Focus on ability to predict separated flow

By Tim Eymann|December 2017

The Fluid Dynamics Technical Committee focuses on the behaviors of liquids and gases in motion, and how those behaviors can be harnessed in aerospace systems.

The study of separated flow and improving the ability to numerically predict it was a major topic of interest for fluid dynamics researchers in 2017. Researchers at NASA’s Langley Research Center in November conducted the NASA Juncture Flow in the 14-by-22-foot (4.3-by-6.7-meter) subsonic tunnel and planned to continue through December. The wind tunnel tests are designed to provide validation-quality data for the onset and progression of a trailing edge separation near the wing-body juncture of an aircraft. This type of separation is particularly difficult for state-of-the-art computational fluid dynamics, or CFD, methods to predict.

Separated flow is a dynamic physical phenomenon that can adversely impact the performance of aerospace vehicles. The mechanisms that lead to and result from separated flow make it difficult to model its effects with the current generation of production CFD software.

The NASA Juncture Flow experiment is providing highly detailed information that can improve the accuracy of the current numerical methods. A unique aspect of the tests is that a laser Doppler velocimetry system is mounted inside the fuselage, taking measurements through windows to capture high-resolution flow data close to the corner region. The tests are the culmination of a significant body of preparation work, including several risk-reduction experiments that identified optimal wing configurations for achieving the desired corner flow separation characteristics.

An additional class of separated flow, smooth-body flow separation, occurs when a boundary layer attached to a solid surface separates after interacting with an adverse pressure gradient generated by a change in the body contour or the presence of a shock. Simulations that rely on the Reynolds-averaged Navier-Stokes equations to simulate the turbulent flow have difficulty predicting the onset and evolution of separated flow and therefore misrepresent its effects on a vehicle.

Scientists at Langley are taking a higher-fidelity approach to modeling the turbulence known as wall-resolved large eddy simulations. The researchers combined this turbulence modeling approach with high-order numerics to simulate a canonical test problem on supercomputers at the Department of Energy using up to 24 billion grid points, which is over 200 times larger than a typical CFD run. From April to October, their work demonstrated that the power of the top supercomputers makes it feasible to perform turbulence simulations at physically relevant conditions. The data generated from these simulations are giving scientists a better understanding of flow separation physics.

At faster speed regimes well above the speed of sound, shock wave-turbulent boundary layer interactions are ubiquitous, occurring along both external and internal flow paths. This high-speed flow phenomena can not only induce flow separation, but also vortical structure formation and pressure losses, as well as amplify heat transfer, all of which can be detrimental to the vehicle operation. While researchers have characterized the physical drivers behind two-dimensional interactions, more realistic three-dimensional interactions have not been studied as extensively. Researchers at Florida State University, Ohio State University, the University of Texas at Austin and Auburn University are working jointly on an Air Force Office of Scientific Research project to fully understand three-dimensional shock/turbulent boundary layer interactions.

Throughout the year, researchers at the partner universities have made measurements on a fin-on-plate and a swept corner configuration using high-fidelity diagnostics. One such test occurred in June at the Florida Center for Advanced Aero-Propulsion Supersonic Wind Tunnel Facility at FSU. The experimenters and CFD researchers are also using outputs generated from small-perturbation inputs to feed large eddy simulations intended to cast regions of the flow as a dynamical system and to numerically study the underlying mechanisms that result in the experimentally observed unsteadiness. The work has revealed the features of the separated flow with remarkable clarity and highlighted the differences between two- and three-dimensional physics. The outcome of this combined experimental and computational research effort will help guide the design of next-generation high-speed flight vehicles.

Contributors: Rajan Kumar, Mujeeb Malik, Christopher Rumsey and Ali Uzun