Professional Documents
Culture Documents
RTICC
Rapid Terrestrial Imaging
CubeSat Constellation
Preliminary Design Report
AA420/421 Space Design
In conjunction with Andrews Space Inc. (SATS)
June 12, 2009
Authors:
Michael Bernhardt, Aaron Borth, Rachel Brennan, Enrique Galgana, Peter Gangar, Austin Kemis, Nikolas
Lutzenhiser, Katie Moravec, Skander Mzali, Zahra Nazari, Josh Ross and Eun-Ju Shin-White
Table of Contents
Table of Contents.......................................................................................................................................... 2
Abstract ......................................................................................................................................................... 3
Mission Design ............................................................................................................................................. 4
1.
OrbitalMechanics ............................................................................................................................. 4
2.
Propulsion ......................................................................................................................................... 4
3.
ImageAcquisition............................................................................................................................ 35
4.
Navigation/ControlSystems ........................................................................................................... 49
5.
Communications ............................................................................................................................. 81
6.
References................................................................................................................................................. 119
Appendices................................................................................................................................................ 127
A. OrbitalMechanics ......................................................................................................................... 127
B. Propulsion ..................................................................................................................................... 144
C. ImageAcquisition.......................................................................................................................... 162
D. Navigation/Control ....................................................................................................................... 164
E.
F.
Abstract
The goal of this project was to develop a detailed mission design for an Earth imaging
application using a low-cost constellation of CubeSats. The requirements for this mission include
coverage between 55 degrees North and South, image resolution of 3 meters, image acquisition
within 5 minutes of command, and download to client within 60 minutes. The mission design
concept calls for a Walker constellation of 33 planes with 10 CubeSats per plane, at an altitude of
520 km. A deployment system is proposed to deliver the CubeSats to each orbital plane via a
system of carriers, with each carrier holding 10 CubeSats. Once deployed, the CubeSats will use
a 90mm Maksutov telescope in combination with a 10MP CCD to capture images covering a
ground area of 5km x 5km with a resolution of 3m. In order to acquire images with a nadir
pointing accuracy of 200m, the attitude determination and control system will use GPS and a
custom designed star tracker to provide high accuracy attitude determination. Supporting the star
tracker will be a arrangement of an IMU, sun sensors, and magnetometers. CMG wheels and
magnetorquers will be used to control the attitude of the CubeSats. Once acquired, the images
will be transmitted via a UHF communications link broadcasted using monopole antennae that
will traverse other satellites in the constellation in order to be delivered to the designated ground
station. Providing power to the systems of the CubeSats will be a combination of solar arrays and
batteries. The computing and data handling system will employ commercial-off-the-shelf
hardware including integrated microcontrollers and custom computer solutions. Accuracy,
weight, cost, and efficiency are the primary concerns of this mission and the devised solutions
will be addressed in the following report.
Mission Design
1. OrbitalMechanics
The tasks for orbital mechanics analysis were to design a constellation, develop a deployment
method, and select a suitable launch vehicle. The mission profile of rapid-response imaging with
maximum earth coverage capability drove the selection of the satellite constellation design. The
deployment method was designed to provide the full deployment of all cubesats within a
reasonable time frame, and with achievable delta-V maneuvers. The launch vehicles were chosen
based on their availability and capability for launching to desired inclinations and altitudes.
1.1.ConstellationDesign
The imaging application of the mission calls for many small satellites in low earth orbit
(LEO) to minimize focal length and telescope diameter to achieve the required image resolution.
Two constellation designs were proposed and evaluated based on their ability to provide
maximum coverage and minimize the total number of satellites. The initial design concept was a
constellation of polar orbits, to create polar streets of coverage (see Figure 1.1.1below).
The primary advantage of this configuration is that it is capable of providing full earth
coverage. There are two disadvantages to this configuration, however. One is that it provides
greater coverage of the poles, where there are less areas of interest for imaging, while it provides
less coverage near the equator. The second difficulty is the large delta-V required to launch to
polar orbit. Due to this high delta-V requirement and the resulting low availability of launches to
polar orbit, the polar streets constellation was ruled out as a feasible option. Choosing an altitude
of 520 km and assuming a launch to 70 would require 20 planes with 20 spacecraft each,
resulting in a delta-V of 2.6 km/s for plane change to 90 plus 45 km/s for 21 plane changes (see
Appendix A.1 for method of calculation). Launching directly into a 90 orbit is also undesirable
as the frequency of launches to polar inclination is around one per year, leading to an
unreasonably long deployment time for 20 planes (see Table 1.3.2 and 1.3.3 below).
The second design proposed was the Walker constellation. In this configuration, the
orbital planes all have the same inclination, but the right ascension of the ascending node for
each plane is equally spaced over 360, creating a crisscrossing pattern of coverage, shown in
Figure 1.1.2 below. Although the cubesats are still spread out near the equator and condensed
near the higher latitudes, they are distributed more evenly than in the polar streets constellation.
The Walker constellation involves a loss of coverage near the poles, but allows a
reduction in the total number of spacecraft because of the better distribution at lower latitudes.
The Walker configuration redistributes coverage by halving the number of spacecraft per plane
and almost doubling the number of planes. The number of spacecraft per plane is reduced if the
spacecraft are phased correctly to alternate their crossing at each ascending and descending node.
The number of planes is increased since the ascending nodes must be spread over 360 of the
entire circumference of the Earth, whereas the orbital planes for polar streets are spread out over
180 of the circumference. The reduced number of spacecraft and better coverage provided by
the Walker constellation, and the high delta-V and low launch frequency to 90 associated with
the polar streets constellation, led the team to choose the Walker constellation for this mission.
Coverage analysis for the Walker constellation was initially conducted over an altitude
range of 400 550 km. This range was set based on the requirements from propulsion and
imaging. The minimum allowable altitude was set at 400 km, below which the orbits would
degrade too fast, threatening loss of the spacecraft and necessitating large delta-V for orbit
boosting. Above 550 km, the total orbit degradation over the systems 1-year lifespan would be
within acceptable range, such that the propulsion for orbit maintenance could theoretically be
disregarded. However, altitudes much higher than 550 would require telescopes with a bigger
diameter to achieve the required resolution. Since the diameter is already pushing the dimension
limits for the cubesat structures, increasing the diameter should be avoided. Thus a range from
400-550 km was within the acceptable propulsion and imaging constraints.
This altitude range and an assumed field of view angle of 45 were used to calculate the
satellites orbital speed, ground track speed, period, and coverage swath. The swath area and the
speeds determined the minimum number of orbital planes and minimum number of spacecraft
per plane to achieve full earth coverage, with a maximum revisit time of 5 minutes. In order to
halve the number of spacecraft per plane, they must be phased such that one crosses from each
orbit at 5 minute intervals, as illustrated in Figure 1.1.3. All calculations and derivations are
explained in Appendix A.1.
The number of planes and satellites per plane, calculated over an altitude range of 400
500 km, allowed choosing a constellation with a reasonable altitude and minimized number of
spacecraft. The total number of cubesats and the number of planes are given in Table 1.1.1 for
nine selected values of altitude and inclination.
Table 1.1.1: Total number of cubesats and planes for selected alititudes and inclinations.
Inclination (degrees)
Altitude (km)
45
55
65
400 km
450 km
500 km
The Walker configuration initially chosen was 500 km altitude with 60 inclination, with
36 planes of 10 spacecraft each. The altitude was chosen to achieve 3 m imaging resolution
while minimizing orbit decay. The inclination was chosen to allow reasonable delta-V (~3 km/s
from 35 to 60) while achieving maximum latitude coverage, since the full earth coverage
requirement must be relaxed to some finite inclination. The full-earth coverage requirement was
relaxed to neglect areas above 60 north latitude and below 60 south latitude since these areas
are of less interest for imaging. The overall configuration was designed to minimize the number
of spacecraft needed and allow some coverage overlap.
Although analysis of the Walker constellation was initially conducted assuming worsecase coverage at the equator, further research showed that the phasing method described above
(with cubesats making alternating passes over nodes along the equator) did not guarantee 5
minute revisit at every point within the coverage band. Various configurations were analyzed for
the fraction of points that are covered with 5 minutes, as well as the maximum time that any
point has to wait for coverage (See Appendix A.1 for method of analysis). The best coverage
occurs at high altitude and low inclination, since high altitude allows a bigger coverage swath
and low inclination requires less area to be covered. A representative set of options for 520 km
altitude and 55 inclination are shown in Table 1.1.2 below.
Table 1.1.2: Total number of cubesats and planes for selected alititudes and inclinations.
Total #
Sats
330
33
# Sats per
plane
10
Phasing
Angle (deg)
19.6
Coverage
Fraction
0.979
Maximum Revisit
Gap (min)
6.12
360
36
10
22.9
0.997
5.379
363
33
11
16.4
0.993
5.626
390
39
10
26.2
1.000
4.856
# Planes
To make the best compromise between minimum planes, minimum satellites, and
maximum coverage fraction, the configuration chosen was a constellation of 33 planes with 10
spacecraft each and 19.6 cubesat phasing between planes, flying at an altitude of 520 km and an
inclination 55. The fraction of points covered within 5 minutes is 0.979 and the maximum
coverage gap for any point is 6.12 minutes. Since the effects of the increase in altitude and
decrease in latitude were considered acceptable by all other subsystems, this configuration was
the final one chosen for this mission.
1.2.DeploymentMethod
In order to deliver the cubesats to their proper locations in the constellation, a system of
carrier vehicles must be designed. The proposed method is to deploy one carrier into each of the
33 planes by launching groups of 6 to differentially precessing orbits, and then have each carrier
deploy its 10 cubesats on its respective orbit using an elliptical phasing orbit.
Several options were considered for deploying the carriers. The first option was to launch
a master carrier to the desired inclination. This master carrier would hold all the individual
carriers and distribute them to their respective planes by executing plane changes at the
northernmost point of the orbit (at the circle of nodes where the Walker orbits cross, shown in
Figure 1.2.1). The master carrier would drop off the first carrier and then execute a plane change
to transport all the remaining carriers to the next orbit. This option would require 32 plane
changes and would involve carrying unnecessary mass to each plane since only one carrier ends
up on the plane. Assuming a master launch at an inclination of 55 to deploy 33 planes, the total
delta-V would be 38 km/s and the total propellant mass for the plane changes (using the specific
impulse of hydrazine) would be 218 million kg. (Please see Appendix A.2 for method of
calculation.) The prohibitively large delta-V and extremely high propellant mass required for this
option ruled it out as a feasible method for carrier deployment.
Figure 1.2.1: Close up of polar ring; plane changes for master carrier occur where orbits cross
The second option is to launch in groups of three carriers. For each launch, one carrier
remains in the launch orbit (which has the desired inclination) and the other two carriers execute
plane changes to split off to the right and left so that the three carriers are deployed onto their
respective planes with the correct right ascensions. Each plane change occurs at the node where
the initial launch orbit crosses the desired orbit, as illustrated in Figure 1.2.2. Assuming 33
planes, the delta-V needed for each plane change would be 1.2 km/s, requiring a total delta-V of
26 km/s to deploy all planes. This option would require 11 launches and take up to 1.5 years if
the carriers were launched as secondary payload.
10
The third option considered was to launch each carrier individually to the desired
inclination. Each launch would be timed so that the right ascension of each orbit would be
correct relative to the others, using the rotation of the Earth to position the launch point correctly.
This option would not require any plane changes, but would take 33 launches and up to 4 years if
each carrier was launched as a secondary payload.
The use of orbit precession was also considered as a possible deployment method, and
was finally chosen as the best option for this mission. All the carriers would be launched to a
particular inclination, and the first carrier would execute a plane change to the desired
inclination. The rest of the carriers would remain on the launch orbit until the orbit had precessed
to the correct right ascension for the next Walker orbit (illustrated in Figure 1.2.3).
11
Assuming a launch to a 50 inclination with a desired inclination of 55, the delta-V for
one 5 plane change is 660 m/s and the total delta-V for 33 planes is 22 km/s. The precession
time between successive planes is 21 days, resulting in a total deployment time of less than 2
years if all the carriers are launched into a single precession orbit. This deployment time could be
significantly reduced by launching the carriers in groups, allowing several launch orbits to
precess around the earth simultaneously. This would eliminate the need for the last carrier to wait
for all the previous carriers to deploy before it reaches its own orbital plane. (Please see
Appendix A.2 for method of calculation.)
Assuming up to 6 carriers are launched together using a Falcon 1e (further discussed in
the Launch Vehicles section below), 5 launches of 6 carriers and 1 launch of 3 carriers will be
required. With a precession time of 21 days between planes, a group of 6 carriers will take 126
days to deploy and the group of 3 will take 62 days. If one group can be launched per month, the
group of 3 carriers will actually deploy before the fifth group of 6 carriers. Thus the time for
carrier deployment is 5 months until the fifth launch plus 126 days until the last carrier is
12
deployed on its orbital plane, or about 9 months total. Based on the lower total delta-V and
potentially much lower deployment time, the precession method with multiple launches was the
chosen carrier deployment method.
Once the carriers are place on their respective orbits, the cubesats must be evenly spaced
out along the orbit. To accomplish this, the carrier will enter an elliptical phasing orbit with a
period slight shorter or longer than that of the cubesat orbit, such that the carrier intersects the
circular cubesat orbit with the appropriate time spacing. The carrier will then provide a delta-V
to enter the circular cubesat orbit, release the cubesat, and the return to its elliptical orbit.
Assuming 10 spacecraft per plane, the time spacing will be 1/10 of the cubesats period. If the
elliptical orbit is designed to provide this much time spacing in a single orbit, the delta-V
required to transfer from the elliptical carrier orbit to the circular cubesat orbit is unreasonably
high (~233 m/s). To reduce this delta-V to a reasonable value (5-10 m/s), the carrier must orbit
an integer number of times to provide the correct time spacing, requiring a lower eccentricity for
the elliptical orbit. This lower eccentricity results in a lower delta-V to transfer from the elliptical
carrier orbit to the circular cubesat orbit, and a longer deployment time to deploy all cubesats.
The elliptical phasing orbit used can either be inside or outside the circular cubesat orbit.
An outer elliptical orbit requires a slightly lower delta-V for transfer and a slightly longer
deployment time than an inner elliptical orbit. For the case of 10 cubesats per plane at 520 km
altitude, choosing an outer carrier orbit with 50 orbital periods between cubesat deployments
results in an orbit with a 6898 km perigee radius and a 6916 km apogee radius. The circular and
elliptical orbits are illustrated in Figure 1.2.4, where the outer red line represents the elliptical
carrier orbit and the inner blue line represents the circular cubesat orbit.
13
The delta-V to transfer between the elliptical carrier orbit and the circular cubesat orbit is
5 m/s. The time between deployments is 3.3 days. Assuming the carrier starts from the cubesats
520 km altitude orbit and carries 10 cubesats, it will need to separate the first cubesat and then
leave and re-enter the circular cubesat orbit 9 times. Each of these 9 times requires two
maneuvers, one to exit the circular orbit (go into an elliptical orbit) and one to enter the circular
orbit (from the elliptical orbit). So the total delta-V is 18 times the delta-V for a single orbitchanging burn (91 m/s), and the total deployment time is 9 times the time between deployments
(30 days). Thus the total deployment time, from the beginning of the first launch to the time
when the last cubesat is deployed on its orbit, is about 10 months.
Since the coverage provided by the constellation depends on its symmetry, every orbit in
the constellation must decay an equal amount. To ensure equal decay on all orbits, all carriers
will wait in their respective orbits until the last carrier is deployed, and then all carriers will
deploy their cubesat simultaneously. The empty carriers will be for communication and data
14
handling, but the specific details of the placement of the carriers after deployment needs further
development. Also, the decommission concept for both the carriers and cubesat has not been
developed, but would probably involve a de-orbit burn to put each spacecraft into an elliptical
orbit where the atmospheric drag will cause the orbit to decay further until the spacecraft burns
up in re-entry.
1.3.LaunchVehicles
This mission required either one or several launch vehicles to boost the carriers into orbit,
depending on the deployment method used. The requirements for the launch vehicles were low
cost, high launch frequency, sufficient payload capability to LEO, and capability to achieve
desired inclination.
Initially several U.S. and Russian launch vehicles were considered since the deployment
method was not yet fully developed and the destination orbit that the launch vehicle was to
achieve was not yet decided. The most common U.S. launch vehicles available for commercial
use for LEO transport were the Delta II and Delta IV launch vehicles. Their performance
capabilities are summarized in Table 1.3.1 below.
Table 1.3.1: U.S. Launch Vehicles
Vehicle
% P/L mass
Launch Site
Inclination
Frequency
Delta II
12.36
$13.6 M
Cape,
Vandenburg
27.8-50
63.8-110
11/yr
Delta IV
medium
7.39
$9.82 M (est)
Cape,
Vandenburg
27.8-50
63.8-110
3/yr
Delta IV
medium +
4.67
$6.44 M (est)
Cape,
Vandenburg
27.8-50
63.8-110
1/yr
Delta IV
Heavy
2.46
$6.25 M (est)
Cape,
Vandenburg
27.8-50
63.8-110
1/yr
15
The Russian launch vehicles meeting our mission requirements included the Kosmos 3M,
Proton K, Proton M, Rokot, Dniepr 1, Soyuz, and Zenit 2. Their performance capabilities are
summarized in Table 1.3.2 below.
Table 1.3.2: Russian Launch Vehicles
Vehicle
% P/L mass
Launch Site
Inclination
Frequency
Kosmos 3M
45.4
$14.00 M
Plesetsk
50.6
1/yr
Proton K
3.2
$5.45 M
Baikonur
5/yr
Proton M
3.0
$5.28 M
Baikonur,
Kazakhstan
7/yr
Rokot KM
33.4
$9.14 M
Plesetsk,
Baikonur
50.6
63, 73, 82, 86.4
6/yr
Dniepr 1
14.1
$4.28 M
Baikonur
2/yr
Soyuz
10.2
$7.23 M
Plesetsk,
Baikonur
52,65,70
16/yr
Zenit 2
4.7
$3.96 M
Baikonur
46.2, 51.4,
63.9, 89.6, 98.9
1/yr
The launch vehicle availability was based on regularly scheduled launches, using the 2009
launch schedule. Using the Delta II and Delta IV launch vehicles would require a minimum of
seven months to launch all 36 carriers in groups of 3 for the clustered deployment method,
whereas using all of the mentioned Russian launch vehicles would require a minimum of three
months. The launch vehicles costs were calculated using launch price per pound and adding on
the percentage of total payload used multiplied by the launch cost. The launch costs for the
Delta IV vehicles are estimated costs and are lower than the actual costs because the launch price
per pound was not available.
16
Based on the current precession-based deployment method, the launch options were
revisited. The three possible launch methods considered were: a single dedicated launch using a
Delta II, multiple dedicated launches using a Falcon 1, or piggy-back rides on Delta IIs using an
ESPA ring. Assuming a cubesat mass of 6 kg, an empty/dry carrier mass of 40 kg, and 50 kg
propellant mass, the total mass of a single carrier would be 150 kg, and the mass of 33 carriers
would be 4950 kg. The cost of the various methods was calculated for comparison, based on
these payload mass estimates. The payload capabilities and costs for each option are summarized
in Table 1.3.3 below. These payload capabilities listed are rough estimates based on LEO around
200 km and a range of launch inclinations available at Cape Canaveral and Vandenburg Air
Force Base.
Table 1.3.3: Launch Vehicles for Precession Deployment
Vehicle
Payload
Capability
(kg)
# Carriers
per Launch
# Launches
Required
Cost per
Launch
Total Cost
Delta II
2700-6100
33
$55 M
$55 M
Delta II
(ESPA)
1020
$10,692/kg
$53 M
Falcon 1
240-420
17
$7.9
$134 M
Falcon 1e
700-1010
$9.1
$54.6 M
Multiple dedicated launches using Falcon 1e vehicles is currently considered the best option
because of its low cost. Although piggyback rides are lower cost based on the estimated cost per
kg, they would result in longer deployment due to the wait time for available rides.
17
2. Propulsion
The propulsion systems required for successful completion of primary mission requirements
are divided into three major sections, based on the sequence of cubesat deployment and
operation. The first stage of deployment consists of the time interval between launch and
insertion of the cubesat carrier into an elliptical deployment orbit discussed in the
constellation section. Propulsion systems reviewed in this section include launch vehicle
selection as well as integrated carrier propulsion. Stage two of deployment consists of
separating the cubesats from the carrier. Stage three consists of station-keeping and orbital
maintenance performed by the cubesats in order to maintain operation for the required
lifetime. A summary of the delta-V requirements and propellant masses required for each
stage of the mission is given in Table 2.1. Delta-V for disposal of the vehicles is not
included due to the fact that orbital decay should lead to automatic disposal, or alternatively a
non-propulsive method such as a tether may be used.
Table 2.1: Mission Delta-V and Propellant Mass Summary
Stage
Carrier
Cubesat
1: LEO to 520km
171 m/s
0 m/s
2: Plane Change
660 m/s
0 m/s
82 m/s
0 m/s
4: Cubesat Launch
10 m/s
5 m/s (imparted)
Total
913 m/s
28 m/s
3: Cubesat
Deployment
18
2.1.CarrierPropulsion
The propulsion system required for the cubesat carriers is sized based on the requirement for
location of the cubesats in the constellation as well as launch vehicle availability. The
original cubesat carrier deployment concept involved the use of two different types of carrier
vehicles, the alpha carrier and the beta carrier. Based on the delta-v trade study conducted, it
was determined that the best solution for deployment was a series of 12 launch vehicles, with
3 carriers per vehicle. Of these, one carrier (alpha type) services the launch plane, while the
remaining pair (beta type) executes ten degree right ascension of the ascending node
(RAAN) plane changes at the intersection nodes of the adjacent planes. The propulsion
requirement for the alpha carriers is far less than that for the beta carriers due to the
extremely high delta-v values required for plane change burns. For the alpha carriers, the
chosen propulsion system consisted of a Northrop MRE-15 Hydrazine Monopropellant
thruster capable of delivering a maximum of 86N of force for approximately 400 seconds,
which was calculated to be sufficient for non-plane change requirements. The beta-type
carrier uses an Aerojet R-42 MMH/NTO bi-propellant thruster rated at 890N. These thrusters
are responsible for raising the carriers from LEO to a 500km altitude circular orbit and
relocating into an elliptical orbit for cubesat deployment. In addition to these requirements,
the beta carriers must complete the aforementioned plane change maneuver. See Table 2.2
for a summary of the total delta-v, thrust required, and engine information for each carrier
vehicle.
19
-Carrier Propulsion
System:
-Carrier Propulsion
System:
Total V
.65 km/s
1.97 km/s
Thrust
Required
Alt 72 N
Plane Change 185 N
Alt 109 N
Plane Change 282 N
Right Ascension 804 N
Propellant
Hydrazine
MMH/NTO
Thrust
86 N at 400 psia
Engine Mass
1.1 kg
4.53 kg
Propellant
Mass
34 kg
127 kg
Total Mass
169 kg
267 kg
Size
Due to the large amount of delta-v and propellant mass required for cubesat deployment for
the Beta type carrier vehicles, the deployment method was revised to use orbital precession
for cubesat deployment. For this revised method, there would only be one type of carrier
vehicle which would position itself in the correct phasing orbit and then the carrier would be
used to accelerate to the required cubesat orbit for deployment of the satellite. The
maneuvers required to place the carrier vehicle in the correct orbit include: boosting the
carrier up to an altitude of 520 km from a starting drop-off altitude of 200 km using a circular
orbit, performing a 5 degree plane change from the launch plane, and finally placing the
carrier vehicle in its desired elliptical phasing orbit. Once the carrier vehicle is in the correct
orbit for cubesat deployment, the carrier vehicle will launch the first cubesat The actual
20
launch method will be a compliant low-acceleration spring similar to that used on the P-Pod
deployment system. Thereafter, the carrier vehicle will be used to impart each cubesat with
the additional required V of 5.076 m/s. Then the carrier will be decelerated by the same
amount for each cubesat deployment to place it back in its original orbit, for a total of 9
cubesat deployments. See Table 2.3 for the details of these maneuvers.
Table 2.3: Carrier Orbital Maneuvers for Hydrazine monopropellant engine, Isp = 230s
Maneuver
Propellant
Mass(kg)
Time (sec)
Plane change (5
deg)
663
26.3
236
200 km-->520 km
171
9.4
85
To perform the cubesat orbit positioning maneuvers, both bipropellant and monopropellant
main engine choices were considered. The bipropellant engine was favored at first for the
high amount of specific impulse available with a bipropellant engine (280 sec) as compared
to a specific impulse of 230 sec for a monopropellant engine. With the higher specific
impulse of a bipropellant engine, the total required fuel and oxidizer mass was calculated to
be 35 kg. A large disadvantage of bipropellant engines is the necessity to use two separate
tanks, one for oxidizer, and one for fuel, along with the added complexity of the propulsion
system as a result. For comparison, the amount of propellant required including an 5%
margin of safety/de-commissioning allocation for a monopropellant engine was calculated to
be 47.3 kg for the cubesat orbit positioning maneuvers (including carrier orbital positioning
and cubesat deployment). The propellant mass was calculated using the rocket equation with
an average mass for each given segment, Equation B.2.1 in the Appendix B.2.
21
Comparing the two values of propellant mass, the mass of the monopropellant engine was
determined to not be significantly more than the bipropellant system mass. The biggest
advantage of using a monopropellant engine over a bipropellant engine is a much lower level
of complexity with only one propellant tank. Since the amount of fuel mass required for a
monopropellant engine was deemed to be reasonable when compared to the extra dry mass
and failure potential of a bipropellant system, a monopropellant engine was chosen. The
proposed monopropellant engine to be used is the Aerojet MR-107S. See Figure B.4.3 in the
appendix for technical information on this engine.
2.1.1.CarrierCubesatDeploymentManeuvers
To perform the required acceleration and deceleration maneuvers for cubesat deployment, the
Reaction Control System (RCS) thrusters will be used. The RCS thrusters will also be used
to correct for thrust misalignment of the main engine with the center of mass of the carrier.
The RCS thrusters were sized based on the thrust from the 250 N main engine assuming
offset angle of 1 degree from the center of mass of the carrier vehicle, resulting in a required
thrust of 2.5 N. In order to perform reaction and attitude control, two 4 N monopropellant
engines will be used. The monopropellant engines will be placed as far from the center of
mass as possible to provide the greatest available torque (See Figure 2.1).
22
A cluster of four RCS thrusters will be used in each group. To perform the cubesat
acceleration/deceleration deployment maneuvers, the required thrust will be divided between
the four thrusters. The time to perform each maneuver was calculated based on an estimated
thrust available, average system mass at that time, and delta v requirement. Table 2.4 shows
the delta v, and estimated time required to complete each maneuver.
Table 2.4. Cubesat Deployment Maneuvers.
Cubesat
Deployments
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
V (m/s)
0
10.1
10.1
10.1
10.1
10.1
10.1
10.1
10.1
10.1
23
0.5
50.5
0.4
47.8
0.4
45.2
0.4
42.5
0.4
39.9
0.3
37.2
0.3
34.6
0.3
32.0
0.3
32.0
2.1.2.Tankage
In order to cause the least disturbance possible in the center of mass during propellant firing,
the total amount of propellant required for the whole mission will be divided evenly into 4
tanks. The propellant tanks were placed at the back of the carrier vehicle in order to have the
propellant tanks close to the main engine and to avoid interference with the placement of
other components on the carrier. To simplify the system tankage, blowdown tanks were
considered. A spherical tank shape was chosen as this is the most efficient structural shape,
which results in the lowest tank mass. A blowdown tank system was chosen for the
simplicity of a blowdown system design. A blowdown system only requires one propellant
tank, with the pressurant enclosed in a small volume above the fuel, which is separated by a
diaphragm. As the propellant is used, the pressure in the fuel tank decreases, resulting in
decreased thrust. To determine the blowdown ratio for the tanks, the operating pressure
range for the main engine (475-150 psi) was considered. The blowndown ratio was
calculated using Equation B.2.2 (see Appendix B.2).
The total volume of the propellant tanks was calculated using the density of hydrazine (1010
kg/m3) and the total propellant mass was calculated for all maneuvers, resulting in a total
propellant mass of 47.3 kg, which includes 5% for contingency propellant for additional
attitude control. The total volume required for fuel in the propellant tanks was calculated to
be 0.047 m3. Using Equation B.2.3 (see Appendix B.2), the pressurant volume was
calculated to be 0.014 m3. The chosen tank pressurant to use was helium. Helium was
chosen for its low molecular weight, to add the least amount of weight possible to the
propellant tank. To reduce the tank size, a pressure regulated tank system was chosen. This
system will be made up of four propellant tanks, for a total volume of 0.044 m3 and four
24
pressurant tanks with a total volume of 0.006 m3. Using a pressure regulated system results
in a constant thrust level for the duration of the engine burn.
Using this system, the carrier engines will be operated at a constant inlet pressure of 350 psi.
The material most commonly used for hydrazine propellant tanks is a Titanium alloy (Ti6Al-4V) because it is strong, lightweight, and nonreactive.
2.1.3.CarrierPropulsionSystemArchitecture
The following is the proposed propulsion system architecture, shown in Figure 2.2.
This diagram includes propellant service valves for filling and draining the propellant, latch
valves to control the propellant flow to the different engines, propellant feed lines, and the
engines. Each engine also includes a flow regulating solenoid valve. Possible valve sources
include the Moog latch and service valves. To feed propellant to the main engine, the two
25
propellant latch valves would be used. At any one time, up to 4 thrusters may be firing, or 2
RCS thrusters will be fired at the same time as the main engine to compensate for off-axis
thrust vectoring. See Figures B.4.5 and B.4.6 for possible propellant feed control valves.
2.2.CubesatDeploymentMethodTradeStudy
The separation of the cubesats from the carrier vehicle falls under the mission requirement
for positioning within the constellation. Secondary requirements derived from the assumption of
a desired one month satellite propagation time interval provided a required delta-v for the system
of 5 m/s per cubesat, or 100N-S of impulse per launch. Design candidates for cubesat
deployment systems included a spring-based system, a magnetic rail-gun or linear actuator
based system, a pneumatic system utilizing a compressed fluid or cold gas to accelerate the
satellite, and a small hobby solid rocket thruster. Of these, the first was discarded from
consideration due to the extremely high spring constant (k = 4000N/m) required to achieve the
required delta-v, as well as the highly nonlinear nature of the applied force and acceleration load
(20g peak) on the cubesat. The spring constant and acceleration values given above result from
the application of the basic particle kinematic equations and Hooke's spring law, assuming a
10kg cubesat to which is imparted a 5m/s Delta-V relative to the carrier over a 10cm distance
(spring length). The solid-booster option was also ruled impractical due to possible damage to
cubesat and carrier resulting from ignition of the booster with the cubesat still inside the carrier.
Additionally, the solid booster issue is complicated by the low duration of the thrust and the
corresponding difficulty in correctly pointing the cubesat in the desired direction of travel using
26
only the onboard CMGs, which may not be able to adjust attitude quickly enough to accurately
position the satellite. Magnetic actuation was favored during the initial phases of the design;
however, the high power requirements (in the kW range) and the prohibitive tube length to get
appropriate delta-V values meant that it too had to be discarded. Currently, the planned
deployment method is to accelerate the carrier into a circular cubesat orbit and to separate the
satellite using a very low delta-V provided by a small spring. Once separation has occurred, the
carrier would then return to its previous elliptical orbit. This method requires an additional 10m/s
of delta-V budget be allocated to the carrier for each cubesat that must be launched. However, it
has none of the logistical and practical concerns of the alternative methods outlined above, and
allows for the mass of the cubesat to be minimized.
2.3.CubesatOnboardPropulsion
Cubesat propulsion system selection for this mission is based on the top level design
constraints explained in the mission profile. More specifically, the need for a propulsion system
is due to the satellite lifetime requirement of one year, as well as maintaining position in the
constellation and tumble recovery. Due to performance constraints on the imaging system of the
satellites stemming from the top-level image fidelity requirement, an altitude range from 400 to
500 kilometers is was initially considered, and it is over this range that the following cubesat
propulsion trade studies were conducted. For satellites orbiting in this altitude band, the major
detriment to orbital lifetime is atmospheric drag. The drag on a satellite is proportional to both its
drag coefficient, which is determined by the shape of the satellite, and its frontal area (defined
for this study as the area projected onto a plane normal to the orbital velocity vector). Integration
of the equations governing atmospheric drag result in a projection of orbital lifetime of satellites
based on insertion altitude, drag coefficient, and frontal area. The functional requirements
27
associated with the predicted "worst case" solar panel frontal area of 0.15m2 and Drag
Coefficient of 2.2 are 7.52N of Drag Force at 500km altitude and 29.68N at 400km altitude,
based on assumed densities of 7.55*10 -12 kg/m 3 at 400 km and 1.80*10 -12 kg/m 3 at 500km .
These density values were taken from Space Mission Analysis and Design, and correspond to
conditions during a solar maximum, which will be appropriate for a launch window around 2012.
These forces acting over the projected satellite lifetime of one year result in a drag delta-v of
23.7 m/s at 500km and 92.1 m/s at 400km. Please see Appendix B for supporting calculations.
Assuming a desired lifetime of 365-400 days, possible insertion altitudes and frontal areas cover
a range of values, which are shown in Figure 2.3.1 below.
Due to power generation requirements, it is likely that the satellites average frontal area
will be 0.15 m2 or less. This value corresponds to the maximum point of the figure along the Yaxis. With this frontal area range, an assumed mass of 10kg, and an assumed attitude averaged
drag coefficient of 2.2, the ballistic coefficient of the satellite ranges from 60.5 to approximately
200. The graph above represents the worst-case scenario of the lowest possible ballistic
coefficient, corresponding to a frontal area of .15 m2 and using the density information taken
from SMAD for the case of a solar maxima. The chosen design altitude is 520km, which is
outside of the range considered in the initial study, indicating that propulsion would not be
28
required to maintain altitude for a one-year period. The original proposed solution to counteract
the orbit degradation was to use a nano-satellite scale low Delta-V propulsion method to
maintain altitude for at least one year. Secondary advantages of such a system include orbit
maintenance and station keeping relative to the other satellites in the constellation. The proposed
propulsion system will be unidirectional and balanced with respect to the center of gravity of the
cubesat. Directional control of the thrust will be accomplished by slewing the satellite using the
included inertial attitude control system. However, following calculations performed to
determine the rate of altitude loss, the propulsion system will be now used exclusively for
station-keeping purposes, as the total altitude loss at the chosen 520km altitude over the course
of a yearlong mission is minimal. The loss is approximately 600m, and is constant for the entire
constellation, so that the arrangement of the cubesats in the constellation will not change with
time from this effect. This altitude drop will not compromise the mission, and so will be
neglected for the propulsion system. Please see Figure 2.3.2 for exact altitude change over the
selected range. Additionally, contributions from solar radiation pressure and the non-spherical
shape of the earth were calculated to total less than 2 m/s of drag based delta-V per year , and
were therefore ignored in the preceding analysis.
29
In order to determine the best type of propulsion system for this mission, several
candidate systems were evaluated based on the following aspects: cost, thruster mass, propellant
type, tanks, supporting hardware, thrust required, power required, and commercial off the shelf
(COTS) availability. After narrowing the possibilities due to mass and power constraints, the
most attractive options from a preliminary perspective are the micro-PPT, monopropellant
hydrazine, and cold-gas thrusters. A summary of the pertinent characteristics of each thruster
candidate is given below in Table 2.3.1.
30
Cold Gas
Bi-propellant
Mono-propellant
Micro-PPT
Specific
Impulse (s)
65-70
280-465
220-235
300
Propellant
N2
Iridium Catalyzed
N2H4
Teflon
Power Required
8W
17 W
15W
2.25 W
Thrust Range
0.02-1 N
4-500N
0.5-400N
20N
Thruster Mass
0.12-0.2kg
0.35kg
0.5kg
0.1kg
Examples
DASA CGT-1
EADS Astrium
S10-13
NorthropMRE-01
AFRL Prototype
Cold-gas is on the edge of what is possible given the packaging constraints of a 2 cubesat
unit by 3 cubesat unit chassis. However, a cold-gas-resisto-jet hybrid using liquid propulsion
suffers from extremely low specific impulse. This leads to an increase in propellant mass that
compromises the mass budget of the spacecraft.
Alternatively, a monopropellant hydrazine arrangement was the second design candidate
considered for this application. This candidate requires far less propellant mass than for the cold
gas option; however the size of the thruster itself becomes an issue. The Northrop Grumman
MRE -01 monopropellant thruster may be taken as a representative case. This thruster features a
nominal thrust rating of some 0.8 N, or 800,000N with a specific impulse of 216s.
Disadvantages of a monopropellant system include the weight of valves and other supporting
hardware, and the relatively high power requirements. The MRE-01 requires 15W of power to
operate which is quite high given the power constraints dictated by the size of the satellite.
31
The third likely option for propulsion application to the cubesat problem is a Micro
Pulsed Plasma Thruster, or Micro-PPT. Currently under development at the Air Force Research
Laboratory, these devices are tiny, self-contained versions of the conventional PPTs that have
been in service since the 1960s. The mass of the AFRL prototype Micro-PPT is 250g, which is
considerably less than the 0.5kg mass of the MRE-01. Supporting system mass for the MicroPPT consists only of the capacitor and control systems, of which the masses are included in the
table above. The valves/plumbing/control systems needed for any liquid propulsion option are
therefore added mass to the values given in the Table 2.3.1. Thrust levels are quite low for
micro-PPTs, in the range of 20N. The Micro-PPT system in use at AFRL uses a nominal power
of 2.25W during cycling, during which the capacitor is fed voltage from the spacecraft bus and
discharges at a rate of 2Hz. Concerns with this design include pricing given the fact that a
proprietary system must be developed and manufactured specifically for this mission. Initial
development costs have been estimated at approximately $10,000/thruster. This is based on an
estimate of the cost of developing the system divided by the proposed number of units to be
manufactured. However, this does not include costs for lifecycle testing, optimization, etc. Costs
may be cut due to the research done by the AFRL from 1999-present on the application of this
technology. This technology is currently at a Technology Readiness Level (TRL) of 6 based on
its current status of prototype systems being tested in a space environment. Additionally,
commercial off the shelf options comparable to the AFRL system are currently available from
Busek, Inc.
Due to the design advantages and disadvantages outlined above, a proprietary system
consisting of two or four Micro-PPTs spaced symmetrically about the satellite's center of gravity
was initially proposed herein as a solution to the station keeping/orbital degradation problem.
32
The reason for using multiple thrusters was that redundancy may be attained at a low mass cost,
since the thruster bodies themselves are so light. Additionally, the required duty cycle of any one
PPT decreases linearly with additional thrusters, so that each thruster would need to be fired less
frequently and would consequently have a longer lifetime. This type of system would be easy to
package due to the nature of these devices, robust (assuming multiple thrusters), and capable of
maintaining the proper orbit for the proscribed lifetime. The packaging of four 7 cm long
thrusters is considerably easier than a single 27cm thruster. The volume dimensions given above
were calculated from the known propellant mass based on an approximate specific impulse of
some 500 seconds and the Tsiolkovsky Rocket Equation, as well as the known density of the
Teflon fuel, 2.2 g/cm 3 . Additionally, if the thrusters are of the self-triggering variety, only a
single Power Processing Unit and Capacitor must be used, further reducing the volume occupied
by the system. This recommendation is of a provisional nature due to the lack of testing and
unknowns regarding cost of development of such a system. If, after a detailed cost analysis is
performed, such a development cycle is deemed to be prohibitively expensive, the fall-back
candidate is a small hydrazine monopropellant thruster such as the MRE-01.
Finally, station-keeping is necessary in order to keep the cubesats in their desired
positions in the constellation. The amount of propellant required for station keeping is
determined by the degree of accuracy of the GPS positioning system. This system has a position
accuracy of 5 meters and a velocity accuracy of 0.1 m/s. Calculation of position correction
behavior is based on a worst-case estimate of drift of the satellite's position using orbit
propagation code written in Matlab. Supporting code may be found in the Appendix C-3. From
this code, the amount of time between correction burns and the impulse needed for each burn
may be determined. The propellant needed for this function is relatively low due to the lack of
33
active forces on the satellite. Assuming a correction burn occurring when a particular satellite
deviates more than 100 meters from its designated position, the amount of propellant required for
one year of operation is 23.71 grams.
Figure 2.3.3: Deviation in Orbital Position vs. Time for GPS Uncertainty 5m, 0.1 m/s
34
3. ImageAcquisition
3.1LensandOpticsSystem
Choosing a lens and optics system for this mission is based on many controlling factors.
The requirements for this mission state that the images acquired should have at least a 3 meter
resolution at the nadir pointing, and be at least 5 km on each side. Furthermore, the satellite
altitude, mass, and volume must be taken into account. Because of the specific resolution
constraints, satellite altitude was the primary driver in this design.
The required diameter of the telescope will play a major role in determining the structural
design of the satellites. A point object imaged by an optical system with aperture diameter D,
produces an image of finite angular spread
where is the wavelength of light. For this application, is chosen to be in the middle of the
optical spectrum ( = 530 nm). To be able to resolve an object that has a finite view angle
from the optical system, we must have
distance of
, so
depending on altitude.
This relation is plotted on Figure 3.1.1. For the chosen altitude of 520 km, a 9 cm diameter
telescope was chosen because of the ready availability of this size off-the-shelf.
35
The focal length of the optical system must be chosen to give the required scale on the
image recording device, or CCD in our case. As explained in the Image Acquisition section, the
image spot produced for a 3 m object should cover at least one pixel. The spot diameter
produced by an object of dimension s at range h is:
that the spot covers 1.5 pixels (
The initial 50 Megapixel CCD that was chosen had a pixel size of
. A plot of the
required focal length vs. altitude h is shown in Figure 3.1.2. For the design altitude
, we must have
readily available Maksutov telescopes, a magnifying element would have to be added behind the
primary mirror in order to extend the effective focal length.
Upon further investigation, it was found that a 10 Megapixel CCD with a 4.75 pixel size
would be sufficient to meet these requirements. Figure 3 shows the effective focal length in
relation to satellite altitude considering the 10 Megapixel CCD. Note in figure 3.1.3, the spot size
is assumed to be 7.125 m, or 1.5 times the pixel size.
37
From this graph, it can be seen that the effective focal length at an altitude of 520 km is
1.23 m. This effective focal length is much closer to the effective focal lengths provided by off
the shelf telescopes which meet our volume constraints.
The CubeSat can be anywhere from 10 - 30 cm long, but the effective focal length is
much longer. In order to reduce this length, a two- or three-mirror telescope can be used. The
three-mirror designs (Three-Mirror Anastigmatic (TMA)), are more compact lengthwise and
account for coma and spherical aberration, as well as stigmatic and chromatic aberration.
However, two-mirror systems require less mass, both in terms of mirror mass and structural
support mass.1 They also have fewer parts, cost less, and are easier to calibrate mirror alignment.
For simplicity, the two-mirror system is preferred.
38
Mission requirements state that the optical system shall have less than or equal to a 3m
resolution (nadir pointing). Figure 3.1.5 illustrates the relationship between the required
resolution of the telescope (3m nadir pointing) and the altitude of the satellite.
39
40
approximately 3.25 m @ 520km altitude) and a mass of 1.68 kg. The advantage of Orions
telescope is its cost. The optical tubes can be ordered separately from unnecessary hardware
such as eyepiece lenses, tripods, and star mapping for a price of $230. The estimated costs for
330 units are approximately $75,900. Due to the extreme cost reduction compared to the
Questar telescope, it is recommended that the resolution requirement be relaxed.
One aspect which should be considered in preserving image quality is the thermal
expansion of materials. Depending on the working range of the telescopes, thermal expansion
could greatly hinder the image quality due to mirror misalignment and defocusing. Choosing
specific materials for the lenses of the telescopes would be unreasonable because the
manufacturing of lenses is the most difficult and expensive part of constructing a telescope.
However, choosing specific materials for the optical tube is not out of the question. Most optical
tubes for telescopes are aluminum. The density of aluminum is approximately 2700 kg/m while
the coefficient of thermal expansion is 23 m/m-K. Assuming a working range of approximately
70 K, this would result in an expansion of approximately .48 mm in the axial direction, and .46
mm in the hoop direction. This increases the telescope diameter by .14 mm, which can interfere
with both image quality and structural design. On the other hand, for approximately $300 per
unit, the optical tubes can be upgraded to carbon fiber tubes. Carbon fiber has a density range of
approximately 1700-2100 kg/m and a coefficient of thermal expansion range of approximately
5-15 m/m-K. This is beneficial because it results in less expansion in the telescope throughout
the working range, and a lighter construction.
41
3.2ImageCapture
The imaging recorder is a critical part for the success of the mission of the spacecraft.
The image recorder will need to be precise in capturing the photo so that the resolution meets the
necessary requirements along with making sure that there is no blur from the speed of the
spacecraft.
For the optics, one of the main factors is the resolution of the photograph. The Request
For Proposal (RFP) states that the photograph must have at least a 3 meter (m) resolution. Thus,
with a digital image recorder such as a Charge Coupled Device (CCD), one pixel will need to be
able to capture an object that is 3m in length and width. For a clear picture, we want to be able to
distinguish a 3m object from that of an object that is smaller. Thus, a 3m object should be able to
be imaged by multiple pixels.
In order to determine the size of the CCD that is needed, the size of the photograph area
will also need to be taken into account. It is stated in the RFP that the photograph must have an
area of 5km x 5km. Using the size of the photograph area and the resolution size, the following
equation is used to determine the amount of pixels needed for the CCD,
N pix =
Eq. 3.2.1
This equation gives the number of pixels along one side of the CCD for a CCD that takes
pictures in black and white. In order to capture a color image (which is requested in the RFP), it
was believed that this number would need to be tripled (red, green, and blue) to allow for the
colored pixels. Thus, it was thought that at least a 30 megapixel CCD was needed for the mission
and our initial choice in a CCD was a 50 megapixel produced by Kodak.
42
In actuality, the calculated number of pixels for the black and white picture was closer to
what would be needed to capture the color picture. After investigating algorithms used to convert
the picture from a mosaic of red, green, and blue, it was found that smaller CCD would be
adequate for our mission.
By looking at a Bayer Demosaicing algorithm
(http://www.cambridgeincolour.com/tutorials.htm), it could be seen for a simple 4 by 4 pixel
arrangement of red, green, and blue pixels, a 3 by 3 picture could be made. This is demonstrated
in the figure below.
The algorithm takes the average of 2 by 2 mosaic pixels to generate one real color pixel. The
picture above shows how a 4 by 4 mosaic of pixels can be split up into 9 total 2 by 2 sections to
yield a 3 by 3 real picture from the 4 by 4 mosaic pixels. This type of algorithm is just one of
many that are in use today but gives a good basis as to the number of pixels that will be needed
for our mission. From this it can be said that at least 1820 pixels per side is required for our CCD
based on
43
N pic = N pix 1
Eq. 3.2.2
This is only one of the parameters that affect the choice of the CCD.
Another parameter for choosing the size of the CCD is the contraint that is placed on the
size of the pixel from the telescope. Because of the dimensionsal constraints based on our
telescope, we need a small focal length in order to be able to see the 5 km by 5 km area as stated
in the requirements. The focal length for the Maksutov telescopes is 1.2 to 1.3 meters and it
would be best if the CCD was able to match these numbers so that no extra magnification
devices would be needed. Using the following calculation
d = f *
Eq. 3.2.3
where d is the size of the image, f is the focal length, and is the angle subtended by the object
(3m), it was found that for a focal length of 1.3 and a phi of 5.8*10-6 (3m/520km), the spot size
on the CCD would be 7.54 micrometers. In order to obtain a clear picture of the 3 meter object,
we would like the spot size to be seen by more than one pixel. Thus if the spot size were to cover
1.5 pixels, the actual size of the pixel would be about 5 micrometers. With the constraint of the
pixel size being less than 5 micrometers and the number of pixels being at least 1820 pixels, we
were able to better choose our CCD.
Based on the constraints listed above, a reasonable choice for the CCD is the Kodak KAI10100 (Figure 3.2.1 below). This CCD is a 10 megapixel CCD that 2840 by 3760 in pixel count.
This CCD also has a pixel size of 4.75 micrometers. Although the cost of this CCD has not been
obtained yet, it should be less than $3,000 per unit (the price of the 50 megapixel CCD). The
CCD has a readout rate of 30 MHz. The noise for the CCD is 10 electrons, which is less than that
44
of the 50 megapixel CCD allowing the shutter times calculated to also be valid for this CCD.
Overall, this CCD meets the requirements of the mission as stated before.
Figure3.2.1:PictureofKodakKAI10100CCD
Along with the CCD, a series of components will work to transfer the output of the CCD
to a digital picture. The first of these components is an analog to digital converter. A 12-bit A/D
converter will be needed based on the dynamic range of the CCD. Analog Devices part AD9949
(website http://www.analog.com/en/index.html) will fit our requirements. The converter has a
low power dissipation, 320 mW, and should fit easily into our power system. The price of one
A/D converter is $7.20.
After the picture has been digitalized, the picture must be transferred from a mosaic to a
picture of real colors. This is done in the manner as explained above. Along with the mosaic
algorithm, the image processor can also implement noise reduction and image compression.
Thus, a separate processor will be needed to do the demosaicing and image reduction. The Texas
Instruments TMS320C6457 is satisfactory for this purpose. This processor is rated to 8,000
MIPS, which was a value given to me by Micheal Bernhardt (computer systems) as an adequate
45
number of iterations needed for processing the image. The processor also uses a low voltage, 3V,
and has a wide temperature operating range (-40C to 100C).
The data size created by the 10 megapixel camera can reach 100 megabits. To be on the
safe side, we would like to be able to save about 8 pictures at a time on the satellites and thus a 4
gigabyte storage capacity will be used on the satellites. Solid state storage is preferred, so that
there is no reaction moment when a hard drive spins up. A four gigabyte flash drive is
satisfactory.
3.3ImageOptimization
In order to determine the shutter time needed to capture the picture, the amount of visible light to
reach the CCD needed to be found. The following equations were used to calculate the intensity
of light on the CCD. The irradiation of earth by the sun on the visible spectrum is:
w
m2
Eq. 3.3.1
because only 40% of solar emission is visible light. The intensity on the CCD is:
I CCD =
Eq. 3.3.2
where is the angle of the, is the angle that the spacecraft is taking the picture at (seen in
Figure 3.3.1), is the Earths albedo and f is the focal length of the telescope. This equation can
be simplified to,
46
I CCD =
I vis * * Alens
*b
*f2
Eq. 3.3.3
where
Figure3.3.1:LightIntensityonSpaceCraft
b = cos( ) cos 2 ( )
The time for saturation of the pixels of the CCD can be calculated as the number of electrons at
saturation over the rate of generation of electrons,
tsat =
N sat e
.
Ne
25, 000
.004
sec
=
6
6.5*10 * b
b
Eq. 3.3.4
The minimum time that the shutter needs to be open was found by taking the minimum number
of electrons needed to exceed the noise over the electron generation rate,
tmin =
N e min
.
Ne
500
.0005
sec
=
6
6.5*10 * b
b
Eq. 3.3.5
Using the equations above and setting the angle that the spacecraft is taking the picture at
to be 45 and the angle of the spacecraft off the equator to be 60, the minimum time for the
shutter to be open was found to be .002 seconds. Using the same values, the saturation shutter
time was found to be .015 seconds. The CCD uses its own electronic shutter and the controls for
47
this will need to be calibrated to fit within this time window. Work will need to be done to
calculate the rotation rate for the satellite so that the ground track does not hamper the resolution
of the photograph.
A focusing mechanism will need to be made for the camera system. A possibility for
moving the CCD very small distances is piezoelectric crystals.
Below (Figure 3.3.2) is a picture that shows a possible lay-up for the telescope and CCD
integration.
Figure3.3.2:SolidworksLayUpofTelescopeandCCDIntergration
48
4. Navigation/ControlSystems
4.1.NavigationSystem
The purpose of the navigation system is to determine the position and velocity of each cubesat in
space and propagate its orbit. Position information relative to the Earth is necessary to determine
which cubesat is closest to the target location commanded for imaging. Orbit propagation is
necessary to predict the future location of the cubesat in its orbit for station-keeping purposes.
NavigationSystem
The suggested navigation system to fulfill these requirements is the Global Positioning System
(GPS). Another option for the navigation system that was explored, but eventually dismissed,
was the use of Ground Dish Antennas. It was decided that this system should not be used for this
mission because it is more expensive, it requires a lot more antennas to match GPS coverage, is
more prone to orbit error, and is more labor intensive for scheduling, collecting, and transferring
data. GPS was chosen for this particular mission because it is the most accurate navigation
device that functions within LEO, which is where the cubesat will be operating. A GPS receiver
is a high-accuracy navigation device that obtains amplified signals from a GPS antenna (which
obtains signals from GPS satellites) and outputs data in coordinate format. The solution of the
GPS receiver includes the cubesats predicted position above the earths surface, velocity vector,
time, and date. The receiver obtains the GPS almanac (GPS satellites positions, velocity
vectors, time, and date) and GPS satellites ephemerides (highly accurate orbital parameters)
from the antenna. From the almanac, the receiver can produce a rough estimation of the
cubesats position. From the almanac and ephemeredes it can produce a very accurate
49
estimation. The receiver then calculates the cubesats present orbit and predicts its future
position in that orbit.
GPSHardware
The hardware chosen for the GPS receiver was the Cornell Cougar GPS receiver, shown in
Figure 4.1.1.
Other receivers that were considered for this mission were the GPS Navigator Receiver and the
SpaceNav GPS Receiver, but were found to be too large to fit within the cubesat. The Cornell
Cougar GPS Receiver operates with a 5 volt DC power supply at 300 mA and uses between 1.5
to 2 Watts of power. It weighs 39 grams and is 9.525cm by 5cm by 1.7cm. Its operational
temperature range is -30C to 70C. Because it was built by a university, it does not have a listed
price. However, another source indicated that a GPS receiver for high accuracy space missions
would cost around $10k. This GPS receiver has an accuracy of 5 meters, which is adequate for
this mission. Though this GPS receiver has not been space tested, simulation tests have been
performed. The simulation tests were run at altitudes from 300 km to 600 km at 7 km/s.
50
A GPS antenna is needed to receive signals from the GPS satellites, amplify those signals, and
send them to the GPS receiver. An option for the GPS antenna is the Synergy Systems SMK-4
GPS antenna, shown in Figure 4.1.2.
Another antenna that was looked into was the Toko DAX Dielectric Patch Antenna, but was
replaced because it did not provide enough signal amplification for the receiver. The Synergy
Systems antenna is 34mm by 25.3mm by 10.9mm and weighs 30 grams. The bandwidth is 2
MHz and the gain is 24dB. It operates at 11mA with a 5 volt power supply. Its operational
temperature range is -30C to 85C and costs only $25. The combined GPS receiver, antenna,
wiring, and screws, weigh about 80 grams.
DataOutput
The navigation solution of this GPS system includes the cubesats position in earth-centered
earth-fixed (ECEF) coordinates, velocity vector, GPS time, GPS week (date), and dilution of
precision. A diagram of ECEF coordinates is shown in Figure 4.1.3.
51
After the receiver is given the GPS satellites positions, velocity vectors, time, week, it sends this
information to the propagator, which is built into the receiver. The propagator calculates a
Keplerian orbit to predict where the cubesat will be at a given future time. It then sends the
cubesats predicted position, velocity vector, time, and week back to the receiver to be sent to
other subsystems. The accuracy of the velocity vector is 0.1 m/s. A block diagram of the data
output is shown in Figure 4.1.4.
52
The first run is called a cold start, where the receiver takes about 10 minutes to download the
GPS almanac (GPS satellite coordinates, velocity vector, time, week) and roughly estimates the
cubesats position. The GPS almanac data is good for several months before it must be
discarded and re-downloaded by the receiver. A warm start is performed every 4-6 hours
where the receiver takes about 3 minutes do download the GPS satellites ephemeredes and
produce a highly accurate estimation of the cubesats position The receiver uses the almanac and
ephemeredes to predict future cubesat positions using Doppler shifts. It listens for Doppler
shifted signals (shifts occurring in the electromagnetic spectrum) about the L1 (1575.42 MHz)
frequency. It outputs a navigation solution at a rate of 0.1 Hz. The output of the receiver is an
ASCII string where each character requires 8 bits. The receiver output rate is 19200 bps. As
functionality measure, the receiver outputs data every 10 seconds, even when it is not being used
or the navigation solution is invalid.
4.2AttitudeDeterminationandControl
The attitude determination and control system (ADCS) is responsible for determining and
controlling the rotational parameters of the spacecraft. It is essential for cubesat imaging,
communications, and other purposes. The ADCS must employ sensors capable of ensuring 200
meters pointing accuracy for the ground target, corresponding to an accuracy of 0.022 degrees,
and actuators that can slew the spacecraft to the desired angle at a rate of 540 deg/min.
Since high accuracy is a requirement of this mission, three-axis control mode is
necessary. Active control with system feedback will make corrections via the control system as
frequently as required to maintain attitude and position accuracy. However, passive control can
provide coarse control. There should be a backup detection system for redundancy. This backup
53
will be independent of the primary system, hence providing a full coverage of attitude detection
regardless of the position of the satellite. The backup system will also rely on analog signals and
preferably introduce minimal additional mass and volume.
Sensors
A summary of typical devices, as well as their performance and physical characteristics, are
given in Table 4.2.1.
Table 4.2.1: Typical ADCS Sensors.
Sensor
Sun Sensors
Star Sensors
(Scanners and Mappers)
Horizon Sensors
Scanner/Pipper
Fixed Head
(Static)
Magnetometer
Mass
Range (kg)
1 to 15
Power
(W)
10 to 200
0.1 to 2
<1 to 5
0 to 3
<2 to 20
1 to 4
0.5 to 3.5
5 to 10
0.3 to 5
0.3 to 1.2
<1
(Wertz and Larson, Space Mission Analysis and design, third edition)
54
Sun sensors are visible-light detectors which measure one or two angles between their
mounting base and incident sunlight. They are popular, accurate and reliable, but require clear
fields of view. They can be used as part of the normal attitude determination system, part of the
initial acquisition and recovery system, or part of an independent solar array orientation system.
3. Star Sensors
Star sensors provide high-precision measurements. Star sensors can be scanners or trackers.
Scanners are used on spinning spacecraft, whereas trackers are used on 3-axis attitude stabilized
spacecraft to track one or more stars to derive 2- or 3-axis attitude information. Even though star
trackers are very accurate, care is required in their specification and use. The Cubesat must be
stabilized to some extent before the trackers can determine where they point. This stabilization
requires alternate sensors. Also, the star tracker is sensitive to the sun, moon, and planets, so that
the sensor is blinded while it is exposed to them. Therefore, even though the star trackers have
very high accuracy, they must be used with an additional sensor type (i.e. IMU or/and sun
sensor). Because the star tracker delivers an accuracy that meets mission requirements, this
option was chosen for the subsystem design for fine pointing.
4. Magnetometers
Magnetometers are simple, reliable, lightweight sensors that measure both the direction and
size of the Earth's magnetic field. When compared to the Earth's known fields, their output helps
establish the spacecraft's attitude. This option was chosen for the subsystem design for coarse
pointing.
55
ProposedHardwareforSensors
In order to meet the accuracy of 0.022 degrees, star trackers must be used. Between the
star tracker data updates, IMU will provide the fine attitude. However, in-house star tracker
needs an initial reference and in case of blocking by the Sun; therefore, magnetometers and sun
sensors are used for coarse pointing.
For the magnetometer, a Honeywell 3-axis magnetic sensor could be used. This sensor
has an ultra-compact, size of 3.0 x 3.0 x 1.4mm3 and a mass of 25.6 milli-grams. It has a threeaxis surface mount sensor which is designed to be very sensitive so that it can measure low
magnetic field. This sensor provides wide magnetic field range of 6 gauss, and has linearity
error of 0.1% in conditions of 1gauss. This choice of sensor is shown in Figure 4.2.1.
Three orthogonally aligned sensors for three-axis measurements will be used. These will
measure the direction and magnitude of the magnetic field. Generally, only direction is required
for attitude determination.
56
For the sun sensor, AeroAstro Medium Sun Sensor could be used. This sensor provides
accuracy of 1 degree and 2-aixs attitude determination, and it has a full angle circular field of
view of 60 degrees. This sensor is shown in Figure 4.2.2.
For fine pointing with rapid update, an IMU will be used. The MICRO-ISU BP3010 from
Bec Navigation Ltd. is currently the best option available. This IMU has a reliable cost with an
accuracy of 0.5 and weights 0.03 kg. It also has a very small size, 35 mm x 22 mm x 12 mm. It
has a worse case drift rate of 0.1 degrees/seconds. This drift rate can be compensated by position
update from GPS and angle update from star tracker. This choice of IMU is shown in Figure
4.2.3.
http://www.becnav.co.uk/imu.html?gclid=CN69nNXXlJkCFRwpawodP2Q8Zw
57
Star trackers will be used for fine pointing because they can provide the attitude accuracy
required: at least 0.022 degrees during image acquisition. The star tracker uses the following
process: the camera takes the image of a field of bright stars and the computer goes through an
algorithm to identify the star pattern with a star catalog in memory to determine the attitude of
the satellite. Most star trackers in the market have volume, weight, cost, or power consumption
that is too large for this Cubesat mission. A comparison of star trackers is shown in Table 4.2.2.
Table 4.2.2: Comparison of startrackers.
Company
Model
AeroAstro
Miniature Star
Tracker
A-STR
Galileo
Avionica
Ball
Aerospace
EADS
Sodern
SunSpace
Size
[mm]
5.4 x 5.4 x 7.6
Weight
[kg]
0.425
Power
[W]
<2
Accuracy
[arcsec]
70
Rate
[Hz]
~1
8.9 ~ 13.5
< 10
10
< 5.5
8~9
< 10
10
SED 16
Dia. 203.2
H = 198.12
170 x 160 x 290
< 3.0
7.5
< 15
10
Sun-Star
1.63
< 3.5
22 (1)
CT-602
As shown in Table 4.2.2, most of the star trackers are not suitable for the cubesat. Only
AeroAstros Miniature Star Tracker meets requirements but has a large cost (~$200,000).
Therefore, independent design of a star tracker is currently under consideration.
DesigningStarTracker
SizingcameralensandCCDchip
The following is a summary of analysis for sizing a camera lens and CCD chip.
Calculations and results are provided in Appendix D.2.3.
With lens equation, the focal length of lens can be determined. For example, if the acquisition
pointing accuracy is 20 arcsecond and the pixel size of 10 m, the focal length of lens would be
58
about 100 mm. Also, the number of pixels on the CCD can be determined with an equation for a
reasonable size of field of view (FOV: the angle of exposure field for the CCD) chosen with
assumption of one pixel corresponding to the accuracy of pointing.
The results indicate that any range of pointing accuracy and field of view would possibly be
satisfied by the pixel size of the existing small digital camera. Therefore, a choice must be made
between picking a large field of view or higher resolution of CCD. These results are summarized
in Figure 4.2.4.
To collect enough starlight, the diameter of the lens must be sufficiently large. The
diameter of the lens can be obtained from the equation of the light power, the light power that
collected by the camera lens from a star with assumption of all photons hit a single pixel and
59
CCD quantum efficiency of 0.25 and the minimum number of electrons of 400 electrons.
Detailed calculations are presented in Appendix D.2.3 Diameter of Lens section. Figure 4.2.5
shows the exposure time versus the diameter of the lens.
As seen in Figure 4.2.5, as exposure time decreases, the diameter of the lens increases.
The lens focal ratio can be obtained with the focal length of the lens and the diameter of
the lens. Figure 4.2.6 shows the pointing accuracy versus the lens focal ratio.
60
As shown in Figure 4.2.6, for higher pointing accuracy shorter exposure time corresponds
to smaller lens focal ratio. The practical focal ratio would be 1.4 because we can obtain a
shorter exposure time at 1/60 sec with a pointing accuracy of 0.013 degrees, a focal length of
45.5mm and a lens diameter of 3 cm, and the total pixels on CCD is 1.42 Megapixels with field
of view of 15.
StarPatternRecognitionAlgorithms
Some researches have been done for the possible choices of algorithms. There are two
similar but different kinds of algorithms to recognize a star pattern: oriented triangles algorithm
and grid algorithm. Both algorithms would provide high accuracy.
The oriented triangles algorithm, first selects a star as a pivot star and two closer stars as
neighbor stars from the image as shown in Figure 4.2.8. The computer initializes the stars in the
image by giving local numbers, and then generates a list of the potential triangles. The stars
given local numbers 1, 2, and 3 become the local triangle, and the surrounding stars become the
reference triangle. Using the distance between stars, the computer algorithm rebuilds the
61
constellation then measures the distance between stars to verify the local number then compare
with star identification number.
(a)
(b)
(Rousseau, Bostel, and Mazari, Star Recognition Algorithm for APS Star Tracker: Oriented Triangles)
(c)
(d)
The grid algorithm is similar to the triangle algorithm but it uses polar coordinate to
construct the pattern. From the image, a reference star is selected as a pivot star and is identified
from the given database. The position of the pivot star becomes the center of the circle with
pattern radius, rp . The surrounding sky is then partitioned over this circle. Then the pivot star is
translated to the center of the FOV, and the related reference stars are translated in the same
distance as the pivot star. Then an alignment star is oriented to the reference frame so that the
related reference stars are rotated at the same angle; then a grid pattern is constructed. The
constructed pattern is compared with patterns in the database for identification. Figure 4.2.9
G. L. Rousseau, J. Bostel, and B. Mazari, Star Recognition Algorithm for APS Star Tracker: Oriented Triangles, IEEE, February
2005
62
shows a basic definition of a coordinate system on the left and the principal of a grid pattern in
polar coordinates on the right.
(LeeandBang,StarPatternIdentificationtechniquebyModifiedGridAlgorithm)
(a) DefinitionofinertialandCCDbodyframes(b)Gridpatternispolarcoordinate
Research has been done by AeroAstro and the MIT space systems laboratory, for
developing a coarse star tracker, showed that this algorithm is more efficient because it has the
potential to reduce the power drawn by processor to factor by 10. This algorithm was claimed to
be more accurate.5
Through research, a technique was found that would increase the accuracy- hyperacuity
technique, called subpixel precision. In a focused image, a star can appear as a point (e.g. one
pixel). If we defocus the image slightly, the star will spread into several pixels, so the
measurement of the center of the star will be more accurate for distance calculation. In this way,
4
5
H.LeeandH.Bang,StarPatternIdentificationtechniquebyModifiedGridAlgorithm,IEEE,VOL.43,NO.3,July2007
R.ZenickandT.J.McGuire,Lightweight,LowpowerCoarseStarTracker,17thAnnualAIAA/USUConferenceonsmallSatellites
63
the determination of the position of a star is more accurate than using one pixel. Figure 4.2.7
shows the hyperacuity technique. However, this technique would take a lot of iterations.
(Liebe,StarTrackersforAttitudeDetermination)
InhouseStarTracker
A prototype in-house star tracker has been designed. The camera chosen is the
EdmundOptics' EO-1312m 1/2" mono CMOS USB Lite. This camera provides resolution of
1280 x 1024 pixels and uses USB 2.0 board. The included software provides the images in JPEG
and Bitmap file format. The best lens for this camera has been chosen to be the EdmundOptics'
25mm Megapixel fixed focal length. The size of camera is 4.4 cm x 4.4 cm x 2.54 cm, and the
lens has diameter of 3.35 cm and height of 3.6 cm. The lens has a field of view of 14.6 degrees,
focal length of 25mm, and focal ratio of 1.4. The focal ratio of 1.4 was chosen for capturing
enough light into CMOS, so it would be wide enough for least 3 magnitude of 5.7 stars. Camera
and lens are shown in Figure 4.2.11.
C.C.Liebe,StarTrackersforAttitudeDetermination,IEEE,June1995
64
The algorithm for the star tracker was designed by Professor A.T. Mattick of the University of
Washington. The 3500 brightest stars are used as a star catalog for the oriented triangles
algorithm. Figure 4.2.12 shows the schematic of in-house star tracker.
As seen in Figure 4.2.12, the camera takes a snapshot of the star field and saves this image as a
bitmap file. Then a program reads the image file and finds the locations and brightnesses of
pixels illuminated by starlight and saves them for the main algorithm program. With initial
reference (right ascension and declination) of pointing axis of camera, the algorithm program
looks through the star catalog to find all the possible stars within the angular uncertainty of 10
degrees. Note that uncertainty of the inputs initial reference should be with in 10 degrees. Then
the program tries to match 3 of these stars to 3 stars from camera image with the brightness, ratio
65
of angles and distance. Then it does a least-squares fit using CCD pointing, minimizing the
difference between the positions of predicted and actual illuminated pixcel positions to determine
the pointing axis of the camera and rotation angle of camera about that axis. This rotation angle
will be used by the control system for pointing acquisition. This algorithm provides an accuracy
of 10 arcsecond in camera pointing axis and about 80 arcsecond in rotation of camera axis.
Figure 4.2.13 shows the inertialandCCDbodyframesofaninhousestartracker.
Figure 4.2.13InertialandCCDbodyframesofinhousestartracker
As seen in Figure 4.2.13, the inertial body frame is using the geocentric frame. is declination
(Dec) in degrees, is right ascension (RA) in hours, and is rotational angle of CCD plane. The
u, v, and w are unit vectors. w is the direction of initial reference. CCD plane is parallel to the
X-Y plane and perpendicular to w. The angles (1,2,3) and the distances (d1, d2, d3) will be
measured and used for the star identification process in the algorithm.
Further tasks would be determining whether it will be sharing a CPU with the main computer or
have a separate one integrated into the control system.
66
Actuators
1. Thrusters
If the orbital maintenance thrusters outlined in the propulsion section are also used for
control, it must be considered that 3 to 10%7 of the total propellant mass would be consumed for
attitude control purposes. In order to control the attitude of the cubesats, a system of thrusters
must provide the attitude maintenance requirements outlined in Table 4.2.3.
Table 4.2.3: Attitude Control Requirements for Space Propulsion
Attitude Control
Acquisition of Sun, Earth, Star
(Wertz and Larson, Space Mission Analysis and design, third edition)
6
Wertz,JamesR.,Space Mission Analysis and design, Space Technology Library, third edition)
67
reaction/momentum wheels is that they saturate rapidly and they need to be desaturated
periodically, which can add complexity and difficulty to the process. Therefore, this option was
not considered as the final solution.
3. Control Moment Gyros
Control moment gyros are single- or double-gimbaled wheels spinning at constant speed.
Control systems with control moment gyros can produce large torques about all three of the
spacecraft's orthogonal axes. However, they require a complex control law and momentum
exchange for desaturation. Almost all the CMGs found so far exceed the weight tolerance of this
mission. A suitable CMG is currently under development at Andrews Space Inc.
4. Magnetorquers
Magnetorquers use magnetic coils to generate magnetic dipole moments. Magnetorquers
can compensate for the spacecraft's residual magnetic fields or attitude drift from minor
disturbance torques. They produce torques proportional and perpendicular to the Earth's varying
magnetic field. Because they use the Earth's natural magnetic fields, they are less effective at
higher orbits.
A summary of possible actuators, as well as their performance and physical
characteristics are given in Table 4.2.4.
Table 4.2.4: Typical ADCS Actuators.
Actuator
Thrusters
Hot Gas (Hydrazine)
Cold Gas
Reaction and
Momentum Wheels
Mass
(kg)
Power
(W)
~2 to 20
10 to
110
>10
Magnetic Torquers
1 to 4000 A.m2
0.4 to 50
90 to
150
0.6 to 16
(Wertz and Larson, Space Mission Analysis and design, third edition)
The moments of inertia for the Cubesat were determined first in order to start the calculations
for sizing the CMGs. For these calculations, the center of mass is assumed to be at the center of
the body with uniform density. In order to find the required moment of inertia for the CMG, the
required slew rate is determined. Since the cubesat will accelerate for half of this angle and
decelerate for half, only half of the angle and time are used in the calculations. The angular
acceleration of the Cubesat can be determined. The calculations were completed for a range of
slew rates from 1 - 15 per second. Then using the moments of inertia of the Cubesat and the
angular acceleration, the required wheel torque is determined, and then the required momentum
of the wheel.
69
The CMG cluster for an x-axis maneuver is shown in Figure 4.2.14. The same convention is
used for the other two directions.
(Lappas,Steyn,andUnderwood,AttitudeControlforSmallSatellitesusingControlMomentGyros)
Once the angular momentum is determined, assuming an angular velocity of 60,000 rpm
for the DC motor (the specifications of the DC motor are presented in the Appendix), the CMG's
required moment of inertia is determined. Note that this DC motor is for the wheel only. A ULT
Applimotion frameless motor (shown in the Appendix) is going to be used for the gimbals.
The required moments of inertia for four cases are of slew rates, 9 deg/sec, 8 deg/sec, 7
deg/ sec, and 5 deg/sec, with three different cases of 6 rad/sec, 10 rad/sec, and 25 rad/sec
maximum gimbal angle rates are calculated and summarized in Appendix D.2.4.
70
As an example, the required moments of inertia for the maximum gimbal angle rate of 6
rad/sec are plotted in Figure 4.2.15.
Figure 4.2.15: Required Moment of Inertia Vs Slew Rates for a Maximum Slew Rate of 6 rad/sec
An important factor in deciding the slew rate is the time it takes for the CMG to reach
that slew rate with assuming a torque of 0.003 N-m, from Andrews, the time in each direction is
determined. The summary results are plotted in Figure 4.2.16.
71
Figure 4.2.16: Slew Rate vs Time it Takes to Reach to that Slew Rate
It is observed that the higher the slew rate, the longer it takes to get to that slew rate.
Then in order to size the wheel and find the diameter that can produce the required slew rate, the
moments of inertia for a brass wheel are calculated for four different combinations, with density
of 8400 kg/m^3 and 8700 kg/m^3 and two thicknesses of 0.005 m and 0.0025m. The diameter is
varied from 0.01m to 0.07 m. As an example, the results obtained from a density of 8400
kg/m^3 is plotted in Figure 4.2.17.
72
Now, the possible moments of inertia obtained from the wheel are compared to the
required moments of inertia, and it can be decided which diameter for the wheel is necessary.
To sum up, from these calculations it is observed that for 9 degrees per second, the
longest time to reach the desired slew rate is about 6 seconds along the 0.1m side. It looks like a
wheel with a diameter of 0.03m and thickness of 0.0025 m can produce the worst case MOI
required.
A simple simulink block diagram can be designed to model the dynamics of the DC
motor, which was chosen to run the CMG wheel, as shown in Figure D.2.5.1. Running Simulink
model with the motor parameters, given in Appendix B.1, and a voltage input of 1 volts, an
output shown in Figure 4.2.18 is obtained. As shown, with 1V input an angular velocity of about
850 rad/sec or 8116.9 rpm is obtained.
73
DesigningtheMagnetorquers:
As mentioned before, the of-the-shelf magnetorquers found in the trade study did not fit
the requirements for this mission. Therefore, the team designed square plates with wires wound
up around it to be used as magnetorquers. For these calculations, the earth magnetic strength is
taken to be about 5*10^-5 Tesla. After studying a range of different areas for the plate, it is
decided to use a plate with an area of 0.003025 m^2. This will require 318 turns of wires to
produce 0.024 mN-m torque, using AWG 26 copper with a diameter of 4.06*10^-4 m and a
resistivity of 1.7*10^-9 ohm-m. For these calculations, current of 1 Amp and power of 1 Watt
are assumed.
A summary of the proposed hardware that has been located so far and their physical
properties is provided in Table 4.2.5.
Table 4.2.5: Proposed Hardware for Attitude Determination and Navigation Systems.
Part
Performance
Mass
[kg]
Size
[cm]
Star Tracker
0.0085Degrees
3-axes (3)
0.110**
IMU
Complete 6 DOF
Max update rate 64Hz
120 aguss
1Degrees
2-axes (2)
Torque of about 0.003Nm
3 wheel
Wheel motor
Gimbal motor
Torque of 0.024mN-m
318 turns of coil
0.030
Magnetometer
Sun Sensor
CMG
Magnetorquer
25.6 x 10-6
0.036
0.5
2V
None
***
0.05 x 3
0.001
0.036
~d=3, t=0.25
0.5
D=~3.5
1
0.08
Power
Required
[Watt]
<2**
74
A summary of the cost of the chosen sensors and actuators for the attitude determination
and navigation systems is given in Table 4.2.6.
Table 4.2.6: Costs Estimates and Companies for Parts used in ADCS and Navigation System.
Type
Star Tracker
Camera
Lens
IMU
Magnetometer
Sun Sensor
GPS
Receiver
Antenna
CMG
Magnetorquers
Hardware
will be designed
EO-1312M 1/2" CMOS Monochrome USB
Lite Edition Camera
25mm Megapixel Fixed Focal Length Lens
MICRO-ISU BP3010
3-Axis magnetic sensor HMC 1043
Medium sun sensor
Cornell Cougar GPS receiver
Toko DAX Dielectric Patch Antenna
Wheels: designed (3)
Wheel motor
Gimbal Motor
Designed (3)
Cost [each]
$725
Company
UW
Edmundoptics
$195
$700
$45
$5640
Edmundoptics
Bec Navigation Ltd.
Honeywell
AeroAstro
<$20000
Cornell Cougar
undetermined
UW
Micromo Electronics
Applimotion, Inc.
UW
undetermined
ADCSDetails
75
Figure 4.2.19: Frame of Reference (Wertz and Larson, Space Mission Analysis and design, third edition)
As shown in the figure above, the ADCS sensors send their inputs to the convertor as
analog signals. The convertor then converts these signals to digital signals. These are sent to the
processor, which contains the control algorithm and process the data. Then the output is sent to
the pulse width modulator and then to the actuators to execute the command.
76
The control model must meet some functional requirements. First it shall fulfill
positioning requirements of solar cells, communications, and payload. Second it shall be able to
counteract the maximum external torque of the environment at the selected orbit at 500 km.
From the calculations, shown in Appendix D.2.1, the worst-case disturbance torques are
determined and summarized in Table 4.2.7.
Table 4.2.7: Worst-Case Disturbance Torques.
Disturbance
Torque [
Note
Gravity-gradient
Solar Radiation
Aerodynamic
Magnetic Field
77
Slew mode shall operate as a transition mode from sun pointing to earth pointing and back again.
Contingency mode shall operate in emergencies if normal mode fails. This mode shall use less
power to meet power or thermal constraints.
The control design is illustrated in Figure 4.2.21.
There are several options for the control law, which relates control torque to the error
signal. One option is proportional control, which is the simplest kind and has the form
Tc = K where Tc is the control torque and K is the system gain and theta is the error signal.
Another option is the Bang-Bang Control, which has the thruster pulse as output and the
direction is determined by the sine of the error signal Tc = Tp sin . Another option is
Proportional-plus-derivative (PD) Control such as
derivative of the error signal provides damping and reduces the angular excursions. The final and
most desirable option is the Proportional-plus-Integral-plus-Derivative (PID) Control
78
In order to design the simulation, the desired attitude is defined via attitude
parameterization methods. The most common attitude parameterization methods are Euler angles
and Quaternion. For space applications, Quaternion is the preferred option in order to avoid
singularities and minimize the use of Sine and Cosine functions. In order to compensate for the
errors, a Full State Feedback controller is used. The error then is presented as shown below.
where matrix A captures the dynamics of the Cubesat, matrix B captures the dynamics of the
actuators, vector u is the input to the actuators, y is the observations made by the sensors, and
matrix C captures the dynamics of the sensors.
Then to make sure that the system is immune to the saturation in the actuators, we need to
check for stability as shown below
where Kp is the gains needed for the Full State Feedback controller to make the system stable
and it also needs to capture the efficiency factor of the controller, control saturation, and slew
rate constraints. For simulation purposes, the linearized controller can be used to simulate the
non-linear system. On the actual Cubesat, there shall be different controllers for different modes
of operations.
In order to fully determine the position and attitude of the Cubesat, the navigation system
must be integrated into the ADCS system. A possible case is shown in Figure 4.2.22.
79
Figure 4.2.22: ADCS Architecture Diagram
As it is observed in the ADCS architecture block diagram, the attitude sensors provide
vector measurements that are passed through an optimal estimator such as the QUEST algorithm
to determine a solution for an attitude estimate. This estimate is then passed to a fine estimator
such as an extended Kalman Filter, along with angular velocity measurements to obtain the
attitude solution. The attitude controller compares the determined attitude with the desired
attitude and calculates appropriate control torque to minimize the error. These torques are sent to
the appropriate actuators to exert moments on the Cubesat. The GPS sensor provides a state
vector to the orbit controller. This position information is compared with the position indicated
from the master controller, which is driven with the mission requirements.
80
5. Communications
The constellation of satellites will work as a system with several individual agents. These
agents include ground stations, CubeSat carriers and the CubeSats within the constellation. The
transmission of data is required for this mission, and the data is transmitted via communication
systems aboard each of the system agents. A properly designed system allows the time from
image request to ground reception to be minimized while preserving image quality and security.
5.1.CommunicationsArchitecture
The communication network consists of a ground station and the constellation of CubeSats.
To optimize the power usage for communication, only predetermined CubeSats will be used in a
relay chain from request to delivery. The ground station will have updated knowledge of the
constellation, including the position and velocity vectors of the CubeSats. With this knowledge,
the ground station will compute the CubeSat node order to be used in the relay chain from the
ground station, to the target CubeSat for data acquisition, and back to the ground station. Figure
5.1.1 shows this communication relay architecture. The node order is then stored into a queue
along with a data request (imaging, position, health status, etc.) and sent to the nearest CubeSat.
This CubeSat, C1, pings a confirmation signal to the ground station, dequeues the first node in
the node order queue, and relays the modified queue. The relayed signal propagates radially to
neighboring CubeSats. Several CubeSats may receive the signal, and they will each execute a
Boolean check to see if they are the intended next node. Only the intended CubeSat, C2, will
satisfy this check, and the rest will ignore the signal. C2 then pings a confirmation signal to C1,
dequeues the first node, and relays the newly modified queue. This process repeats as necessary
81
until the target CubeSat receives the signal. It then collects the requested data, appends the
collected data to the queue, and relays the newly modified queue along the predetermined return
path. The ground station then receives the queue with the requested data but stripped of the
communication node path.
C3
C1
C2
CN
Ground
Station
5.2.CommunicationsOverview
The primary factors of the communications system are the unit costs, range, and power
requirements. The system hardware consists of antennas, receivers, and transmitters. The
receivers and transmitters are often combined into a transceiver. Before analyzing the hardware,
82
the system constraints must be defined. These include constraints from the geometry of the
constellation design as well as constraints from data requirements.
There are two communication links analyzed in detail: crosslink and downlink. For crosslink
communication, the orbit geometry is derived from the current constellation design; a Walker
constellation with an inclination of 55 and an altitude of 520 km. There are four
communication paths that need to be analyzed within the crosslink communication, shown in
Figure 5.2.1.
(a)
(b)
(c)
Figure 5.2.1. Four crosslink communication paths: (a) In-plane (r1 = 4000 km), (b) Cross-plane
parallel (r2 = 1375 km, r3 = 3200 km), (c) Cross-plane ascending-descending (r4 = 5400 km)
With 10 CubeSats per plane the in-plane spacing is approximately 4000 km. There are two
scenarios for cross-plane communications: ascending/descending and parallel communication.
For both scenarios the greatest distance is at the equator. To communicate between an ascending
and descending CubeSat as they cross the equator requires initializing communication at a
distance of approximately 5400 km. There are two possible communication links for parallel
cross-plane communications, as shown above in Figure 5.2.1. As one CubeSat C1, crosses the
equator the adjacent plane will have one CubeSat, C2, phased 19.6 above the equator and one
CubeSat, C3, phased 10.4 below it. The distance from C1 to C2 is approximately 3200 km and
that between C1 and C3 is approximately 1400 km.
83
For downlink it is assumed that the ground station is a on-meter parabolic dish with tracking
capabilities to 10 above the horizon, as shown in Figure 5.2.2.
84
the image quality. Utilizing this JPEG format would decrease the required data transfer rate to
12 kbps, significantly decreasing the load requirements of the communications system.
Downlink communication operates with constraints independent of crosslink. For a circular
orbit with an altitude of 520 km, the ground velocity is approximately 7.6 km/s. For a tracking
range of 160 (10 above the horizon), this results in a maximum in-view time of 450 seconds.
Allowing for two minutes for the ground station to initiate communication with an overhead
CubeSat and two minutes for the CubeSat to transfer a single image, there are over three and a
half minutes left of access time to allow for additional communication (impulsive system health
check, navigation corrections, etc.) or other interruptions such as tracking errors. For one image
to be transmitted in two minutes requires a transfer rate of 84 kbps, seven times the data rate of
crosslink communication.
The well-known Link equation provides a relationship between the required power of the
transmitter and the constraints discussed above as shown in Appendix E. Typical values for
several terms in the equation are assumed at this stage of analysis, and are also summarized in
Appendix E. Using these estimated values, the Link equation is simplified to relate the design
parameters, i.e. transmitter power, frequency, and antenna gains. The transmitter power is
estimated to be half of the input DC power, which is targeted to be no more than 5W for
crosslink and no more than 1W for downlink. The antenna types are chosen using analysis via
the Link equation and by inspection of the necessary gain patterns for each of the links in the
communication architecture.
85
5.3.CommunicationsHardware
The driving factors for the communication hardware on the CubeSat are mass, volume and
cost. For antennas, these are met using simple designs. The uplink communication is not
constrained by these same factors and was not analyzed. Figure 5.3.1 shows the general gain
patterns that provide the desired communication performances. The crosslink communication
requires the ability to communicate with any of the neighboring CubeSats, according to the
queue of nodes. This suggests the use of monopole antennas, which have toroidal gain patterns
about the monopole axis. With ambiguity in the relative attitudes of neighboring CubeSats, a
more isotropic gain pattern is desirable. This is easily accomplished using multiple monopole
antennas arranged orthogonal to one another. The downlink communication is towards the
ground station, which only requires a unidirectional gain pattern. Though antennas such as the
patch antenna have such unidirectional gain patterns, the added hardware and system complexity
does not make them the most viable option. Instead, by arranging three monopole antennas to be
mutually orthogonal to each other, the attitude of the CubeSats becomes a nonissue for any of the
communication links, up, cross or down. Additionally, the idea to use the CubeSat carrier
vehicles as dedicated downlink nodes has been proposed. With this approach, a dedicated highgain antenna will be used along with accurate ADCS pointing to allow greater access time and a
higher data transfer rate for downlink communications. Further research and analysis will be
required for this approach.
86
U/L:Singletarget,directionalsignal
X/L:Multipleneighbors,ideallyisotropicsignal
D/L:Singletarget,maximizeflybyaccesstime
87
Figure 5.3.2. Power requirements for crosslink comm using half wave monopole antennas.
From Figure 5.3.2 it can be seen for a maximum transmit power of ~2.5W (input power
~5W) the wavelength must be greater than about 0.8 m (375 MHz) for the ascending/descending
cross-plane communication, but only about 0.6 m (500 MHz) for in-plane communication, both
of which correspond to the UHF frequency band. Similarly, Figure 5.3.3 shows the relationship
between inter-CubeSat distance, wavelength and transmission power for downlink
communication using the half wave monopole antenna. From this plot it can be seen that the
downlink antenna requirements can be met using the same antenna size as the crosslink.
88
Figure 5.3.3. Power requirements for downlink comm using half wave monopole antennas.
A particular transceiver that may meet the mission requirements is the RF DataTech
LRT470 radio module. It operates at frequencies of 406-475 MHz, which equates to a
wavelength range of 63-72 cm. Using this, however, prevents the ascending/descending crossplane communication at the equator. Figure 5.3.4 shows the communication power requirements
as a function of communication range for crosslink using the extremities of the transceivers
range. Higher frequencies help transfer data quicker but require higher power, as shown in
Figure 5.3.4. To conserve power, the low frequency, 406 MHz, will be used for communication.
As this will be the primary communication frequency, the antennas will be appropriately sized at
15.75 cm. These antennas will be tape-spring antennas, deployed from a coiled position. The
downlink communication is assumed to be the transfer of a large data packet to ground station
and will, by default, use the high frequency of 475 MHz. The power requirements as a function
of communication range for downlink communication is shown in Figure 5.3.5.
89
Figure 5.3.4. Crosslink communication power requirements using the LRT470 transceivers
maximum and minimum frequencies.
Figure 5.3.5. Downlink communication power requirements using the LRT470 transceivers
maximum frequency.
90
5.4.CentralProcessing
A central processor provides the satellite with the needed processing for navigation,
control, communication, and general system tasks. The processing throughput required for the
central processor was estimated using Table 16-13 Size and Throughput Estimates for Common
Onboard Applications in Space Mission Analysis and Design by James R. Wertz and Wiley J.
Larson. The estimated throughput required for navigation, control, communication, and general
system tasks is less than 2 MIPS. However, this estimate is subject to change as individual
algorithms for tasks are identified. An estimate of 5 MIPS to account for intensive processes
such as the star tracking algorithm provides a conservative value for analysis.
The throughput required for image processing was many orders of magnitude greater than
the throughput required for navigation, control, communication, and general system tasks.
Instead of having the central processor perform all tasks in the satellite, a separate processor will
perform the image processing. This configuration would allow the central processor to have
relatively low capabilities and, consequently, consume less energy. The image processor would
be turned off or in a sleep state when not needed. This configuration would decrease the energy
consumed at the expense of increased complexity.
The hardware chosen to satisfy a throughput of 5 MIPS is manufactured by Pumpkin, Inc.
Pumpkin, Inc. hardware was chosen because of the ease of incorporation into the CubeSat
structure. The central processor is the MSP430 series microcontroller by Texas Instruments.
The MSP430 series microcontroller has a throughput up to 25 MIPS and consumes less than 60
mW. Also, it can operate at temperatures between -40C and 85C. Pumpkin, Inc. incorporates
the microcontroller into a pluggable module. The price of the Pumpkin, Inc. pluggable processor
91
92
6. SupportSystems
6.1. Power
PowersystemsarerequiredtosupporttheothercomponentsoftheCubeSat.DisplayedbelowinTable
6.1.0isthesupportedhardwareforthemission,theiroperatingvoltage,andtheirdutycycles.Adash
indicatesunknownorinapplicablevalues.Thesumofpowerrequirementstotal20.82Watts,butthe
powerrequiredfortheentiresystemisestimatedat10Wattsonaverageasaconservativeestimate;
theCubeSatswillnotbeacquiringimagesfortheentiretyofthemission,norwilltheybeconstantly
readjustingtheirorientationortheirorbits.
SolarPower
Thepreliminaryanalysisinvolvesthedeterminationofthetypeofpowersystemrequiredforthe
mission.Themostpracticaloptionamongradioisotope,electrochemical,andsolarpowerwillbea
combinationofsolarpowerusingphotovoltaiccellsforprimarypowerduringlightsideoperationand
electrochemical(batterycells)providingpoweronthedarksideoftheCubeSatorbits.
Severalsolarcellsemiconductorsareavailableforthismission,includingGe,Si,GaAs,ortriplejunction
cellsusingGaInP2,listedbelowinTable6.1.1.
Thedeterminingfactorinmaterialselectioninthiscaseisefficiency,sothetriplejunctioncellhasbeen
chosenforthisapplication.PreliminaryresearchshowstheSpectrolabNeXtTripleJunction(XTJ)Solar
cellstogivethebestperformanceatAM0(standardspace)conditions.
93
Table6.1.0CubeSatSystemsPowerRequirements
System
Subsystem
Component
Powerrequired
Operating
Dutycycle
GNC
Navigation
GPS
2Watts
V
l
5Volts
ADCS
DCmotor(CMG)x3
0.5Watts(ea.)
7.5Volts
.01
Magnetorquerx2
1Watt(ea.)
StarTracker
1Watt
5Volts
Magnetometer
20mW
2.0Volts
IMU
0.5Watts
SunSensor
Propulsion
MicroPPT
2.5Watts
5.5Volts
CarrierPropulsion
Valves,valveheaters,
72Watts(max)
28Volts
Imaging,CDH
CDH
FM430Flight Module
50mW
5Volts
Communications
Transceiver
5Watts
1015Volts
Imaging
CCD
3Watts
15Volts
.01
A/DConverter
153mW
3Volts
.1
SupportSystems
Power
Battery(charging)
3.1Watts
8.4Volts
Thermal
Structures
Total
20.82Watts
10Watts
EstimatedMaxPowerUsed
94
Table6.1.1SemiconductorMaterialProperties
Semiconductor
Bandgap
Wavelength, Responsivity,
Efficiency,
Ge
.66
1.9
.57
6%
Si
1.1
1.1
.41
15%
GaAs
1.43
.85
.32
19%
GaInP2
1.9
.65
Triplejunction
(GeGaAs
GaInP2)
29%
Table6.1.2TripleJunctionSolarCellProperties
TripleJunctionCell
Vmp
(V)
Jmp
Powerper20
(mA/cm2) cm2cell(W)
Efficiency,
.81
Baremass
per20cm2
cell(g)
1.68
SpectrolabNeXtTripleJunction1
(GaInP2/GaAs/Ge)
SpectrolabUltraTripleJunction2
(GaInP2/GaAs/Ge)
SpectrolabImprovedTripleJunction3
(GaInP2/GaAs/Ge)
2.333 17.32
2.35
16.3
.77
1.68
28.3%
2.27
16
.73
1.68
26.8%
29.9%
Iterationsofanalysishaveshownthata9Vbuswillbesatisfactory,asthesolararraybusvoltagemust
beslightlyhigherthanbatterychargevoltageandthesolarcellsprovidediscreetincreasesinvoltage.It
isplannedforthesolarcellstocoverallsidesexceptforthetopandbottommostside,showninFigure
6.1.3.Thetopsurfaceisassumedtohave100cm2ofsolarpanelarea,leavingroomforastartracker.
95
Table6.1.3SimplifieddiagramofCubeSat,withonesolarwing
Thesystemwillrequireasimplepowerconditioningschemeusingchargecontrollers,andDCDC
convertersforhardwarethatrequirestheseadditions.Currenttechnologiesoflowvoltage,lowpower
DCDCconvertersshowarangeof9098%efficiency.Figure6.1.4,shownbelow,illustratesthe
preliminarydesignofthepowerarchitecture.Solararraysprovidepowertothesystemandchargethe
battery,adjustedbyacontroller.
96
Figure6.1.4Preliminarypowerarchitecture
Althoughthetotalpowerfromhardwaresumsto20.82Watts(Table6.1.0),itisestimatedthat10Watts
maximumwillbeusedatanygiventime.Addingasafetymarginof20%resultsin12Wattsofsolar
powergenerationrequired.Toachievesolarpowergenerationof12W,320cm2ofsolarpanelsdirectly
facingthesunisrequired.Costsforthesolararraysystem,withanestimateof$250/W4forcurrent
technologies,totalaround$5350(2700cm2totalcellareaat1366W/m2and29%efficiency).
Themassofthebaresolarcellsis84mg/cm2.For2700cm2ofsolarcells,thisamountsto227gof
mass.However,thisfigurerepresentsonlythebarecellsoasafetymarginof20%isaddedtoaccount
forotherstructuralsupportthatmayberequired,producing272gofsolarpowersystemmass.Each20
cm2cell(assumed4x5cm)canprovide2.33Vand346mA(17.32mA/cm2)atmaximumpower
generation.
Toattaina9Vbus,4ofthesemustbeplacedinseries,andtoachievetherequiredcurrent,theremust
be6stringsof4cells,totaling24cells.
97
Todeterminethesolarwingarearequired,ifatall,fortheCubeSat,analysiswasdonebasedon
averagingthepowergeneratedforallpossibleorientations.Additionally,thispowergenerationis
weightedbythefractionoftimeinlighttothetotalorbitperiod(worstcaseof56minutesoutofa94
minuteorbit).Ithasbeendetermined(seeAppendixF1.3)thatorientationformaximumpoweryields
16.6Wwith400cm2ofextrasolarpanelareaperpendiculartothelargestCubeSatfaces(asshownin
Figure6.1.3)withthepreviousfactorsaccountedfor.Duringlightsideoperation,themaximumpower
attainablewithactivepointingis27.9Wwitha400cm2solarwing.Figure6.1.4,below,showsthe
relationshipbetweenthesolarwingareaandtheaveragepowergeneratedduringanorbitusingactive
maximumpowerpointing,theaveragepowerrequired,andthepowergeneratedbasedontheaverage
ofallorientations.
Table6.1.4Averagepowergeneratedduringanorbit
98
Carrierpowerhasnotbeenfullyexplored,asthesupportedhardwarehasnotbeenyetidentified.
However,similaranalysiswasdone(seeAppendixF1.3)toshowthatbycovering4sidesofthecarrier
(60cmonaside,14400cm2total,12kg),itispossibletogenerateanaverage70Wofpower(basedon
theaverageofallorientations),or120Wofpowerusingmaximizedpointing.Currentpower
requirementestimatesrequire80Wattsofpowerforpropulsionduringtheinfrequentorbitchanges.
Itmustbeemphasizedthatmanyoftheseanalyseshavetakentheworstcaserequirementsand
expandedonthembyextramarginsofsafety.Theorbitsattheirspecificinclinationandaltitude(55
inclinationand520kmaltitude)experienceaworstcaselightsidephaseof56minutesina94minute
orbit.
ElectrochemicalPower
Thedesignparametersofthemissionrequireapowersourcewhensolarpowerisnotavailable,namely
onthedarksideoftheCubeSatorbits.Theoperationalrequirementsfordarksidepoweris
minimizationofmassandvolumeandtheabilitytowithstandtheapproximate5600chargingcyclesof
lighttodark(with94minuteorbitperiodsat520kmaltitude)duringtheoneyearlifetimeofthe
mission.Liionbatterieswerechosensincetheyprovidehighspecificenergy(ontheorderof180W
h/kg),havelittletonoselfdischarge,arelowvolume,andhavelongcyclelives.Withanestimateof5W
ofrequiredpower,a7.4Vbus(duetothecombinedbatteryvoltage),andaworstcaseestimateof36
minutesofdischargetimeduringthedark,2Liioncellswillberequired,havingatleast405mAhof
capacityeach.Asanexample,theSanyoLiionbatteryUF634042Fholds1200mAhofcapacityper
batterycellandcanchargeathighcurrents(1230mA).Duetounknowncyclelifecharacteristicsof
thesebatteries,ahigherenergycapacityisusedtodecreasetherequireddepthofdischarge.Usingthis
batteryasanexample,thedepthofdischargeis34%.Additionally,therechargetimeestimateof58
99
minutesrequires3.1Wofadditionalsolarpower(seeAppendixF1.1)at731mAand4.2V.Eachcell
weighs25g.
Forthecarriervehicle,anestimateof50Wattsofpowerrequiredthroughouttheorbitanda
28Vbus,64batterieswillberequired(8inseries,8stringsinparallel).Thisprovides30Whofenergy
andonlya10.6%(126.7mAh)depthofdischargeperbattery.Eachbatterydischargesatanaverage
rateof211mAduringthedarkperiodoforbitandchargesatanaveragerateof914mAduringthelight
periodoforbit.Theaveragepowerrequiredduringthelightperiodoforbittochargethebatteriesis
30.7W,andthetotalmassofallbatteriestogetheris1.6kg.
6.2. Thermal
Thermalcontrolisnecessarytoprovideanoperationaltemperatureforthevarious
componentsonthesatellite.Methodsforthermalcontrolarecategorizedinpassiveand
activesystems.Passivethermalcontrolsystemsincludesurfacefinishes,insulation,and
radiators.Heatersareconsideredanactivethermalcontrolsystem.Thesatelliteswillattempt
tomainlyutilizeapassivesystemtomaintainanoperationtemperature.
Requirements
Thethermalrequirementsforvariouscomponentsofthesatellitewerecompiledfrom
Table1143.ExamplesofTypicalThermalRequirementsforSpacecraftComponentsinSpace
MissionAnalysisandDesignandvariouscomponentmanufacturers.Table6.2.1summarizes
thethermalrequirementsforthesatellitescomponents.
100
Table6.2.1ThermalRequirementsforSpacecraftComponents
Component
OperationalTemperatureRange(C)
Batteries
0to60
Optics
30to60
StarTracker
30to60
CMG
30to70
CCD
50to70
ComputerHardware
40to80
SunSensor
40to90
Antennas
100to100
SolarPanels
150to110
Basedontheoperationaltemperaturerangesofeachcomponent,anaveragethermal
equilibriumfrom0Cto20Cofthecomponentsinsidethesatellitesstructurewouldbe
sufficient.Thebatteriesarethemostsensitivetotemperaturebecausethechargeand
dischargevoltagevarieswithtemperature.
OrbitEnvironment
Theenvironmentenvelopingthesatelliteduringitsorbitsisinfluencedbythe
inclination,longitudeoftheascendingnode,andaltitudeofthesatellite.Radiation
encounteredbythesatelliteoriginatesfromthesun,earth,andreflectedsunlightofftheearth.
Thethermalparametersforanaltitudeof520kmand55inclinationarecalculatedinAppendix
F.2.1andsummarizedinTable6.2.2.
101
Table6.2.2Thermalfluxencounteredbythesatellite
SolarRadiation
EarthRadiation
ReflectedSolar
Radiation
1,370W/m
200W/m
175W/m
Thermalfluxfromsolarradiationisinthedirectionawayfromthesunandthermalfluxfrom
earthradiationandreflectedsolarradiationisinthedirectionawayfromtheearth.
Theperiodofdarknessduringthesatellitesorbitatanaltitudeof520kmand55
inclinationisshowninFigure6.2.1.Thecoordinatesystemusedtodeterminetheperiodof
darknessiswiththexaxisandyaxisintheeclipticplane.Theyaxisisinthedirectionofthe
sunandthexaxisisparalleltothevelocityoftheearth.Thelongitudeoftheascendingnodeis
measuredfromthexaxis.
Figure6.2.1Darkperiodversuslongitudeoftheascendingnode
102
Thesatellitewillexperiencebetweenapproximately22and36minutesofdarknessduringits
lifetime.Theaveragedarkperiodisabout30minutes.
SatelliteGeometryandMaterial
Thegeometryofthesatelliteaffectstheamountofheatabsorbedandradiatedfrom
thesatellite.ThegeometryoftheimagingandcarriersatellitesisshowninFigure6.2.2.The
imagingandcarriersatelliteswereassumedtomaintainanattitudeofnadirpointingduringan
orbit.Theprojectedareawasaveragedoveranorbitandwasassumedconstant.The
materialscoveringbothsatellitesaremainlyphotovoltaiccellsandaluminum.
Figure6.2.2Imagingandcarriersatellite
103
Thegeometryandmaterialparametersfortheimagingandcarriersatellitesaresummarizedin
Table6.2.3.
Table6.2.3.GeometricandMaterialParameters
GeometricParameters
ImagingSatellite
CarrierSatellite
Mass
6kg
60kg
SurfaceArea
0.38m
0.80m
SunProjectedArea
0.07m
0.18m
EarthProjectedArea
0.10m
0.15m
PhotovoltaicCellCoverage
80%
40%
MaterialParameters
Absorptivity
Emissivity
PhotovoltaicCell
0.90
0.85
Aluminum
0.13
0.065
Thephotovoltaiccellswereassumedtogatherenergyfor50%ofthetimeduringthelightside
oftheorbit.
TemperatureVariations
UsingthefinitedifferencemethodasdescribedinAppendixF.2.2,thetemperatureof
thesatellitescanbeapproximated.Thetemperatureofeachsatelliteisassumedtobethe
samethroughouttheentiresatellite.Also,eachsatellitesisassumedtohavetheparameters
describedbeforeandaspecificheatcapacityof500J/kgK.Thesimulationwasstartedat20C
104
andusedtheaverageorbitdarkperiodaccordingtoFigure6.2.1.Figure6.2.3andFigure6.2.4
showsthetemperaturevariationsoftheimagingandcarriersatellites.
Figure6.2.3Temperatureversusdurationoftheimagingsatellite
105
Figure6.2.4Temperatureversusdurationofthecarriersatellite
Theimagingsatellitereachedanequilibriumtemperatureof1415Cafterabout7hours.The
carriersatellitereachedanequilibriumtemperatureof82Cafterabout35hours.The
sensitivityoftheequilibriumtemperatureduetovariationofdarkperiodandsolarconstantare
summarizedinTable6.2.4.
Table6.2.3.EquilibriumTemperatureSensitivity
ImagingSatellite
CarrierSatellite
Variationofdarkperiod
7.7C
7.0C
Variationofsolarconstant
1.5C
1.4C
Mostofthecomponentsaboardthesatelliteswouldoperateunderthesetemperature
conditions.However,thebatterieswouldnotperformadequatelyunderthesetemperatures.
Hardware
Thethermalcontrolforthemajorityofcomponentsispassiveandnospecialtreatment
isneededforsurfacestomaintainoperationaltemperature.Thebatterieswillrequirepatch
heaterstoraisetheirtemperaturetoperformwithintheirspecifications.Thebatteriesshould
bethermallyinsulatedtoreduceheattransfersbetweenthebatteriesandothersurfaces.This
willreducetheamountofpowerconsumedbythepatchheaters.
Temperaturesensorsprovideinformationonthetemperatureofcomponents.
Temperaturesensorsplacedthroughoutthesatellitewillprovideatemperaturemapofthe
satelliteandwilldetermineifacomponentisapproachinganonoperationaltemperature.
106
Commoncommercialtemperaturesensorsareinexpensiveandareaccuratetowithinadegree
Celsius.
Conclusion
Theequilibriumtemperaturesoftheimagingandcarriersatelliteswillprovide
operationaltemperaturesforthemajorityofcomponents.Thebatteriesineachsatellitewill
requirepatchheaterstomaintainanoperationaltemperature.Themajorassumptionsmade
wereperfectthermalconductionbetweencomponentsandconstantprojectedareasduringan
orbit.Furtheranalysisneedstobeconductedtosizethepatchheatersforthebatteriesand
investigateinternalheattransfersinthesatellites.
6.3Structures
ThestructuraldesignandinternallayoutoftheCubeSatsandcarriervehiclesisimportanttoestablishing
acosteffectiveconstellation.ThestructuralaspectofthismissioniscomprisedoftheCubeSatdesign
andthecarrierdesign.Thesecomponentshavedifferentsetsofgoverningrequirementsandarebuiltto
bothsimilaranddifferentstructuralobjectives.ThedesignprocessesforboththeCubeSatandcarrier
aredetailedbelow.
107
CubeSatDesign
ThedesignoftheCubeSatswasbasedonthefollowingprimaryparameters:
AccommodatingthelargesizeofthetelescopeandattachedCCD
Balancingthelayoutofcomponentstoallowalignmentbetweenthepropulsionandthe
centerofmass
Withstandingtheforcesexperiencedduringlaunch
Providingprotectionfromlaunchandspaceenvironments
InordertoaccomplishthiswhileminimizingthevolumeofeachCubeSat,astartingdesignwasselected
thatusedthelargedimensionsofthetelescopeandCCDastheoveralllengthofthesatellite.The
componentdimensionsshowninTable6.3.1werethenusedtodeterminetheadditionalCubeSatunits
neededtoaccommodatethenecessaryhardware.
Table6.3.1:CubeSatinternalhardwaredimensions
System
Subsystem
Component
GNC
Nav
ADCS
Propulsion
GPSboard
DCMotorforCMG
CMGWheels(3wheels)
3Magnetorquers
Magnetometer
StarTracker
Camera
Lens
IMU
SunSensor
PPTUnit
Imaging,CDH
SupportSystems
CDH
Communications
Imaging
Power
Motherboard
Transceiver
CCD
Telescope
2Batteries
108
Volume
Dimensions(cm)
9.525x5x1.7
2x2x3
r=2
2.54x2.54x1.9
0.3x0.3x0.14
4.4x4.4x2.54
3.35dia,3,6length
3.5x2.2x1.2
2.43dia,3.49length
2x1x8
9x9x2
7.8x5.2x2
2x1.5x0.75
26.3long,10.6diameter
3.9x4.2x0.06
BasedonthevolumeandlayoutestimationsoftheCubeSathardware,aconfigurationcomprisinga3U
structureattachedinparalleltothetelescopemountedina1Ucubewaschosen.Figure6.3.1showsa
dimensionedCADrenditionofthisdesignwithattachedsolarwings.
Figure6.3.1:CubeSatdesign
AnexplodedviewoftheCubeSathardwareisshowninFigure6.3.2.
109
Figure6.3.2:CubeSatexplodedview
ThetotalmassofeachCubeSatis4.6kgincludingallofthehardwareshowninFigure6.3.2.Allowingfor
amarginoferror,theCubeSatmassusedinallcalculationswas6kg.Basedonthearrangementof
components,thecenterofmassisshowninFigure6.3.3withtheprincipleaxisofinertiashown.
110
Figure6.3.3:CubeSatcenterofmass
ThemomentsofinertiaabouttheprincipalaxisshowninFigure6.3.3areasfollows:
Ix=193000g*cm2
Iy=523000g*cm2
Iz=605000g*cm2
AllexteriorsurfacesoftheCubeSatexceptthebasearecoveredwithsolarcells.Thewingconfiguration
enablestheCubeSattoemploygreaterrotationalfreedomwithmaximumpowercollectioncapability
andnotbeconstrictedtoanattitudeatnadirpointing.Itisstilltobedeterminedwhatthemost
efficientdefaultattitudewillbeorwhetheritwouldbepossibletogeneratesufficientpoweratalltimes
withoutreturningtheCubeSattonadirpointingaftereachpictureistaken.
CarrierDesign
Thelayoutofthecarrierswasdesignedaroundthefollowingparameters:
Accommodatingthe10CubeSatsrequiredtopopulateeachplane
Containingthepropulsionandinstrumentationneededtoensureproperpositioningofthe
carriers.
Balancingthelayoutofcomponentstoallowalignmentbetweenthepropulsionandthecenter
ofmass
Withstandingtheforcesexperiencedduringlaunch
Providingprotectionfromlaunchandspaceenvironments
111
Inordertoaccomplishthiswhileminimizingthevolumeofeachcarrier,astartingdesignwasselected
thatheldthe10CubeSatsinagridfashion,withonesideastheejectionsideforallCubeSats.Figure
6.3.4showsadimensionedCADrenditionofthisdesignwithaCubeSatprotrudingfromitsslot.
Figure6.3.4:Frontviewofcarrierdesign
AnillustrationoftheCubeSatcontainmentandreleasemechanismsisshowninFigure6.3.5.The
CubeSatisheldinitspigeonholebyaspringloadedcatch.WhentheCubeSatisreleased,the
compressedspringshownisreleasedtodelivertherequireddeltaVtotheCubeSattotransferitinto
orbit.
112
Figure6.3.5:CubeSatheldincarrierslot
AdditionalvolumewasaddedtotherearoftheCubeSatholdinggridtoincludethepropulsionand
instrumentationrequiredtopositionthecarrierstolaunchtheCubeSatsintotheircorrectorbits.An
artistrenditionofthepropulsionsystemisshowninFigure6.3.6.
Figure6.3.6:Rearviewofcarrierdesign
Inordertodeliverthecarrierstotheirrespectiveorbits,Falcon1elaunchvehicleswillbeusedtocarry6
carriersatatimeinanarrangementshowninFigure6.3.7.
113
Figure6.3.7:Falcon1enoseconepackingarrangement
Anadditionaloptionconsideredfordeployingthecarrierswastolaunchthemasasecondarypayload
onaDeltamodellaunchvehicle.TheEvolvedExpendableLaunchVehicle(EELV)SecondaryPayload
Adapter(ESPA)ringshowninFigure6.3.8wouldbeusedtoattachthecarrierstothelaunchvehicle.
114
Figure6.3.8:ESPAring
ESPAisajointprogramdevelopedbytheDoDSpaceTestProgramandtheSpaceVehiclesDirectorateof
theAirForceResearchLaboratory.Itisdesignedtocarrysixsmallsatellitesassecondarypayloads,each
weighingupto170kg.TheESPAringisfittedbetweenthelaunchvehicleandtheprimarypayloadas
showninFigure6.3.9.
Figure6.3.9:ESPApayloadconfiguration
115
Advantages of using this launch system include ejecting secondary payloads into orbits independent of
the primary payload and ease of accessibility of the secondary spacecraft after encapsulation through
fairing access doors. This provides opportunity for checking spacecraft batteries, health, and removing
inhibit plugs within only several days of launch.
The significant disadvantage to using the ESPA ring is that it mounts the carriers in a cantilever type
fashion. Unlike a typical launch system where the thrust axis is parallel to the separation system axis and
produces a compressive force on the spacecraft, a cantilever-mounted structure generates significant
bending moment. This imposes more stringent structural requirements on the carrier. The dimensions and
mass of the carrier are also limited when using the ESPA ring. The ESPA payload requirements are
shown in Table 6.3.2
Dimension
Width
Height
Length(fromflangemount)
Requirement
60cm
60cm
96cm
C.G.(fromflangemount)
48cm
FlangeMountDiameter
Mass
38cm
170kg
ThecarriershavebeendesignedtofitwithintherequirementsoftheESPAringtobeabletoconsider
thisasaviableoption.
116
StructuralSupport
InorderforthecarriersandCubeSatstosurvivethelaunchenvironment,aforceprofilemustbe
generatedbasedontheinformationdetailedinthePayloadPlannersGuide.UsingaDeltaIVHeavyas
anexamplelaunchvehicle,Figure6.3.6showsthestopsigngraphforthedesignloadfactorsfor
dynamicenveloperequirements.
Figure 6.3.6: Delta IV Heavy Design Load Factors for Dynamic Envelope Requirements
CarrierandCubeSatloadmodelsmustbegeneratedandevaluatedbasedontheaccelerationdynamic
envelope.Thesemodelswillbeusedtodeterminethestructuralrequirementsofthesatellites.
ThemostviablematerialtouseforthestructuralelementsofbothCubeSatsandcarriersisthe
aluminumalloy6061T6.Thisisattributedtoitscommercialavailability,easeofmanufacture,and
relativelylowcost.Thematerialpropertiesfor6061T6aluminumaredetailedinTable6.3.3below.
117
Table6.3.3:Propertiesofaluminumalloy6061T6
MaterialProperty
Magnitude
ModulusofElasticity
68.9GPa
TensileYieldStrength
276MPa
PoissonsRatio
0.33
Density
2700kg/m^3
CoefficientofThermalExpansion(20100C)
23.6x10^6
Compositematerialsmaybeaviablealternativetoaluminuminselectlocations.Compositeshavethe
valuablebenefitsofhavingamuchlowerdensitythanmetalsandyieldstrengtharound10timesthe
yieldstrengthofaluminum.Despitetheiradvantages,compositesareprohibitivelyexpensive,requiring
thatthebenefitsoutweightheaddedcost.Duetothesequalities,theprimarycomponentsunder
considerationforcompositeconstructionarefueltanksandstructuralmembersrequiringhighstrength
characteristics.Additionalstudiesmustbeconductedtodeterminetheplausibilityofusingcomposite
materialsintheconstructionoftheCubeSatsand/orcarriers.
118
7. SystemsOverview
Thedesignofthismissionasoutlinedinthisreporthassatisfiedtherequirementsoutlined.Inaddition,
theconstellationhasremainedwithinthemassandcostlimitsimposedonthemission.Followingare
thedetailsofthesemassandcostallocations.
Satellitemass
ThemassrequirementperCubeSatis10kg,andbasedonthevaluesshowninTable7.1,theactual
massofeachCubeSatwillbesignificantlylessthanthelimit.Thiswillbeadvantageousforevaluating
launchcostandstructuraldesign.
Table7.1:CubeSatmassanalysis
Subsystem
Component
Mass
Imaging
CCD
50g
Telescope
1.6kg
Navigation
GPS
80g
Propulsion
PPT
320g
ADCS
2StarTrackers
220g
100g
CMG(Wheelsand
Motor)
3Magnetorquers
IMU
30g
2SunSensors
80g
CDH
Computing
100g
Communications
200g
Power
Batteries
40g
SolarArrays
300g
Structures
4CubeSatUnits
600g
Total
3.9kg
AllowingforUncertainty
119
240g
6.0kg
UsingalargepercentageofthesamecomponentsasintheCubeSatsaswellastheadditional
hardwarerequiredforeachcarrier,thetotalmassofeachcarrierisdetailedinTable7.2below.
Table7.2:Carriermassanalysis
Subsystem
Component
Mass
Navigation
GPS
80g
Propulsion
MainEngine
1.01kg
Valves
4.56kg
16Thrusters
5.28kg
Fuel/Pressurant
44.2kg
ADCS
StarTracker
220g
IMU
30g
2SunSensors
80g
CDH
Computing
100g
Communications
200g
Power
Batteries
2.5g
SolarArrays
10kg
Structures
StructuralSupport
600g
Cargo
10CubeSat
Satellites
Total
30kg
AllowingforUncertainty
120
160kg
170kg
MissionCost
Withamaximummissioncostof$200million,thecostanalysiswasbrokenintoindividualCubeSat
cost,individualcarriercost,andtotalCubeSat,carrier,andlaunchcost.
TheindividualCubeSatcostdetailedinTable7.3isbasedonthehardwarecurrentlyunder
consideration.Additionalsystemdevelopmentisstillrequired,sotheCubeSatcomponentcostsmay
changeasthedesignisperfected.
Table7.3:CubeSatcostanalysis
Subsystem
Component
Cost
Imaging
CCD+Processor
$10K
Telescope
$1K
Navigation
GPS
$10K
Propulsion
PPT
$10K
ADCS
2StarTrackers
$2K
3SunSensor
$17K
CMG
$2K
IMU
$1K
CDH
Computing
$4K
Communications
$3K
Power
Batteries
$1K
SolarArrays
$7K
Thermal
Heaters,MLI
$1K
Structures
4CubeSatUnits
$16K
Total
$85K
121
TheindividualcarriercostdetailedinTable7.4isbasedusingthesamehardwarearchitectureasfor
theCubeSats.Duetothesizedifference,componentssuchasthepropulsionsystemandattitude
controlsystemwillneedtobescaledup.
Table7.4:Carriercostanalysis
Subsystem
Component
Cost
Propulsion
Engine/Propellant
$200K
Navigation
GPS
$20K
ADCS
StarTracker2pcs
$2K
SunSensor3pcs
$15K
IMU
$2K
CDH
Computing
$4K
Communications
$3K
Power
Batteries/SolarCells
$50K
Structures
$80K
Total
$426K
ThetotalmissioncostestimateshowninTable7.5calculatesthetotalmissioncostbasedonthe
individualCubeSatandindividualcarriercosts,andtheaverageestimatedlaunchcostper3carriers
fromTablesA.3.1andA.3.2.Asshown,ourestimatedtotalmissioncostiswellunderour$200million
limit,allowingreasonableroomforadditionalcostrequirements.Apointtobetakenintoaccountis
thatthecostsofengineeringandlaborarenotincludedinthetotalmissioncostestimate.Thismay
considerablyaffectthetotalcost.
122
Table7.5:TotalMissionCost
Stage
#Units
Cost
CubeSats
330
$30M
Carriers
33
$15M
Launch
$100K
Total
$145K
Fromouranalysisoverthelasttwoquarters,wehavemetourrevisedrequirements;aconstellation
designhasbeenestablishedthatwillprovide5minutecoverageintervals,animagingsystemthatis
capableof3mnadirresolution,andacommunicationsarchitecturehasbeendesignedthatwillallow
imagestobedeliveredwithin1hourofacommand.Theproblemsencounteredwhiledesigninga
systemtomeetthesecriteriahavebroughtaboutmanyinventivesolutions.Additionalperfectingmust
bedoneinordertobringthismissionintoreality,butsignificantprogresshasbeenmadetowardthat
goal.
123
References
http://www.mae.cornell.edu/cubesat/
http://www.colorado.edu/geography/gcraft/notes/gps/gps.html
Hopkins, Joshua B, et. al. International Reference Guide to Space Launch Systems. 3 ed.
Reston, VA: AIAA, 1999.
DeltaIV. Encyclopedia Astronautica. 1 Mar 09.
<http://www.astronautix.com/lvs/deltaiv.htm>
Space Transportation Costs: Trends in Price Per Pound to Orbit 1990-2000
124
September 6, 2002
<http://www.futron.com/pdf/resource_center/white_papers/FutronLaunchCostWP.pdf>
Caffrey,Robert,GaryMitchell,ZenoWahl,andRayZenick.ProductPlatformConceptsApplied
toSmallSatellites:ANewMultipurposeRadioConceptbyAeroAstroInc.Aug.2002.
AeroAstroInc.31Jan.2009<http://www.aeroastro.com/publications/SSC02X8.pdf>.
"Dipoleantenna."Wikipedia,thefreeencyclopedia.29Jan.2009
<http://en.wikipedia.org/wiki/Dipole_antenna>.
Klofas,Bryan,JasonAnderson,andKyleLeveque.ASurveyofCubesatCommunicationSystems.
18Apr.2008.CaliforniaInstituteofTechnology.02Feb.2009
<http://atl.calpoly.edu/~bklofas/Presentations/DevelopersWorkshop2008/CommSurvey
Bryan_Klofas.pdf>.
"Patchantenna."Wikipedia,thefreeencyclopedia.29Jan.2009
<http://en.wikipedia.org/wiki/Patch_antenna>.
Pedtke,Dan,MarkLofquist,andKimberlyKohlhepp.TheModularSBandRadioSuite.Aug.
2004.AeroAstroInc.31Jan.2009<http://www.aeroastro.com/publications/SSC04V
4.pdf>.
"Transceivers."SatelliteCommunicationsEquipment.Ed.MarkTermondt.SatcomServices.03
Feb.2009<http://www.satcomservices.com/transceivers.htm#ac>.
Wetrz,JamesR.,andWileyJ.Larson,eds.SpaceMissionAnalysisandDesign.NewYork:
SpringerLondon,Limited,1992.
125
126
Appendices
A. OrbitalMechanics
A.1. ConstellationDesign
The following constants and equations were used to analyze the coverage of each constellation
configuration considered.
The radius Rearth of the Earth is
(Eqn.A.1.1)
The gravitational constant earth of the Earth is
(Eqn.A.1.2)
The radius of the satellites orbit Rsat is defined as the sum of Earths radius Rearth and altitude h.
(Eqn.A.1.3)
The velocity Vsat of the satellite in orbit is calculated as follows.
(Eqn.A.1.4)
The period of the resulting orbit Torbit is calculated from the circumference and velocity of the
orbit.
(Eqn.A.1.5)
The velocity of the ground track Vground can be determined from circumference of the Earth and
the period of the orbit.
(Eqn.A.1.6)
As the cubesat travels above the earth, it sweep out a certain area that is within its field of view.
This swept out area is called its coverage swath, illustrated in Figure A.2.1 below.
127
Figure A.2.1: Side and top view of coverage swath for each cubesat
The swath width wswath is calculated based on the altitude h and the maximum slew angle from
nadir.
(Eqn.A.1.7)
The necessary swath length lswath is determined from the ground track velocity Vground and the
required revisit time trevisit .
(Eqn.A.1.8)
The above parameters are used to determine the number of planes and satellites per plane
necessary for a polar constellation with full earth coverage. The number of planes Nplanes is
determined from the circumference of the Earth and the swath width wswath. Half the
circumference is used since each orbital plane provides coverage over two opposite points on the
Earth in ascension and descension.
(Eqn.A.1.9)
The number of satellites per plane Nsats/plane is determined from the total circumference of the
Earth and the swath length lswath.
(Eqn.A.1.10)
For a Walker constellation, the number of planes is determined differently due to the inclination.
The spacing snodes of the right ascension nodes depends on wswath and the sine of the inclination i.
This is illustrated in Figure A.2.2 below for i = 60and wswath = 1000 km.
128
(Eqn.A.1.11)
The entire circumference of the Earth is used to calculate the number of planes since the
ascending nodes must be spread over 360 for a Walker constellation.
(Eqn.A.1.12)
For a Walker constellation, each location where two orbits cross will have two spacecraft passing
over within the revisit time (5 min), allowing the revisit time on an individual orbit to be twice as
long (10 min). At the equator, each ascension node is paired with a descension node. Assuming
worst-case coverage occurs at the equator, these ascension/descension nodes pairs must be
spaced closely enough to eliminate any coverage gap between orbits. Therefore, the number of
satellites per plane can be reduced to half the number needed for a polar constellation.
(Eqn.A.1.13)
To ensure that one spacecraft always crosses the ascension/descension node within the revisit
time, they must be phased such that they cross alternately at that point, as illustrated in Figure
A.2.3. This figure shows one crossing node, with a 10-minute revisit time on a single orbit and a
5-minute revisit time at the crossing node.
129
This phasing is best explained by example. Assume 36 planes and 10 spacecraft per plane. The
10 spacecraft with have true anomalies spacing of 36 around the single orbit. The two orbits that
cross at given point along the equator will have a right ascension that differs by 180. The
spacecraft on the second of these orbits must cross halfway in between the spacecraft of the first
orbit, so their true anomalies will be spaced out by 36, but must be offset by 18.
Orbit 0:
Orbit 18:
= 0
= 180
Thus the phasing varies 18 over 18 planes, so each plane must have its spacecraft phasing by 1
from the previous plane. This phasing angle is based on the assumption of worst-case coverage at
the equator, but since this assumption is incorrect, full coverage at the equator actually does not
guarantee full coverage everywhere else. Thus other phasing options must be considered.
Polar Streets Analysis
The MATLAB code presented below was used to conduct analysis for the polar streets
constellation, to find the number of planes and satellites per plane needed for a given altitude, as
well as the associated delta-V required to deploy the constellation.
% AA 421 Polar Streets Analysis
% Peter Gangar
% 04/13/09
clear all
close all
clc
% Earth constants
Re=6378;
mu=398600;
Ce=2*pi*Re;
g0=9.81;
Ihyd=325;
%
%
%
%
%
% Orbital parameters
h=520;
fov=45;
tr=5;
% Derived parameters
rs=Re+h;
vs=sqrt(mu/rs);
cs=2*pi*rs;
130
T=cs/vs/60;
vg=Ce/(T*60);
ws=2*h*tan(fov*pi/180);
ls=vg*tr*60;
%
%
%
%
np=ceil(Ce/2/ws);
ns=ceil(Ce/ls);
npc=np-1;
% integer # of planes
% integer # of sats per plane
% integer # of plane changes
%
%
%
%
kg,
kg,
kg,
kg,
mass of cubesat
mass of carrier
extra mass for supporting structure
total mass to orbit
i1=70;
i2=90;
di=(i2-i1)*pi/180;
dvpc=2*vs*sin(di/2);
ra=2*pi/np;
dvra1 = 2*vs*sin(ra/2);
MR= exp(-(dvra1*1000)/(g0*Ihyd));
dvtot=dv1+dv2+dv3;
Walker Analysis
The FORTRAN code presented below was used to conduct analysis for the Walker constellation,
to find the fraction of points covered with the specified revisit time, and the maximum revisit
time gap for given configurations of the constellation.
131
C WALKERA.FOR
C ANALYSIS OF WALKER CONSTELLATIONS.
C USER SPECIFIES ALTITUDE, INCLINATION, AND FIELD-OF-VIEW OF SPACECRAFT
C [FOV IS MAX ANGLE OF SPACECRAFT FROM NADIR POINTING]
C USER SPECIFIES PARAMETER RANGE FOR WALKER CONSTELLATIONS TO ANALYZE:
C MIN AND MAX VALUES OF # PLANES, NP, MIN AND MAX # SATS/PLANE, NS,
C AND MAX # OF SATS IN CONSTELLATION, NSATS.
C USER SPECIFIES A "TARGET" REVISIT TIME (FOR COMPUTING FRACTION OF
C COVERAGE AREA MEETING THAT TIME).
C THIS ROUTINE RUNS THROUGH NP AND NS VALUES.
C AT EACH NP,NS, IT RUNS THROUGH ALL ALLOWABLE VALUES OF PHASE INTEGER NF.
C AT EACH NP/NS/NF (A SPECIFIC WALKER CONSTELLATION), IT RUNS THROUGH ALL
C LATITUDES FROM 0 TO ORBIT INCLINATION, AND A RANGE OF LONGITUDES.
C AT EACH LAT/LONG IT FINDS MAX TIME THAT THIS LOCATION IS NOT IN VIEW OF
C A SPACECRAFT.
C FOR A GIVEN NP/NS IT REPORTS THE "BEST" NF - THE ONE WITH MINIMUM
C VALUE OF MAX OUT-OF-VIEW TIME WITHIN COVERAGE AREA.
C
C OUTPUT FILES:
C WTIME.TXT: MAX-TIME-OUT-OF-VIEW(MINUTES) NSATS NP NS NF(BEST)
C WFRAC.TXT: FRAC NSATS NP NS NF(BEST)
C WHERE FRAC=FRACTION OF COVERAGE AREA HAVING OUT-OF-VIEW TIME
C
LESS THAN TARGET VALUE SPECIFIED BY USER.
C INFO: WALKER CONSTELLATION HAS NP PLANES, EACH AT SAME INCLINATION.
C PLANES ARE EQUALLY SPACED IN RA. THERE ARE NS SATS IN EACH PLANE,
C SPACED EQUALLY AROUND PLANE. PHASE DIFFERENCE BETWEEN PLANES SPECIFIED
C BY INTEGER NF. PHASE DIFFERENCE (FROM ASCENDING NODE) BETWEEN ADJACENT
C PLANES GIVEN BY D_PHASE(DEG)=NF*360./NSATS, NSATS=NP*NS.
C NF CAN TAKE ON VALUES 0 TO NP-1.
PROGRAM WALKERA
IMPLICIT REAL (A-H,Q-Z)
DIMENSION TLV(400),TGV(400),TLO(400),TGO(400)
DIMENSION IUSE(400)
DIMENSION AL(0:100),P(0:100,0:100)
DIMENSION NPV(10000),NSV(10000),NFV(10000)
DIMENSION FRAC(10000)
DIMENSION NPL(1000),NSL(1000),NFL(1000)
DIMENSION TM4(1000)
1
FORMAT(A,\)
RE=6380E3
WRITE(*,1) ' ENTER ALTITUDE IN KM: '
READ(*,*) HKM
H=HKM*1000
R=RE+H
XMU=4E14
Z=1
PI=4*ATAN(Z)
PI2=2*PI
TP_SEC=PI2*SQRT(R**3/XMU)
TP=TP_SEC/60
W=PI2/TP
132
KTOT=0
DO 600 NP=NPMIN,NPMAX
NSLAST=NSMAX
NS2=NTMAX/NP
IF(NS2 .LT. NSMAX) NSLAST=NS2
DO 500 NS=NSMIN,NSLAST
NT=NS*NP
DTHETA=PI2/NS ! SAT-SAT ANGLE IN A PLANE
DRA=PI2/NP
! PLANE-PLANE ANGLE IN RA
DA_LONG=DRA/NLONG ! LONGITUDE SAMPLE INCREMENT
PU=PI2/NT
TMIN4=1000
DO 400 NF=0,NP-1
ATRY=0
ARV=0
DPHI=PU*NF
133
10
20
DO 20 N=0,NP-1
AL(N)=DRA*N
DO 10 M=0,NS-1
P(N,M)=(N*DPHI+M*DTHETA)
CONTINUE
CONTINUE
TOFF3=0
DO 300 ILAT=0,LINC+5
ALAT_DEG=ILAT
ALAT=ALAT_DEG*PI/180
SLAT=SIN(ALAT)
CLAT=COS(ALAT)
IF(ILAT .EQ. 0) THEN
AREA=.5
ELSE
AREA=CLAT
ENDIF
TOFF2=0
DO 200 ILONG=0,NLONG
ALONG=ILONG*DA_LONG
TOFF1=0 ! MAX OFF TIME AT THIS LAT AND LONG
L=0
DO 95 N=0,NP-1
DO 94 M=0,NS-1
QC=CLAT*COS(ALONG-AL(N))
QS=CLAT*CI*SIN(ALONG-AL(N))+SLAT*SI
Q2=QC*QC+QS*QS
IF(Q2 .LE. CA2) GOTO 94
Q=SQRT(Q2)
AV=ACOS(CA/Q)
QSA=ABS(QS)
QCA=ABS(QC)
IF(QSA .LT. QCA) THEN
E=ATAN(QSA/QCA)
ELSE
E=PI/2-ATAN(QCA/QSA)
ENDIF
IF(QC .GT. 0) THEN
IF(QS .GT. 0) THEN
B=-E
ELSE
B=E
ENDIF
ELSE
IF(QS .GT. 0) THEN
B=E+PI
ELSE
B=PI-E
ENDIF
ENDIF
P0=-P(N,M)-B
134
TL=(P0-AV)/W
TG=(P0+AV)/W
92
93
94
95
CONTINUE
CONTINUE
IF(L .EQ. 0) THEN
TOFF1=TMAX
GOTO 180
ENDIF
C ORDE
110
115
135
120
CONTINUE
IF(TLO(1) .GT. .0001) THEN
DT=TLO(1)
IF(DT .GT. TOFF1) TOFF1=DT
TON=TLO(1)
ELSE
TON=0
ENDIF
TOFF=TGO(1)
140
180
DO 140 I=2,L
IF(TLO(I) .LE. TOFF) THEN
IF(TGO(I) .GT. TOFF) TOFF=TGO(I)
ELSE
TON=TLO(I)
DT=TON-TOFF
IF(DT .GT. TOFF1) TOFF1=DT
TOFF=TGO(I)
ENDIF
CONTINUE
CONTINUE
IF(TOFF1 .GT. TOFF2) TOFF2=TOFF1
IF(ILAT .LE. LINC) THEN
ATRY=ATRY+AREA
IF(TOFF1 .LE. TRV) ARV=ARV+AREA
ENDIF
200
300
CONTINUE
IF(ILAT .LE. LINC) THEN
IF(TOFF2 .GT. TOFF3) THEN
TOFF3=TOFF2
ILAT_TOPS=ILAT
ENDIF
ENDIF
CONTINUE
IF(TOFF3 .LT. TMIN4) THEN
TMIN4=TOFF3
NFMIN=NF
ENDIF
KTOT=KTOT+1
NPV(KTOT)=NP
NSV(KTOT)=NS
NFV(KTOT)=NF
FRAC(KTOT)=ARV/ATRY
400
C
CONTINUE
WRITE(1,2) NP,NS,NT,NFMIN,TMIN4
WRITE(*,2) NP,NS,NT,NFMIN,TMIN4
LTOT=LTOT+1
NPL(LTOT)=NP
NSL(LTOT)=NS
NFL(LTOT)=NFMIN
136
TM4(LTOT)=TMIN4
500
CONTINUE
600 CONTINUE
2
FORMAT(1X,4I4,F8.4)
OPEN(1,FILE='WTIME.TXT')
DO 900 I=1,LTOT
TMN=1000
JMN=0
DO 850 J=1,LTOT
IF(TM4(J) .LT. 0) GOTO 850
IF(TM4(J) .LT. TMN) THEN
TMN=TM4(J)
JMN=J
ENDIF
850 CONTINUE
IF(JMN .EQ. 0) GOTO 901
NP=NPL(JMN)
NS=NSL(JMN)
NF=NFL(JMN)
NT=NP*NS
WRITE(1,6) TM4(JMN),NT,NP,NS,NF
TM4(JMN)=-1
900 CONTINUE
901 CONTINUE
CLOSE(1)
OPEN(1,FILE='WFRAC.TXT')
DO 700 I=1,KTOT
F=0
JMX=0
DO 650 J=1,KTOT
IF(FRAC(J) .GT. 2) GOTO 650
IF(FRAC(J) .GT. F) THEN
F=FRAC(J)
JMX=J
ENDIF
650 CONTINUE
IF(JMX .EQ. 0) GOTO 710
NP=NPV(JMX)
NS=NSV(JMX)
NF=NFV(JMX)
NT=NP*NS
WRITE(1,6) FRAC(JMX),NT,NP,NS,NF
FRAC(JMX)=5
DO 680 J=1,KTOT
IF(NPV(J) .EQ. NP .AND. NSV(J) .EQ. NS) FRAC(J)=5
680 CONTINUE
700 CONTINUE
710 CONTINUE
CLOSE(1)
6
FORMAT(1X,F7.3,1X,I5,2X,3I3)
STOP
END
137
A.2. DeploymentMethod
Carrier Deployment
The MATLAB code presented below was used to conduct analysis of the various deployment
methods once a specific Walker constellation configuration had been chosen. It is used to find
delta-V and propellant mass for the master carrier method, delta-V for the clustered launch
method, and delta-V and deployment vs. launch inclination for the precession method.
%% AA 421 Walker Analysis
% Peter Gangar
% 05/13/09
clear all
close all
clc
% Earth constants
Re=6378;
mu=398600;
Ce=2*pi*Re;
g0=9.81;
Ihyd=325;
%
%
%
%
%
% Orbital parameters
h=520;
fov=45;
tr=5;
i=55;
%
%
%
%
% Derived parameters
rs=Re+h;
% km, radius of sat orbit
vs=sqrt(mu/rs); % km/s, velocity of sat orbit
cs=2*pi*rs;
% km, circumference of sat orbit
T=cs/vs/60;
% min, period of sat orbit
vg=Ce/(T*60);
% km/s, ground speed of sat
ws=2*h*tan(fov*pi/180); % km, swath width
ls=vg*tr*60;
% km, swath length
np=33;
ns=10;
% integer # of planes
% integer # of sats per plane
138
me=5*np;
% kg, extra mass for structure supporting carriers
mt=me+np*(mc+ns*ms);
% kg, total mass to orbit
% Spacing distances and angles
ra=360/np; % deg, spacing of right ascension
sa=Ce/np;
% km, spacing of right ascension
so=sa*sin(i*pi/180);
% km, spacing of planes (orthogonal to flight path)
dt=so/Ce*360;
% deg, angle for plane change at polar node
% Deployment Methods
%% Master Carrier
npc=np-1;
dvp1 = 2*vs*sin(dt*pi/180/2);
node
MR= exp(-(dvp1*1000)/(g0*Ihyd));
mptot = m(npc)-(ns*ms)-(np*mc)-me;
%% Clustered Launches
nps=np*2/3;
dvc1=dvp1;
dvctot=nps*dvc1;
%% Precession
i0=[45:0.1:i-1 , i+1:0.1:65];
for n=1:length(i0)
thtpc = abs((i-i0(n))*pi/180);
dv1(n) = 2*vs*sin(thtpc/2);
dvtot(n) = np*dv1(n);
j2 = -1.083*10^-3;
di(n) =
((3*j2*(2*pi/(T*60))*6378^2)/(2*rs^2))*cos(i*pi/180)*(3600*24)*(180/pi);
di0(n) =
((3*j2*(2*pi/(T*60))*6378^2)/(2*rs^2))*cos(i0(n)*pi/180)*(3600*24)*(180/pi);
139
end
figure(1)
plot(i0,dv1)
grid
xlabel('Launch Inclination (degrees)')
ylabel('Delta-V for single plane change (km/s)')
figure(2)
plot(i0,dvtot)
grid
xlabel('Launch Inclination (degrees)')
ylabel('Total Delta-V for all plane changes (km/s)')
figure(3)
plot(i0,t1)
grid
xlabel('Launch Inclination (degrees)')
ylabel('Precession Time Between Planes (days)')
figure(4)
plot(i0,ttot)
grid
xlabel('Launch Inclination (degrees)')
ylabel('Total Precession Time (days)')
figure(5)
plot3(i0,dv1,t1)
grid
xlabel('Launch Inclination (degrees)')
ylabel('Delta-V for single plane change (km/s)')
zlabel('Precession Time Between Planes (days)')
Cubesat Deployment
The cubesats must be spaced out evenly along their orbit. The time spacing Tspacing within a
single orbital is the number of cubesats Nsats divided by the period of the cubesat orbit Tsat.
(Eqn.A.2.1)
140
To deploy each cubesat with the correct phasing, the elliptical orbit of the carrier must be
designed to return to the circular cubesat orbit either ahead or behind the cubesat orbit by the
amount Tspacing. (For this analysis, the case of a shorter period is assumed since the required
delta-V does not vary significantly between the two cases.)
The carrier must orbit several times between each deployment so that the elliptical orbit is close
to circular and the delta-V at the deployment is achievable. Assuming an integer number Norbs of
carrier orbits, the period of the elliptical carrier orbit is as follows.
(Eqn.A.2.2)
The time T1 to deploy one cubesat, and the time Tall to deploy all cubesats can now be found.
(Eqn.A.2.3)
(Eqn.A.2.4)
The semi-major axis a, the perigee and apogee radius rp and ra respectively, and the eccentricity
e are then calculated as follows, where rsat is the radius of the circular cubesat orbit.
(Eqn.A.2.5)
(Eqn.A.2.6)
(Eqn.A.2.7)
(Eqn.A.2.8)
The velocity at the deployment point can then be calculated by finding the angular momentum h,
using it to find the velocity at apogee va of the carrier orbit.
(Eqn.A.2.9)
(Eqn.A.2.10)
The resulting delta-V is the difference between the velocity vc of the circular cubesat orbit and
the velocity at apogee va of the elliptical carrier orbit.
141
(Eqn.A.2.11)
The MATLAB code presented below was used to conduct analysis of the possible options for
elliptical carrier orbits to be used to deploy all cubesats on a given orbital plane with the correct
spacing. It is used to calculate the perigee or apogee radius, single-maneuver delta-V, total deltaV, single-cubesat deployment time, and total deployment time corresponding to the integer
number of orbits allowed between deployments.
% AA 420 Cubesat Deployment
% Peter Gangar
% 03/09/09
clear all
close all
clc
% Analysis for eliptical carrier orbit for cubesat deployment
% Physical constants
mu=398600;
re=6378;
% Set parameters
h=520;
Ns=10;
% Cubesat orbit
rc=re+h;
vc=sqrt(mu/rc);
Tc=2*pi*rc^(3/2)/sqrt(mu);
142
143
B. Propulsion
B.1References
B.2Equations
(Note: Specific ranges of variables studied are given in the Matlab code)
Carrier Propulsion
The following set of equations were used to size the propellant and pressurant tankage system for
carrier propulsion.
The propellant mass required for each maneuver, based upon the initial mass and delta v required
can be calculated using Equation 1:
(Eqn.B.2.1)
For a blowdown tankage system, the blowdown ratio can be cacluated based upon the operating
pressure range of the engine using Equation 2:
(Eqn.B.2.2)
The pressurant tank volume can be calculated, based upon the propellant tank volume and the
blowdown ratio, as given in Equation 3:
(Eqn.B.2.3)
Satellite Lifetime:
To determine the satellite lifetimes without propulsion based on a range of insertion altitudes
(Figure 8), the following expressions were used. First, utilizing Newtons Second Law, we state
that the change in kinetic energy of an orbital body is equal to the negative product of its velocity
and the drag force acting upon it.
144
(Eqn.B.2.4)
Next, from orbital mechanics we state that the energy of a body in a circular orbit is equal to
minus the quotient of the gravitational parameter mu divided by twice the radius of the orbit.
Differentiating, we arrive at an alternate expression for the drag power.
(Eqn.B.2.5)
(Eqn.B.2.6)
Once we have this expression, we can take as an assumption an exponential scaling of density
with altitude, and also use the definition of the drag coefficient and the circular velocity equation
from orbital mechanics to define all of the component terms in the equation immediately above.
(Eqn.B.2.7)
(Eqn.B.2.8)
(Eqn.B.2.9)
Equating the two sides of the preceding expression and cancelling terms, we attain an expression
for the change in radius with time. This expression may then be rearranged and integrated to give
the lifetime of the satellite, based on the starting height, density, frontal area, and drag
coefficient.
(Eqn.B.2.10)
(Eqn.B.2.11)
(Eqn.B.2.12)
Propulsion Requirements
Determining the thrust required to hold orbit at the injection altitude begins with equating the
thrust force needed to the drag force on the satellite. We can then simply use the result from the
definition of the drag coefficient above to determine this required force.
(Eqn.B.2.13)
145
To arrive at the total delta-v of the mission per year, we simply assume a constant deceleration
from the drag force in the absence of any thrust, and multiply this acceleration by the time
interval of interest.
(Eqn.B.2.14)
Once the delta-v for one year has been calculated, we can use this result to determine the amount
of propellant required for our design cases, as follows: initially, the rocket equation is used to
calculate the required mass fraction, which is then multiplied by the wet mass of the cube-sat to
give the total propellant mass.
(Eqn.B.2.15)
(Eqn.B.2.16)
The mass of the propellant, as well as its volume, are used as metrics for evaluating the
applicability of a design candidate. The volume of the propellant is calculated using the ideal gas
law, as follows.
(Eqn.B.2.17)
Where R is the specific gas constant, and tank temperature and pressure were assumed to be
300K, and 10 MPa. For the case of solid propellants such as the Teflon used in the PPTs, the
propellant volume was calculated by multiplying the known propellant mass by the density of the
material. The length of the stick required was then calculated by assuming a one-centimeter
diameter cylindrical propellant charge.
%Inputs
m = 10; % Mass in kg
Cdmin = 2.2; % Minimum Estimated Drag Coefficient
Cdmax = 2.75; % Maximum Estimated Drag Coefficient
Amin = 0.02; % Minimum Frontal Area (Incl. Solar Panels) in m^2
Amax = 0.15; % Maximum Frontal Area in m^2
146
147
% Propellant Data
Icg = 65; %Cold Gas Specific Impulse
Ibi = 450; %Bipropellant Specific Impulse
Ippt = 500; %Pulsed Plasma Thruster Specific Impulse
Iion = 1000; %Ion Thruster Specific Impulse
Ihyd = 240; %Hydrazine Monoprop Specific Impulse
%Delta-V Requirement
for y = 1:length(h0)
for k = 1:length(h0)
Drag(k) = 0.5*rho(k)*((v(k)*1000)^2)*Area(y)*Cdmin;
Adrag(k) = Drag(k)/m;
T(k) = 2*pi*sqrt((r(k)^3)/mu);
N(k) = (3600*24*365)/T(k);
vf(k) = v(k)-(Adrag(k)*T(k));
dv(k) = v(k)-vf(k);
deltav(y,k) = dv(k)*N(k);
end
end
%Propellant Mass, Volume and Tankage Info.
appt = pi*(rppt/1000)^2;
for z = 1:length(h0)
for x = 1:length(h0)
mpcg(z,x) = m-m/(exp(deltav(z,x)/(g0*Icg)));
mpppt(z,x) = m-m/(exp(deltav(z,x)/(g0*Ippt)));
mpion(z,x) = m-m/(exp(deltav(z,x)/(g0*Iion)));
mphyd(z,x) = m-m/(exp(deltav(z,x)/(g0*Ihyd)));
rtcg(z,x) =
100*((3/(4*pi))*((mpcg(z,x)*(8314/28)*300)/(2*10^7)))^(1/3);
rthyd(z,x) =
100*((3/(4*pi))*(mphyd(z,x)*(8314/32)*300)/(2*10^7))^(1/3);
148
%III: X-Y Station Keeping Monte Carlo Calculation - commented so that I can
%understand it later
clear all; close all;
%AA420 Stationkeeping Worst Case Code
%Modified 4/20/09
%Inputs
m = 10; %Cubesat mass (kg)
h0 = 500; %Assumed Initial Starting Altitude
i0 = 60 * (pi/180); %Assumed Initial Inclination in Degrees
149
150
om0 = acos(N0(1)/sqrt(N0(1)^2+N0(2)^2+N0(3)^2));
om1 = acos(N1(1)/sqrt(N1(1)^2+N1(2)^2+N1(3)^2));
ap0 = acos(dot(N0,e0)/(sqrt(N0(1)^2+N0(2)^2+N0(3)^2)*e0mag));
ap1 = (2*pi)-acos(dot(N1,e1)/(sqrt(N1(1)^2+N1(2)^2+N1(3)^2)*e1mag));
nu00 = acos(dot(e0,R0)/(e0mag*r0));
nu10 = (2*pi)-acos(dot(e1,R1)/(e1mag*r1));
p0 = 2*pi*sqrt(a0^3/mu);
p1 = 2*pi*sqrt(a1^3/mu);
E0 = acos((e0mag+cos(nu00))/(1+e0mag*cos(nu00)));
E1 = acos((e1mag+cos(nu10))/(1+e1mag*cos(nu10)));
M00 = E0-e0mag*sin(E0);
M10 = E1-e1mag*sin(E1);
%Propagate Orbits
pl0 = a0*(1-e0mag^2);
pl1 = a1*(1-e1mag^2);
for i = 1:numsteps
n0 = (mu/a0^3)^0.5;
n1 = (mu/a1^3)^0.5;
M0(i) = M00+ n0*(tstep*i);
M1(i) = M10+ n1*(tstep*i);
nu0(i) = M0(i)+2*e0mag*sin(M0(i))+1.25*e0mag^2*sin(2*M0(i));
nu1(i) = M1(i)+2*e1mag*sin(M1(i))+1.25*e1mag^2*sin(2*M1(i));
if nu0(i) > (2*pi)
nu0(i) = nu0(i)-(2*pi);
end
if nu1(i) > (2*pi)
nu1(i) = nu1(i)-(2*pi);
end
%Re-Build Orbital State Vectors
151
numvect = [1:numsteps];
figure (2),
plot3(Rx0,Ry0,Rz0,'b',Rx1,Ry1,Rz1,'g',xearth,yearth,zearth,'ok')
xlabel('X Position(km)','FontSize',12)
ylabel('Y Position (km)','FontSize',12)
zlabel('Z Position (km)','FontSize',12)
title('\it{Position Trace of Optimum(Blue) and Deviated(Green)
Orbits}','FontSize',16)
figure (3), plot(numvect,nu0,'g',numvect,nu1,'b')
xlabel('Elapsed Time of Flight (min)','FontSize',12)
ylabel('True Anomaly (radians)','FontSize',12)
title('\it{True Anomaly}','FontSize',16)
figure (4), plot(numvect,M0,'g',numvect,M1,'b')
xlabel('Elapsed Time of Flight (min)','FontSize',12)
ylabel('Mean Anomaly (radians)','FontSize',12)
title('\it{Mean Anomaly}','FontSize',16)
152
Carrier Code: This code calculates the required delta-v for each carrier type
and determines the amount of propellant mass required for orbital maneuvering
as a function of the specific impulse. Discrete values are then calculated
for several key points.
%AA420 Cube-sat Carrier Propulsion
%LEO from SMAD p.730 - 185km
%Inputs
mc = 10; %Cube-sat Mass in kg
ns = 10; %# of cube-sats in carrier
nc = 1; %# of carriers per launch
ms = 50; %Carrier Structural Mass in kg
leo = 185; % Low Earth Orbit in km
153
154
155
end
Delta_V_45_400
Delta_V_45_450
Delta_V_45_500
Delta_V_55_400
Delta_V_55_450
Delta_V_55_500
Delta_V_65_400
Delta_V_65_450
Delta_V_65_500
=
=
=
=
=
=
=
=
=
ddvt(1,1)
ddvt(1,2)
ddvt(1,3)
ddvt(2,1)
ddvt(2,2)
ddvt(2,3)
ddvt(3,1)
ddvt(3,2)
ddvt(3,3)
156
ThrusterSpecificationSheets
NorthropGrummanMRE01MonopropellantThruster
For satellite attitude and velocity control.
Technical Data
Propellant: Hydrazine
Thrust at maximum operating
Pressure: 1.0 N at 350 psia
Thrust at 275 psia inlet pressure: 0.8 N
Steady state specific impulse at
275 psia inlet pressure: 216 seconds
Operating pressure range: 5-600 psia
Life (demonstrated)
Maximum throughput: 34 kg
Maximum cycles: 370,000
Thrust valve power at 28 Vdc: 15 W
Weight (STM/DTM): 0.5 kg/0.9 kg
Envelope (width x length): 114 mm x 175 mm
Spacecraft Programs
Chandra X-ray Observatory, DSP, STEP4
157
Figure B.4.2.Spec sheet for Aerojet R-42.
158
159
160
Figure B.4.5 Moog Latch Valve.
161
Figure B.4.5. Moog Service Valve.
C. ImageAcquisition
C.1.Optics
The RFP states that the image resolution shall be less than 3m, with an image dimension of at
least 5000m per side. Thus, equation (1) relates the object distance, image distance, and focal
length.
(Eq. C.1.1)
Since the object distance is much greater than the image distance, the image distance is
essentially the effective focal length.
162
The relationship between the angle of the telescope to the image, the wavelength of light, and the
diameter of the lens is:
(Eq. C.1.2)
The wavelengths for visible light range from 400-750 nm. Since green is in the middle of the
light spectrum, a wavelength of 530 nm was used to estimate the diameter of the telescope.
C.2.ImageCapture
In order to figure out the shutter time needed to capture the picture, the amount of visible light to
reach the spacecraft needed to be found. The following equations were used to calculate the
intensity of light on the CCD,
I vis
.4 * I sun = 400
w
m2
Eq. C.2.3
because only 40% of total light is visible light, which is what we are going for.
I CCD =
Eq. C.2.4
where is the angle of the spacecraft off the equator, is the angle that the spacecraft is taking
the picture at, is the rate of reflection and f is the focal length of the telescope. This equation
can be simplified to,
I CCD =
I vis * * Alens
*b
*f2
Eq. C.2.5
where
b = cos( ) cos 2 ( )
Eq. C.2.6
The time for saturation of the pixels of the CCD can be calculated as the number of pixels over
the fraction of electrons that hit the CCD over time,
tsat =
N pixel
Ne
40, 000
.006
sec
=
6
6.5*10 * b
b
Eq. C.2.7
163
The minimum time that the shutter needed to be open was found by taking the minimum number
of electrons needed to hit the CCD to overcome noise over the fraction of electrons that hit the
CCD over time,
tmin =
N min
500
.0005
sec
= 6
=
10 * b
Ne
b
Eq. C.2.8
D. Navigation/Control
D.1. NavigationSystem
Nomaterialposted
D.2. AttitudeDeterminationandControl
; R is
orbit radius (m), is the maximum deviation of the Z-axis from local vertical in radians, and
and are moments of inertia about z and y (or x, if smaller) axes in
.
For
Dimensions
164
Eq. (D.2.1.3)
Eq. (D.2.1.4)
Solar Radiation
Solar radiation pressure,
surface is either transparent, absorbent, or a reflector, but most surfaces are a combination of the
three. Reflectors are classed as diffuse or specular. In general, solar arrays are absorbers and the
spacecraft body us a reflector. The worst case solar radiation torque is
Eq. (D.2.1.5)
where
and
,
is the center of gravity, q is the reflectance factor (ranging from 0 to 1), and is the angle of
incidence of the Sun.
Assumption: the reflectance factor,
center of mass,
Eq. (D.2.1.6)
Eq. (D.2.1.7)
Magnetic Field
Eq. (D.2.1.8)
where
is the magnetic torque on the spacecraft; D is the residual dipole of the vehicle in
, and B is the Earths magnetic field in tesla. B can be approximated as
for a polar orbit to half that at the equator. M is the magnetic moment of the Earth,
, and R is the radius from dipole (Earth) center to spacecraft in
Aerodynamic
Atmospheric density for low orbits varies significantly with solar activity.
165
Eq. (D.2.1.10)
where
Assumption:
and
Eq. (D.2.1.11)
Eq. (D.2.1.12)
];Print["TgX = ",TgX]
TgX = 1.2251310-8
TgY=(3
];Print["TgY = ",TgY]
TgY = 7.6570710-9
Solar Radiation
Fs=1358;c=3*108;Amax=0.15;Amin=0.02;q=0.5;i=0;
cps-cg=0.05;
the center of
166
Fmax=0.5 (
Ta_max = 3.9121510-7
Cd Amin V2);Tamin=Fmin (0.025);Print["Ta_min = ",Tamin]
Fmin=0.5 (
Ta_min = 5.216210-8
D.2.3. Sizing a camera lens and CCD chip for Star Tracker
Accuracy
Use lens equation to calculate focal length, the length between lens and CCD.
Eq. (D.2.3.1)
where
is pixel size,
167
The assume that the accuracy of pointing is the field angle corresponding to 1 pixel and the pixel
size is 10m. Table D.2.3.1 shows the range of the pointing accuracy with corresponding focal
length of lens.
Table D.2.3.1: Range of pointing accuracy and corresponding lens focal length..
Pointingaccuracy
[indegrees]
Pointingaccuracy
[inradians]
Pointingaccuracy
[inarcsec]
focallengthoflens[mm]
0.0057
1E04
20.6
100.0
0.0069
0.00012
24.8
83.3
0.0080
0.00014
28.9
71.4
0.0092
0.00016
33.0
62.5
0.010
0.00018
37.1
55.6
0.012
0.0002
41.3
50.0
0.013
0.00022
45.4
45.5
0.014
0.00024
49.5
41.7
0.015
0.00026
53.6
38.5
0.016
0.00028
57.8
35.7
0.017
0.0003
61.9
33.3
0.018
0.00032
66.0
31.3
0.019
0.00034
70.1
29.4
0.020
0.00036
74.3
27.8
0.022
0.00038
78.4
26.3
0.023
0.0004
82.5
25.0
The pointing accuracy is calculated for range of 50 m to 200 m of the ground distance and the
altitude of 500 km.
Field of view
A reasonable size of field angle should be chosen, and it can be used to determine the number of
pixels on the CCD using the following equation for one direction, with the assumption of one
pixel corresponding to
168
Eq. (D.2.3.2)
where Nx is the number of pixels in x-direction, field is the field of view, and
is the accuracy
of pointing. If both directions have the same number of pixels, the total pixels on CCD will be
. Table D.2.3.2 shows various pixel sizes for the CCD, with corresponding pointing
accuracy and field of view.
Table D.2.3.2: Total pixels on CCD vs. pointing accuracy for selects fields of views.
Accuracyof
pointing
[indegrees]
0.0057
0.0069
0.0080
0.0092
0.010
0.012
0.013
0.014
0.015
0.016
0.017
0.018
0.019
0.020
0.022
0.023
* assumption:
TotalpixelonCCD[Megapixels]
10
3.05
2.12
1.55
1.19
0.94
0.76
0.63
0.53
0.45
0.39
0.34
0.30
0.26
0.24
0.21
0.19
11
3.69
2.56
1.88
1.44
1.14
0.92
0.76
0.64
0.55
0.47
0.41
0.36
0.32
0.28
0.26
0.23
12
4.39
3.05
2.24
1.71
1.35
1.10
0.91
0.76
0.65
0.56
0.49
0.43
0.38
0.34
0.30
0.27
13
5.15
3.58
2.63
2.01
1.59
1.29
1.06
0.89
0.76
0.66
0.57
0.50
0.45
0.40
0.36
0.32
14
5.97
4.15
3.05
2.33
1.84
1.49
1.23
1.04
0.88
0.76
0.66
0.58
0.52
0.46
0.41
0.37
15
6.85
4.76
3.50
2.68
2.12
1.71
1.42
1.19
1.01
0.87
0.76
0.67
0.60
0.53
0.48
0.43
16
7.80
5.42
3.98
3.05
2.41
1.95
1.61
1.35
1.15
1.00
0.87
0.76
0.67
0.60
0.54
0.49
17
8.80
6.11
4.49
3.44
2.72
2.20
1.82
1.53
1.30
1.12
0.98
0.86
0.76
0.68
0.61
0.55
18
9.87
6.85
5.04
3.86
3.05
2.47
2.04
1.71
1.46
1.26
1.10
0.96
0.85
0.76
0.68
0.62
19
11.0
7.64
5.61
4.30
3.39
2.75
2.27
1.91
1.63
1.40
1.22
1.07
0.95
0.85
0.76
0.69
20
12.1
8.46
6.22
4.76
3.76
3.05
2.52
2.12
1.80
1.55
1.35
1.19
1.05
0.94
0.84
0.76
Diameter of Lens
The diameter of lens can be obtained from the equation of the light power from a star, where D is
lens diameter in meters.
169
Eq. (D.2.3.3)
This power is the light power collected by the camera lens. The equation for the intensity of
visible light from a star is as follows, where M is magnitude of brightness of star.
Eq. (D.2.3.4)
Substituting the intensity equation into light power equation and change unit of diameter to
centimeter gives the following.
Eq. (D.2.3.5)
The light power from a star can be related to the number of photons per second by the following
equation, where
Eq. (D.2.3.6)
Assuming that all photons hit a single pixel, the number of electrons generated per second, is
given below, where
A typical value of
becomes
Eq. (D.2.3.8)
Therefore, the number of electrons generated during exposure can be calculated by the following
equation, where t is exposure time.
170
Eq. (D.2.3.9)
The equation for the number of electrons generated during exposure becomes
Eq. (D.2.3.10)
The minimum number of electrons to have a good star image is about 400 electrons.
Therefore,
Solving for D,
Table D.2.3.3 shows the different diameter along the exposure time.
Table D.2.3.3: Exposure time vs. lens diameter.
ExposureTime[sec]
1/2
1/4
1/8
1/15
1/30
1/60
1/125
1/250
1/500
* assumption: the magnitude of brightness of star, M = 5.5
Diameter[cm]
0.56296
0.79614
1.12591
1.54172
2.18032
3.08344
4.45056
6.29404
8.90111
The lens focal ratio can be obtained with the following equation, where
171
Accuracyofpointing
[indegrees]
0.0057
0.0069
0.0080
0.0092
0.010
0.012
0.013
0.014
0.015
0.016
0.017
0.018
0.019
0.020
0.022
0.023
1/2
17.76
14.80
12.69
11.10
9.87
8.88
8.07
7.40
6.83
6.34
5.92
5.55
5.22
4.93
4.67
4.44
1/4
12.56
10.47
8.97
7.85
6.98
6.28
5.71
5.23
4.83
4.49
4.19
3.93
3.69
3.49
3.31
3.14
1/8
8.88
7.40
6.34
5.55
4.93
4.44
4.04
3.70
3.42
3.17
2.96
2.78
2.61
2.47
2.34
2.22
ExposureTime[sec]
1/15
1/30
1/60
6.49
4.59
3.24
5.41
3.82
2.70
4.63
3.28
2.32
4.05
2.87
2.03
3.60
2.55
1.80
3.24
2.29
1.62
2.95
2.08
1.47
2.70
1.91
1.35
2.49
1.76
1.25
2.32
1.64
1.16
2.16
1.53
1.08
2.03
1.43
1.01
1.91
1.35
0.95
1.80
1.27
0.90
1.71
1.21
0.85
1.62
1.15
0.81
1/125
2.25
1.87
1.60
1.40
1.25
1.12
1.02
0.94
0.86
0.80
0.75
0.70
0.66
0.62
0.59
0.56
1/250 1/500
1.59
1.12
1.32
0.94
1.13
0.80
0.99
0.70
0.88
0.62
0.79
0.56
0.72
0.51
0.66
0.47
0.61
0.43
0.57
0.40
0.53
0.37
0.50
0.35
0.47
0.33
0.44
0.31
0.42
0.30
0.40
0.28
Then the moments of inertia of the body with assuming uniform density and center of mass is to
be at the center of the body as calculated as shown below.
172
1
m( w2 + d 2 ) 0.0417[kg.m 2 ]
12
1
I w = m(h 2 + d 2 ) 0.108[kg.m2 ]
12
1
I d = m(h 2 + w2 ) 0.083[kg.m2 ]
12
Ih =
Eq. (D.2.4.1)
In order to find the required moment of inertia for the CMG, first the required slew rate is
determined. Since the cubesat will accelerate for half of this angle and decelerate for the half,
only half of the angle and time are used in the calculations. Using the equation below, the
angular acceleration of the cubesat can be determined. Please note that calculations were
completed for a range of slew rates from 1 - 15 per second.
1
2
ii
ii
= t2 =
2
t2
Eq. (D.2.4.2)
Then using the moments of inertia of the cubesat calculated earlier and this angular acceleration,
the required wheel torque is determined with the following equation.
ii
N wreq = I s
Eq. (D.2.4.3)
where Nw-req is the required torque needed to achieve the specified maneuver, Is is the spacecraft
ii
173
Then using the equation below to determine the momentum of the wheel.
Nx
cos cos
Eq. (D.2.4.4)
where Nx is the wheel torque on x-axis maneuver, h is angular momentum of the CMG,
is the
gimbal angles rate, and angles of and are corresponding to Figure D.2.4.2. The CMG cluster
for an x-axis maneuver is shown in Figure D.2.4.2. The same convention is used for the other
two directions. Note that for these calculations, is assumed to be 60 degrees and , which is the
initial gimbal angle position, is assumed to be zero for all directions. The calculations are carried
out for three cases with maximum gimbal angle rates of 6 rad/sec, 10 rad/sec, and 25 rad/sec.
174
Once the angular momentum is determined, assuming an angular velocity of 60,000 rpm for the
DC motor, the following equation can be used to determine the CMG's required moment of
inertia. Note that this DC motor is for the wheel only. We will use a ULT Applimotion frameless
motor for the gimbals.
h0 = I CMG I CMG =
h0
Eq. (D.2.4.5)
The required moments of inertias for four cases are of slew rates, 9 deg/sec, 8 deg/sec, 7 deg/
sec, and 5 deg/sec, with all three different cases of maximum gimbal angle rates are calculated
and summarized in Table D.2.4.2.
Table D.2.4.2: Required MOI for Four Slew Rates using Three Different Maximum Gimbal Angle Rates
175
An important factor in deciding the slew rate is the time it takes for the CMG to reach that slew
rate, determined from the equation below.
t=
Eq. (D.2.4.6)
Assuming a torque of 0.003 N-m, from Andrews Space, the time in each direction is determined.
The results are summarized in Table D.2.4.3.
176
Then in order to size the wheel and find the diameter that can produce the required slew rate, the
moments of inertia for a brass wheel are calculated for four different combinations, with density
of 8400 kg/m^3 and 8700 kg/m^3 and two thicknesses of 0.005 m and 0.0025m. The diameter is
varied from 0.01m to 0.02 m. The formulas used are as follows
Iz =
mr 2
2
1
I x = I y = m(3r 2 + h 2 )
12
177
Eq. (D.2.4.7)
Dia.
[m]
Volume[m3]
Mass[kg]
I_z
3
(thicknessof0.005m) (8400kg/m ) [kgm2]
I_x=I_y
Mass[kg]
I_z
2
[kgm ] (8700kg/m^3) [kgm2]
I_x=I_y
[kgm2]
0.01
3.93E07
3.30E03
0.012
5.65E07
4.75E03
0.014
7.70E07
6.47E03
0.016
1.01E06
8.44E03
0.018
1.27E06
1.07E02
0.02
1.57E06
1.32E02
3
Dia.
Volume[m ]
Mass[kg]
[m] (thicknessof0.0025m) (8400kg/m3)
4.12E08
8.55E08
1.58E07
2.70E07
4.33E07
6.60E07
I_z
[kgm2]
2.75E08
3.42E03 4.27E08
5.26E08
4.92E03 8.86E08
9.27E08
6.70E03 1.64E07
1.53E07
8.75E03 2.80E07
2.39E07
1.11E02 4.48E07
3.57E07
1.37E02 6.83E07
I_x=I_y
Mass[kg]
I_z
2
[kgm ] (8700kg/m^3) [kgm2]
2.85E08
5.45E08
9.60E08
1.58E07
2.47E07
3.70E07
I_x=I_y
[kgm2]
0.01
0.012
0.014
0.016
0.018
0.02
8.25E08
1.71E07
3.17E07
5.40E07
8.66E07
1.32E06
4.47E08
9.04E08
1.65E07
2.79E07
4.44E07
6.73E07
4.63E08
9.37E08
1.71E07
2.89E07
4.60E07
6.98E07
7.85E07
1.13E06
1.54E06
2.01E06
2.54E06
3.14E06
6.60E03
9.50E03
1.29E02
1.69E02
2.14E02
2.64E02
178
6.83E03
9.84E03
1.34E02
1.75E02
2.21E02
2.73E02
8.54E08
1.77E07
3.28E07
5.60E07
8.97E07
1.37E06
179
In order to model the DC motor, Newton's equations and Kirchhoff's laws are combined to
obtain the equations below.
i
Eq. (D.2.5.1)
= (b + KI )
Eq. (D.2.5.2)
1
J
i
1
I = (v K RI )
L
Eq. (D.2.5.3)
where is the angular velocity, J is the moment of inertia, b is the rotational friction, K is the
motor constant, I is the current, L is constant matrix, v is voltage, R is the motor resistance, and
is the angular position.
0
i
= 0
i
I
0
1
b
J
K
L
180
K
0
=
v
J
I
R 1
L
L
Eq. (D.2.5.4)
y = ( 0 1 0)
I
Eq. (D.2.5.5)
where the voltage is input and the angular velocity is the output. Then from these, the transfer
function can be obtained as shown below.
s
1
b
T (s) = ( 0 1 0 ) 0 s +
J
K
0
L
T ( s) =
K
J
R
s+
L
0
0
1
L
Ks
JLs + ( JR + Lb)s 2 + ( K 2 + bR)s
3
Eq. (D.2.5.6)
Eq. (D.2.5.7)
This can be modeled in Simulink for further simulations, as shown in Figure D.2.5.1.
Figure D.2.5.1: Simulink Block Diagram with Voltage Input and Angular Velocity Output
181
E. Communications
TheLinkEquationisgivenas
Eb Pt Ll Gt Ls La Gr
=
N0
kTN R
(Eq.E.1)
whichcanberearrangedtoisolatePtandsubstituteefficienciesforthelosstermstogive
Pt
Eb kTN R (1 C ) 4 r 2
N0
Gtt
r Aeff
(Eq.E.2)
Theratherthana=istoassurethatthepoweratthereceivingendisatleastasgreatas
theminimalpowerrequired.IntheLinkEquation,Aeffistheeffectiveareaofthereceiving
antenna.Thisiscommonlyapproximatedas
(Eq.E.3)
Inthis,Gristhegainofthereceivingantenna.Fortheonboardantennas,quarterwave
monopoles,thepeakgainis1.6andforthegroundstationsantenna,a1meterparabolicdish,
thepeakgainis25.
Themodulationchosenisthedifferentialphaseshiftkey(DPSK).Thismodulationcanbe
eitherbinaryorquadriphased,butthishasnotbeenanalyzedintoatradestudyasofyet.The
primarybenefitoftheDPSKmodulationisitisnotsusceptibletophasedisturbancesinthe
transmission.However,thiscomesatthecostofsensitivitytonoise.Toachieveabiterror
182
rate(BER)ofnomorethan105,thesignaltonoiseratiomustbeapproximately10,asfound
inFigure139intheSMADbook.
Thecrosslinkrawdatarateisfoundusingtheimaginghardware,a10megapixelCCDwith10
bitresolution,amissionrequirementof60imagesperday,andaninherentlimitationofonly
imagingduringthelightsideoftheCubeSatorbits.Theorbitsareinthelight61%ofthetime
onaverage,whichlimitsthetransmissionofthe60imagesto14.4hours.Withthis,theraw
datarateinbits/secondisfoundas
(Eq.E.4)
Thedownlinkrawdatarateisfoundbytakingthesizeofasingleimageandtheaccessible
rangeoftheflybytime.At520kmthegroundspeedofacircularorbitis7.6km/s.The
groundstationisassumedtobeabletotracktowithin10ofthehorizon,oraspanof160.
Thisleadstoamaximumaccessibledistanceof1820kmandamaximuminviewtimeof
approximatelysevenminutes.Itwasassumedthatcommunicationinitiationwouldtaketwo
minutesandthetransmissionofasingleimagewasrestrictedtotwominutes,leavingthree
minutesofmargin.Asforthecrosslinkrawdatarate,theimagesizeforthedownlinkis
dictatedbythe10megapixelCCDwith10bitresolution.Fromthis,theresultingdownlinkraw
datarateis
(Eq.E.5)
183
Tocompressthedata,JPEGimagecompressionwillbeused.Imagescanbecompressedupto
compressionratiosof98%usingJPEGcompression,whichmeansthatthefinalsizeisonly2%
oftheinitialsize.However,atsuchhighcompressionratiostheimagequalitysuffers
drastically.Acompressionratioof90%isconsideredlosslessanddoesnotsacrificequality.
ThegeneralvaluesusedintheLinkEquationanalysisaresummarizedinTableE.1below.The
crosslinkspecificvaluesaresummarizedinTableE.2andthedownlinkspecificinE.3.
Table E.1. Link Equation Values
Symbol
Definition
AssumedValue
Eb/N0
Signaltonoiseratio
10
Ll
Lineloss
N/A
Ls
Spaceloss
N/A
La
Transmissionpathloss
N/A
k
Boltzmannsconstant
1.38x1023[(m2kg)/(s2k)]
TN
Systemnoisetemperature
300[k]
C
Datacompressionratio
90%
t
Transmitantennaefficiency
0.8
r
Receiveantennaefficiency
0.8
Wavelength
63[cm]
Wavelength
74[cm]
184
F. SupportSystems
F.1. Power
F1.1 Chargepowerrequired
Withapowerrequirementestimateof5Wduringthedarkperiodofeachorbitanda
dischargetimeof36minutes,therequiredtotalenergyisdefinedas
Edisc=Preqtdark
(Eq.F.1.1)
Withachargingtimeof56minutes,thepowerrequiredtochargeisdefinedby
Pcharge=Edisc/tcharge
(Eq.F.1.2)
Thechargepowernecessaryinthiscaseis3.1Waveragethroughoutthelightsideperiodof
orbit.However,tochargetheSANYOUF634042Fbatteries,thecurrentmustbecontrolled
specificallyat1230mAaccordingtothespecificationssheetandreducedgraduallyafterabout
50minutes(ifneeded).Thechargetimefor810mAhofdischargedcapacityat1230mAis40
minutes.
F1.2 OrientationPowerAnalysis
Todeterminetheaveragepowergeneratedduringanorbitandtocomparethisfiguretothe
averagepowerrequired,thefollowingequationususedtofindthefrontalareashowntothe
sunfromanyarbitraryorientation,with"extra"areaaddedduetosolarwings.
185
Aarb=300cos()cos()+(100+extra)sin()sin()+600sin()cos()
(Eq.F.1.3)
FigureF.1.4belowillustratestheassumedaxesofrotation.Eachnumberrepresentsthearea
ofsolarcellsassumedoneachfaceoftheCubeSat,with100cm2allottedtothetopportionto
leaveroomforsensorsandexternalequipment.
FigureF.1.4OrientationofCubeSat
Tofindtheaverageareaexposedtothesun,thetripleintegralisfoundwithlimitsfrom0to
/2andthendividedby3/8,theintegralwithoutthefunction.Theaverageareaexposedto
thesunisthenfoundtobe
Aavg=4(1000+extra)/2
186
(Eq.F.1.5)
Usingthisareaandthegivensolarpanelpropertiesandaveragingthetimeinlightoverthe
totalorbittime,theFiguresF.1.6andF.1.7areproduced.Onethingtonoteisthatthepowers
calculatedformaximumpowergeneration
FigureF.1.6Powerpropertiesofgivenorientation
187
FigureF.1.7Orientationrequiredtoachievemaximumpower
Toobtainthemaximumpower,Matlabsfminsearchfunctionwasusedonthereciprocalof
Eq.F.1.3tofindthemaximumareashowntothesun.Matlabsfminsearchfunctionrunsa
minimizationcodetofindtheabsoluteminimumofnvariablefunctionsbasedoninitial
conditionstodeterminethevalueofthevariablesthatproduceaminimum.Itcanbeseenin
thesecondfigurethatpastacertainamountofextrasolarwingarea,theorientationdirectly
facesthelargestareaoftheCubeSat.Asthewingareagrows,themaximumpowergeneration
isachievedfromincreasinginclinationtowardfacingthewingarea.
Similaranalysiswasdoneforeachcarriervehicle,andthepowergeneratedfromthe
sunandmaximumpowerorientationcanbeseeninthefollowingfigures:
188
FigureF.1.8Powerpropertiesofcarriervehicle
189
FigureF.1.9Orientationrequiredtoachievemaximumpower
F1.3 PowerAnalysisMATLABCode
CubeSatpowerrequirementsanalysis
%Number of strings
M_str = ceil(I_bus_guess/I_cell);
I_bus = I_cell*M_str;
N_cell = N_str.*M_str;
P_bus = N_cell*P_cell;
A_array = N_cell*A_cell/.0001; %cm^2
Vmp = 2.33;
Jmp = 17.32/1000; %A/cm^2
Imp = Jmp*A_cell/.0001; %A
P_cell = Imp*Vmp;
if length(P_req_guess) > 1
figure(1)
plot(P_req_guess,N_cell)
grid on
axis equal
xlabel('Power required by spacecraft
(W)');
ylabel('Total number of solar cells
required');
N_cell_guess = ceil(P_req_guess/P_cell);
A_array_guess = N_cell_guess*A_cell/.0001;
%cm^2
V_bus_guess = 9; %A guess
V_cell = Vmp;
N_str = ceil(V_bus_guess/V_cell);
190
q_level = 1-dod;
E_bat_Sanyo = 187*.0246; %W-h/kg * kg
N_bat = floor(V_bus_solar/V_bat);
V_bat_total = N_bat*V_bat;
figure(2)
plot(P_req_guess,A_array)
grid on
xlabel('Power required by spacecraft
(W)');
ylabel('Area of solar panels required by
spacecraft (cm^2)');
title('Area of solar cells required to
generate power required');
end
V_bus_bat = N_bat*V_bat;
E_disc = P_req_dark*t_dark/3600; %watt-hour
Cap_disc = E_disc/V_bat; %Amp-hour
E_bat = E_disc/(N_bat*q_level); %watt-hour
Cap_bat_req = Cap_disc/(N_bat);
Cap_bat = Cap_bat_req/q_level;
P_charge = 3600*E_disc/t_charge; %Power
needed to charge
costrange = 240*[20];
massrange = 1.68*39*2*[1 2.5]/1000; %kg
% Batteries
T = 94/60; %Orbital period, hours
cycles = 365*(24/T); %cycles of light-todark
Orientationanalysis
ylabel('Power (W)');
legend('Average power with pointing',
'Average power', 'Required power',
'Location', 'SouthEast');
title('Power generation as a function of
solar wing area')
clear all
close all
theta = 0:.1:90;
phi = 0:.1:90;
psi = 0:.1:90;
angles = (pi/180)*[theta; phi; psi];
extra = 0:10:1000;
figure(2)
plot(extra, maxangles(1, :), extra,
maxangles(2, :), extra, maxangles(3, :))
legend('Angle theta at max power', 'Angle
phi at max power', ...
'Angle psi at max power',
'Location', 'SouthEast');
xlabel('Area of solar panel wings (cm^2)')
ylabel('Angle in degrees')
title('Orientation of spacecraft at maximum
power')
for i = 1:length(extra)
[maxang(:,i), p1(i)] = fminsearch(@(x)
powerval(x,extra(i)), [pi/5, pi/3, pi/3]);
maxpower(i) = (56/94)*1/p1(i); %Max
power averaged over a worst-case orbit. 56
min out of 94
maxangles(:,i) = (180/pi)*maxang(:,i);
end
%
To change the top base area of solar
panels, change the number next to extra
%
from 1000 to 1100 if going from a top
area of 100 cm^2 to 200 cm^2
%
(also in powerval function)
avgpower =
(56/94)*.29*1366*(1/(100^2))*(4*(1000 +
extra))/(pi^2);
% Carriers
extra1 = 0:60:3600;
%extra2 = 0:50:2500;
for j = 1:length(extra1)
[maxang2(:,j),p2(j)] = fminsearch(@(x)
powerval2(x,extra1(j),0), [pi/5, pi/3,
pi/3]);
maxpower2(j) = (56/94)*1/p2(j);
maxangles2(:,j) = (180/pi)*maxang2(:,j);
end
figure(1)
plot(extra, .9*maxpower, extra, .9*avgpower,
extra, 12*ones(length(extra)), 'r')
xlabel('Area of solar panel wings (cm^2)');
191
figure(4)
plot(extra1, maxangles2(1, :), extra1,
maxangles2(2, :), extra1, maxangles2(3, :))
legend('Angle theta at max power', 'Angle
phi at max power', ...
'Angle psi at max power',
'Location', 'Best');
xlabel('Area of extra solar panel wings
(cm^2)')
ylabel('Angle in degrees')
title('Orientation of carrier at maximum
power')
axis([0 3600 -10 100])
P_req = 50;
figure(3)
plot(extra1, maxpower2, extra1, avgpower2)%,
extra1, P_req*ones(length(extra1)), 'r')
xlabel('Area of extra solar panel wings
(cm^2)');
ylabel('Power (W)');
legend('Average power with pointing',
'Average power', 'Location', 'SouthEast')
title('Power of carrier based on extra solar
panel area from wings')
Cubesatorientationfunction
function p = powerval(x0, extra)
theta = x0(1);
phi = x0(2);
psi = x0(3);
Carrierorientationfunction
function p = powerval(x0, extra1, extra2)
%Carrier numbers
theta = x0(1);
phi = x0(2);
psi = x0(3);
A = (3600+extra1)*cos(theta)*cos(psi) + (3600+extra2)*sin(psi)*sin(phi);
p = 1/(.29*A*1366*(1/(100^2)));
Carrierpowerrequirementsanalysis
clear all
close all
P_req_guess = 48;
e_cell = .2861;
V_cell = Vmp;
N_str = ceil(V_bus_guess/V_cell);
Vmp = 2.33;
Jmp = 17.32/1000; %A/cm^2
Imp = Jmp*A_cell/.0001; %A
P_cell = Imp*Vmp;
%Number of strings
M_str = ceil(I_bus_guess/I_cell);
I_bus = I_cell*M_str;
N_cell = N_str.*M_str;
P_bus = N_cell*P_cell;
A_array = N_cell*A_cell/.0001; %cm^2
N_cell_guess = ceil(P_req_guess/P_cell);
A_array_guess = N_cell_guess*A_cell/.0001;
%cm^2
192
if length(P_req_guess) > 1
figure(1)
plot(P_req_guess,N_cell)
grid on
axis equal
xlabel('Power required by spacecraft
(W)');
ylabel('Total number of solar cells
required');
title('Number of cells required by power
required')
figure(2)
plot(P_req_guess,A_array)
grid on
xlabel('Power required by spacecraft
(W)');
ylabel('Area of solar panels required by
spacecraft (cm^2)');
title('Area of solar cells required to
generate power required');
end
% Carriers
P_req_dark = 50; %W, a guess for power
required in the dark
t_dark = 36*60; %36 minutes worst case, in
seconds
V_bat = 3.7;
V_charge = 4.2;
t_charge = (94.6-36)*60;
dod = .5; %depth of discharge
q_level = 1-dod;
E_bat_Sanyo = 187*.0246; %W-h/kg * kg
N_bat_series = floor(V_bus_solar/V_bat);
V_bat_total = N_bat_series*V_bat;
V_bus_bat = N_bat_series*V_bat;
E_disc = P_req_dark*t_dark/3600; %watt-hour
M_bat_parallel = 8;
N_bat = M_bat_parallel*N_bat_series;
Cap_disc = E_disc/V_bat; %Amp-hour
E_bat = E_disc/(N_bat*q_level); %watt-hour
Cap_bat_req = Cap_disc/(N_bat);
Cap_bat = Cap_bat_req/q_level;
P_charge = 3600*E_disc/t_charge; %Power
needed to charge
I_disc = P_req_dark/V_bat_total;
I_disc_bat = I_disc/M_bat_parallel;
mass_bat_req = E_bat/175; %Batteries over
specific energy
mass_bat_total = N_bat*mass_bat_req;
P_req = P_charge + P_req_guess;
I_charge = P_charge/V_charge;
I_charge_bat = I_charge/M_bat_parallel;
Cap_Sanyo = 1.2; %A
dod_true = Cap_bat_req/Cap_Sanyo;
193
F.2. Thermal
F.2.1. OrbitEnvironment
FromEquations(11.17)and(11.18)inSpaceMissionAnalysisandDesign,
(Eq.F.2.1)
(Eq.F.2.2)
Parametersforobitsnear520kmaltitudeand55inclination,
=231W/m
=1367W/m
=0.86
=0.43
=0.30
F.2.2. TimeDependentHeatTransfers
Thetotalthermalpowerbeingabsorbedbyasatelliteisasfollows.
(Eq.F.2.3)
(Eq.F.2.4)
194
Breakingtheorbitintothelightanddarkperiodresultsinthefollowingequations.
(Eq.F.2.5)
(Eq.F.2.6)
Therelationshipbetweenthechangeofanobjectstemperaturetoachangeinabsorbedheat
isasfollows.
(Eq.F.2.7)
Startingataninitialtemperature,thechangeinabsorbedheatduringthelightordarkphaseof
theorbitcanbeapproximatedbythefollowing,wheretisafinitetimestep.
(Eq.F.2.8)
Theresultingtemperatureafterafinitetimestepisasfollows.
(Eq.F.2.9)
Repeatingtheprocessesandshiftingbetweentheequationforthelightanddarkperiodofthe
orbitswhenappropriateresultsinanapproximationforthesatellitestemperatureasthe
satelliteorbits.
195
G. BiographicalSketches
Michael (Mike) Bernhardt
Michael Bernhardt is a senior in the Aeronautics and Astronautics department at the University
of Washington. He has an amateur interest in rocketry and space. He has participated as a
consultant for high school rocket competitions. He possesses rudimentary skills in computer
programming and electronics. Michael Bernhardt currently spends his free time making
computer renders and animations. Michael Bernhardt investigated and researched thermal
control systems and computer hardware solutions. He wrote feasibility and cost analyses for the
thermal and computer subsystems.
Aaron Borth
My name is Aaron Borth and I am part of the imaging group for the proposal. Last summer as an
intern I had the opportunity to work at a small design company, PCSI Design. I was able to do
design work on aerospace products as well as consumer products (handheld firestarter). I spent a
lot of time doing design work in Solidworks and also transfering engineering data from CATIA
V5 to the Solidworks environment (737 door seals). As for classes, on top of the classes required
for the A&A Undergraduate, I have taken Advanced Propulsion, Systems Engineering, FEA
Analysis, and am currently taking Heat Transfers. For the proposal, I wrote the section about the
image recording device (CCD) and shutter time and helped set up MATLAB codes for
calculating the shutter time and focal length of the telescope.
Rachel Brennan
Peter Gangar
Peter Gangar is a senior in the Aeronautics and Astronautics department at the University of
Washington. He graduated from Bellevue Community College with an Associate of Arts and
Sciences degree in 2005. Besides the foundational aerospace courses, his relevant coursework
includes orbital mechanics, controls, and systems engineering. In addition to aerospace
engineering, Peter is pursuing a Bachelor of Arts degree in Classics to be completed in 2010. He
joined the University of Washington's 2007-08 Design Build Fly Team as part of the structures
team, manufacturing control surfaces and wings. He has also served as treasurer for AIAA
student branch since early 2008. During the summer of 2008, Peter worked as systems
engineering intern for Phantom Works, assisting on a project to implement RFID technology on
aircraft parts.
Peter served under the navigation/controls team for the 2009 senior design project. During the
proposal stage, he was responsible for constellation design, coverage analysis, and orbital
mechanics. He contributed the constellation design section to the proposal.
In his non-existent free time, Peter enjoys reading, playing the violin, and spending time with his
family. He hopes to earn a private pilot license, and dreams of becoming an astronaut. Whether
or not he achieves this dream, his ultimate goal is to serve God in whatever he does.
197
Nikolas Lutzenhiser
198
propulsion system design of both the satellite constellation and the satellite carriers. He enjoys
video gaming, motorsports, and soccer.
Katie Moravec
The relevant coursework I have taken includes Orbital and Space Flight Mechanics, Control in
Aerospace Systems, Propulsion, Advanced Propulsion, Structural Vibrations, Structural Analysis
I & II, and Aerospace Lab I & II. The course I enjoyed the most was controls, which was also
my strongest subject area. For the past three years, I have had summer-long internships at
Boeing. The first two summers I spent in Flight Test Engineering, and the third summer I spent
in Product Development Weights Engineering. During my third internship, I worked in a team
that came up with design improvements for the 777 freighter. During our Structural Analysis II
course, I was part of a team that designed a wooden wingbox. In this proposal, I was responsible
for the GPS system. I spent a number of hours researching GPS systems that would best fit the
requirements for this project. After gathering information on cost, accuracy, power
requirements, weight, size, and electrical interface, I found a possible GPS system that meets the
needs of this project.
Skander Mzali
Skander Mzali is a Senior in Aeronautics and Astronautics and the University of Washington,
pursing a Bachelor's of Science in Aeronautic and Astronautic Engineering at 19 years old. He
hopes to continue his Aerospace studies in Graduate School to attain a Master's degree. Skander
worked on the research and analysis of power systems for the spacecraft.
Zahra researched about sensors and actuators, to be used in attitude determination and navigation
systems. She studied the performance, physical characteristic, and costs for each of these
devices. She also prepared a summary of her and Eun-Ju 's collected information for the
proposal. In the write-up, she discusses the cons and pros of the devices and suggests a possible
control mode and navigation system. During autumn quarter she worked on the Spring-Mass
System project in professor Ly's control lab. The system consisted of four sliders attached to
each other via three springs. The objective was to move the furthest slider from the motor for one
inch in the shortest time. She successfully modeled the system and designed a PID controller to
199
improve the response of the system. Zahra is now taking AA448, which is a control systems
class on sensors and actuators.
Josh Ross
Josh is a senior in the Aeronautics and Astronautics department at the University of Washington.
He is focusing his studies on spacecraft dynamics, trajectory determination laws, and distributed
space system communications and applications. He currently works in the Distributed Space
Systems Lab. Coursework relevant to the mission includes orbital mechanics, spacecraft
dynamics, plasma physics, and aerodynamics of bodies in rarefied flow.
Josh is in the Communications group and is responsible for designing a communication
algorithm that optimizes time, power, and computation. Josh works with Miguel Carrion in
researching the hardware to be used on the Cubesat communication system. Outside of
communications, Josh has worked with Peter Gangar for constellation design. Josh originally
proposed a Walker constellation, and then worked with Professor Mattick and Peter to determine
the impracticality of such a complex constellation.
Eun-Ju researched about sensors and actuators, which can be used for attitude determination and
navigation. She studied their performance, physical characteristics, and costs. She calculated the
worst-case disturbance torques and reported them on the proposal. During the autumn quarter,
she worked on the Magnetic Ball Levitation project in professor Ly's control lab. She succeeded
in designing a PID controller in order to improve the model, so that the magnetic ball would
ascend while under the influence of the magnetic field. Eun-Ju is now taking AA448, which is a
control systems class on sensors and actuators.
200