- Decide final dimensions and mass of CubeSat to begin structural design.
- Produce structure concept designs
- Ensure that concepts meet CubeSat specifications
- Aim is to reduce weight while maintaining strength
- Incorporate any structural changes as a result of the mission objective (eg. Deployable balloon storage)
- Create CAD model of structure
- Produce design concepts
- Define CAD model (PTC Pro/E is generally used for this purpose)
- CAD models of Pumpkin structures are available from (http://www.cubesatkit.com/content/design.html) for reference.
- Analyse CAD model
- Export CAD model to FEM software
- ANSYS, ANSYS Workbench, Pro/E Mechanica are all capable of performing the necessary analyses
- Export CAD model to FEM software
- Setup simulation
- Apply launch loads
- Worst case launch loads may be determined by referring to NASA GEVS documentation
- Define boundary conditions
- Apply launch loads
- Select material for construction
- Aluminium 7075 has been used previously by Strathclyde for CubeSat structures
- Manufacture structure
- Material must be anodised along rails
- Apply finishes for flight
- Experimental testing
Regulations are applied to CubeSats to enable them to be stored in a rocket and then ejected using a P-POD. As structures with external parts or different dimensions from those of the P-POD may get trapped or damaged in the P-POD.
All materials should be approved for space to prevent using low quality materials that outgas.
From the general description of a CubeSat the following outer dimensions must be observed when designing a CubeSat structure:
|CubeSat Size||Height (mm)||Width (mm)||Depth (mm)||Mass (kg)|
Structural analysis may be broken down into two categories: acceptance testing and qualification testing. Acceptance testing is conducted on a finished, physical model of the flight-ready structure at 100% of the worst case loading scenario. Qualification testing is usually conducted using an engineering model and is performed at 150% of the worst case loading scenario. This results in a design safety factor of at least 1.5.
The structure may be simulated using any of the FEM packages mentioned above in section 1.1 (ANSYS, ANSYS Workbench, Pro/E Mechanica, etc.)
The CAD model may be considered as one solid piece of a given material during analysis in order to simplify modelling.
Features such as small extrusions or holes may be omitted so as to reduce the computational load of the simulation. This also reduced the possibility of large stress concentrations within the analysis results.
For best results, it is recommended that no proposed structure has a natural frequency under 150Hz. This is a common design requirement of Pumpkin structures and is, therefore, a sensible value to accept as a limitation.
(See below for justification of terms)
The forces acting on the base of the CubeSat structure may be assessed manually using:
Equation 1 F=g×(m_Total-m_Structure )×(NASA Worst Case Factor)
The stress that results from this loading may be calculated using:
Equation 2 E=εσ=(L/δ)(F/A)
Where all symbols have their usual meanings.
Application of loads
The loads that the structure will experience during launch will be larger than any others experienced during its lifetime. It is therefore necessary to simulate the behaviour of the structure during the takeoff of its launch vehicle.
The loads that must be applied to the model during analysis will differ depending on the launch vehicle, and it is often difficult to know which vehicle will eventually be used at such an early stage of the design process.
As such, it is useful to consider a general worst case scenario, and design the structure accordingly. NASA has developed such a scenario for CubeSat structural analysis, known as the GEVS profile.
The loads that may be derived from this worst case and the specifics of a particular mission should be multiplied by 1.5 in order to satisfy the conventional minimum safety factor provision.
The loads may then be applied to the CAD model.
Assistance with “ASD”, or “Acceleration Spectral Density” in the above graph may be found at
http://pdf.cloud.opensystemsmedia.com/vmecritical.com/EquipmentReliability.Dec05.pdf (a copy of this PDF should also be included in the same folder as this document)
The maximum gravitational loads that should be applied to the model are 10g for acceptance testing and 14.1g for qualification testing.
A CAD model must be constrained in some way during analysis.
For a CubSat structure, these constraints will depend upon the orientation of the satellite on the P-POD during launch:
The face of the CubeSat that is facing the ground should be fully constrained (Δx = 0, Δz = 0, Δy = 0, Δϴx = 0, Δϴz = 0, Δϴy = 0). This will simulate the floor of the P-POD accelerating as the launch vehicle takes off.
The face of the CubeSat that is facing upwards should be constrained in all directions other than the vertical direction (Δx = 0, Δz = 0, Δϴx = 0, Δϴz = 0). This will allow the structure to warp slightly during takeoff, but will still simulate the rigid sides of the P-POD.
Sideways oriented faces should remain unconstrained (as they will have to warp slightly in response to the boundary condition applied to the upper face).
The static analysis is conducted in order to find the maximum and minimum stresses and displacements that the model will experience.
The modal analysis is conducted to find the mode shapes of the structure and the frequencies at which they occur.
Mode shapes occur when the structure oscillates at one of its natural frequencies.
The buckling analysis is conducted in order to find the ultimate load factors under which the model will buckle and subsequently fail.
NB Small features of the CAD model such as holes or extrusions cause large stress concentrations and will affect the buckling analysis if they are omitted.
The design of the structure should be altered with each analysis in order to optimise the structure for the purposes of the mission.
The standard materials for CubeSat structures are aluminium 7075 and 6061.
Aluminium 7075 is reportedly difficult to shape manually as it is heat treated and quite brittle.
Alternatives may be researched based upon their overall strength and the ease of working them manually. If aluminium 7075 or 6061 is not used, a DAR (deviation waiver approval request) form must be filled out prior to certification of the structure for flight.
[[collapsible show="+ Manufacturing" hide="- Hide Manufacturing"]]
Structure should be manufactured from flat sheet metal using either hand tools or a CNC cutting machine to produce the panels and their respective lightening holes.
Panels should then be bent into shape and the sides bonded together
In most cases, the material must be heated and softened in order to facilitate bending.
Bend radius is constrained by the CubeSat specifications to ≤1mm
Final flight finishes are applied to the engineering model
Sufficiently smooth surface finished is required (surface roughness no greater than 1.6 μm)
Must be anodized along the corner rails
[Edited extract from “Design, Manufacture and Testing of a CubeSat structure for the UKube-1 mission” thesis by Gordon Mackenzie]:
The structure was manufactured in three parts and consisted of the main body and two end plates. The main body was attached to several small tabs on the end plates with small bolts to complete the structure.
The first step after submission of the drawings was for the design to be cut from single sheets of Aluminium 7075 by a CNC router. This created the cut-outs in the side panels and end plates. It also cut the perimeter, including the tabs on the end plates. DXF (AutoCAD drawing) files are formatted in a way that the CNC machine could understand and implement.
Then the main body had to be folded and secured. Before the bending could be done however, the material had to be annealed as it was initially too strong to bend without cracking. This was carried out by marking out the fold lines and then rubbing these areas with soap. The soap was used to indicate the correct temperature during the annealing process.
The properties of the soap meant that when the correct temperature was reached, the soap would combust, turning a black/brown colour. An acetylene torch was used to heat the areas to be folded until this occurred and the parts were then left to cool enough to remove from the area and work with. When cooled, the parts were quickly taken and folded. The annealing warped the plates slightly, but this was corrected for the most part when the material was folded.
The flat sheet was then folded 4 times, once along the middle of each of the three rail sections and then once along the fold over tab which was used to bolt the body together. In a flight model, this would be done with rivets, so as to make the fixture permanent and to reduce external and maximise internal dimensions. The end plates were clamped in place so that each of the tabs protruded from the edges of the rig and was then hammered until flush with the edge of the bending rig.
Once the three constituent parts were completed, they were fastened with small bolts. Although these would be unacceptable on a flight model due to space requirements, they were deemed sufficient in this case as no internal payload would be present at any point during testing. For a full flight model, the bolts would have to be counter-sunk and use special clinch nuts that are much slimmer than the traditional style. This is to comply with the stringent requirements on the dimensions.
The bend radius of each fold should have been equal to the thickness of the material on each of the bends. This was not achieved as the simple box bending device did not allow for much accuracy in the bend and the warping of the sheets due to annealing further decreased the quality of each fold. The small offset tab was also inaccurately created as the four bolts did not allow sufficient pressure to be created to produce the two bends accurately.
Testing the structure in vibration to see how it would react on its expected launch, using sinusoidal sweep and random vibration tests.
Sinusoidal sweep tests
Sinusoidal sweep tests are used to assess the natural frequencies of the structure using experimental means (a mechanical vibration table).
Acceptance testing usually runs between 5g and 10g (accelerations as small as 0.05g may be utilised in the event of equipment constraints).
Results are conventionally graphed with vibration frequency against acceleration.
Random vibration tests
The random vibration tests are utilised in order to assess the ability of the structure to cope with its takeoff loads.
This test is also carried out using a mechanical vibration table, but unlike the sinusoidal sweep tests, a signal generator is used to generate the random oscillations for the experiment.
Random frequency variations are recorded in total root mean square Gs (Grms). A recommended range for random vibration testing is from 20Hz to 2000Hz.
For a flight model, additional tests including shock and acoustic experiments must be carried out to assess the worthiness of a CubeSat structure.
- High competency in CAD software (ANSYS Workbench, Pro/E, etc.)
- High competency in FEM software (ANSYS, Pro/E Mechanica)
- Rudimentary understanding of structural design, stress concentrations, material properties/selection, etc.
- Experience with manual working of metals
- Knowledge of experimental structural testing
- Excellent communication with subsystem teams
- Access to computers with appropriate software (University computer labs or at home via VPN)
- Lab technicians (allow access to tools, workshops, testing equipment, etc.)
- CNC cutting machine (optional, but very useful)