Putting the "FUN" in Fundamental Aerodynamics Research
New algorithms added to NASA's FUN3D software are dramatically reducing the time and cost of designing safe, efficient, and affordable aircraft and space vehicles
Curiosity's flawless landing on Mars...the Kepler mission's continued discovery of new planets in our galaxy...the space shuttle's voyages to deliver astronauts to the International Space Station (ISS)...Apollo 11 and Neil Armstrong's first steps on the moon. With NASA's unparalleled record of scientific and engineering achievements in aeronautics, you'd think it would be easy to design safe, efficient, and affordable aircraft and space vehicles. Think again.
Even with NASA's collective brainpower, software tools, and supercomputing resources, its aeronautics and astronautics experts are still working to understand the complex physical forces and flow fields encountered by vehicles in flight. For the last 20 years, scientists have been conducting fundamental research to develop and mature the FUN3D computational fluid dynamics (CFD) software, a suite of unstructured-grid CFD codes used for high-fidelity aerodynamic analysis and design.
"We put the fun in FUN3D," jokes Eric Nielsen, a software developer at NASA Langley Research Center, Hampton VA., who understands that, for many of us, it's hard to grasp how his work affects real-world issues. "The bottom line is that using computational tools like FUN3D to simulate and design complex vehicle configurations can speed up the design cycle and cut down on expensive wind tunnel experiments and flight testing—saving time and money."
Nielsen explained that aircraft designers are often interested in the sensitivity of vehicle performance to hundreds or even thousands of design parameters such as wing thickness and fuselage diameter. A conventional CFD approach requires an extensive series of additional simulations to quantify the impact of each individual parameter on performance. But since just one such simulation may require a million CPU hours to perform, traditional techniques are out of the question.
To address this problem, Nielsen's team has developed a unique, adjoint-based design methodology that produces efficient sensitivity analysis and is designed to take full advantage of the power of today's supercomputing architectures. The approach reduces the cost associated with CFD-based design by orders of magnitude — reducing potentially billions of hours of computing time down to the cost of a single, affordable simulation.
These adjoint-based methods, implemented on the 11,776-node Pleiades supercomputer at the NASA Advanced Supercomputing (NAS) facility at Ames Research Center, allow design engineers to formally tackle a broad array of complex optimization problems, including conventional aircraft, rotorcraft, launch vehicles, supersonic concepts, biologically-inspired micro-air vehicles, and wind energy devices. Real-world impacts include:
- Rotorcraft performance — Industry researchers and the U.S. Department of Defense use FUN3D to tackle the highly challenging multidisciplinary problems associated with modern rotorcraft configurations. FUN3D's adjoint-based approach will enable the systematic design of rotorcraft concepts with vastly improved performance and efficiency.
- Aircraft design — U.S. airframe designers across the industry use FUN3D to develop a new generation of vehicles with improved performance and reduced aircraft noise near airports. Potential drag reductions on future designs can offer tremendous savings in fuel cost and emissions when applied across an entire fleet of aircraft.
- Launch vehicles — FUN3D is being used to analyze and design the nation's next generation of launch vehicles. One example is the recent success of SpaceX in demonstrating the ability of commercial companies to access low-Earth orbit and the ISS; FUN3D is used extensively by SpaceX engineers. NASA also performed numerous FUN3D simulations in preparing for the recent Mars Science Laboratory mission.
- Energy production and efficiency — Although developed with the aerospace industry in mind, FUN3D has also been used for analysis and design of wind energy devices and micro-air vehicles. Designers in the long-haul trucking industry have also benefitted from its unique capabilities.
Recently, the adjoint-based capabilities of FUN3D have also been used to enable NASA and its industry partners to formally address the impact of sonic booms. Nielsen's team is working with designers to minimize the noise felt by observers located many miles beneath a supersonic vehicle traveling overhead.
The recent addition of new SGI Sandy Bridge nodes containing Intel E5-2670 processors, plus a customized parallel I/O capability at the NAS facility, allow the terabytes of data generated by high-fidelity simulations to be transferred and stored efficiently. Further expansions to Pleiades, along with planned upgrades to FUN3D, will enable aerospace engineers to continue to meet new challenges as they design the vehicles of the future.
Related SC12 Demo
Contact Information:
Eric J. Nielsen
NASA Langley Research Center
eric.j.nielsen@nasa.gov
(757)864-2239
More Information:
"The bottom line is that using computational tools like FUN3D to simulate and design complex vehicle configurations can speed up the design cycle and cut down on expensive wind tunnel experiments and flight testing — saving time and money." — Eric Nielsen, NASA/Langley
