NASA Challenge: Ultralight Starshade Structural Design

Large-Diameter Application Starshade (LDAS)

Joe Taylor
 INTRODUCTION AND BACKGROUND - 1/24 axisymmetric model for the inflatable
torus and composite tubes
The Large-Diameter Starshade (LDAS) is a design based on the - Shell elements were utilized in place of solid models.
existing JPL HabEx concept. There are two areas that were This is an appropriate simplification since all structures
identified for improvement from the heritage design: the have thin walls that will not have large changes in stress
starshade petals and the inner perimeter truss structure that across the thickness of the wall.
attaches to the petals. - The petals are modeled as a remote mass attached to the
structure.
LDAS Truss Structure Design Changes
- Spring elements are used to model the steel cables
The HabEx baseline design uses a truss structure that is inspired attaching the structure to the cylinder hub.
by the Astromesh antenna reflector. This design is selected for its
low mass, high stiffness, and relatively small stowed volume.
However, when scaled to 100m, this design falls short of the
design challenge requirements for mass and stowed volume.

Figure 2. Simplified FEM Models used for Initial Trade Study

Using the FEM models, the designs were evaluated for weight
and stiffness in relation to the specified design loads. Each design
was evaluated using a range of wall thicknesses and strut
diameters to more accurately evaluate the design space.

The results of the trade study are plotted below:

Figure 1. Simplified Baseline JPL Habex Model (Ref Only)

Two alternative designs were considered for replacements over

the Astromesh design for LDAS:

- An inflatable torus inspired by NASA’s Inflatable

Antenna Experiment (IAE). Inflatable space structures
can achieve very high packaging efficiency and are well
suited for large, light-weight space structures.
Rigidizable thermoplastics can be used to allow for
flexibility during packaging while maintaining structural
integrity once deployed. This would mitigate the risk of
leaks that could occur due to potential micrometeorite
impacts.
- An extreme modification of the existing truss structure. Figure 3. Weight vs. Stiffness Sensitivity Study
By removing truss members from the structure, it is
possible to reduce mass and stowed volume, at the The ideal structure would minimize deformation under the design
expense of reduced stiffness of the deployed structure. loads while also minimizing total weight of the structure. Based
on this criteria, the composite tube structure demonstrates
A trade study was performed to compare the performance of potential as an improvement over the existing Astromesh design.
these two design concepts. Three models were created: a baseline The composite tube structure is able to achieve this by using a
model representing the Astromesh structure used in the HabEx larger cross-sectional diameter that increases the moment of
design, a model that uses an inflatable torus, and a model that inertia of the individual truss members. This diameter will have
uses a composite tube structure that is connected by hinges. Due to be traded against dimensional constraints for stowed volume.
to limited computational resources, a number of simplifications Mass savings is achieved by the use of fewer truss members than
were made to the models for this trade study: the Astromesh design. Although the inflatable structure shows
potential as a light-weight alternative to the Astromesh structure
- Symmetry boundary conditions were utilized
as well, it provides less stiffness for a given mass allocation
- 1/12 axisymmetric model for the baseline
compared to the other two designs. Based on the results of the
Astromesh structure
trade study, the composite tube structure was selected for LDAS.
 By adding curvature to the petal in the out-of-plane direction,
additional stiffness can be achieved without the need for separate
hinged rib structures.

LDAS DESIGN SUMMARY

LDAS replaces the Astromesh truss structure with a hinged

composite tube structure, attached to the center spacecraft
cylinder by stainless steel cables. The petals use curvature in the
out-of-plane direction to minimize the need for additional support
structures and reduce petal mass.

Figure 4. Composite Tube Inner Disk Structure

LDAS Petal Design Changes

The HabEx petal design consists of a lightweight planar carbon

fiber composite that is supported by carbon fiber rods and a
hinged truss structure that provides out-of-plane stiffness.

Given the looser edge shape tolerances and the need for a Figure 6. Large Diameter Application Starshade (LDAS)
light-weight design for the 100 m diameter Starshade, several
design changes are suggested to the petal design. In place of the LDAS Key Dimensions
hinged rib that supports the HabEx petal design, the new petal
design uses out-of-plane curvature in the composite structure to LDAS has been scaled to meet all dimensional requirements in
provide additional stiffness to the petal. By removing the hinged the deployed configuration. The petal-to-petal diameter is 100 m
rib, additional mass savings can be achieved in each starshade and the central disk has a radius of 50 m. The central cylinder has
petal, while maintaining stiffness in both the in-plane and a 2 m diameter and is reserved for the spacecraft bus and
out-of-plane directions. The petal will be sufficiently thin to allow propellant tanks.
the petal to be wrapped during deployment, similar to how a tape
measure is stored in the retracted position while maintaining
stiffness when deployed.

Similarly to the HabEx design, LDAS will also incorporate an

optical edge formed by a thin metal attachment that is chemically
etched to produce a sharp beveled edge.

Figure 5. Petal Design with Out-of-Plane Curvature Figure 7. Starshade Key Dimensions
LDAS Mass Estimate

The system mass is estimated in Table 1 below. The central

cylinder is not included in the mass budget.

Table 1. LDAS Mass budget

Weight (kg) Source

Composite Tube 121.68 CAD

Structure

Petal Assemblies 488.40 FEM calculated from

shell model

Tensioning Cables 66.24 CAD

Other support 300.00* Estimated Figure 9. Starshade Stowed Configuration

hardware*
LDAS is deployed in a similar sequence to the HabEx design, by
Total 976.32 unfurling the petals and then rotating the struts and petals radially
using torsion springs in the truss hinges, until the structure is fully
*Note: the weight of the hinges, springs, and other support deployed. The composite tube structure and petal deployment is
hardware is estimated. This hardware has not been optimized for controlled via a steel cable that is reeled in using a motor located
weight, but can be updated as the design matures. on the center cylinder.

LDAS Cast Shadow

The support structure and petal geometry meet the volumetric

restrictions and will only expose the sharp edges to the sun
throughout the angles specified. The requirement was verified
using the CAD model.

Figure 10. Spring-Actuated Hinge Joints

To avoid creating a large moment about the hinge during

deployment, the steel cables are tensioned after the structure is
fully deployed. After the structure is fully deployed, the cables
tension the structure symmetrically and cancel out the moment on
the torsion spring (see figure below). A locking feature is used to
Figure 8. Petal Geometry Stayout Zones further ensure the structure does not collapse after initial
deployment.
Stowed Configuration

The petals are stowed by wrapping around the stowed inner disk
structure. The volume of the Starshade in the stowed
configuration is 4.5 m x 4.5 m x 10 m, in order to meet the
cylindrical package diameter required by figure 12-10 of the
Falcon Payload User's Guide.
 STARSHADE STRESS ANALYSIS

Finite element analysis (FEA) was used to demonstrate

compliance to the specified loads. LDAS consists of 24 petals that
are equally spaced around the inner disk structure. This allows for
an axisymmetric finite element analysis to be performed, provided
the applied loads are also axisymmetric. For this design challenge,
3 load cases were considered.

Table 2. Limit Load Factors

X-dir Y-dir Z-dir

Load Case 1 ± 0.03 g 0 0

Figure 11. Tension Cable Angles During Deployment Sequence
Load Case 2 0 ± 0.03 g 0

Due to the way the design deploys, adjacent petals will be Load Case 3 0 0 ± 0.03 g
unfolded in opposite directions. This helps mitigate the risk of
edge contact during deployment. The deployment sequence is
pictured below. The LDAS coordinate system is pictured in the below figure.

Figure 12. LDAS Deployment Sequence Figure 13. LDAS FEM Coordinate System

Between the three load cases, only Load Case 3 is truly

axisymmetric. The axisymmetric model can still be used for the
two in-plane load cases, to approximate the deformation for the
worst-case petal (i.e. the petal that sees 100% of the load in the
radial and tangential directions).

The inner disk optical shield does not provide any additional
stiffness and was not included in the model.
Material Assignments 𝐺π𝑑
4
Torsional Spring Rate (N-m/rad) : κ = 32 𝐿
The material property assignments are specified below. All
material properties are S-basis material properties at ambient All symmetry boundaries were constrained using a “frictionless”
temperatures. symmetry plane. This allows in-plane deformation in the
symmetry plane, but zero deformation out of the symmetry plane.
Table 3. Material Property Assignments The petal assembly is bonded to the composite tube inner disk at
the hinge joints.
Material Yield Ultimate Notes
Strength Strength
The g-load was applied globally to all geometry in the simulation.
(MPa) (MPa)
FEA Results
Petal Carbon Fiber - 345 Carbon Fiber with
(Hexcel phenolic sandwich core The resulting deformations in the 3 directions are pictured in the
M55)
below figures and summarized in Table 3. The max edge
Optical Stainless 172 482 displacement was determined by probing the optical edge for total
Edge Steel 304L displacement.

Central Aluminum 276 310 Aluminum honeycomb

Hub 6061-T4 structures have been
used for this application

Inner Carbon Fiber - 345 Carbon Fiber with

Disk (Hexcel phenolic sandwich core
M55)

Boundary Conditions and Mesh Information

The shell model was meshed with tetrahedral elements. A total of

31008 elements were used in the model. Local mesh refinement Figure 15. Load Case 1 Deformation (X-Direction, In-Plane)
was used on the optical edge to accurately capture critical edge
dimensions.

Figure 16. Load Case 2 Deformation (Y-Direction, In-Plane)

Figure 14. Boundary Conditions and Mesh

A fixed boundary condition was applied to the central hub. The

tensioning cables are modeled as spring elements. Spring rates for
the cables were calculated using the following equations:

𝐸𝐴
Linear Spring Rate (N/m) : 𝑘 = 𝐿 Figure 16. Load Case 3 Deformation (Z-Direction, Out-of-Plane)
 Table 3. Deformation Results Summary Due to time and computing constraints, a transient analysis was
not performed.
Petal Edge Max Requirement
SCALING LAWS
Displacement (mm) (mm)
A smaller scale physical model can be used to demonstrate
X-Dir Loading 15.02 50
feasibility of the LDAS concept. All deformations in the structure
are linear elastic. For a smaller scale model, the starshade
Y-Dir Loading 30.48 50
assembly can be approximated as a series of cantilever beams
under a uniform distributed load (g-loads in X, Y, and Z).
Z-Dir Loading 377.64 1000
Deflections for the composite tube structure and the petal
Von Mises stresses were also calculated for the above load cases assembly can be calculated separately using the following
and are summarized in Table 4 below. All stresses show positive equation for cantilever beams.
margin to the limit load cases.
4
𝑤𝐿
δ𝑚𝑎𝑥 =
Stress concentrations were observed at the nodes where the spring 8𝐸𝐼
elements attached to the inner disk structure. These stresses are a
singularity due to the way the cables are modeled and are not Note that for the composite tube structure, the moment due to the
indicative of actual stresses in the structure (spring elements exert mass of the petal assembly should also be considered and can be
force on one node, i.e. an infinitely small area, and will have calculated using the following equation.
infinitely high stresses as the mesh is refined). Hand calculations
2
that consider a larger area for the cable tension to act on show that δ𝑚𝑎𝑥 =
𝑀𝐿
2𝐸𝐼
stresses in these locations are low.

Margin of Safety is calculated as: Assuming linear elastic deformations, test results from a small
scale physical model can be extrapolated to the full-scale 100m
𝐴𝑙𝑙𝑜𝑤𝑎𝑏𝑙𝑒 𝑆𝑡𝑟𝑒𝑠𝑠 model using the above equations. Extrapolations should also
𝑀. 𝑆. = -1
𝐹.𝑆. 𝑥 𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑆𝑡𝑟𝑒𝑠𝑠 account for any effects due to gravitational effects on earth.

Where factors of safety are 1.25 (yield) and 1.4 (ultimate) for REFERENCES
metallic structures and 1.4 (ultimate) for composite/bonded
structures. “Habitable Exoplanet Observatory (HabEx).” NASA, NASA,
https://www.jpl.nasa.gov/habex/.
Table 4. Stress Results Summary

Max Stress Location M.S.

(MPa) Ultimate

X-Dir 12.5 Petal 18.7

Loading

Y-Dir 38.8 Petal 5.4

Loading

Z-Dir 89.3 Petal 1.8

Loading

Transient Simulation

A transient simulation can be performed to confirm the Starshade

can recover its zero-g shape within 10 seconds or less. Since there
are no dampers or isolators included in the design, the main
sources of damping are Rayleigh damping due to energy
dissipation in the materials and damping due to welds and bolted
connections.