Скачать 154.46 Kb.
It was determined that the prototype deorbiting system was to be an inflatable device. Using rigid structures to increase the surface area would be impractical due to the added mass of the rigid structure, actuators, gears, motors etc. In addition, inflation can occur as a passive action. Once the valve to the pressurant gas is opened no further electrical or mechanical input is necessary. With the rigid structures enough battery reserve must be maintained to run the electric motors/actuators to drive the gears and tracks. Electrically intensive experiments may not provide enough end-of-life power margins to accommodate that. There are also fewer moving parts with the inflatable. For example launch loads may accidentally misalign gears and tracks or electric motors may seize with the rigid structure whereas the lid and valve are the only moving inflatable parts and are inherently simple.
As mentioned previously, the primary materials required for the successful demonstration of the deorbiting technology include: aluminum 6061-T6, Elastosil S36 adhesive, Suva 236fa refrigerant, and Kapton polyimide film. During the course of designing and fabricating the prototype, as with all projects, substitutions and compromises had to be made. These compromises were approached from the standpoint that the purpose of the project was to demonstrate our ability to communicate with the CubeSat bus and trigger the deployment of an accurately sized and operational deorbiting device.
While NASA regulations require that structural components for CubeSats be manufactured from certain alloys of aluminum, the current prototype would not actually be utilized on a launch vehicle. With that being said the decision was made to work with the machine shop and utilize the materials they had on hand. The design constraint was that the deorbit device enclosure must still fit within the size restrictions as originally outlined. The added benefit of this decision was that the container was fabricated from stock, effectively eliminating the cost of procuring aluminum which was originally estimated to be $15-$20. Costs for aluminum were subsequently verified upon contact with several suppliers throughout the greater Hampton Roads area as well as nationwide. Prices ranged from as low as $4 per sheet for a couple square feet to as much as $146 for 48 ft2. Substituting stock material for aluminum 6061-T6 was deemed acceptable as it will not impact the overall success or failure of the prototype demonstration.
For demonstration purposes, the inflatable device must be activated similarly to how it would be when actually deployed on the satellite. Developing the prototype as an integrated whole and also incrementally in stages requires many test inflations. Utilizing Suva 236fa for each test firing was determined to be impractical; each time a test was to be performed the small cartridge would have to be charged with new refrigerant and sealed. The lab in which testing and integration is carried out provided an option to eliminate the cylinder recharging from the routine. A metered air supply from a large reservoir is fed to the lab via air line, which has a built in pressure gauge. Air is available in pressures ranging from just barely above atmospheric to as much as 100 psi. This pressure range is more than viable for the inflatable. On orbit, the Suva would be providing a pressure of 84.7 psi for inflation into a vacuum. To compensate for sea level the pressure on the air supply can be metered to 99.4 psi mimicking the Suva and also compensating for ambient conditions. Using the lab air supply was a conclusion based on both necessity and practicality: necessity because attempts to contact a Suva 236fa supplier were unsuccessful, and practical for the reasons outlined above. Success of the prototype demonstration will not be impacted by this substitution and the results will actually be improved. Improvement will come from the fact that the pressure supplied to the inflatable can be more readily controlled with the air supply as opposed to a small gas cylinder whose pressure is sensitive to temperature fluctuations. Also, the effects of under pressurization and over pressurization can be easily studied by simply increasing the flow from the air supply. A benefit in project cost is seen in the elimination of purchasing refrigerant, obtaining a suitable gas cylinder, and EPA regulations.
A further challenge was encountered due to the specification of the Elastosil S36 adhesive. Wacker Chemie AG was contacted in an attempt to obtain a price quote for small quantities of this specific product. It was discovered that they had no knowledge of Elastosil S36, and were certain they did not produce it. Going back to the drawing board the base specifications for an adhesive applicable to the CubeSat environment were determined as tabulated in Figure 6 below.
Figure : Specifications for CubeSat adhesive
Upon returning to Wacker Chemie with these specifications they suggested another one of their products, Elastosil S691. The 691 series is a two part RTV silicone rubber approved for space applications in terms of being both low outgassing and radiation resistant. It has a slightly narrower temperature range however, from -110 oC to +148 oC. Another area for concern was that the low range of the tensile strength of S691 is 0.5 N/mm2 below the NASA mandated minimum tensile strength of 4.5 N/mm2. This is not a terribly large difference, but it should be kept in mind as a non-ideal curing or mixing environment may cause the cured adhesive to lean towards the low end of the strength range. In light of this development and due to the fact that Wacker Chemie can only quote bulk quantities of adhesive, it was concluded that alternative adhesives would be explored for the deorbiting prototype.
As mentioned in the discussion of the aluminum and Suva this prototype would not see actual flight and so some liberty could be taken in terms of materials when practicality dictated otherwise. Since only one prototype was to be constructed and the ultimate goal was to demonstrate the ability to initiate and perform the inflation commonly available adhesives were chosen to seal the edges of the inflatable, which also further reduced cost. A consumer contact cement was chosen as an acceptable adhesive for the thin film media in use.
One aspect of the material selection that remained unchanged was the Kapton polyimide film. Kapton had been accrued in the previous semester in sufficient quantity so that no further product need be purchased. However, sixty feet of 2 mil Mylar film was procured for testing purposes. This will provide a media similar to the Kapton so that more testing can be conducted. Kapton has impressive material properties for a thin film. It has a Young's Modulus of 2.5 GPa (370,000 psi), a density of 1.42 g/cc, and a Poisson's ratio of 0.34.
First round testing was conducted with the intent of determining the feasibility of the current inflatable folding scheme. This is a critical aspect of the inflatable deorbiting device. How the bag is stowed prior to deployment helps determine the volume that must be allotted to contain it. The less efficient the folding technique the more mass and space will be required to stow it.
The test utilized the air line in the lab for pressurizing purposes. Pressure was read from the pressure gauge supplied with the line. For this first test the inflation pressure was regulated manually to 14.7 psig. This provided a slow enough deployment to ensure that all the seams of the test inflatable were securely fastened and there were no leaks prior to higher pressure testing. The dummy inflatable was constructed from waste bin liner, as this material is both elastic and flexible, much like the Kapton material. Two dummy inflatables were fabricated to test two different adhesives. On one bag the seam was sealed with Loctite brand adhesive while Mod Podge brand adhesive was utilized on the other. To initiate inflation a slit was cut in one corner of the bag where the air line was inserted and then sealed. Figure 7 below is a diagram of the test set up.
After curing, both bags were folded using a simple folding technique similar to that outlined in the base research, detailed in Figure 7 below (7).
Figure : Corner Folding Method (7)
Both bags were initially tested at 14.7 psig. It was found that the Mod Podge adhesive had not adequately cured, rupturing the seal. The Loctite sealed bag was then admitted to round two of testing in which the inflation pressure was increased to 60 psig with the same simple folding. The importance of folding technique was demonstrated by this test, as even the elevated pressure was not enough to completely expand the inflatable. The problem is believed to be that one seam of folding was creased more tightly than it had been in the previous test and the pressure from the adjacent expanding area actually acted to compress the fold further. To ensure the sealing properties the inflatable was manually unfolded and the inlet pressure increased until the seam burst at approximately 115 psig. As a result of this testing two conclusions were drawn: 1) the Loctite adhesive was adequate for testing purposes and 2) the folding technique previously prescribed must be further analyzed and optimized. A photo of the test is shown in Figure 8 below.
Figure : Test Photo
The choice of valve was a critical step for the implementation of the deorbit device. The sourcing for a space ready solenoid valve that meets the volume requirements proved to be quite difficult. It has been determined that a custom valve will need to be designed for the purposes of the space qualified solenoid valve.
Continuing the practice of using readily available suppliers for the prototype, a small solenoid valve was purchased. The Clippard Company manufactures a number of small pneumatic and hydraulic components. It was determined that an EM series 2 – 12 solenoid valve will meet the requirements of testing. At this point several local suppliers were contacted. Quotes were obtained from Hydraulic Pneumatic Systems, Inc. and Carolina Fluid Components, LLC. The prices were compared with the Clippard factory supplier and a purchase from the factory was the most economical.
The physical properties of the EM-2-12 solenoid valve will permit the use of the lab supplied air supply. The valve operates on 12 VDC and has a temperature range of 0 to 180 degrees Fahrenheit. The physical properties of the valve are quite compact, roughly 1 inch tall and 0.75 inches in diameter. The small size of this valve limits the flow rate of the system. The maximum flow rate at the pressure we will be using is 0.6 scfm (10).
The valve is connected to the lab air supply via a machines steel manifold. This was selected from the available list from Clippard. The chosen manifold was the 15491-2 inline manifold. This extended the total length of the apparatus by 0.8 inches. The manifold has 1/8-27 NPT and 10-32 threaded connections for connection to the air supply (10). The assembled valve and manifold can be seen below in Figure 9.
Figure : Valve and Manifold Assembly
For this project it was absolutely essential to have a reliable astrodynamics program capable of performing the complex calculations necessary to monitor the decay of an on-orbit satellite. The ideal program would have to be versatile enough to allow for all orbital parameters to be variable, would have to take into account solar cycles for solar heating and atmospheric density variation, and also industry standard so that the results would be able to be verified. All of these characteristics and more are represented in Satellite Tool Kit (STK) developed by Analytical Graphics, Inc.
STK allows for the simulation of all manner of vehicle: satellites, surface ships, submarines, and land vehicles. STK is able to link the time dependent dynamic positions of all of these various vehicles along with over 40,000 ground stations worldwide. With this information mission planners can determine when a certain satellite in a specific orbit will be within communications view of specific targets. In addition to this another powerful utility of STK is that it can solve complex orbital equations. For problems that deal with real orbits the assumption of a single point mass orbiting a much larger point mass is not enough. This is referred to as the classical two-body problem and is extensively covered in the literature (11). In actual orbits there are several sources of orbital perturbation. Perturbation is defined as a deviation from the norm, or in this case, a force that causes deviation from a prescribed orbit. The fact is that the Earth is not perfectly round. Earth's ellipsoidal character is known as oblateness and is one of the primary sources of orbital perturbation. Another source of orbital perturbation is the presence of the Moon. In fact every celestial body exerts a force on each other; this includes the solar system and the galaxy. A precise orbit determination is dependent upon all of these forces. Each applicable force is rolled up into a single equation that gives the acceleration of the orbiting body of interest in relation to the coordinate system. This equation is
The above equation is a second order, nonlinear, vector differential equation that has yet to be solved analytically (11). As such, numerical solutions must be found to accommodate the complexities just in the summation of the forces, which include perturbations, gravitational forces, and of course drag forces, which are all functions of velocity, position, and time. This also happens to be the equation governing the present problem of orbital decay. STK provides such numerical computing power. Specific geometries of the deorbiting system can be supplied to STK and the program can calculate the time dependent orbital parameters. NASA and international regulations requiring that orbital objects have a life of 25 years can be satisfied by the present inflatable deorbiting device, and this can be verified using industry standard STK software. Licenses were obtained both for this project and for the Mechanical and Aerospace Engineering department at Old Dominion University.
The Air Force Research Lab provided the plug and play development kit for prototyping purposes. Most of the architecture is done for handling the base health functions of the bus. However the integration of sensors is done by coding in C language. The challenges are compounded by the compiling which is done off site by the kit developers. In this manner, the software development is a two part operation. One part is done on site at Old Dominion University, and a second part is done automatically on offsite servers. This dependency has limited the use of this setup for the testing of the deorbit device.
LabVIEW is being used to test and operate the prototype setup in the laboratory. LabVIEW has the interactive GUI setup that can be used to actuate the relay and valve for inflation. The user must also have LabVIEW’s multifunctional Data Acquisition (DAQ) device. The DAQ device contains analog and digital inputs/outputs, allowing one to interface with different systems. In order to initiate inflation, the user must first create a Boolean program. This allows one to control the program through a “true” or “false” statement. When inflation is desired, the user simply switches the Boolean statement. The change in statement commands the program to send a small signal (1 to 5 Volts) to the DAQ device. The device outputs the signal to a relay, where the small voltage is used to close the relay contacts. The relay is used to switch the 12 V power supply for the solenoid valve. Inflation can thus be actuated with one click on LabVIEW’s user interface.
The LabVIEW setup can also be used to collect data from measuring devices in the inflation sequence. It will be useful to collect pressure data inside and around the inflatable. This data can be compared to a video of the inflation. This may provide valuable information of the effects of pressure on folding technique.
The remote activation of the deorbit device will be conducted using an AVR microcontroller. This device uses the same C coding language as the Air Force Research Lab kit. The software chosen to write and compile the code is AVR Studio 5. This is the proprietary software developed by Atmel for coding AVR microcontrollers.
AVR Studio 5 uses a version of Microsoft Visual Studio 2010 to develop and compile the code. The benefit of using AVR Studio is the ability to easily flash the new code onto a physical microcontroller. AVR Studio has a database of microcontrollers and the proper instructions for interfacing between computer and hardware.
The Old Dominion University machine shop was used to construct portions of this project. The machine shop built an actual 1 unit CubeSat shape for volume purposes Figure . The most important work done by the machine shop was a replica of the volume allowed for the entire deorbit device. The sketch passed to the machine shop is shown in Figure 10. The finished piece is made of aluminum and shown in Figure 12. These parts were necessary to build the prototype to the design constraints. The physical limitations of the part in Figure and Figure have proven very useful. There was no explicit cost associated with these parts. This was due to the use of scrap materials from the machine shop.
Figure : Machine Shop Drawing
Figure : 1U CubeSat Shape
Figure : Volume Requirement
To demonstrate that the deorbit device will function in the space environment, a vacuum chamber must be used. One type of vacuum chamber that can be used for demonstration is a bell jar, shown in Figure 13. A bell jar consists of a glass casing that is placed on top of a base. Then, air is pumped out of the chamber by a standard vacuum pump. Bell jars are low-cost vacuums that can typically be pumped down to less than one millibar. For this demonstration, 1 mb is sufficient for space simulation, in comparison to the standard atmospheric pressure of 1013.25 mb.
Figure : Bell jar (with metal bell jar guard) for simulating space conditions
Inside the vacuum chamber, demonstration can be executed with more plausible pressures that the device would experience in space. The low pressure atmosphere will affect many aspects of the device performance, such as how the membrane tears and how quickly and easily the inflatable material unfolds. Also, with the introduction of the vacuum chamber, lab airflow need not be used; the device can be inflated internally and isolated, presenting a more realistic demonstration. Bell jars can be equipped to allow for the passing of electrical feedthroughs to the inside of the chamber. Consequentially, LabVIEW can be used for initial testing. Remotely controlled demonstrations follow.
Once the inflatable is evacuated, folded, and placed in the container there must be a way to restrain and protect it. This dictates the use of a container lid. At the same time, once the CubeSat has arrived on orbit the lid must not restrict the inflatable from deploying, as this would cause a failure of the system and violate the deorbiting regulations. The lid must be both strong enough to restrain the device when stowed but passive enough to allow inflation. Puncturable membranes allow this flexibility. Another option that was explored was a mechanical door held in place by a latch. When this latch was remotely activated it would release the door, which would move out of the way of the inflatable with a spring. While simple, this solution presents the challenges of having to package the spring, hinges, latch, and latch servo. Space is already a premium inside the inflation container. Mass is also a consideration. While taken together, the accoutrements necessary for a mechanical door could add up to a large fraction relative to the overall inflatable mass (especially the servo), which is restricted to 150 g. Measures would also have to be taken to ensure that the door would not swing back and strike the inflatable once deployed. In light of these challenges, the puncturable membrane becomes a practical option.
Suitable materials for the container lid for the actual device have many of the same requirements as the inflatable itself. A thin material such as Kapton can provide the rigidity necessary for storage and installation purposes. To increase ease of deployment, serrations can be etched into the surface of the thin film lid, creating stress concentrations so that the film tears more easily due to internal pressure. Estimating the pressure necessary to cause tearing of the membrane is important to ensure the pressure of the inflatable device can accomplish the task. For a flexible membrane the increase in curvature due to a steady internal pressure is analogous to a blister test in the thin film industry. This pressure dependent curvature is given as
where P is the internal pressure, a is the semi-major axis of the subsequent ellipsoidal shape, and Y is defined as
where E is the Young's Modulus of the material (2.5 GPa for Kapton) and ν is Poisson's ratio (0.34 for Kapton). This pressure dependent evolution is plotted in Figure 14, and the geometry of the problem is shown in Figure .
Figure : Plot of semi-minor axis as a function of internal pressure
Figure : Geometry of blister test problem
The stress developed in the material is a balance to the internal pressure exerted by the inflatable. For the CubeSat deorbiting device this pressure varies from a minimum of 0 psi to a maximum pressure of 84.7 psi (584 kPa). The radius of curvature of the ellipse that the pressure will act over is
Wherein b is the semi-minor axis of the ellipse (h), φ varies from 0o to 90o, and e is the eccentricity defined as
The circumference per unit length that the pressure acts on is given by the integral
This can be solved to produce what the ASME (12) defines as a torispherical shell. The shell thickness required to resist and internal pressure P is given from the above equations as
r is the radius of interface between the torisphere and its base and E is the joint efficiency factor which is modified depending on the manner in which the shell is joint to its base. For first approximations E is normally assumed to be 1. In the present case there is no interface and so r goes to infinity, giving M a value of 0.75. Figure gives semi-minor axis and burst thickness as functions of internal pressure. Burst thickness is the thickness the membrane must be less than in order to rupture. Assuming a pure hoop stress results in an average error of 33.31%.
Figure : Values for semi-minor axis and burst thickness
Requirements for proof of concept testing in the lab were not as stringent as would be for the actual launch-ready device. The requirements for testing are that it be capable of constraining the evacuated inflatable until inflation is initiated, and that upon initiation it not impede the inflation of the device. Attempts were made to utilize Kapton for this purpose in the lab, but due to limited quantity and resources an alternate was sought. The substitute material would have to be very flexible, able to hold its shape due to its own inherent rigidity, but also tear easily. It was decided that aluminum foil possessed all of the qualities, and at the same time was readily available for minimal cost. Multiple sections of foil were cut to fit the container opening, approximately 10cm x 10cm.
In order to determine if stress concentrations were necessary to allow efficient tearing, the burst pressure of the aluminum foil had to be determined. A hollow cylinder approximately 4 inches in height was sealed at one end and fitted with a tap and check valve along the circumference so that air could be fed into cylinder. The open end was covered with the aluminum foil, without stress concentrations. Metered lab air was then allowed to flow through the check valve into the cylinder until the aluminum foil lid was compromised. Failure occurred at approximately 59.8 kPa gauge (8.67 psig) at the circumference of the lid. This is well below the inflation pressure of the inflatable device. To further ensure failure of the foil small sections were cut out of the lid used for final deployment testing, as shown in Figure . An added benefit of the stress concentrations was that they allowed access of the air line to the inflatable.
Figure : Container lid for deployment testing, showing cut-outs for stress concentration
The use of a digital microcontroller was a necessary step to simulate a remote activation sequence. The microcontroller chosen for the deorbit device simulation was an AVR ATMega32. The ATMega32 is an 8-bit programmable flash microcontroller with 32K bytes of memory. This microcontroller is capable of executing commands well over 1 MHz speed. The pin diagram for the ATMega32 can be seen in Figure .
Figure : ATMega32 Pinout
The microcontroller purchased was already integrated with a development board. This development board greatly assisted the ease of interfacing the components with the microcontroller. The development board had several sockets that could be used as digital I/O as well as onboard electronics. Several useful devices were preinstalled to ports on the development board. These included an infrared detector (IR) and a buzzer. These two devices were used to input commands and device states.
The microcontroller was the core processor for the activation sequence of the deorbit inflation. The activation was triggered by a handheld remote control IR transmitter. This transmitter is similar to a common television remote control. The remote control sent a coded burst of IR radiation to a receiver on the microcontroller development board. This IR detector was configured to one of the ports on the microcontroller as an input signal.
Once the microcontroller received the IR signal, processing was executed internally to change the state of other ports on chip. On port was configured to activate buzzer located on the development board. The buzzer would activate for 1 second when the state of the IR detector changed. The second port was configured to control the solenoid valve relay. The relay was controlled by sending a 5 volt signal to the coil portion of the relay. This allows the relay to close and complete the 12 volt circuit for the solenoid valve. The coding logic for the microcontroller port was chosen to be an exclusive XOR logic gate. This allowed the port to be toggled on or off depending on the signal received from the IR detector. The coding for the microcontroller can be found in Appendix III.
Several different sources of power were required to complete the activation of the solenoid valve using the microcontroller. The actual valve was designed to operate on 12 volts direct current power. This voltage was obtained by using a commercially available 12 volt regulated power supply. The microcontroller development board was design to optimally operate on 5 volts direct current. This was obtained by connecting the development board to a standard USB computer port. However, no communication took place between the computer and the microcontroller. Figure below shows a complete view of the equipment used during testing.
Figure : Lab electronics equipment