
19-foot aluminum fishing skiffs at least twice daily. Served as skiff
captain
daily while
mother
ship was
at
anchor,
stood
wheel watch every second day while underway
and
anchoring.