Autonomous Satellite Docking System Extends The Life Of Micro-Satellites With On-Orbit Repair Capability.

Simulation Software Provides Zero G Test Environment for Autonomous Satellite Docking System.

The number of satellites that can be built and launched is limited by high costs, making life extension for satellites an aerospace engineering priority. One solution is to design satellite architecture for on-orbit servicing using an automated payload handling system to replace damaged or obsolete systems and components. A standardized autonomous docking system is an essential system for on-orbit repair facilities. But like any other component used in space, developing an automated docking system requires proving it will function in a zero G environment, which is time-consuming and expensive. Working with the Defense Advanced Research Projects Agency (DARPA) and the Air Force Research Laboratory, Space Vehicles Directorate Small Business Innovative Research (SBIR) Programs, and Microcosm, Inc. of El Segundo, California, Michigan Aerospace Corporation kept costs manageable by creating a zero G test environment with system simulation and analysis software validated by physical testing.

The Air Force Research Laboratory became interested in adapting Michigan Aerospace's autonomous docking system for use with micro-satellites, which weigh 50 to 100 lbs., because its simple design could be scaled for use with large or micro-satellites. Because the system would operate in zero gravity, it would have to be tested in zero gravity to ensure it would work, which is a process too costly for most small firms. Pete Tchoryk, Executive Vice President said, "Conventional methods of prototype construction and testing are costly and inefficient. The inefficiency is magnified when designing a system for a micro-gravity environment because of the extreme cost involved in performing even a single test. In most engineering applications, simple, relatively inexpensive laboratory testing models can be developed prior to constructing a functional prototype for testing the basic concept of the device. For space systems, in particular space mechanical systems such as a satellite docking mechanism, the usefulness of laboratory concept test models is extremely limited by the presence of gravity. We turned to computer simulation with MSC.ADAMS dynamic analysis software because of its versatility in simulating the kinematic reactions between two spacecraft in micro-gravity and therefore mitigating a great deal of risk without requiring a large amount of prototype testing."

With an on-orbit capability to replenish fuel and other consumables restricted by weight and launch constraints, the usefulness of a satellite can be extended. For example, a critical factor in redirecting satellite orbit is fuel. With the ability to replenish fuel, orbits can be redirected, reducing vulnerability to attack, allow it to observe new crisis areas and reduces the need for multiple satellites to provide global coverage.

Autonomous Satellite Docking System

Adapted from a docking mechanism developed over a decade ago, the autonomous docking system allows a high degree of tolerance and is very forgiving when aligning servicing connectors and ports. The approach utilizes a simple, mechanized method for orbital replacement unit (ORU) transfer, utilizing fluid, and electrical ports located around the outside edge of the mechanism. This allows flexibility in the number and type of connectors that are used, while minimizing the complexity of the connection process.

Mr. Tchoryk said, "Our design uses a cone and a probe on the chaser spacecraft and probe on the target micro-satellite. A flexible cable extends from the probe of the chaser spacecraft to the bottom of the cone on the target micro-satellite. When it reaches the bottom, it automatically latches providing a soft dock. A hard dock between the two spacecraft is made by reeling in the cable. With only two mechanisms, it's about as simple as you can get in terms of active mechanisms. An actuator extends the cable and a mechanism at the end of the cable provides the latch." The docking sequence is shown in Figure 1.

Docking Sequence:

1. Station-Keeping. Using inputs from the docking sensor, the Autonomous Guidance Navigation Control (AGNC) guides the chaser vehicle to a station-keeping distance from the target spacecraft, maintaining anywhere from 1 meter to tens of centimeters separation.
2. Soft-Docking. The flexible cable extends toward the target spacecraft until the cable latches onto the bottom of the cone on the target vehicle. With only minimal forces applied to the target spacecraft, the two spacecraft are soft docked.
3. Hard-Docking. As the cable retracts, the target is pulled into contact with the three auto-alignment posts. The two spacecraft are hard docked when the probe head is fully seated inside the cone.
4. Servicing. When the two spacecraft are rigidly docked and aligned, fluid and electrical connections are mated with the target vehicle and servicing commences.

Figure 1 Docking Sequence

Simulation and Testing

One of the goals was to create a dynamic model of the entire rendezvous and docking process. A model for the simulation was developed by incorporating theoretical models of probe/cone performance, as well as empirical (real-world) data from Michigan Aerospace's air bearing table, flat floor physical tests at Marshall Space Flight Center's Flight Robotics Laboratory and zero G flight tests on a KC-135A.

Using geometry created in SolidWorks, the simulation provided enhanced visualization of docking scenarios. The model consisted of a target satellite located at a user-specified stand-off distance and orientation with respect to the chaser spacecraft. Parameters and components of the probe/cone mechanism were modified to analyze the performance of the mechanism under various spacecraft configurations and mission scenarios. The model for the docking mechanism used in the KC-135 testing is shown in Figure 2.

Figure 2 ADAMS Model of Docking Mechanism Tested on the KC-135

Each spacecraft was modeled in MSC.ADAMS as an independent rigid body with a finite mass in a zero-gravity environment, shown in transparent form in Figure 3. The docking mechanism interfaces were modeled as independent bodies attached to the spacecraft solids to aid in simplifying the analysis of the model, since only the contact surfaces needed to be analyzed for contact reactions by MSC.ADAMS' 3D contact solver.

Figure 3 Transparent Version of Model

The soft-dock cable was modeled as a series of discrete rigid-body segments, joined together by spherical joints, as shown in Figure 4. The discretized cable model has the physical characteristics of a chain, and approximates the behavior of a flexible cable. To emulate the stiffness of a real cable, each link in the cable model was joined to the next link by a short one-dimensional beam element with a finite stiffness value. This allowed stiffer or more flexible cables to be modeled simply by changing a single global stiffness value for the cable segments.

Figure 4 Flexible Cable Model

The resulting model was then constrained to 2-dimensional movement with a simulated gravitational field and compared to the results of the 2D flat floor testing at Marshall Space Flight Center to validate its performance. Once the model had been developed to perform in a satisfactory manner, it was used to predict and modify the design of the micro-satellite docking system that was tested on the KC-135.

Mr. Tchoryk said, "We are trying to simulate operation in an environment in which we will not get there until we produce the final product. We don't have the option of going into the plant and pulling it off the assembly line or changing the assembly line. In this respect, the simulation is essential. We didn't have this option in the past."

Simulation focused on examination of the behavior/loading of the cable/latch subassembly used to capture, position and dock the target satellite. To develop a comprehensive understanding of behavior model parameters were varied, such as the initial satellite relative positions, masses, cable/latch stiffness and cable/probe deployment speeds.

Mr. Tchoryk said, "The key is to obtain high fidelity models that represent reality. We've made a number of progressions to increase the fidelity of this model enabling us to predict how this mechanism is going to perform. It would have been very difficult to do without MSC.Adams."

2D Physical Tests

A physical prototype of the docking mechanism was tested on the flat floor facility in Marshall Space Flight Center's Flight Robotics Laboratory, shown in Figure 5.  This test was designed to simulate the effects of the docking mechanism on free-floating bodies in microgravity, in a 2-dimensional scale. The facility contains a 3,800-square-foot floor space, perfectly flat to within 0.002-inch across its entire surface.

 
Figure 5 Flat-Floor Testing at Marshall Space Flight Center (Chaser Vehicle on Left, Target Vehicle on Right)

A pair of mobility bases (large structures), floating on cushions of compressed air, were used for simulating two spacecraft in proximity operations. Each base was a large frame with precision air-bearing casters, which contained on-board control and data collection computers for communicating with a ground station via short-range radio frequency network connection; a battery bank; high-pressure compressed-air bottles; and a set of maneuvering air-jet thrusters.

Because the bases floated on air cushions on a nearly-perfectly flat floor, the effects of friction on their movement are negligible and allow a 2D simulation of microgravity maneuvers. The thrusters were used with guidance software written by SAIC and Marshall Space Flight Center's VGS (Video Guidance System) position/orientation sensor to simulate proximity operations between two docking satellites as it would occur during an Earth orbit. The chaser end of the docking mechanism was mounted to one of the bases and the target end to the other base.

"Based on flat floor testing at Marshall Space Flight Center, we were able get a good look at how the model performed in a 2D frictionless environment," said Mr. Tchoryk. "When the cable was deployed, gravity made it bow, which wasn't very realistic. It would still work, but it wouldn't give us a very realistic picture of how the cable would work in space. Still, the knowledge that we gained helped us with improving our model and refining our zero G simulated test environment to provide very close correlation between the simulated model and the physical tests."

Zero G Physical Tests in KC 135A

The KC135A is similar to the Boeing 707 with a cargo bay test area approximately 60 feet long, 10 feet wide, and 7 feet high. A typical mission lasts 2 to 3 hours and consists of 40 to 50 parabolas. Basically, the KC 135A delivers a zero G test environment for 20 seconds during a controlled parabolic maneuver.

The KC 135A enabled the following tests:

1. Determine soft dock capability under the full compliance range of the docking mechanism
2. Determine auto-alignment and hard dock capability under the full compliance range of the docking mechanism
3. Determine undocking capability and stability
4. Determine cable assembly behavior in zero-g
5. Validate ADAMS dynamic models
6. Assess extensibility to Orbital Express

The experiment was contained in a cage about 2-1/2 feet by 2-1/2 feet by 5 feet with all the appendages. Within the cage was a free flying target satellite represented by an 18" cube, large enough to contain the docking cone. Attached to the outer frame was the probe, which acts as the chaser spacecraft. This test setup is shown in Figure 6.

Figure 6 KC-135 Docking Mechanism Experiment

In zero G, the entire structure lifts off the floor. The target satellite, held in place with pressurized cylinders, is released. Both the outer frame and inner target satellite are floating. The cable is extended and latches onto the target satellite, which is reeled in, auto aligned and rigidized. A typical zero-g docking sequence on the KC-135 is shown in Figure 7.

Figure 7 Typical Zero-G Docking Sequence on the KC-135

"We were able to test our entire docking sequence during the 20 second zero G period," said Mr. Tchoryk. "Throughout four days of testing, we tried various conditions, where the target satellite was offset in position and angle. We tried different types of cables with different flexibilities and different extensions and speeds. Almost all the dockings were successful. We did have some dockings where the target was pushed away, but this was caused by the aircraft's drift, causing the target to drift away before it could be latched."

A model of the docking system that correlated very well with one of the cases on the KC 135A was identified. With this model, a review was made of all the offset cases, including the cases that didn't work, such as when the docking system pushed off without locking. By reproducing the initial conditions, using the same model, it was determined if the model would work with all the other anomalies.

"Not only did we have a model that works with a few conditions, but we have a large enough data set that we also predict other anomalies and offset cases," said Mr. Tchoryk. "When we can look at the model and determine if it will work under all these conditions, we've achieved our goal."

The alternative to simulating the complete docking system is to make assumptions about the mechanism and its performance in terms of components. It's difficult to extend those equations to the docking mechanism as a system without being able to simulate the system and its motion in a dynamic environment. All these calculations would have to be made using software that outputs numbers without any visual representation of the system. A few calculations would have to be done on the initial impact of the probe and the cone, and then validated in the laboratory.

"It wasn't impossible to test the docking system without MSC.Adams, but not with our level of funding," said Mr. Tchoryk. "Without MSC.Adams, we wouldn't get the whole system perspective or the variations to see how it performs under different conditions. Without visual and quantitative feedback, the data has to be interpreted which is not easy, especially with a complicated interface, such as the flexible cable and the cone."

Customer
Michigan Aerospace Corporation
Ann Arbor, Michigan
United States
http://www.michiganaero.com/

Software involved
MSC.ADAMS®, Solid Works