An Architecture For Self-Fabricating Space Systems
An Architecture For Self-Fabricating Space Systems
Robert P. Hoyt,* Jesse
I.
Cushing,
Jeffrey
T.
Slostad,
Greg
Jimmerson,
Todd
Moser,
Greg
Kirkos, Mark
L.
Jaster,
Nestor
R.
Voronka
Tethers Unlimited, Inc., Bothell WA, 98011, USA
On-orbit fabrication of spacecraft components can enable space programs to escape the
volumetric limitations of launch shrouds and create systems with extremely large apertures
and very long baselines in order to deliver higher resolution, higher bandwidth, and higher
SNR data. This paper will present results of efforts to investigated the value proposition and
technical feasibility of adapting several of the many rapidly-evolving additive manufacturing
and robotics technologies to the purpose of enabling space systems to fabricate and integrate
significant parts of themselves on-orbit. We will first discuss several case studies for the value proposition for on-orbit fabrication of space structures, including one for a starshade designed to enhance the capabilities for optical imaging of exoplanets by the proposed New
World Observer mission, and a second for a long-baseline phased array radar system. We
will then summarize recent work adapting and evolving additive manufacturing techniques
and robotic assembly technologies to enable automated on-orbit fabrication of large, complex, three-dimensional structures such as trusses, antenna reflectors, and shrouds.
Nomenclature
= material mass density
D = beam diameter
E = material modulus
l = beam length
m = the mass per unit length of a beam
I. Introduction
HE SpiderFab effort, funded by NASAs Innovative Advanced Concepts (NIAC) program, has investigated the
value proposition and technical feasibility of radically changing the way we build and deploy spacecraft by enabling space systems to fabricate and integrate key components on-orbit. Currently, satellites are built and tested on
the ground, and then launched aboard rockets. As a result, a large fraction of the engineering cost and launch mass
of space systems is required exclusively to ensure the system survives the launch environment. This is particularly
true for systems with physically large components, such as antennas, booms, and panels, which must be designed to
stow for launch and then deploy reliably on orbit. Furthermore, the performance of space systems are largely determined by the sizes of their apertures, solar panels, and other key components, and the sizes of these structures are
limited by the requirement to stow them within available launch fairings. Current State-Of-the-Art (SOA) deployable technologies, such as unfurlable antennas, coilable booms, and deployable solar panels enable apertures, baselines, and arrays of up to several dozen meters to be stowed within existing launch shrouds. However, the cost of
these components increases quickly with increased size, driven by the complexity of the mechanisms required to
enable them to fold up within the available volume as well as the testing necessary to ensure they deploy reliably on
orbit. As a result, aperture sizes significantly beyond 100 meters are not feasible or affordable with current technologies.
CEO & Chief Scientist, 11711 N. Creek Pkwy S., D113, Bothell WA 98011, and AIAA Member.
Aerospace Engineer 11711 N. Creek Pkwy S., D113, Bothell WA 98011.
V.P. & Chief Engineer, 11711 N. Creek Pkwy S., D113, Bothell WA 98011.
V.P. & Chief Technologist, 11711 N. Creek Pkwy S., D113, Bothell WA 98011.
1
American Institute of Aeronautics and Astronautics
On-orbit construction and 'erectables' technologies can enable deployment of space systems larger than can fit in
a single launch shroud. The International Space Station is the primary example of a large space system constructed
on-orbit by assembling multiple components launched separately. Unfortunately, the cost of multiple launches and
the astronaut labor required for on-orbit construction drive the cost of systems built on the ground and assembled
on-orbit to scale rapidly with size.
A. The SpiderFab Solution
The SpiderFab architecture seeks to escape these size constraints and cost scaling by adapting additive manufacturing techniques and robotic assembly technologies to fabricate and integrate large space systems on-orbit. The
vision that has motivated this effort is that of creating a satellite chrysalis, composed of raw material in a compact
and durable form, software DNA assembly instructions, and the capability to transform itself on-orbit to form a
high-performance operational space system. Fabricating
spacecraft components on-orbit provides order-ofmagnitude improvements in packing efficiency and
launch mass. These improvements will enable NASA,
DoD, and commercial space missions to escape the volumetric limitations of launch shrouds to create systems
with extremely large apertures and very long baselines.
Figure 1 provides a notional illustration of the value
proposition for SpiderFab relative to current state of the
art deployable technologies.
The larger antennas,
booms, solar panels, concentrators, and optics created
with SpiderFab will deliver higher resolution, higher
bandwidth, higher power, and higher sensitivity for a
wide range of missions. Moreover, on-orbit fabrication
changes the cost equation for large space systems, ena- Figure
1.
SpiderFab
Value
Proposition.
On-orbit
fabrication
bling apertures to scale to hundreds or even thousands of of
spacecraft
components
enables
higher
gain,
sensitivity,
meters in size with providing order-of-magnitude im- power,
and
bandwidth
at
lower
life-cycle
cost.
provements in system performance-per-cost.
In this paper we will first describe a concept architecture for a system designed to fabricate and integrate large
spacecraft components on-orbit. We call this architecture "SpiderFab" because it involves a robotic system that
builds up large, sparse structures in a manner similar to that in which a spider spins its web: by extruding highperformance structural elements and assembling them into a larger structure. We will then evaluate the value proposition of this on-orbit fabrication architecture for several classes of spacecraft components, including antennas and
starshades. Next, we will detail concept solutions for the technical capabilities required to realize the proposed architecture, and describe proof-of-concept testing performed to establish technical feasibility of these solutions. Finally, we will describe an incremental development approach to enable maturation of these capabilities to mission
readiness.
2
American Institute of Aeronautics and Astronautics
Figure
2.
Samples
fabricated
using
FFM.
On
Earth,
slumping
due
to
gravity
limits
the
element
dimensions
of
sparse
structures
to
centimeter
scales,
but
this
limit
will
not
be
present
in
microgravity.
3
American Institute of Aeronautics and Astronautics
Slumping due to gravity in the lab limited the free-standing lengths of the elements to roughly a centimeter, but in zero-g the element lengths would be limited
only by the reach of the fabrication tool.
A second technical challenge for on-orbit additive manufacturing is the vacuum and thermal environment of space. Our preliminary testing of FFF processes
in vacuum has indicated that the lack of an atmosphere is likely not an impediment, but the absence of conductive and convective cooling will require careful
design of any process that involves thermal processing of materials so that printed
structures cool and solidify in the desired manner. Furthermore, temperatures and
temperature gradients can vary greatly depending upon the solar angle and sunlit/eclipse conditions, and methods for controlling these temperatures will be necessary to prevent undesired stresses from distorting structures under construction.
Although current 3D printing processes such as FFF can now handle a wide
range of thermoplastics, and EBF3 can work with metals, the structural performance of these materials is still not optimal for large sparse space structures. If
we are to pursue the construction of kilometer-scale systems, we must utilize materials with the highest structural performance available. Additionally, the speed Figure
3.
TUI's
FFF
machine
print-
of current 3D printing processes are not suitable for creating large space systems. ing
a
sparse
truss
structure.
A typical FFF machine requires an entire afternoon to print an object the size of a coffee mug. For these reasons, we
are pursuing an approach that fuses the flexibility of FFF with the performance and speed of another additive manufacturing process: automated fiber layup. Essentially, we are working to develop a capability to rapidly '3D print'
composite structures using high-performance fiber-reinforced polymers. This method will enable a robotic space
system to build up very large, sparse structures in a manner similar to that in which a spider spins a web, extruding
and pultruding structural elements and assembling them in 3-dimensional space to create large apertures and other
spacecraft components. For this reason, we have termed this method the "SpiderFab" process. The incorporation
of pultrusion into the 3D printing process is particularly important, because it enables structural elements to be fabricated with high-modulus, high-tenacity fibers aligned in directions optimal for the service loads the structure must
sustain.
The materials used in this process must be suitable for the space environment. In particular, they must be able to
withstand the temperature extremes, UV light, radiation, and atomic oxygen that may be present in their operational
orbit. Furthermore, low outgassing characteristics are necessary to prevent outgassed volatiles from contaminating
optics, solar panels, and other components. In this work, we have focused on the use of Carbon Fiber reinforced
Polyetheretherketone (PEEK) thermoplastics. These CF/PEEK composites have excellent structural performance,
very high temperature tolerance, and very low outgassing characteristics. Although these materials are challenging
to process due to the high melting temperature of PEEK, in this and other parallel efforts we have made excellent
progress in developing techniques to perform thermoforming, pultrusion, and Fused Filament Fabrication with these
materials. Although our work to date has focused on CF/PEEK composites, we should note that the SpiderFab process is readily adaptable to other composite choices, and we have also performed initial development with fiberglass-PET composite materials.
2. Mobility & Manipulation
In order for a robotic system to fabricate a large structure, it will require means to move itself relative to the
structure under construction, as well as to distribute the raw materials from the launch volume to the build area on
the structure. Additionally, it will require the capability to manipulate structural elements to position and orient
them properly and accurately on the structure. There are multiple potential solutions for both requirements. In developing the SpiderFab architecture, we have focused on the use of highly dexterous robotic arms because, serendipitously, under a separate contract effort we are currently developing a compact, dexterous robotic arm for nanosatellite applications. In our concept implementations, one or more such robotic arms will be used to position fabrication
heads, translate the robot across the component under construction, and position structural elements for assembly.
3. Assembly & Joining
Once the robot has created a structural element and positioned it properly on the spacecraft structure, it will require means to bond the element to the structure. This bonding could be accomplished using welding, mechanical
fasteners, adhesives, and other methods. Because our SpiderFab efforts have focused upon the use of fiberreinforced thermoplastics, we can take advantage of the characteristics of thermoplastics to accomplish fusionbonding using a combination of heat and pressure.
4
American Institute of Aeronautics and Astronautics
4. Thermal Control
A significant challenge for fabricating precise structural elements, managing structural stresses in the elements,
and reliably forming fusion bonds between the elements will be managing the temperature of the materials in the
space environment, where both mean temperatures and temperature gradient vectors can vary dramatically depending upon the direction to the sun and the position in orbit. In the SpiderFab implementations we propose to use additives or coatings in the fiber-reinforced thermoplastics to cold-bias the materials and minimize their thermal fluctuations under different insolation conditions, and use contact, radiative, and/or microwave heating to form and bond
these materials.
5. Metrology
Automated or tele-robotic systems for constructing large components will require capabilities for accurately
measuring the component as it is built. This metrology will be needed at two scales: macro-scale metrology, to
measure the overall shape of the component to ensure it meets system requirements, and micro-scale metrology, to
enable accurate location of material feed heads with respect to the local features of the structure under construction.
Technologies currently in use in terrestrial manufacturing processes, such as structured-light scanning and stereoimaging, can be adapted to provide these functionalities.
6. Integration of Functional Elements
Once the SpiderFab system has created a base structure, it will also require methods and mechanisms to integrate
functional elements such as reflective membranes, antenna panels, solar cells, sensors, wiring, and payload packages
into or onto the support structure. Because most of these components can be packaged very compactly, and require
high precision in manufacture and assembly, in the near term it is likely to be most effective to fabricate these components on the ground and integrate them on-orbit. In the long-term, it may be possible to implement additive manufacturing methods capable of processing many materials so that some of these components could be fabricated insitu, but nonetheless it will only be advantageous to do so if on-orbit fabrication provides a significant improvement
in launch mass or performance. The techniques for automated integration of functional elements onto a space structure will depend upon the nature of the element. Reflective membranes and solar cells can be delivered to orbit in
compact rolls or folded blankets and unrolled onto a structure using thermal bonding, adhesives, or mechanical fasteners to affix them to the structure. Sensors, payloads, and avionics boxes can be integrated onto the structure using mechanical fasteners. Wiring can be unspooled and clipped or bonded to the structure, and attached to payload
elements using quick-connect plugs.
C. Concept Implementations
1. SpiderFab Truss-Fabricator for Large Solar Array Deployment
Figure 4 illustrates a concept for on-orbit fabrication of support structures for large solar arrays. In this concept,
three SpiderFab "Trusselator" heads will extrude continuous 1st order trusses to serve as the longerons, and a fourth
fabrication head on a 6DOF robotic arm will fabricate and attach cross-members and tension lines to create a truss
support structure with 2nd-order hierarchy. As it extends, the support structure will tension and deploy a foldable/rollable solar array blanket prepared on the ground. To create the structural elements forming the truss-oftrusses, this system will process a Continuous Fiber Reinforced Thermoplastic (CFRTP) yarn consisting of highmodulus fibers co-mingled with thermoplastic filaments. This yarn can be wound in a highly compact spool for
launch and then processed to create stiff composite structures. Figure 5 shows a proof-of-concept demonstration of
a Trusselator mechanism creating long truss structures. The spool shown on the left of Figure 5 holds enough yarn
Truss%Structure%for%Golay03%
Sparse%Aperture%
5
American Institute of Aeronautics and Astronautics
6
American Institute of Aeronautics and Astronautics
III.
The Phase I effort evaluated the value proposition for on-orbit fabrication of space systems using the SpiderFab
architecture, by first considering the trade-offs between building components on the ground versus building them on
orbit, and identifying two key advantages that on-orbit fabrication can provide. We then reviewed NASA's Technology Roadmaps to identify Technology Areas and future NASA missions where SpiderFab could provide significant advantages. Then we developed performance metrics to quantify the potential advantages that SpiderFab could
provide for several space system components, including high-power solar arrays, phased array radars, optical occulters, and antenna reflectors. In each case, we found that SpiderFab can enable order-of-magnitude improvements
in key performance metrics; in this proposal we will present the value proposition analyses for optical occulters and
antenna reflectors, and refer the reviewer to our Phase I final report for details on the other case studies.
A. Build-on-Ground vs. Build-on-Orbit
On-orbit fabrication of a space system can free the system design from the volumetric constraints of launch vehicles and reduce the mass and engineering costs associated with designing the system to survive launch. Additionally, an on-orbit fabrication capability enables repair and reconfiguration after launch, reducing risks due to design
errors and increasing mission flexibility. However, these advantages must be traded against the additional cost and
complexity of enabling these components to be fabricated and integrated in an automated manner in the space environment. Furthermore, whereas in the conventional approach components are fabricated, integrated, and tested prior
to launch, a program using on-orbit fabrication must commit and expend the costs associated with launch before
these parts are created and integrated. Consequently, although our far-term goal is to enable fabrication and integration of essentially all of a spacecraft on-orbit, we must approach this goal incrementally, focusing initial investment
on classes of components where our current technology capabilities can provide a significant net benefit. Satellites
and other spacecraft are typically composed of a number of subcomponents, ranging from bulk structures to actuated
mechanisms to complex microelectronics. All of these components could, in theory, be fabricated on-orbit, but investing in developing the capability to do so can only be justified if on-orbit fabrication can provide a dramatic net
improvement in performance-per-cost. On-orbit fabrication can provide benefits primarily in two ways: launch
mass reductions, and packing efficiency improvements.
B. Mass Optimization
Fabricating a space structure on-orbit can reduce system mass because the design of structural components can
be optimized for the microgravity loads they must sustain in the space environment, not for the 100's of gravities
shock and vibrations they would experience during launch. Additionally, large structures built on-orbit do not require the hinges, latches, and other complex mechanisms needed by deployable structures, reducing the 'parasitic'
mass of the structure and enabling it to be fully optimized for its design loads. Building a structure on-orbit, rather
than designing it for deployment, also enables its geometry to be varied and/or tapered in an optimal manner
throughout the structure, which for very large structures supporting well-defined loads can result in significant mass
savings. Furthermore, it enables creation of structures with cross-sections that would be too large to fit in a launch
shroud, taking advantage of geometric optimizations that can provide large improvements in structural performance.
For example, the bending stiffness of a longeron truss increases as the square of its diameter D:
!"
! ! !
=
D ,
(1)
!
! !!
where is the material mass density, m is the mass per unit length of the beam, E is the material modulus, and is
a constant accounting for battens, cross members, and joints.1 Whereas a deployable truss designed to stow within a
launch shroud will typically have a maximum diameter on the order of a meter, trusses fabricated on orbit can readily be built with diameters of several meters or more, providing an order of magnitude improvement in stiffness per
mass. Moreover, large structures can be built with 2nd or higher-order hierarchical geometry, enabling an additional
30-fold increases in structural performance.2
C. Packing Efficiency Improvements
The second manner in which on-orbit fabrication can enable significant improvements is the packing efficiency
of large components. Figure 10, adapted from Reference [1], compares the packing efficiency of deployable trusses
(flown) and erectable trusses (proposed). Existing deployable technologies fall one to two orders of magnitude short
of ideal packing efficiency (ie - 95% to 99% of their stowed volume is "wasted"). Proposed erectable technologies,
in which individual structural elements such as longerons and struts are launched in tightly packed bundles and then
assembled on-orbit to fabricate large sparse structures, may be able to improve the packing efficiency somewhat,
7
American Institute of Aeronautics and Astronautics
One measure of packaging efficiency, is the amount of volume the packaged beam requires to achieve a specific
bending stiffness, EI. The volume of material in a beam can be determined from the equation shown on figure 4
by using the expression that M = V. The resulting equation is shown on figure 6 for the same composite
reference values as used in figure 4. In this expression, a parameter is introduced to account for the amount of
packaged volume greater than the material volume. For example, = 1 for perfect packaging and the packaged
volume is equal to the material volume. To provide other references, lines for = 10 and 100 are shown. In the
lower right corner the two lower values are erectable structures. As would be expected, they have quite a good
packaging factor. Also, the two squares at the left of the figure are the composite coilable beams (Murphy, 2005).
Technology
Area
Starshade
(occulter)
Large
Deployable
Antennas
Deployable
Boom/Mast
High
Power
Solar
Array
Radiators
Large
Solar
Sail
Solar
Concentrator
Large
Aperture
Tele-
scope
Need
30-100m,
0.1m shape accuracy
10-14m
20 Gbps from 1AU
20-500m
30-300kW
0.5-1
kW/kg
multi-MW
2
Example Mission/Program
Reference
>1000
m
1 g/m2
85-90%
concentrator
LEO
Cargo
Tug;
LEO-GEO
Tug;
efficiency
Extremely
Large
Space
Tele-
2
50m
aperture
scope
(EL-ST),
TPF-C
Counts/sec
Blue Sky
O3 edge
0.3
0.4
0.5
0.
~80,000 km
Figure
11.
New
Worlds
Observer
starshade
concept.
A starshade positioned between a distantexoplanets
star and a telescope attenu4m UV-Visible
TELESCOPE
ates
light
from
the
star
to
allow
the
telescope
to
image
planets
orbiting
that
star.
[
Images
f
rom
Ref
3]
Provides large collecting area and high
resolution for both exoplanet science
starshade
designed by the NWO team, illustrated
and The
generalNWO
astrophysics.
Carriesspacecraft
5 instruments tobooms
enable a to
wide
array ofan
UVopaque metalized Kapton blanket with folded
deployed
unfurl
and Visible astrophysics
in Figure 12, uses several radiallyrigid edge pieces. Using the largest
available Delta-IVH launch shroud, this SOA deployable design could enable a starshade with a diameter of 62 m.
Instrument
Primary Use
FOV
Bandpass
The
mass of Detecting/
the starshade
component of the system (not including the spacecraft bus), was estimated by the NWO
NWO uses a 3-step alignm
ExoCam
Imaging Exoplanets 26 x 26 0.25-1.7 mm
the starshade and telescop
team
kg. of Exoplanets
ExoSpecto be 1495
Spectroscopy
10 x 3
0.25-1.7 mm
Deployable Apodized STARSHADE
Shadow Sensor
WF Camera
UVSpec
N/A
10 20
<1
1.7-3 mm
0.4-0.9 mm
0.12-0.5 mm
L2 Orbit
6 months staggered launch
~800,000 km radius
7.7$m$
~180 day period
NWO Observatory Mass
Telescope Spacecraft
Starshade Spacecraft
CBE
Cont.
Alloc. CBE
Cont.
Alloc.
Spacecraft dry mass
4077
30%
5300
2710
30%
3523
Propellant Mass (bi prop)
448
0%
448
476
0%
476
Propellant Mass (Xenon)
n/a
n/a
n/a
1220
0%
1220
Spacecraft Wet Mass
4525
27%
5748
4406
16%
5219
4.5$m$
62$m$ 114 5%
Payload Adapter Fairing
120
114
5%
120
49 structure
5%
System
Figure
12.
SOA
Deployable
NWO
Starshade
Design.
The
NWO
SSeparation
tarshade
design
folds
t49o
fit
a
5%
62m
d51
iameter
w51
ithin
Total Launch Mass PI: W. Cash
4688
26%
5919
4570
26%
5390
All mass in kg
Bac
Figure
13.
Notional
Comparison
of
Support
Structures
of
the
NWO
Deployable
Starshade
and
a
SpiderFab
Starshade.
On-
orbit
fabrication
enables
creation
of
structures
with
variable
dimensions
and
geometries
optimized
to
the
operational
loads
in
the
microgravity
environment.
Figure 13 presents a notional comparison between the NWO deployable starshade's structural design and the
structures enabled by SpiderFab on-orbit fabrication. The NWO starshade's opaque membrane is deployed and supported by 16 radial spoke telescoping booms made of glass-reinforced polymer composite. The diameter of these
booms is limited by packaging concerns to be less than a meter. Once deployed, these booms must support the
opaque membrane against thrusts and torques applied by the central spacecraft. The lower half of Figure 13 illustrates the kind of structure made possible by SpiderFab. We created this structure using ANSYS tools, using esti9
American Institute of Aeronautics and Astronautics
mates of the torques and thrusts the structure must support and assuming the use of high-performance carbon fiber
composites. Freed from the constraints of launch shroud dimensions and the requirement for a structure to be unfoldable or unfurlable, the support structure for the starshade could be made with a variable cross-section and variable geometry. The structure could be several meters deep in the middle and taper out towards the periphery, and the
concentration and geometry of the structural elements can be varied so as to optimize its strength to the operational
loads. As illustrated in Figure 14, our analyses indicate that with the same amount of mass allocated for the SOA
deployable starshade, a SpiderFab process could create a starshade structure of twice the diameter - four times the
area. In this case the SpiderFab starshade mass estimate included an allocation of 250 kg + 150 kg margin for the
robotic system required to fabricate the support structure (based upon the mass of our KRAKEN robotic arm and
estimates derived from past experience on the Mars Polar Lander misson), and for the opaque membrane, we assumed the same total thickness of Kapton film (125 m) used in the NWO design. In addition to increasing the size
of the starshade that could be deployed with a given launch mass, SpiderFab also enables a 30-fold reduction in
stowed volume, from 120 m3 for the SOA deployable approach down to 4 m3 for the on-orbit fabrication approach.
This volume estimate assumed an 80% packing efficiency for the carbon fiber composite source material for the
support structure (readily achievable with yarns or flat tapes) and included 2 m3 allocated for the SpiderFab robotic
system) This reduction in stowed volume could enable the Starshade component of the NWO mission to launch on a
Falcon-9 rather than a Delta-IVH, reducing its launch cost by a roughly a third.
Figure
14.
Size
increase
achievable
with
SpiderFab.
SpiderFab
enables
dramatic
increases
in
aperture
size
with
equal
launch
mass
and
significantly
smaller
stowed
volume.
Doubling the size of the starshade would enable the NWO telescope to resolve planets 2 times closer to a star.4
This closer inspection would increase the number of potential Earth-like targets within the star's habitable zone by a
factor of 8. Additionally, doubling the occulter size would double the maximum wavelength at which the starshade
would provide sufficient attenuation, from 1 to 2. This larger wavelength window would bring the system into
the range where the James Webb Space Telescope (JWST) can operate, potentially enabling the JWST to be used as
part of the NWO system, or at least as part of a pathfinder demonstration of the NWO architecture. By reducing the
number of launches required to deploy a NWO system from two Delta-IV Heavies to one Falcon-9, and by increasing the number of planets the system could resolve, the SpiderFab approach could enable a net benefit of providing a
16-fold increase in the number of Earth-like planets the NWO mission could discover per life-cycle cost. More succinctly, SpiderFab enables NASA to discover 16X more Earth-like planets per dollar.
F. Value Proposition for Large Antenna Reflectors
Fundamentally the majority of NASA, DoD, and commercial space systems deliver one thing to their end-users:
data. The net quality of this data, whether it is the resolution of imagery, the bandwidth of communications channels, or the signal-to-noise of detection systems, is largely driven by the characteristic size of the apertures used in
the system. Deployable antennas reflectors therefore represent a very important potential market for application of
on-orbit fabrication technologies.
We can compare the potential performance of SpiderFab for large antenna reflectors by comparing it with stateof-the-art deployable antennas such as the Astromesh reflectors produced by Northrop Grumman's Astro Aerospace
10
American Institute of Aeronautics and Astronautics
subsidiary, and the unfurlable antennas produced by Harris Corporation. The Astromesh reflectors use a tensegrity
design in which a hoop-shaped truss deploys to spread open a conductive mesh, and a system of tension lines strung
across the hoop serve to hold the mesh in the desired parabolic configuration. The Harris antennas typically use
several radial spokes that unfold like an umbrella to spread apart and shape a conductive mesh. These tensegritybased SOA deployables are exceptionally efficient in terms of mass, and we believe it is unlikely that an on-orbit
fabrication approach can provide a significant improvement in launch mass. However, these deployables are not
optimum from the perspective of stowed volume and cost, and therefore there is substantial opportunity for an onorbit fabrication architecture such as SpiderFab to provide significant capability improvements by enabling much
larger apertures to be deployed within the constraints of existing shrouds.
600"
Image$Scale$Indicates$
RelaDve$Size$or$Performance$
Est.%Aperture%Cost%($M)%
500"
Astromesh$
Largest$Deployable$
That$Fits$in$Current$
Launch$Shrouds$
100m$
SpiderFab$
Dish$
400"
300"
Harris$
rFab$
Spide
200"
Monocoque$
100"
Data$Source:$$Barnhart,$D.,$
Phoenix$ID$Brieng$2/11/13$
0"
0"
100"
200"
300"
Aperture%Mass%(kg)%
400"
500"
600"
Figure
15.
Mass
and
Cost
Scaling
of
Deployable
Antenna
Reflectors.
On-orbit fabrication of antenna apertures using SpiderFab can change the cost equation for apertures, enabling deployment of very large apertures at lower cost than
conventional deployable technologies.
Figure 15 plots the mass and estimated cost of current SOA deployable antennas.5 The size of the antenna images used in the plot indicate the relative size and/or performance of the antenna. The plot demonstrates that the cost
of these deployables increases rapidly with the size of the aperture reaching costs on the order of several hundred
million dollars for apertures of a few dozen meters. The cost scaling is exponential with size due to the complexity
of the additional folding mechanisms required as well as the facility costs needed to assemble and qualify very large
components. Furthermore, because these deployable antennas are limited in terms of how compactly they can fold
up, the largest aperture that can be deployed with these SOA technologies is on the order of several dozen meters.
SpiderFab changes the cost equation for large antennas. For an antenna fabricated on-orbit, the cost will primarily
be driven by the cost of building, launching, and operating the robotic system needed to construct it. In this analysis, we have estimated the recurring cost of such a robotic system at $25M-$75M, based upon use of an ESPA-class
microsat bus such as the ~$20M Space Test Program Standard Interface Vehicle (STP-SIV) as well as estimates for
the robotic systems based upon the Mars Polar Lander (MPL) robotic arm and the DARPA Phoenix mission. This
'base' cost may make SpiderFab non-competitive for small apertures. However, once that robotic system is paid for,
the incremental cost for creating a larger antenna is primarily the cost for launching the required material and operating the robotic system for a longer duration. In particular, we can eliminate the facility costs for assembling and
testing very large antennas. As a result, the antenna life cycle cost will scale much more gently with aperture size,
making antennas with diameters of hundreds of meters affordable.
IV.
These SpiderFab concepts require capabilities for: (1) Processing Suitable Materials to Create Space Structures, (2)
Mechanisms for Mobility and Manipulation of Tools and Materials, (3) Methods for Assembly and Joining of Structures, (4) Methods for Thermal Control of Materials and Structures, (5) Metrology to enable closed-loop control of
the fabrication process, and (6) Methods for Integrating Functional Elements onto structures built on-orbit.
11
American Institute of Aeronautics and Astronautics
Figure
17.
Handheld
SpiderFab
tool
and
samples
of
composite
lattice
structures
fabricated
with
the
tool.
Pultrusion
of
CFRTP
elements
can
enable
free-form
fabrication
of
large,
sparse
composite
structures
with
excellent
structural
performance.
assembly applications. A developmental model of the 7DOF KRAKEN arm is shown in with a notional SpiderFab
feed head mounted on a 3DOF 'carpal-wrist' gimbal.
C. Assembly & Joining
To enable a robotic system to construct complex sparse lattice structures, we developed a concept design for a specialized Joiner Spinneret end effector that uses Fused Filament Fabrication (FFF) techniques to join tubular truss
elements. This tool, illustrated in Figure 18, is designed to approach the new tubes to be joined from the side (radially), clamp onto the tube, and then use a rotary stage to reach 360 degrees around the end of the tube, while allowing the end effector to approach and retract radially from the side of the tube. As illustrated in Figure 18, a finger
with 3 independently cable-driven joints allows the spinneret print head to reach every spot and every angle needed
to print a uniformly filleted joint, even when it requires reaching between tubes at tightly angled orientations to each
other. The smaller scale motion stages built into the finger allow the new tube to be fixtured by the same robotic
arm that is performing the joining, which simplifies the accuracy and obstacle avoidance schemes required in generating the tool paths. Figure 19 shows a multi-element joint fabricated with optimized geometry using 3D printing,
assembled with carbon composite tubes. The joiner spinneret can also be used to add brackets, bolt-holes, and other
features to enable mounting of payloads and functional elements, as illustrated notionally in Figure 20.
Figure
18.
Conceptual
Tube-Joining
Process
Using
Fused
Filament
Fabrication.
The
Spinneret
uses
a
FFF
head
on
the
joining
tool
to
fashion
a
joint
between
the
element
and
the
existing
structure.
13
American Institute of Aeronautics and Astronautics
D. Thermal Control
Thermoforming and bonding of fiber-reinforced thermoplastics requires control of the temperature of both the material being processed and the structure it is being applied to in order to ensure reliable bonding and minimize stresses
and distortions in the structure. This will be a significant challenge in the space environment, as temperatures and
thermal gradients can vary dramatically depending upon solar angle and eclipse/sunlit conditions. Terrestrial highprecision FDM 3D printing machines typically house the entire workspace and material processing tools within a
thermally-controlled enclosure to minimize warping of parts due to coefficient of thermal expansion (CTE) behavior. This solution will not be practical for building very large space structures. To address this challenge, we propose to pursue a method combining low-CTE material combinations, surface coatings to minimize temperature variations, and local spot-heating to ensure the temperatures necessary for reliable bonding. To ensure a joint is at the
proper temperature to enable reliable fusing of new material to it, we can use spot-heating with IR radiators, lasers,
RF heaters, or conductive-contact heaters. Figure 21 illustrates a concept approach to using an IR laser pre-heating
areas onto which the tool will 3D print material, and Figure 22 shows a photo of an initial test of using a high-power
IR laser to spot-heat a section of a 3D-printed joint. The initial testing indicated that this approach is feasible, but
further work will be required to develop a reliable and controllable process. An additional method that may be feasible would be for the SpiderFab Bot to use positionable shades (such as the gimbaled solar panel shown in Figure
9) and/or reflectors to control insolation conditions within the work volume
E. Metrology
On-orbit construction of large space system components in an automated or telerobotic manner will require capabilities for measuring the component as it is built in order to ensure its final form meets the requirements for it to perform its functions. This metrology will be required on both the global scale to measure overall shape quality, for
instance to ensure a parabolic antenna dish has the required surface quality, and on the local scale, to enable the fabrication tool to position itself and new components relative to the structure under build. A number of technologies
currently in use in the manufacturing and construction industries are applicable to this challenge, including structured light mapping, LIDAR, and imaging photogrammetry. Each has relative advantages and disadvantages. In
order to establish the basic feasibility of the required metrology capabilities, we worked with a vendor of a structured light scanner technology, GOM Systems, and performed a test in which we used a GOM scanner to measure
the as-built shape of at truss fabricated in the lab with the an early version of our Trusselator mechanism. We then
used this as-built data to design and 3D print a notional mounting bracket shaped to mate perfectly with the truss.
This exercise was a relatively simplistic demonstration, but establishes a basic proof-of-concept for metrology-based
control of the SpiderFab fabrication process.
14
American Institute of Aeronautics and Astronautics
Current%SOA%
Technology*
Element*
Trusselator%
Flight%Demo%
RF%Aperture%
Demo%
SpiderFab%
Mission%Demo%
SpiderFab%Handheld%Tool%
Large%Composite%
Structures%
Trusselator%
SBIR%Eort%
Solid%Surfaces%&%
Components%
(eg.%mounGng%
brackets)%
Desktop%3D%Printer%
Membranes%&%
Photovoltaics%
Thin%Film%
Antennas%
Flexible%PVs%
(a)
RoboGc%Assembly%
%
(c)
(b)
(d)
Future Work
Several improvements are necessary to automate the
steps in this proposed design process by converting
the output of more traditional electronics PCB CAD.
One of these improvements is the ability to project a
circuit design onto a multi-curved surface. The
capability does not yet exist in the currently
implemented CAD software that does not distort the
soon-to-be three-dimensional shape of our circuit.
Inclusion of this feature will greatly reduce the
amount of time spent between circuit design and
three-dimensional circuit conversion. This work
FREND%
Arm%
NIAC%Ph2%
SpiderFab%
SelfZFabricaGng,%%
SelfZAssembling%
Satellite%
DARPA%PHOENIX%
TUI/NRL%
KRAKEN%Arm%
Integrated%
Circuitry%
(a)
AM%+%DP%ConducGve%Inks%
(b)
(c)
Surface%
CoaGngs%
%
Trusses%
Sparse%
Apertures%
Solar%Arrays%
Plasma%Spray%CoaGngs%
NIAC/SBIR*
Antennas%
Radiators%
Shrouds%
Reectors%
Solar%Sails%
Trusses%
Sparse%
Apertures%
Solar%Arrays%
GCD/SST*
OpGcs%
Antennas%
Radiators%
Shrouds%
Reectors%
Solar%Sails%
Trusses%
Sparse%
Apertures%
Solar%Arrays%
TDM*
10%
Mission*Programs*
Figure
23.
SpiderFab
Capability
Maturation
Plan.
Implementation
of
the
SpiderFab
systems
is
amenable
to
an
incremental
development
program,
with
affordable
CubeSat
and
hosted
demonstrations
building
capabilities
towards
demonstrating
con-
struction
of
large
apertures
and
eventually
a
fully
self-fabricating
space
system.
15
American Institute of Aeronautics and Astronautics
VI.
Conclusion
The SpiderFab effort has investigated the value proposition and technical feasibility of radically changing the way
we build and deploy spacecraft by enabling space systems to fabricate and integrate key components on-orbit. We
began by developing an architecture for a SpiderFab system, identifying the key capabilities required to fabricate
large spacecraft components on-orbit, and developed two concept implementations of this architecture, one specialized for fabricating support trusses for large solar arrays, and the second a more flexible robotic system capable of
fabricating many different spacecraft components, such as antenna reflectors and optical occulters. We then performed several analyses to evaluate the value proposition for on-orbit fabrication of spacecraft components, and in
each case we found that the dramatic improvements in structural performance and packing efficiency enabled by onorbit fabrication can provide order-of-magnitude improvements in key system metrics. To establish the technical
feasibility, we identified methods for combining several additive manufacturing technologies with robotic assembly
technologies, metrology sensors, and thermal control techniques to provide the capabilities required to implement a
SpiderFab system. We performed lab-based, proof-of-concept level testing of these approaches, in each case
demonstrating that the proposed solutions are feasible, and establishing the SpiderFab architecture at TRL-3. Further maturation of SpiderFab to mission-readiness is well-suited to an incremental development program. A pair of
initial low-cost flight demonstrations can validate key capabilities and establish mission-readiness for modest applications, such as long-baseline interferometry. These affordable small demonstrations will prepare the technology
for full-scale demonstration in construction of more ambitious systems, such as an Arecibo-scale antenna reflector.
This demonstration mission will unlock the full game-changing potential of the SpiderFab architecture by flight
qualifying and validating an on-orbit fabrication and integration process that can be re-used many times to reduce
the life-cycle cost and increase power, bandwidth, resolution, and sensitivity for a wide range of NASA Science and
Exploration missions.
VII. Acknowledgments
This work was supported by NASA Innovative Advanced Concepts (NIAC) Grant NNX12AR13G.
16
American Institute of Aeronautics and Astronautics
References
1.
Mikulas,
M.M.,
et
al.,
"Truss
Performance
and
Packaging
Metrics,"
NASA
Technical
Document
20060008916.
2.
Murphey,
T.W.,
Hinkle,
J.D.,
"Some
Performance
Trends
in
Hierarchical
Truss
Structures,"
AIAA-2--3-
1903.
3.
Cash,
W.,
et
al.,
"Astrophysics
Strategic
Mission
Concept
Study:
The
New
Worlds
Observer,"
24
April
2009.
4.
Cash,
W.,
personal
commun.,
4Feb2013.
5.
Data
extracted
from
DARPA/TTO
Phoenix
Program
Industry
Day
Briefing,
11Feb13.
17
American Institute of Aeronautics and Astronautics