Skip to main content
NASA Advanced Supercomputing (NAS) Division World-class modeling, simulation, and supercomputing in support of the NASA mission

Using NAS-Developed Tools to Quiet the Boom of Supersonic Flight

Click or Touch Image to Zoom and Pan Visualization of the X-59 aircraft concept
A flow visualization of NASA’s X‑59 supersonic, low-boom aircraft concept. This computational fluid dynamics simulation, produced on the Pleiades supercomputer using the NAS-developed Launch Ascent and Vehicle Aerodynamics (LAVA) code, will be used to help determine which components contribute to the pressure signatures below the aircraft that create the sonic boom in flight. Red and blue represent high and low pressure, respectively; the symmetry plane—which bisects the vehicle along its center line—is colored by Mach number. James Jensen, NASA/Ames

NASA’s Quiet SuperSonic Technology (QueSST) X‑59 is an advanced low-boom aircraft concept that will be used to emulate and test noise generated by future supersonic over-land transport designs. As part of this project, researchers in the Computational Aerosciences Branch in the NASA Advanced Supercomputing (NAS) Division are using agency-developed tools to run simulations of various design iterations of the aircraft and verify its supersonic performance.

However, designing and testing new aircraft concepts is expensive, both in time and costs, so QueSST project engineers have been working with computational fluid dynamics (CFD) experts at NAS to develop and validate designs using complex computer models. While performance optimization of traditional aircraft primarily focuses on maximizing the lift-to-drag ratio, designing the X‑59 also requires the prediction of sonic boom noise. To achieve this, thousands of CFD simulations using NAS’s in-house-developed Launch Ascent and Vehicle Aerodynamics (LAVA) and Cart3D tools were performed for mission planning and rapid-testing of design modifications on the Pleiades, Electra, and Endeavour supercomputers at the NAS facility, located at Ames Research Center in Silicon Valley.

Modeling Sound

Predicting ground noise is challenging, and with no experimental data for the X‑59—which is currently under assembly at Lockheed Martin’s Skunk Works factory in Palmdale, California—it is crucial to verify the CFD solutions across as many solvers as possible in order to build confidence in the predictions. Teams at NASA’s Ames and Langley Research Centers worked together to ensure that the various CFD codes used across the project predicted similar loudness values. At NAS, the team compared the results from the two solvers, Cart3D and LAVA.

In a typical Cart3D simulation—which solves Euler equations on a Cartesian mesh and requires about six hours when using one node of either Electra or Pleiades—a nearfield pressure signature is extracted from the simulated flowfield around the X‑59. To estimate the boom noise at ground level, the pressure signature is propagated from a flight altitude of 55,000 feet to the ground using an atmospheric propagation solver. Additional tools are used to provide uncertainty estimates of the pressure signatures and loudness, validation that is needed due to the varied operating conditions and vehicle configurations being tested over the course of the project.

Photo of NASA’s X-59
Computational schlieren image (similar to those produced by the schlieren technique in experiments) from a high-resolution Cart3D simulation. The dark and bright regions represent shockwaves and expansions, respectively. The complex shock system of NASA’s X-59 is shown, including the conical shocks behind the aircraft. The simulation includes propulsion effects, air probes, and secondary-air-system ducts. Significantly weaker shocks propagate from the lower surface of the aircraft, quieting sonic booms to sonic thumps on the ground. Marian Nemec, Michael Aftosmis, NASA/Ames

“The large capacities of Pleiades and Electra are critical for handling the thousands of cases required for aerodynamic performance databases, shape optimization, and uncertainty quantification of the X‑plane,” said Marian Nemec, an aerospace engineer within NAS’s Computational Aerosciences Branch and the primary developer of Cart3D’s adjoint-based mesh refinement module and shape optimization framework. “We use the Electra system for most production runs, where we can predict the entire boom footprint of the aircraft overnight when using several hundred cores. Our shape optimization cases run on the shared-memory Endeavour system, where complex optimizations require several days on 512 cores.”

These simulations contributed to numerous design improvements throughout the QueSST project, such as reducing the noise generated by the nose of the aircraft, instrumentation probes, and secondary-air-systems inlets. Cart3D was also used by the team at NAS to support the agency’s supersonic wind-tunnel tests, including using CFD simulations to verify the design of a new wind-tunnel strut that holds the scale X‑59 aircraft model in position, ensuring it does not interfere with pressure signatures during testing.

The ground level noise predictions from Cart3D simulations compare very well with those produced for the X‑59 using LAVA, which solves steady Reynolds-Averaged Navier-Stokes (RANS) equations on a polyhedral mesh. Recent improvements to the LAVA solver have significantly decreased turnaround time, with a typical X‑59 simulation running on 1,500 cores of the Pleiades supercomputer completing in only an hour or two, leading the QueSST team to choose LAVA as the main code to be used to develop a large supersonic performance database. The database will be used to feed a flight planning tool that will train pilots how to fly the X‑59 and hit target noise levels during flight tests. It will need to cover all possible combinations of control surface settings and flight conditions that might be encountered during the test campaign—a massive undertaking for the team at NAS.

“Without high-performance parallel computing, supersonic database generation for the X‑59 would not be possible,” said James Jensen, one of the NAS aeronautics engineers working on the QueSST project. “The thousands of LAVA and Cart3D simulations required to perform mission planning for the X‑59 make high-performance computing an indispensable part of this program.”

The supersonic performance data for pilot training will also be used by the flight test program to provide a variety of high-quality data for the Federal Aviation Administration (FAA) and the International Civil Aviation Organization (ICAO) to help determine the future of civilian supersonic flight.

The Future of Supersonic Flight

At the end of 2020, the construction of the X‑59 reached its halfway point, with final assembly scheduled for 2021. The aircraft will then undergo a multitude of tests to ensure the integrity of its structure and that all of its components all work correctly. Data from CFD simulations will be used during this period to help predict flight performance, as well as to develop training for pilots.

Community response testing, which is set to begin in 2024, will include flights over a wide variety of environments and locations to build up a database of noise levels and acceptability metrics. This database will be used by the FAA and ICAO to set noise level requirements for supersonic over land flight, replacing the blanket ban that is currently in effect. While the X‑59 itself won't carry passengers at supersonic speeds, it is the first great leap toward developing standards that will lead to faster, efficient air travel around the world.

—Katharine Pitta, NASA/Ames Research Center