Advanced Computing in the Age of AI | Friday, March 29, 2024

NASA Teams up with TACC to Simulate Space Junk 

<img style="float: left;" src="http://media2.hpcwire.com/dmr/NASALDEF.jpg" alt="" width="95" height="62" border="0" />While debris might seem innocuous enough to those grounded on Earth, even pebble-sized junk in orbit poses a serious problem that calls for serious compute power. In order to engineer sturdier shields, NASA has joined forces with the Texas Advanced Computing Center (TACC) to simulate the collisions between space junk and spacecraft...

While debris might seem innocuous enough to those grounded on Earth, even pebble-sized junk in orbit poses a serious problem. While you may imagine abandoned screws and bolts floating serenely past the windows of the International Space Station, the reality is that these objects are actually racing along at up to 33,500 miles per hour—about ten times faster than a speeding bullet.

At that speed, even dust is enough to rip insulation from a ship or throw it off course if it’s able to accumulate into dust clouds.

And unfortunately there’s no shortage of space junk surrounding our planet, because just about every time we launch into orbit, we leave something behind—whether it’s fragments of aluminum, nylon, or even liquid sodium from Russian satellites. According to NASA, that debris adds up to more than 21,000 pieces of ‘space junk’ larger than a baseball with an additional 500,000 golf ball-sized pieces filling in the gaps.

Yes, space is vast, and there is by no means a barricade of junk separating Earth from the rest of space. And our spacecraft have been designed to withstand the impact of small objects. But as Eric Fahrenthold, professor of mechanical engineering at The University of Texas at Austin, explains, this problem calls for some serious research, and some serious compute power.

“If a spacecraft is hit by orbital debris it may damage the thermal protection system,” said Fahrenthold. “Even if the impact is not on the main heat shield, it may still adversely affect the spacecraft. The thermal researchers take the results of impact research and assess the effect of a certain impact crater depth and volume on the survivability of a spacecraft during reentry.”

As you might guess, only a small number of collisions that occur in orbit can be accurately simulated in an Earth-bound laboratory, because the objects are too large, small, dangerous or fast to reproduce. Instead, NASA has turned to running simulations on the Ranger, Lonestar and Stampede supercomputers at the Texas Advanced Computing Center, where Fahrenthold and his students have been working to establish ballistic limit curves that predict at what size and speed a piece of debris will break through a spacecraft’s shielding.

The simulations were able to model both how the projectile would break apart on impact, as well as the thermal and mechanical loads that the target surface would experience on impact. This means that after researchers provide information about the material’s strength, flexibility and thermal properties, the supercomputers are capable of capturing the cascade of events that unfold as layers of fabric fray, capturing some of the projectile’s fragments while giving way to others.

The work was then tested against NASA’s real-world testing, which involved light gas guns launching centimeter-sized projectiles at 10 kilometers per second. For a comparison, the space junk in orbit travels anywhere from 5 to 15 kilometers per second, on average.

"We validate our method in the velocity regime where experiments can be performed, then we run simulations at higher velocities, to estimate what we think will happen at higher velocities," Fahrenthold explained. "There are certain things you can do in simulation and certain things you can do in experiment. When they work together, that's a big advantage for the design engineer."

The team presented their results at the April 2013 American Institute for Aeronautics and Astranautics’ meeting, which were later published in Smart Materials and Structures and International Journal for Numerical Methods of Engineering. There, they detailed how variations in speed, impact angle, and size of the debris would affect the depth of the crater it would produce in ceramic tile thermal protection systems.

But this won’t be the only application to benefit from Fahrenthold’s research. With the help of Moss Shimek, a graduate student, Fahrenthold have stretched the research to study the impact of projectiles on body armor as well. This extension of the team’s research was made possible because body armor uses many of the same materials used for orbital debris protection in space, like Kevlar. What they found is that more flexible weaves can increase protection offered by body armor.

"Currently body armor normally uses the plain weave, but research has shown that different weaves that are more flexible might be better, for example in extremity protection," Shimek said.

"Future body armor designs may vary the weave type through a Kevlar stack," Shimek said. "Maybe one weave type is better at dealing with small fragments, while others perform better for larger fragments. Our results suggest that you can use simulation to assist the designer in developing a fragment barrier which can capitalize on those differences."

While these outcomes had to be verified by experiments in the lab, Farhenthold said that supercomputer-enabled simulations ultimately led to faster, cheaper research, whose findings could expand beyond the range of what’s possible in the lab.

"We are trying to make fundamental improvements in numerical algorithms, and validate those algorithms against experiment," Fahrenthold concluded. "This can provide improved tools for engineering design, and allow simulation-based research to contribute in areas where experiments are very difficult to do or very expensive."

EnterpriseAI