? Massachusetts Institute of
Technology
? Space Systems Laboratory
? Lockheed Martin Corporation
? Advanced Technology Center
Electromagnetic Formation Flight
NRO DII Final Review
Friday, August 29, 2003
National Reconnaissance Office
Headquarters
Chantilly, VA
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
Motivation
? Traditional propulsion uses propellant as a reaction mass
? Advantages
– Ability to move center of mass of spacecraft
(Momentum conserved when propellant is included)
– Independent (and complete) control of each spacecraft
? Disadvantages
– Propellant is a limited resource
– Momentum conservation requires that the necessary propellant
mass increase exponentially with the velocity increment (?V)
– Propellant can be a contaminant to precision optics
DII EMFF Final Review Aug. 29, 2003
Question I:
? Is there an alternative to using propellant?
? Single spacecraft:
– Yes, If an external field exists to conserve momentum
– Otherwise, not that we know of…
? Multiple spacecraft
– Yes, again if an external field exists
– OR, if each spacecraft produces a field that the others can
react against
– Problem: Momentum conservation prohibits control of the
motion of the center of mass of the cluster, since only
internal forces are present
DII EMFF Final Review Aug. 29, 2003
Question II:
? Are there missions where the absolute position of the center of mass of
a cluster of spacecraft does not require control?
? Yes! In fact most of the ones we can think of…
– Image construction
? u-v filling does not depend on absolute position
– Earth coverage
? As with single spacecraft, Gravity moves the mass center of the
cluster as a whole, except for perturbations…
– Disturbance (perturbation) rejection
? The effort to control perturbations affecting absolute cluster
motion (such as J2) is much greater than that for relative motion
? Only disturbances affecting the relative positions (such as
differential J2) NEED controlling to keep a cluster together
– Docking
? Docking is clearly a relative position enabled maneuver
DII EMFF Final Review Aug. 29, 2003
Example: Image Construction
? Image quality is determined by the point spread function of aperture
configuration
()
2
1
1
)(
2
exp
sin
sin
)cos1(
,
?
?
?
?
?
?
?
?
?
?
?
?
∑ ?
?
?
?
?
?
ψ+ψ
λ
π
?
?
?
?
?
?
?
?
?
?
?
?
?
λ
θπ
?
?
?
?
?
?
λ
θπ
?
?
?
?
?
?
λ
θ+π
=ψψ
=
N
n
njniji
yx
i
D
D
J
D
I
? The geometry dependence can be expanded into terms which only depend on
relative position
Aperture dependence Geometry dependence
I ψ
i
()= I
Ap
ψ
i
()N + cos
2π
λ
ψ
i
(x
1
? x
2
)
?
?
?
?
+ cos
2π
λ
ψ
i
(x
1
? x
3
)
?
?
?
?
+ ...
?
?
?
?
?
?
DII EMFF Final Review Aug. 29, 2003
Comparison - Golay Configurations
PSFs for the Golay configurations shown here will not change if
the apertures are shifted in any direction
DII EMFF Final Review Aug. 29, 2003
Question III:
? What forces must be transmitted between satellites to allow for all relative
degrees of freedom to be controlled?
– In 2-D, N spacecraft have 3N DOFs, but we are only interested in
controlling (and are able to control) 3N-2 (no translation of the center of
mass)
– For 2 spacecraft, that’s a total of 4:
? All except case (4) can be generated using axial forces (such as electrostatic
monopoles) and torques provided by reaction wheels
? Complete instantaneous control requires a transverse force, which can be
provided using either electrostatic or electromagnetic dipoles
1
2
3
4
DII EMFF Final Review Aug. 29, 2003
What is it NOT good for?
? Orbit Raising
? Bulk Plane Changes
?De-Orbit
? All these require rotating the system angular
momentum vector or changing the energy of the orbit
? None of these is possible using only internal forces
DII EMFF Final Review Aug. 29, 2003
Forces and Torques: Conceptual
NS S N NS
S
N
B
A
A B
? In the Far Field, the dipole field structure for electrostatic and
electromagnetic dipoles are the same
? The electrostatic analogy is useful in getting a physical feel for
how the transverse force is applied
? Explanation …
DII EMFF Final Review Aug. 29, 2003
EMFF Vehicle Conceptual Model
? In the Far Field, Dipoles add as vectors
? Each vehicle will have 3 orthogonal
electromagnetic coils
– These will act as dipole vector
components, and allow the magnetic
dipole to be created in any direction
? Steering the dipoles electronically will
decouple them from the spacecraft
rotational dynamics
? A reaction wheel assembly with 3
orthogonal wheels provides counter
torques to maintain attitude
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
– Calibration
–Movie
? Space Hardware Design
Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
Magnetic Dipole Approximation
? The interaction force between two arbitrary magnetic circuits is given by the
Law of Biot and Savart
I
1
I
2
O
? In general, this is difficult to solve, except for cases
of special symmetry
? Instead, at distances far from one of the circuits, the
magnetic field can be approximated as a dipole
[]
011 1
11
53 3
()
??? ? 3 ()
44
o
r
Br r
rr r
μμμμ μ
μ μ
ππ
?
????
=?=??
??
??
????
nullnull null
null
null
where its dipole strength μ
1
is given by the product of the total current around
the loop (Amp-turns) and the area enclosed
DII EMFF Final Review Aug. 29, 2003
Dipole-Dipole Interaction
? Just as an idealized electric charge in an external electric field can be assigned
a scalar potential, so can an idealized magnetic dipole in a static external
magnetic field, by taking the inner product of the two
? Continuing the analogy, the force on the dipole is simply found by taking the
negative potential gradient with respect to position coordinates
? In a similar manner, taking the gradient with respect to angle will give the torque
experienced by the dipole
? Since the Force results from taking a gradient with respect to position, and the
Torque does not, the scaling laws for the two are given as
2
UBμ=??
null
null
22
()
rr r
F UB Bμμ= ?? =? ? = ??
null null
null null
2
TU B
θ
μ=?? = ×
null
null
12
0
4
3
~
2
F
s
μ μ
μ
π
12
0
3
3
~
4
T
s
μ μ
μ
π
DII EMFF Final Review Aug. 29, 2003
How Far Apart Will They Work?
? Writing the force in terms of the coil radius (R), separation distance (s) and
total loop current (I
T
), the force scales as
? We see that for a given coil current, the system scales ‘photographically’,
meaning that two systems with the same loop current that are simply scaled
versions of one another will have the same force
? For design, it is of interest to re-write in terms of coil mass and radius, and
physical constants:
? The current state-of-the-art HTS wire has a value of
And the product of coil mass and radius becomes the design parameter.
4
2
0
3
~
2
T
R
FI
s
π
μ
??
??
??
2
2 4
72
0
4
331
~(10)()
22 2
CC C
CC
MI IR
FMR
R ss
π
μ
πρ
?
??
????
=
???? ??
????
??
14,444 /
C
I
Amkg
ρ
??
=?
??
??
DII EMFF Final Review Aug. 29, 2003
More of ‘How Far apart will they work’?
Force vs. Separation Distance
1.0E-04
1.0E-03
1.0E-02
1.0E-01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1 10 100
Separation (m)
Fo
r
c
e
(
N
)
MR = 0.1
MR = 0.3
MR = 1
MR = 3
MR = 10
MR = 30
MR = 100
MR = 300
kg-m
EMFORCE Testbed
Force vs. Separation Distance
1.0E-04
1.0E-03
1.0E-02
1.0E-01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
11010
Separation (m)
Fo
rc
e
(
N
)
MR = 0.1
MR = 0.3
MR = 1
MR = 3
MR = 10
MR = 30
MR = 100
MR = 300
kg-m
With further simplification:
2
4
1
~31.2( )
CC
FMR
s
The graph to the right shows a family of
curves for various products of M
C
and R
C
2
3
-7 C
2
I3m
(10 ) = 31.2
2 kg-sρ
??
??
??
2
3
-7 C
2
I3m
(10 ) = 312
2 kg-sρ
??
??
??
Example:
? 300 kg satellite, 2 m across, needs
10 mN of thrust, want M
C
< 30 kg
? EMFF effective up to 40 meters
DII EMFF Final Review Aug. 29, 2003
Far Field/Near Field Comparison
? The far field model does not work in the
near field
? (Separation/Distance)>10 to be within 10%
– Some configurations are more accurate
? A better model is needed for near-field
motion since most mission applications will
work in or near the edge of the near field
– For TPF, (s/d) ~ 3 - 6
β =10o
β =30o
β =45o
β =60o
β =90o
β =15o
β =30o
β =45o
β =60o
β =90o
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
2-D Dynamics of Spin-Up
? Spin-up/spin-down
– Spin-up from “static” baseline to rotating cluster for u-v plane filling
– Spin-down to baseline that can be reoriented to a new target axis
? Electromagnets exert forces/torques on each other
– Equal and opposite “shearing” forces
– Torques in the same direction
? Reaction wheels (RW) are used to counteract EM torques
– Initial torque caused by perpendicular-dipole orientation
– Reaction wheels counter-torque to command EM orientation
– Angular momentum conserved by shearing of the system
EM Torque
RW Torque
N
S
S N
DII EMFF Final Review Aug. 29, 2003
2-Satellite Spin-up
2
x centripical
F ma m rω==
0
y
Fmrω
β
=
=
null
0
4
0
4
0
3
3
(2cos cos sin sin )
4
3
(cos sin sin cos )
4
1
(cos sin 2sin cos )
4
AB
x
AB
y
AB
z
F
d
F
d
T
d
μμμ
α βαβ
π
μμμ
α βαβ
π
μμμ
α βαβ
π
=?
=? +
=? +
d
N
S
α
N
S
β
?x
?y
524
0
2
1
24
32 4
3
cos
4
AB
mrπ ωω
μμ
μ
ω
α
ωω
?
+
=
??
=?
??
+
??
null
null
? 6 DOF (4 Translational, 2 Rotational)
? 4 DOF (2 Translational, 2 Rotational)
? 2 Reaction wheels control 2 Rotational
DOF
? 2 dipole strengths and 2 dipole angles
to control 2 translational degrees of
freedom (relative motion)
– 2 extra degrees of freedom.
– Allows for many different spin-up
configurations
– Allows for different torque
distribution
– Become more complex with more
satellites
– Must solve non-linear system of
equations
DII EMFF Final Review Aug. 29, 2003
Torque Analysis
S
N
S
N
A
B
s
Magnet
A
Magnet
B
1
2sin
cos
53cos(2)
β
α
β
?
??
=± ±
??
+
??
? Shear forces are produced when the dipole axes
are not aligned.
? Torques are also produced when the shear
forces are produced (Cosv. of angular mom.)
? The torques on each dipole is not usually equal
– For the figure to the right
? Even for pure shear forces, (F
x
= 0) one can
arbitrarily pick one of the dipole angles.
1
2
A
B
τ
τ
=
DII EMFF Final Review Aug. 29, 2003
Satellite Formation Spin-Up
? Spin-up of complex formations can be achieved by utilizing magnetic dipoles.
? There are a number of possible combinations of magnet strengths and dipole
configurations to achieve a given maneuver.
? These different configurations cause different distribution of angular
momentum storage.
1 2 3 4
Time
20000
40000
60000
80000
100000
120000
Magentic Strengths
Outer
Inner
Center
1 2 3 4
Time
100
200
300
400
Acc . Ang . Mom .
Center
Outer
Inner
Magnetic Strength (A m
2
)
Angular Momentum (N m s)
600 kg s/c, 75m diameter formation, 0.5 rev/hr
DII EMFF Final Review Aug. 29, 2003
Steady State Rotations
? Spin-up of formations
are not restricted to
linear arrays
? Configurations of any
shape can be spun-up
? Shown here is a SPECS
configuration of 3
satellites in an
equilateral triangle.
Initial position
Steady state position
DII EMFF Final Review Aug. 29, 2003
Solving the EOM
? For a given instantaneous force profile, there are (3N-3) constraints (EOM),
and 3N variables (Dipole strengths).
– This allows us to arbitrarily specify one vehicle’s dipole
– Allows the user the freedom to control other aspects of the formation
especially angular momentum distribution
– For a specific choice of dipole, there are multiple solutions due to the
non-linearity of the constraints
? To determine the required magnetic dipole strengths
– Pick the magnetic dipole strengths for one vehicle
– Set the first equation equal to the desired instantaneous force and solve
for the remaining magnetic dipole strengths.
– There will be multiple solutions. Pick the solution that is most favorable
DII EMFF Final Review Aug. 29, 2003
Multiple Solutions
DII EMFF Final Review Aug. 29, 2003
3D Formations
? We also have the ability
to solve for complex 3D
motion of satellites.
DII EMFF Final Review Aug. 29, 2003
3D Formations
? Here is another example of a 3D configuration
DII EMFF Final Review Aug. 29, 2003
Choosing the Free Dipole
? Choose the free dipole such that a cost function is optimized
– Angular momentum distribution
– Dipole strength distribution
– Currently using Mathematica’s global minimization routine
? Simulated Annealing
? Genetic algorithms (Differential Evolution)
? Nelder-Mead
? Random Search
? Choose the free dipole based on a specific algorithm
– Aligning with the Earth’s magnetic field
– Favorable angular momentum distribution
DII EMFF Final Review Aug. 29, 2003
Angular Momentum Management
? Being able to control the angular momentum gained by the
individual satellites is crucial to the success of EMFF
? Because the torques and forces generated by EMFF are
internal, there is no way to internally remove excess angular
momentum from the system
– Angular momentum can be transferred from one spacecraft
to another
? Since EMFF systems do not employ thrusters, other innovative
methods must be used to remove the excess angular
momentum
– The formation must interact with its environment
? Using the Earth’s magnetic field
? Using differential J2 forces
DII EMFF Final Review Aug. 29, 2003
Earth’s Magnetic Field
? Many formation flight missions will operate in LEO.
? The electromagnets will interact
with the Earth’s magnetic field
producing unwanted forces and
torques on the formation
? The Earth’s magnetic field can be
approximated by a large bar
magnet with a magnetic dipole
strength of 8*10
22
? (EMFF Testbed ~ 2*10
4
)
DII EMFF Final Review Aug. 29, 2003
Earth’s Magnetic Field
? The Earth’s Magnetic field produces an insignificant
disturbance force, but a very significant disturbance
torque, due to the scaling of force and torque
~3*10
4
-3*10
-2
Nm~2*10
1
NmT ~ μ
0
(μ
1
μ
2
) / r
3
~1*10
4
-2*10
-3
N~1*10
-5
NF ~ μ
0
(μ
1
μ
2
) / r
4
2-100 m> 6,378,000 md
5*10
5
Am
2
8*10
22
Am
2
μ
Another Sat.Earth
DII EMFF Final Review Aug. 29, 2003
Possible Solutions for Dealing with the
Earth’s Magnetic Field
? Ignore the disturbance forces from the Earth’s magnetic field on the
formation as a whole
– This frees up the arbitrary dipole, but disturbance forces are still
accounted for.
? Periodically alternate the magnetic dipole directions, so that the
accumulated torques average to zero
? Turn off all the satellites but one, and use the electromagnets like
torque rods to dump the angular momentum
? Choose the arbitrary dipole wisely so that the total acquired angular
momentum on the formation is zero
? Choose the arbitrary dipole wisely so that you can use the Earth as
a dump for angular momentum.
DII EMFF Final Review Aug. 29, 2003
The Earth as a Momentum Dump
? Use the Earth’s dipole to our advantage by transferring
angular momentum to the Earth
– Already done for single spacecraft using torque rods
– Can be expanded for use with satellite formations
? Strategy:
– Pick a satellite to dump momentum
– Turn up its dipole strength to maximum
– Align the dipole to optimize momentum exchange
– Solve the remaining dipoles for the required
instantaneous forces
– Once the required momentum has been dumped, pick
another satellite that needs to dump momentum
DII EMFF Final Review Aug. 29, 2003
Simulation Results
? Satellites are undergoing a specific forcing profile in the presence of
the Earth’s magnetic field
– This way the satellites that are not dumping momentum are still being
disturbed by the Earth’s magnetic field.
? Each satellite starts off with excess angular momentum
? The satellite with the most excess momentum is selected for angular
momentum dumping
? The formation is then maintained to have H<100
DII EMFF Final Review Aug. 29, 2003
EMFF Simulator
? Currently designing a software simulator to test different angular
momentum control schemes
? Built in Mathematica, it has the ability to provide MatLab style outputs
? It will have the ability to test
control algorithms in the
presence of the Earth’s
magnetic field or under the
influence of the J2 disturbance
force.
? Currently being used to verify
angular momentum dumping
algorithms in the presence of
the Earth’s magnetic field.
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
? Motivation:
– Dynamic analyses must be performed to verify the stability and
controllability of EMFF systems.
? Objective:
– Derive the governing equations of motion for an EMFF system:
? Analyze the relative displacements and rotations of the bodies.
? Include the gyroscopic stiffening effect of spinning RWs on the
vehicles.
– Linearize the equations, and investigate the stability and controllability of
the system.
– Design a closed-loop linear controller for the system.
– Perform a closed-loop time-simulation of the system to assess the model
dynamics and control performance.
– Experimentally validate the dynamics and control on a simplified
hardware system.
Objective : Multi-Vehicle EMFF Analysis
DII EMFF Final Review Aug. 29, 2003
? Two-spacecraft array
– Each has three orthogonal electromagnets
? EM pointing toward other spacecraft
carries bulk of centripetal load; others
assist in disturbance rejection
– Each has three orthogonal
reaction wheels, used for system
angular momentum storage and
as attitude actuators
? State vector:
x r φΨα
1
α
2
α
3
β
1
β
2
β
3
r
·
φ
·
Ψ
·
α
·
1
α
·
2
α
·
3
β
·
1
β
·
2
β
·
3
T
=
3-D Dynamics of 2-S/C EMFF Cluster
r
Z
Y
φ
ψ
e
?
r
e
?
φ
e
?
ψ
e
?
x
e
?
y
e
?
z
X
Spacecraft A
Spacecraft B
r
e
?
r
e
?
φ
e
?
ψ
Spacecraft A
Z
YX
Piece of an imaginary
sphere centered at
the origin of the
X, Y, Z frame
DII EMFF Final Review Aug. 29, 2003
3-D Nonlinear Translational Equations
? Translational equations of motion for spacecraft A:
?In e
r
, e
φ
, e
ψ
components:
? And the forcing terms are of the form:
( )
2/21/22/11/1/
11
BABABABABA
FFFF
m
F
m
r
nullnullnullnullnull
nullnull
null
+++==
=r
nullnull
null
r
··
rψ
·
2
– rφ
·
2
ψ
2
cos–
2r
·
φ
·
ψcos rφ
··
ψcos 2rφ
·
ψ
·
ψsin–+
2r
·
ψ
·
rψ
··
rφ
·
2
ψsin ψcos++
??
??
??
??
=
3μ
0
μ
A
μ
B
64πmr
4
-----------------------
sα
1
cα
2
sβ
1
cβ
2
2cα
1
cα
2
cβ
1
cβ
2
sα
2
sβ
2
+–
cα
2
cβ
2
sα
1
cβ
1
sβ
1
cα
1
+()
cβ
1
–cβ
2
sα
2
cα
1
cα
2
sβ
2
–
??
??
??
??
=
m
F
BA 1/1
null
DII EMFF Final Review Aug. 29, 2003
3-D Nonlinear Rotational Equations
? Rotational equations of motion for spacecraft A:
? Torques:
θ
·
x
θ
·
y
θ
·
z
??
??
??
??
??
??
??
A
α
·
3
α
·
1
sα
2
–
α
·
2
cα
3
α
·
1
cα
2
sα
3
+
α
·
1
cα
2
cα
3
α
·
2
sα
3
–
??
??
??
??
??
=
θ
··
x
θ
··
y
θ
··
z
??
??
??
??
??
??
??
A
α
··
3
α
··
1
sα
2
– α
·
1
α
·
2
cα
2
–
α
··
2
cα
3
α
·
2
α
·
3
sα
3
α
··
1
cα
2
sα
3
α
·
1
α
·
2
sα
2
sα
3
– α
·
1
α
·
3
cα
2
cα
3
++–
α
··
1
cα
2
cα
3
α
·
1
α
·
2
sα
2
cα
3
– α
·
1
α
·
3
cα
2
sα
3
– α
··
2
sα
3
– α
·
2
α
·
3
cα
3
–
=
I
rr s,
I
rr w,
+00
0 I
rr s,
I
rr w,
+0
00I
zz s,
θ
··
x
θ
··
y
θ
··
z
??
??
??
??
??
??
??
A
0 ?
zw,
I
zz w,
0
?
zw,
I
zz w,
–0
000
θ
·
x
θ
·
y
θ
·
z
??
??
??
??
??
??
??
A
+=
10 0
0cα
3
sα
3
0sα
3
–cα
3
cα
2
0sα
2
–
01 0
sα
2
0cα
2
cα
1
sα
1
0
sα
1
–cα
1
0
001
T
r
T
φ
T
ψ
??
??
??
??
??
A
2/21/22/11/1/ BABABABABA
TTTTT
nullnullnullnullnull
+++=
T
r
T
φ
T
ψ
??
??
??
??
??
A
μ
0
– μ
A
μ
B
32πr
3
----------------------
sα
2
sβ
1
cβ
2
sα
1
cα
2
sβ
2
–
cα
1
cα
2
sβ
2
2sα
2
cβ
1
cβ
2
+
cα
1
cα
2
sβ
1
cβ
2
2sα
1
cα
2
cβ
1
cβ
2
+
??
??
??
??
=
1/1 B
DII EMFF Final Review Aug. 29, 2003
Linearization: Nominal Point
? Conservation of Angular Momentum:
? Nominal State Trajectory:
I
zz w,
?
zw,
φ
·
0
+()I
zz s,
mr
0
2
+()φ
·
0
+0=
I
zz w,
?
zw,
mr
0
2
φ
0
·
+0=≈
x
0
r
0
φ
0
Ψ
0
α
10,
α
20,
α
30,
β
10,
β
20,
β
30,
r
·
0
φ
·
0
Ψ
·
0
α
·
10,
α
·
20,
α
·
30,
β
·
10,
β
·
20,
β
·
30,
T
=
r
0
φ
0
t()0 0 0 0 0 0 0 0 φ
·
0
0 0 0 0 0 0 0
T
=
?
DII EMFF Final Review Aug. 29, 2003
10000 000 0
0 r
0
00 000 0
00r
0
00 000 0
00 0I
zz s,
0 000 0
00 0 0 I
rr s,
I
rr w,
+ 000 0
00 0 0 0 I
rr s,
I
rr w,
+00 0
00 0 0 0 0 I
zz s,
00
00 0 0 0 0 0 I
rr s,
I
rr w,
+0
0000 000I
rr s,
I
rr w,
+
?r
··
?φ
··
?ψ
··
?α
··
1
?α
··
2
?α
··
3
?β
··
1
?β
··
2
?β
··
3??
??
??
??
??
??
??
??
??
??
??
02r
0
φ
·
0
–0000000
2φ
·
0
0000 000 0
00000 000 0
00000 000 0
00000mr
0
2
φ
·
0
00 0
0000m– r
0
2
φ
·
0
00 0 0
00000 000 0
00000 000mr
0
2
φ
·
0
00000 00m– r
0
2
φ
·
0
0
?r
·
?φ
·
?ψ
·
?α
·
1
?α
·
2
?α
·
3
?β
·
1
?β
·
2
?β
·
3??
??
??
??
??
??
??
??
??
??
??
+
5φ
·
0
2
– 00 0 000 00
000 c
1
00c
1
00
00r
0
φ
·
0
2
0 c–
1
00 c
1
–0
000 2c–
0
00c–
0
00
000 0 2c–
0
00 c–
0
0
000 0 000 00
000 c–
0
002– c
0
00
000 0 c–
0
00 2– c
0
0
000 0 000 00
?r
?φ
?ψ
?α
1
?α
2
?α
3
?β
1
?β
2
?β
3
??
??
??
??
??
??
??
??
??
??
??
+
c
2
00c
2
0 0 000000
0
c
2
2
-----–00
c
2
2
----– 0 000000
00
c
2
2
-----–00
c
2
2
-----–000000
02c
3
00c
3
0 K
T
00000
00 2c
3
–00c
3
–0K
T
0000
00 0 00 0 00K
T
000
0 c
3
002c
3
0000K
T
00
00 c
3
–002c
3
–0000K
T
0
00 0 00 0 00000K
T
?μ
A1
?μ
A2
?μ
A3
?μ
B1
?μ
B2
?μ
B3
?i
RW A1,
?i
RW A2,
?i
RW A3,
?i
RW B1,
?i
RW B2,
?i
RW B3,??
??
??
??
??
??
??
??
??
??
??
??
=
Motion in X-Y Plane: Linearized
Equations
c
0
mr
0
2
φ
·
0
2
–
3
------------------≡
c
1
r
0
φ
·
0
2
–
2
--------------
≡
c
2
φ
·
–
0
3μ
0
32πmr
0
3
-------------------
≡
Gyroscopic
Stiffening Terms
DII EMFF Final Review Aug. 29, 2003
Full-state system (n=18) has eigenvalues:
— Several poles on the
imaginary axis and
one unstable pole
— λ
7,8
at +/- array spin-rate
— Poles move away from
origin as increases
EMFF Stability
λ
123456,,,,,
0= λ
78,
φ
·
0
±= λ
910,
iφ
·
0
±=
λ
11 12,
i
r
0
φ
·
0
I
rr s,
I
rr w,
+()
-------------------------------- mmr
0
2
I
rr s,
I
rr w,
+
3
---------------------------+
??
??
±=
λ
13 14,
i
r
0
φ
·
0
I
rr s,
I
rr w,
+()
--------------------------------
mmr
0
2
I
rr s,
I
rr w,
++()±=
λ
15 16,
ir
0
φ
·
0
m
3I
zz s,
-------------±=
λ
17 18,
ir
0
φ
·
0
m
I
zz s,
---------±=
0
φ
null
DII EMFF Final Review Aug. 29, 2003
Current system is in 2
nd
order form:
Place in 1
st
order form:
Form controllability matrix:
System is fully controllable because C has full rank
EMFF Controllability
21
[B AB A B A B]
n
C
?
= …
n : number of states
uxx BA +=
null
T
]
~~
[ xxx
null
=
A
0 I
M
1–
K– M
1–
C–
= B
0
M
1–
F
=
rank C
()
18 n==
FuxKxCxM =++
~~~
nullnullnull
DII EMFF Final Review Aug. 29, 2003
EMFF Linear Controller Design
From state-space equation of motion:
Form the LQR cost function:
Choose relative state and control penalties:
— ?r : 10 : 5 ?φ : 10
-15
: 3 ?ψ : 1 : 1
— All Euler angles and their derivatives : 1
— All electromagnets, all reaction wheels : 1
The cost, J, is minimized when:
R
xx
: state penalty matrix R
uu
: control penalty matrix
[]
∫
∞
+=
0
dtRRJ
uu
T
xx
T
uuxx
xxu KPBR
T
uu
?=?=
?1
PBPBRPAPAR
T
uu
T
xx
1
0
?
?++=
Algebraic Ricatti Equation (A.R.E.)
uxx BA +=
null
[ ] xxuxx
CL
ABKABA =?=+=
null
rnull? φ
null
? ψnull?
DII EMFF Final Review Aug. 29, 2003
Simulation of EMFF Dynamics
Closed-loop time simulations were performed of both the
nonlinear and linearized equations of motion
— Both employ the same linear feedback controller
“Free vibration” response was investigated
— Initial condition : deviation from nominal state of one or more
degrees of freedom (?r in the results shown here)
— Closed-loop response to perturbed initial condition is simulated
— Perhaps offers more insight than simulating response to random
disturbances
Results demonstrate:
— the range in which the linearized equations are valid
— the range in which linear control is sufficient
— the importance of the relative control penalties chosen for the
various degrees of freedom
DII EMFF Final Review Aug. 29, 2003
Simulation of EMFF Dynamics Results
(I)
Initial conditions: 1% deviation
from nominal array radius
— Nonlinear and linear simulations
diverge
— System remains stable in both
simulations
Initial conditions: 0.001% deviation
from nominal array radius
— Simulations of nonlinear and
linearized equations are identical,
except for small numerical error in
angles ?ψ, ?α
2
, ?α
3
, ?β
2
, ?β
3
— Both use linear controller
DII EMFF Final Review Aug. 29, 2003
Simulation of EMFF Dynamics Results
(II)
Initial conditions: 4% deviation
from nominal array radius
— Divergence of radial separation shows
linear control not sufficient in this case.
Redesign with greater penalty on ?r?
Investigate nonlinear control techniques?
— Linear simulation does not capture
divergence of dynamics.
Initial conditions: 3% deviation
from nominal array radius
— Radial separation remains stable
— Elevation angle of array may go unstable
(probably numerical error).
— Check by increasing relative penalty on
?ψ and redesigning controller.
DII EMFF Final Review Aug. 29, 2003
xa
cent
2
?=
4
2
2
16mx
c
xx
μ
??=
nullnull
? 1-D simplification of linearized 3-D dynamics
? Constant spin rate for data collection
? Relative radial position maintenance: disturbance rejection
π
μμ
2
3
,
)2(
0
4
2
== c
x
c
F
EM
m
F
ax
EM
cent
?=?
nullnull
N SN S
a
centrifugal
F
EM
F
EM
?
2x
a
centrifugal
Simplified System: Steady-State Spin
DII EMFF Final Review Aug. 29, 2003
avg
x
x
x
x
μ
δμδδ
2
0
2
0
2
??=??
nullnull
Perturbation Analysis:
Perturbed Dynamics of Steady-State
Spin
δμμμδ +=+=
avg
xxx ,
0
()
( )
()
4
0
2
0
2
0
16 xxm
c
xxxx
avg
δ
δμμ
δδ
+
+
?+?=+
nullnullnullnull
c
xm
avg
5
0
2
2
16 ?
=μ
Steady-State Control Perturbation Equation
Use binomial formula
to expand terms
Neglect H.O.T.
Solve for S.S. Control
when 0=xnullnull
?±=
2,1
s
Unstable dynamics:
* Same as λ
7,8
in 3-D EMFF
system analysis
uxx BA +=
null
avg
x
x
x
x
x
x
x
x
μ
δμ
δ
δ
δ
δ
?
?
?
?
?
?
??
+
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
=
?
?
?
?
?
?
?
?
?
?
?
?
2
0
0
2
0
0
2
0
0
10
nullnullnull
null
DII EMFF Final Review Aug. 29, 2003
? Follow same control design
process as for full-state, 3-D system:
? Select state and control penalties:
? Solve the A.R.E. analytically by
enforcing
that P must be positive semidefinite:
? The displacement and velocity
feedback gains are then:
Linear Control Design
[]
∫
∞
+=
0
dtRRJ
uu
T
xx
T
uuxx
xxu KPBR
T
uu
?=?=
?1
ρ=
uu
R
?
?
?
?
?
?
=
00
0λ
xx
R
0
2212
1211
≥
?
?
?
?
?
?
=
PP
PP
P
[]
2212
2
1
2
PPPBRK
T
uu
ρ
?
==
?
DII EMFF Final Review Aug. 29, 2003
? Now solve for the closed-loop dynamic matrix, where:
? Evaluate as increases from 0 ?
? The closed-loop poles for
the most efficient controller
lie along this curve.
State-Space Analysis
[]xxuxx
CL
ABKABA =?=+=
null
xu K?=
ρ
λ
∞
DII EMFF Final Review Aug. 29, 2003
? Nearly frictionless
1-dimensional airtrack
? Can be set up in a
stable or unstable
configuration, depending
on the tilt angle
? Unstable mode has dynamics
nearly identical to a 1-DOF
steady-state spinning cluster!
– Closing the loop on the unstable configuration will demonstrate an
ability to control systems such as the steady-state spinning cluster.
Experimental Validation: 1-D Airtrack
Ultrasound
displacement sensor
Free magnet
on “slider”
Fixed
electromagnet
DII EMFF Final Review Aug. 29, 2003
Free Permanent
Magnet
Fixed
Electromagnet
5
0
0
2,1
6
mx
is
p
π
μμ
±=
Stable poles:
Similar linearization, state-space
analysis, and
LQR control design
to
steady-state spin system
Open-loop step response
— Very light damping means poles are nearly on
the imaginary axis, as expected
Closed-loop step response has reduced
overshoot and increased damping
[ ]33.474.11=K
Optimal gains:
Pole?Zero Map
Real Axis
Imag Axis
?2 ?1.5 ?1 ?0.5 0 0.5 1 1.5 2
?2
?1.5
?1
?0.5
0
0.5
1
1.5
2
0 10 20 30 40 50 60
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
Step Response: LQR Control of Stable Airtrack System
Time [seconds]
Separation Distance [meters]
Closed Loop
Open Loop
Experimental Results: Stable Airtrack
DII EMFF Final Review Aug. 29, 2003
0 10 20 30 40 50 60
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Time [seconds]
Separation Distance [meters]
LQR Control of Unstable Airtrack System
Closed Loop
Open Loop
Open Loop
Pole?Zero Map
Real Axis
Imag Axis
?2 ?1.5 ?1 ?0.5 0 0.5 1 1.5 2
?2
?1.5
?1
?0.5
0
0.5
1
1.5
2
Unstable dynamics:
[ ]88.056.2?=K
Optimal gains:
Fixed
Electromagnet
Free Permanent
Magnet
5
0
0
2,1
6
mx
s
p
π
μμ
±=
Experimental Results: Unstable
Airtrack
Similar dynamics and
control design
to
steady-state spin and stable-airtrack
Open-loop response is divergent
Closed-loop response is stable!
Stabilizing this system means we should
be able to perform steady-state control
and disturbance rejection for a spinning
cluster!
DII EMFF Final Review Aug. 29, 2003
Open-loop response is divergent.
— Constant current is applied to EM
— Magnets diverge from steady-state
separation distance
Fall apart if disturbed one way
Come together if disturbed the
other way
Closed-loop response is stable!
— Oscillates at about ~0.2 Hz
— Maximum displacement from
steady-state location is ~1 cm
— Performance limitations due to
model uncertainty and amplifier
saturation
Video: Control of Unstable Airtrack
DII EMFF Final Review Aug. 29, 2003
? Modeled the dynamics of a two-vehicle EMFF cluster
– Nonlinear, unstable dynamics
– Linearized dynamics about a nominal trajectory (steady-state spin)
– Stability: 3-D system has six poles at the origin, ten poles along the
imaginary axis, and a stable/unstable pair of poles at the array spin-rate
– Controllability: System is fully controllable with 3 electromagnets and 3
reaction wheels per vehicle
? Simulated two-vehicle EMFF closed-loop dynamics
– Demonstrated stabilization of unstable nonlinear dynamics using linear control
– We can investigate for future systems:
? whether linear control is sufficient for a given configuration
? what the “allowable” disturbances are from the nominal state
? how the relative state and control penalties may improve the closed-loop
behavior
? Validated EMFF dynamics and closed-loop control on simplified hardware system
– Airtrack: stable and unstable configurations (1-DOF)
– Demonstrated stabilization of an unstable system with dynamics similar to an
EMFF array undergoing steady-state-spin
Control Summary
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
EMFF Roadmap
DII EMFF Final Review Aug. 29, 2003
EMFF Applications in 10-20 Years
-3000
-1500
0
1500
3000-3000
-1500
0
1500
3000
-1500
-750
0
750
1500
Velocity vector (m)
Cross Axis (m)
Z
eni
t
h
-
N
ad
i
r
(
m
)
Docking
Sparse Apertures
Cluster Reconfiguring
Image from 1999 TPF Book
EMFF Secondary Mirrors
DII EMFF Final Review Aug. 29, 2003
EMFF Applications in 30-40 Years
Reconfigurable Arrays & Staged Deployment
Reconfigurable Artificial Gravity
Space Station
Protective magnetosphere
DII EMFF Final Review Aug. 29, 2003
Additional Mission Applications
Distributed Optics
Non-Keplerian Orbits
Primary
mirror
Primary
mirror
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
‘Stationary’ Orbits
? For telescopes and other observation missions with
an extended look time, holding an fixed observation
angle is important
? Satellite formations in Earth’s orbit have an intrinsic
rotation rate of 1 rev/orbit
? EMFF can be used to stop this rotation and provide
a steady Earth relative angle.
? Using Hill’s equations…
2
2
32
2
x
y
z
xnxny a
ynxa
znza
=++
=? +
=? +
nullnull null
nullnull null
nullnull
22
? ?3f xmn x zmn z=? +
null
2
4
3
yz
xf mn xz
xy
ττ
??
??
=× = ?
??
??
null
nullnull null
? Unless the required force vector aligns with the
position vector, torques are produced
– Zero torque solutions are
? Holding a satellite in the nadir direction
? Holding a satellite in the cross-track
direction
? For other pointing angles, torques will be
produced
? Any angular momentum buildup can be removed
by:
– Moving to an opposite position.
– Interacting with the Earth’s magnetic field
DII EMFF Final Review Aug. 29, 2003
Rotating Linear Array: 2 vs. 3 Spacecraft
Mission Efficiency metric:
Bm
B
ain
F
toto
2
4
422
2
1
2
3
ωπμ ==
arrayo
mc
w
J =
totarray
mm 2=
Bm
B
ininin
aF
tot
oo
B
ooii
o
2
4
22
4
2
4
2
1
)(2
3
ωπμ =
?
?
?
?
?
?
?
?
+=
Two Spacecraft
75.2
2
3
=
J
J
5
3
B
c
o
o
πμ
=
Three Spacecraft
inner
outer
tottotarray
mmm +=2
? Adding combiner almost triples mission efficiency
? Trend Continues ? adding more spacecraft increases
mission efficiency
DII EMFF Final Review Aug. 29, 2003
RLA: Mission Efficiency Trends
? Normalized Mission Efficiency
4
1
1
1
n
i
n
i
n
J
n
?
?
=
??
??
?
??
=
∑
2
0
o array
nia
J
cm
=
? Comparing J
3
/J
2
, then
J
4
/J
3
, J
5
/J
4
, J
6
/J
5
, etc.
? Diminishing returns of
adding S/C
DII EMFF Final Review Aug. 29, 2003
3 S/C RLA: EMFF System Trades
array
inner total
MMγ=
1
2
array
outer total
MM
γ?
=
?Define Mass Fractions:
?Identical or Mother-Daughter Configuration
Center Spacecraft experiences no
translation ? no mass penalty ?
suggests larger center spacecraft
?Identical Configuration is non-optimal
?Higher rotation rate for mother-daughter
configuration for fixed masses
DII EMFF Final Review Aug. 29, 2003
Case Study: Sparse aperture (TPF)
sc dry sa core coil
mm mm m=+++
? Compare total system mass for various
propulsion options with EM option for
the TPF mission (4 collector and 1
combiner spacecraft)
? Array is to rotate at a fixed rotation rate
(ω = 1rev/2 hours)
? All collector spacecraft have same EM
core and coil design
? All spacecraft have the same core
? Force balancing equations:
s/3 s/6 s/6 s/3
NSNSNSNS NS
EM mass components
Superconducting wire (m
sc
)
Density (ρ
St
) 13608 kg m
-3
Copper coil (m
coil
)
Density (ρ
Cu
) 8950 kg m
-3
Resistivity (ρ)1.7x10
-8
?m
Solar Array (m
sa
)
Power to mass conv (P
conv
)25 W kg
-1
*
Source: TPF Book (JPL 99-3)
TPF spacecraft
*
(m
dry
)
Collector Spacecraft
Dry 600 kg, 268 W
Propulsion 96 kg, 300 W
Propellant 35 kg
Combiner Spacecraft
Dry 568 kg, 687 W
Propulsion 96 kg, 300 W
Propellant 23 kg
1 12131415
cent M M M M
F FFFF=+++
221232425
cent M M M M
F FFFF=? + + +
1 2 3
4 5
DII EMFF Final Review Aug. 29, 2003
Case Study: Sparse aperture (TPF-2)
? Cold Gas - Low I
sp
, high propellant requirements
– Not viable option
? PPTs and Colloids - Higher I
sp
– still significant propellant over mission lifetime
? FEEPs – Best for 5 yr mission lifetime
– Must consider contamination issue
– Only 15 kg mass savings over EMFF @ 5 yr mark
? EM coil (R = 4 m) (M
tot
= 3971 kg)
– Less ideal option when compared to FEEPs even for
long mission lifetime
? EM Super Conducting Coil (R = 2 m) (M
tot
= 3050 kg)
– Best mass option for missions > 6.8 years
– No additional mass to increase mission lifetime
– Additional mass may be necessary for CG offset
? Estimated as ~80 kg
? Massachusetts Institute of
Technology
? Space Systems Laboratory
? Lockheed Martin Corporation
? Advanced Technology Center
EMFF Testbed
DII EMFF Final Review Aug. 29, 2003
EMFFORCE Project Overview
? Goal: Demonstrate the feasibility of electromagnetic
control for formation flying satellites
? Design and build a testbed to demonstrate 2-D formation flight
with EM control
– Proof of concept
– Traceable to 3-D
– Validate enabling technologies
? High temperature superconducting wire
From Design to Reality
Metrology and
Comm
Gas supply
tank
Magnet and
cryogenic
containment
Electronics
boards
Batteries
Base and
gas
carriage
Reaction
wheel
DII EMFF Final Review Aug. 29, 2003
Testbed Overview
? Functional Requirements:
– System will contain 2 vehicles
– Robust electromagnetic control will replace thrusters
– Each vehicle will be:
? Self-contained (no umbilicals)
? Identical/interchangeable
? Vehicle Characteristics
– Each with 19 kg mass, 2 electromagnets, 1 reaction wheel
? Communication and processing
– 2 internal microprocessors (metrology, avionics/control)
– Inter-vehicle communication via RF channel
– External “ground station” computer (operations, records)
? Metrology per vehicle
– 1 rate gyro to supply angular rate about vertical axis
– 3 ultrasonic (US) receivers synchronized using infrared (IR) pulses
DII EMFF Final Review Aug. 29, 2003
Electromagnet Design
? Coil wrapped with alternating layers of wire and Kapton
insulation
– 100 wraps
– Radii of 0.375m and 0.345m
? Toroid-shaped casing: Insulation & Structural component
? Operable temperature at 77 K
? Surround by liquid nitrogen
? American Superconductor Bi-2223 Reinforced
High Temperature Superconductor Wire
– Dimensions
? 4.1 mm wide
? 0.3 mm thick
? 85 m length pieces
– Critical Current
? 115 amps, 9.2 kA/cm2
– Below 110 K
DII EMFF Final Review Aug. 29, 2003
Containment System Design
? Requirements:
– Keep the wire immersed in liquid
nitrogen.
– Insulate from the environment the wire
and the liquid N
2
.
? Non-conductive material.
– Stiff enough to support liquid N
2
container and its own weight.
? Material: Foam with fiber glass wrapped
around it.
Top view of half of the Container
Section of
container
DII EMFF Final Review Aug. 29, 2003
Power Subsystem
? Coil & Reaction Wheel Power:
– Rechargeable NiMH D-cell batteries
– MOSFET controller – uses H-bridge
circuit to control current through gates
– 20 minute power duration
? Coil: 100 Amps @ 5 Volts
Electromagnet
1
4
2
3
- 3.6 V
+
35 Amps
35 A
35 A
> 100 Amps
DII EMFF Final Review Aug. 29, 2003
Air Carriage and Reaction Wheel
? Reaction Wheel
? Store angular momentum
– Provide counter-torques to electromagnets
– Provide angular control authority
– 0.1 Nm Torque at 10 Amps
? Flywheel Requirements:
– non-metallic ? Urethane Fly Wheel
– Maximum wheel velocity at 7000 RPM
? Motor tested in EM field with no variation in performance
? 2-D Friction-less environment provided by gas carriage
– allows demonstration of shear forces, in concert with
reaction wheel
– Porous Membrane, Flat air bearings provide
pressurized cushion of gas
–CO
2
gas supply: rechargeable compressed gas
tank, 20 minute duration
DII EMFF Final Review Aug. 29, 2003
Model Calibration
0.1 0.2 0.3 0.4 0.5 0.6 0.7
5
10
15
20
25
B
axial
vs. Axial Distance
Axial Distance [m]
B [gauss]
Gauss meter
Inner
coil
stack
B
axial
vs. Radius
Radius [m]
B [gauss]
0.1 0.2 0.3 0.4 0.5 0.6
-10
10
20
30
Gauss
meter
Inner
coil
stack
B
radial
vs. Radius
Radius [m]
B [gauss]
0.1 0.2 0.3 0.4 0.5 0.6
10
20
30
40
Gauss
meter
Inner
coil
stack
? B-field measurements
– Axial and Radial B-field
measurements taken at varying
radii.
– Inner coil stack
? .67 m inner diameter
? 40 turns
? I = 30 amps
DII EMFF Final Review Aug. 29, 2003
Degrees of Freedom Validation
? Initially we had problems demonstrating shear forces
? The reaction wheel is designed for small shear forces
? Vehicle tends to ‘stick’ to table, so larger forces are needed to move the
vehicle
? Larger shear forces produce larger torques
? The torque generated would cause the vehicle to rotate
? As the vehicle rotated, the dipoles aligned causing the vehicles to attract
? Used Vehicle’s ability to steer the dipole to compensate
NS
S
N
B
A
DII EMFF Final Review Aug. 29, 2003
EMFF: Validation of Degrees of Freedom
DII EMFF Final Review Aug. 29, 2003
Testbed Future Work
? Control Testing
a. One vehicle fixed – disturbance rejection
b. One vehicle fixed – slewing, trajectory following
c. Both vehicles free – disturbance rejection
d. Both vehicles free – slewing, trajectory following
e. Spin-up
? Vehicle Design
– Containment system redesign: Plastic or copper tubing
– Reaction Wheel
? Motor is too weak to counteract high torque levels
? Reaction wheel is also possibly undersized
? Three vehicle Control Testing
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
Cryogenic Containment
? Significant research concerning maintaining
cryogenic temperatures in space
– Space Telescope Instrumentation
– Cryogenic propellant storage
? Spacecraft out of Earth orbit can use a sunshield
that is always sun-pointing to reflect radiant energy
away
? For Earth orbit operation, this won’t work, since
even Earth albedo will heat the ‘cold’ side of the
spacecraft
? A cryogenic containment system, similar in concept to that used for the EMFF testbed
must be implemented, using a combination of a reflective outer coating, good insulation,
and a cyo-cooler to extract heat from the coil
? Using a working fluid to carry heat around to the cry-cooler will be explored, or possibly
using the wire itself as the thermal conductor
DII EMFF Final Review Aug. 29, 2003
Efficient High Current Supplies
? The existing controller for the testbed was based on a
pulse width modulated controller found for use with
radio controlled cars and planes
? An H-bridge is used to alternate applied potential to the
coil, with the next current delivered dependent on the
amount of time the voltage is applied in a given direction
? The drawback is that current is always flowing through
the batteries, which both provide a power sink as well as
dissipate heat
? One solution may be to incorporate very high Farad
capacitor instead of a batter, to reduce the internal
resistance
? Alternatively, a method of ‘side-stepping’ the storage
device altogether may be employed, allowing the
current to free-wheel during periods of low fluctuation
DII EMFF Final Review Aug. 29, 2003
High B-field effects: Findings in literature
? NASA reports, Lockheed Martin reports, other contractors (when
available), IEEE journal articles
? Nothing for very high fields (0.1 T and above)
? Effects of earth’s magnetic field (0.3 gauss or so)
? Effects of on-board field sources such as
– Magnetic latching relays
– Traveling wave tubes
– Tape recorders
– Coaxial switches
– Transformers
– Solenoid valves
– Motors
DII EMFF Final Review Aug. 29, 2003
Vulnerable equipment
? All these fields are much smaller than what is being projected
for magnetic steering coils
? Equipment traditionally known to be susceptible to magnetic
effects:
– Magnetometers
– Photomultipliers
– Image-dissector tubes
– Magnetic memories
– Low-energy particle detectors
– Tape recorders
? Digicon detectors in Hubble FOS were found to be vulnerable to
magnetic effects
? Quartz-crystal oscillators ditto (AC fields)
DII EMFF Final Review Aug. 29, 2003
High field concerns
? Other effects may come into play that are negligible at low field
strengths
– Eddy currents in metal harnesses
– Hall effects in conductors
– Effects in semiconductors?
? Most EMI requirements hard to meet
? Shielding requirement translates into a mass penalty
? Pursuing more literature results, but this is effectively a new
regime – may require testing
DII EMFF Final Review Aug. 29, 2003
Shielding Considerations
? Attenuation of a DC magnetic field resulting from an enclosure scales approximately as
? Where μ is the permeability, ? is the thickness of the material, and R is the characteristic
radius of enclosure
? Some high permeability materials:
A =
μ
2
?
R
? Reducing a 600 G (0.06 T) field to ambient (0.3 G) requires an attenuation of
2x10
3
, or a minimum ?/R of 0.01
? This is .1 mm thickness for each 10 cm of radius enclosed
Material Density (lbs/cu-in) Permeability Saturation (G)
Amumetal 0.316 400000 8000
Amunickel 0.294 150000 15000
ULCS 0.0283 4000 22000
DII EMFF Final Review Aug. 29, 2003
Further Shielding Considerations
? Geometry
– Shielding acts to divert field lines around components
– Gentle radii are better for re-directing field lines than sharp corners
?Size
– Smaller radii are more effective, so shielding should envelop the component
to be protected as closely as possible
? Continuity
– Separate pieces should be effectively connected either mechanically or by
welding to insure low reluctance
? Closure
– Components should be completely enclosed, even if by a rectangular box to
shield all axes
? Openings
– As a rule, fields can extend through a hole ~5x the diameter of the hole
? Nested Shields
– In high field areas, multiple shield layers with air gaps can be used very
effectively. Lower permeability, higher saturation materials should be used
closer to the high field regions
DII EMFF Final Review Aug. 29, 2003
Shielding with Auxiliary Coils
? In addition to high permeability materials, shielding can be achieved locally using
Helmholtz coils
? An external field can be nullified with an arrangement of coils close to the region
of interest
? The small coil size requires proportionally smaller amp-turns to achieve nulling of
the field
– Will not significantly affect the main field externally
DII EMFF Final Review Aug. 29, 2003
Outline
? Motivation
? Fundamental Principles
– Governing Equations
– Trajectory Mechanics
– Stability and Control
? Mission Applicability
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? Mission Analyses
– Sparse Arrays
– Filled Apertures
– Other Proximity Operations
? MIT EMFFORCE Testbed
–Design
–Calibration
–Movie
? Space Hardware Design Issues
– Thermal Control
– Power System Design
– High B-Field Effects
? Conclusions
DII EMFF Final Review Aug. 29, 2003
Conclusions
? There are many types of missions that can benefit from propellantless
relative control between satellites
– Provides longer lifetime (even for aggressive maneuvers)
– Reduces contamination and degradation
? Angular momentum management is an important issue, and methods are
being developed to de-saturate the reaction wheels without using thrusters
? Preliminary experimental results indicate that we are able to perform
disturbance rejection in steady state spin dynamics for multiple satellites
? Optimal system sizing has been determined for relatively small satellite
arrays. Currently larger formations are being investigated
? Although low frequency magnetic interference data is difficult to find,
shielding against the relatively low fields inside the coils appears to be
possible
? Preliminary validation with the MIT Testbed has been achieved, and more
complex maneuver profiles will be accomplished with future upgrades
