Unformatted text preview: Electromagnetic Formation Flight
Progress Report: July 2002 Submitted to: Lt. Col. John Comtois
Technical Scientific Officer
National Reconnaissance Office Contract Number: NRO00002C0387CLIN0001 MIT WBS Element: 6893087
Submitted by: Prof. David W. Miller
Space Systems Laboratory
Massachusetts Institute of Technology DESCRIPTION OF THE EFFORT
The Massachusetts Institute of Technology Space Systems Lab (SSL) and the Lockheed
Martin Advanced Technology Center (ATC) are collaborating to explore the potential for an
ElectroMagnetic Formation Flight (EMFF) system applicable to Earthorbiting satellites flying
in close formation. PROGRESS OVERVIEW
At MIT, work on electromagnetic formation flight (EMFF) has been pursued on two
fronts: the MIT conceive, design, implement and operate (CDIO) class, and the MIT Space
Systems Lab research group, as described in the April 2002 progress report.
The CDIO class has completed its first semester performing trades on and preliminary
design of a sixdegreeoffreedom electromagnetic formation flight testbed, called
“ElectroMagnetic Formation Flight of Rotating Clustered Entities,” or “EMFFORCE.”
EMFFORCE will utilize electromagnets to control the size and attitude of a cluster of bodies.
The MIT Space Systems Lab research staff is supporting the CDIO class with guidance in trades
analysis and design.
Recent work has focused on trade analyses for the design of the electromagnets that will
be used in the EMFFORCE testbed. These electromagnets will act as actuators, controlling the
relative degrees of freedom of the bodies that compose EMFFORCE. The following report
summarizes recent progress on the trades performed in the design of the EMFFORCE
actuation system. Preliminary Trade Analysis for the Electromagnet
Design for the EMFFORCE Testbed
1 Actuation System Overview
The actuation system includes the electromagnet and reaction wheel subsystems. The
electromagnet is designed to provide the electromagnetic force and torque used to induce the
spinup of the vehicles in formation, as well as to reject disturbances in the system during steadystate spin. The reaction wheel assembly is designed to provide torque to counteract the torque
induced by the electromagnet. Figure 1.2 shows two dipoles at rest in a perpendicular
orientation at a separation distance 2m. The arrows represent the direction of the induced forces
and torques on the dipoles. Figure 1.3 depicts the spinup of two dipoles from an initially
perpendicular orientation. As current is applied to the dipoles, they begin to move in the
direction of the induced forces and torques. The reaction wheel is then used to apply torque to
counteract the induced torque. This applied torque is then decreased until the dipoles are in
steadystate rotation, aligned along the same axis. S
S N N
2m Figure 1.2: Perpendicular Dipoles
Figure 1.3: Spinup 2 Purpose of Electromagnet
The amount of force induced by the electromagnet, made up of a solenoid coil and core
material, is dependent upon the strength of its magnetic field. The solenoid coil produces a
magnetic field when current is applied to it. When the solenoid is wrapped around a
ferromagnetic material, this core has the effect of multiplying the strength of the magnetic field. 3 Design Trade Analysis
For the electromagnet, trade analyses were performed to select the core material, core
configuration, and mass of the core, coil, and system for the EMFFORCE testbed. 3.1 Core Material
One of the main focuses of the EMFFORCE project is to optimize the amount of force
generated by the electromagnet while minimizing its mass. This force depends upon the strength
of the induced magnetic field. Therefore, the core material will be selected based on magnetic
properties that provide the maximum induced magnetic field at a minimum mass.
When current is applied to a solenoid wrapped around a ferromagnetic core, a magnetic
field, or Bfield, is induced in the core. As the current is increased, the strength of the Bfield
increases until it reaches a strength that cannot be exceeded with increasing current. This is
known as the saturation value, or Bsaturation, for a given material. Table 3.1 compares the Bsaturation
values and densities of several ferromagnetic materials. Table 3.1: Properties of Magnetic Materials
Material
Composition
Iron
99.91% Fe
Steel
98.5% Fe
45 Permalloy 54.7% Fe, 45.0% Ni
78 Permalloy 21.2% Fe, 78.5% Ni
Supermalloy
15.7% Fe, 79.0% Ni, 5.0% Mo Bsaturation [Tesla]
2.15
2.10
1.60
1.07
0.80 Density [g/cm3]
7.88
7.88
8.17
8.60
8.77 Iron and steel have the highest values for Bsaturation of 2.15 and 2.10 Tesla, respectively.
Both iron and steel have the same density, 7.88 g/cm3, which is the lowest among the magnetic
materials considered. High Bsaturation values combined with lower densities are desirable
properties to increase the strength of the induced magnetic field while keeping the mass as low as
possible. EM Core Material
Induced Field vs. Applied Field 2.5 B [Tesla] 2
1.5
1
0.5
0
0 5000 10000 15000 20000 25000 30000 35000 H [Am ps/m ]
AISI 1010 steel Remko soft pure iron Figure 3.1: Iron and Steel  Induced Field vs. Applied Field
Figure 3.1 is a plot of induced versus applied field for iron and steel. For higher values of
the applied field, both iron and steel reach a Bsaturation point at approximately 2.1 Tesla. However,
for lower values of the applied field, this plot shows that iron has higher induced field values
than steel. This is important because it indicates that iron will induce a larger magnetic field, and
therefore higher levels of force, than steel at low current levels.
Another magnetic property that must be considered is the hysteresis effect. When a
magnetic material is magnetized by an applied field, it will not completely demagnetize when the
applied field is reduced to zero. In order to bring the induced field back to zero, a field must be
applied in the opposite direction. The strength of this applied field is called the coercive force.
The coercive force for iron and steel are 79.6 and 143.2 Amps/m, respectively. Therefore,
hysteresis is less of an issue with iron since it requires less of an applied force in the opposite
direction to drive the induced field back to zero.
Recently, another possibility for the electromagnet core was discovered. Hiperco 50A, a
magnetic alloy made up of 48.9% Fe and 49.0% Co, exhibits better magnetic properties than
iron, which was previously the best choice for core material. Table 3.2 shows that Hiperco 50A
has a higher value of maximum induced field and requires the same amount of applied field to
counter the hysteresis effect. Table 3.2: Hiperco 50A vs. Iron
Coercive Force [Amps/m]
Material
Bsaturation [Tesla]
2.40
79.6
Hiperco 50A
2.15
79.6
Iron Typical D.C Magnetization Curves  Hiperco Alloy 50A vs. Electrical Iron Hyperco 50A 24 Iron Induction, B, Kilogausses 20 16
12 8 4 0 0.1 1.0 2 4 6 8 10 100 1000 Magnetizing Force (H, Oersteds)
Hiperco alloy 50A strip, .035" (.89 mm) thick, 1600o F
(871o C), 2hr., dry H2 .
Hiperco alloy 50A bar, 1875o F (1010o C), water quenched
plus 1600o F (871o C), 2 hr., dry H2 .
Hiperco alloy 50A bar, 1600o F (871o C), 2 hr., dry H2 .
Hiperco alloy 50A bar, 1533o F (820o C), 2 hr., dry H2 .
Electrical Iron bar, 1550o F (843o C), 4 hr., wet H2 , FC. Figure 3.2 Hiperco 50A and Iron – Induced Field vs. Applied Field (Adapted from
www.cartech.com.) Figure 3.2 shows the relationship between induced field [Tesla] and applied field
[Oersteds] for Hiperco 50A and iron. Hiperco 50A has significantly higher values for induced
field for a given applied current than iron.
At this time, more analysis is required to determine whether Hiperco 50A is a better choice
than iron for the electromagnet core material. Initial analysis indicates that while Hiperco 50A
exhibits better magnetic properties than iron, it will be extremely costly to the project and may
exceed the electromagnet budget allocation. There are very few distributors of Hiperco 50A, and
it will be considerably more expensive to purchase and machine this material than iron. 3.2 Core Configuration
Several preliminary options for the electromagnet configuration on the EMFFORCE
testbed have been considered. These include a basic dipole, a Ypole (3 legs), an Lpole (2 legs
connected at their ends), and an Xpole (4 perpendicular legs). All are shown in Figures 3.33.6. Figure 3.3: Dipole Figure 3.6: YPole
Before discussing the details of each configuration, it is important to consider the
requirements of the system. The reasoning will come in the following section, but for now we
assume that our system is driven by its mass. Thus, for a given configuration, if two magnets are
at distance between zero and two meters, the mass should be minimized for a given force.
Also, the selected configuration needs to produce the least amount of torque for a given
force, since this excess torque must be transferred to the reaction wheel. The reaction wheel’s
mass will be proportional to the amount of torque that needs to be stored in it, so any
unnecessary torque on the system should be minimized.
Finally, there is a requirement that the design of the electromagnet should minimize the
effort in transitioning from a twodimensional testbed to one that operates in three dimensions.
This means that the direction of the poles in the magnet should be able to change for the selected
configuration.
In order to determine which configuration will satisfy these goals, preliminary calculations
were run and a modeling software called “MagNet” was used. From these calculations, the
following table was constructed. For each configuration, the performance in several categories is
rated between 1 and 5, with 5 being the best performance. The “total” performance of each
configuration is summarized in the rightmost column, and it appears the dipole is quite favorable. Table 3.3: Trades Evaluation for the EMFFORCE Electromagnet Configuration Trades Transition to 3D Minimum mass Minimum torque Totals
Config.
for a given force for a given force
YPole
4
3
2
LPole
4
2
3
Dipole
2
5
4
XPole
4
1
3 9
9
11
8 In addition to the most important trades summarized in the table, there are some additional
arguments against using the Lpole and the Xpole configurations.
The Lpole has a center of mass outside of its own body and would produce adverse torque
that would have to be countered by the reaction wheel. Thus using an Lpole would increase the
complexity in controlling a body’s rotation, since the forces on the system would not be acting
through the system’s center of mass.
The Xpole was eliminated due to the fact that the mass would have to be distributed to 4
legs. Energizing one of these legs would only be utilizing ¼ of the mass that would be used in
the single dipole situation and would not produce a large enough force, relatively.
As mentioned before, these are preliminary calculations, and they must be supported with
more modeling and testing. However, the favored configurations at this point are the dipole and
the Ypole. In order to determine whether to use a dipole or a Ypole, it is necessary to compare
them closely, with the help of the modeling software.
Two simulations were carried out on the modeling software. In both simulations each
configuration has the same volume and mass. The first simulation was with two Ypoles
separated by a distance of 2 meters; one leg on each Ypole was energized with a current of 6
amps through 1200 turns of wire. The energized legs were perpendicular to each other. See
Figure 3.7. 2 Figure 3.7: YPole simulation
from MagNet For the dipole simulation, two dipoles were used, separated by a distance of 2 meters and
perpendicular to each other. Both were energized with a current of 6 amps through 1200 turns of
wire. See Figure 3.8.
From the results, the following conclusions can be made:
It was originally believed that the dipole would generate a greater force for a given mass,
because for a set core mass, it utilizes the entire core. Also, the Ypole would not energize the
entire core mass at one time and thus it would “waste” some mass. Since the force is what drives
the satellites to spin up, this was a key factor.
After running the simulation, it turns out that the Ypole magnetizes the remaining ends
that are not energized and the amount of force generated is greater. This can be explained
through Figure 3.7, which shows that the legs on the first Ypole (that are not energized) are
closer to the second Ypole legs (that are not energized) than are the two dipoles’ main cores. 2m Figure 3.8: Dipole simulation
from MagNet For the requirement of easy transition to three dimensions, it was determined that the dipole
cannot change the direction of its poles to an arbitrary direction. The Ypole, on the other hand,
is more versatile and more transitions more easily to threedimensional applications. This is due
to the fact that the Ypole magnetic field direction can vary by changing the amount of current
energizing each of its legs. This would decrease the time it would take to rotate an
electromagnet to the desired alignment using a reaction wheel. The Ypole configuration,
however, also generates a greater adverse torque than the dipole, which would have to be
counteracted by the reaction wheel, making it more massive.
In summary, the Ypole simulation shows a greater force than the dipole configuration, yet
produces more unnecessary torque. The Ypole performs better in transition to three dimensions
and has less mass for a given force than the dipole configuration. However, the torque produced by the Ypole is greater than that produced by the dipole; in other words, the dipole performs
better in terms of minimizing the amount of torque for a given core mass.
Although it seems that the Ypole would be the best configuration, this cannot be
determined yet. For example, the torque on the electromagnets could be more important than the
force on the electromagnets, in which case the performance of the Ypole and the dipole could
change. Since this is unknown, more tests will be run to determine the relations between these
factors.
Another possible configuration is a dipole mounted on a rotating disk, so that the direction
of the poles could be changed as easily as with the Ypole. While the dynamics of this system
could be more complicated, it will be considered in future trade analyses. 3.3 Core and System Mass
The driving parameter for this design problem is mass. Increases in the total system mass
quickly ripple through the system, causing increases in the size of the electromagnet, power
system, and structure. Therefore the design of the electromagnet was conducted with the
intention of working at the minimum system mass for a given system performance (i.e.
separation distance and spin rate).
In order to determine the magnet size, a mathematical model of the system was developed.
The total mass of the system is given as:
mtot = mcore + mcoil + m power + mo Equation 3.1: Total Mass Equation
where the terms are the total mass of the system, the mass of the magnetic core, the mass of the
coil of the magnet, the mass of the power system, and the initial system mass, respectively.
Consider the steadystate spin mode, shown in Figure 3.9, where the electromagnets must
provide a force that is equal in magnitude but opposite in direction to the centripetal force of the
system. The magnetic force on the leftmost vehicle resulting from the other two is given by: Figure 3.9: Steady State Mode 3µ o µ A µ B 3µ o µ A µ C
+
s4
2π ( s ) 4
2π ( )
2
Equation 3.2: Magnetic Force for Three Vehicles
Fmag = where s is the separation distance between the two outermost vehicles; µa, µb, and µc are the
magnetic moments the three electromagnets, respectively; and µo is the permeability of free
space. The magnetic force must be set equal to the centripetal force defined by: s
Fcent = Ω 2 ( )mtot
2
Equation 3.3: Centripetal Force
where Ω is the spin rate of the system and mtot is the total mass of a single vehicle.
The magnetic moment of an electromagnet can be approximated as the product of the
induced magnetic field (B), the core mass, and the inverse of the core density. Equating Equation
3.2 and 3.3 with this substitution for the magnetic moment and solving for the total mass of the
system yields: mtot = 51 Bmcore 2
(
)
µ o s 5 ρ coreΩ Equation 3.4: Total mass in terms of mcore
Finding the mass of the other components in Equation 3.1 in terms of the mass of the core
yields: mcoil ρ coil (πrcoil ) 2 4mcoreα 2 2
=
(
)3 H
iα
ρ coreπ
Equation 3.5: Mass of the coil where the mass of the coil is expressed in terms of the coil radius, the coil density, the
current (i), the core aspect ratio (α, defined as the length divided by the diameter), and the
applied magnetic field (H). The expression for the battery mass is defined based on the available
battery parameters. These equations define the necessary parameters to model the system. The
only effect that is neglected in the model is the effect of system mass on the mass of the required
structure. This is incorporated into mo, which has been assumed to be a constant for the current
analysis. This model has been incorporated in a Mathematica script, which can be run to find the
optimum operating masses for the system. Before the results of this design are finalized there are
a few more factors that must be incorporated. First, it is important to obtain detailed material
properties for the core and coil, including BH curve data for the core material. Also, the model
does not yet include an accurate trade between the current and the number of wire wraps. This
study will essentially trade between coil mass and battery mass for a given amount of applied
magnetic field. Incorporating this trade into the script is anticipated to be a simple step.
Preliminary calculations yield a core mass of ~6.5 kg and a total system mass of ~15 kg.
The goal is to reduce the mass as the system design optimization continues. 4 Conclusion
The electromagnet will be composed of a coil and a core. The favored core materials are
currently Iron and Hiperco 50A, and the favored configurations are the fixed dipole, Ypole, and
steerable dipole. Further analyses and simulations will be performed in order to downselect the
electromagnet materials and configuration, and the optimal number of coil wraps will be
determined at that time. All design decisions will be based on the assumption that the success of
the system depends largely on reducing mass while still meeting the necessary requirements. ...
View
Full
Document
This note was uploaded on 11/08/2011 for the course AERO 16.810 taught by Professor Olivierdeweck during the Winter '07 term at MIT.
 Winter '07
 OlivierdeWeck

Click to edit the document details