Simulating Supersonic Parachute Inflation for Future Mars Landings
On February 18, 2021, NASA’s Perseverance Mars Rover entered the thin Martian atmosphere at hypersonic speeds of about 12,100 miles (12,500 kilometers) per hour. The shape of its space capsule generated enough drag to slow its descent to about 1,500 mph—or twice the speed of sound—but in order to slow down enough to land safely, the capsule needed to deploy a parachute. Perseverance’s supersonic parachute, 65-feet in diameter and only three thousandths of an inch thick, fully opened in less than half a second, undergoing aerodynamic forces of over 30 thousand pounds.
Understanding the inflation process of parachutes at supersonic speeds is a challenging problem NASA has been tackling since the 1960s through decades of flight and wind tunnel tests. To reduce the reliance on expensive flight tests, the agency is developing numerical methods to simulate supersonic parachute design and inflation. In addition to reducing costs, this approach could increase mission confidence by simulating how parachutes perform in a wide range of environments and trajectories that would be difficult to test in the real world.
The Launch Ascent and Vehicle Aerodynamics (LAVA) team in the NASA Advanced Supercomputing (NAS) Division was tasked by the agency’s Entry Systems Modeling project to assess NASA’s modeling and simulation capabilities for this complex problem. Using their LAVA computational fluid dynamics (CFD) software and the power of the Pleiades and Aitken supercomputers at the NAS facility, the team developed a CFD framework to accurately and efficiently simulate the complex aerodynamics at play during supersonic parachute inflation.
Solving Parachute Physics
The physics of parachute inflation include several complicating factors, including unsteady turbulent wakes; the interaction of these wakes with the shockwave caused by the spacecraft at supersonic speeds (known as a bow shock); supersonic flow through openings in the parachute; and the motion and deformation of the parachute body itself.
To simulate this chaotic environment, the LAVA team simulated several flight tests performed at supersonic conditions and compared results between the simulation and experimental data in order to validate the coupled LAVA solvers’ accuracy.
These flight test simulations included NASA’s Advanced Supersonic Parachute Inflation Research Experiment (ASPIRE) project—which helped engineers at the Jet Propulsion Laboratory select a design for the Perseverance rover’s parachute—and the coupled LAVA solvers were able scale to simulate very large fluid dynamics problems. To improve simulation accuracy, the team also developed software modules to simulate the permeability of the parachute’s broadcloth material, enabling them to effectively model the complex contact mechanics of soft materials such as fabric.
Focus on Validating LAVA Accuracy
“We are actively developing the methods and models that go into this tool to ensure that it provides the required level of fidelity to be make accurate predictions,” said Cetin Kiris, Computational Aerosciences Branch chief in the NAS Division and LAVA team lead. “So we are focused primarily on validating the tool’s predictions with existing ASPIRE data, and learning as much as we can about what engineering simplifications are appropriate.”
The team looked into many factors, including how the starting shape of the parachute—how it unfurls from its folded state—can influence the aerodynamic loads on the simulated structure. Verifying the parachute’s starting shape from the ASPIRE tests is incredibly difficult, and its variability is one of the primary points of uncertainty for the LAVA team’s models.
Their work in the last year focused on a study of the first of three ASPIRE flight tests that launched between October 2017 and September 2018 (each with progressively higher momentum at opening and different canopy materials and construction), which will act as a benchmark as the team continues to improve the accuracy and capabilities of the LAVA solver for future missions.
High-Performance Computing is a Must
“Without high-performance computing clusters such as the Pleiades and Aitken supercomputers at the NAS facility, these simulations would not be possible,” said Jonathan Boustani, one of the researchers at NASA’s Ames Research Center who worked on the project.
Without high-performance computing clusters such as the Pleiades and Aitken supercomputers at the NAS facility, these simulations would not be possible.
A typical simulation of the half-second parachute inflation period ran for a week and used more than 2,500 of Aitken’s Cascade Lake cores, producing several dozen terabytes of data. A parametric study such as this—where many simulations are run at varying descent speeds, inflation points, cable line length, and parachute material—required tens of thousands of cores to complete and generated several hundred terabytes of data.
Future work will shift from refining the physics of the simulation to focusing on high-performance computing performance and user friendliness. “Once the tool can reproduce the flight test accurately enough, the next step is to optimize the process: to reduce the turnaround time of these simulations enough to be used within the design cycle of a robotic mission to another planet,” said Kiris.
One of the team’s long-term goals is to ensure that LAVA will be flexible enough to predict the behavior of subsonic parachutes as well, like those that will be used by the Orion crew module—which is set to launch with astronauts aboard Artemis II missions in late 2023—on its descent from the Moon back to Earth.
—Katharine Pitta, NASA/Ames Research Center