The Aerospace Engineering Computational Fluid Dynamics Laboratory (AECFDL) at North Carolina State University is concerned with developing novel algorithms for solving the equations governing fluid flow and applying such algorithms toward the solution of complex engineering problems. Three faculty members, Drs. Jack R. Edwards, Hassan A. Hassan, and D. Scott Mcrae, (see Faculty page) direct research performed within the laboratory. Current research thrusts are centered in four major areas: (see Research page)

• Transition and turbulence modeling

• Chemically-reacting and multi-phase flows

• Simulation of flows within high-speed propulsion devices

• Meteorological flows

Each of these efforts mentioned above leverages CFD algorithms and modeling techniques developed within the AECFDL. These include:

Low diffusion flux-splitting schemes (LDFSS).

The LDFSS are high-resolution upwind-differencing strategies valid for real fluid flows at all speeds. Extensions suitable for multi-component, multi-phase flows have been derived by Dr. Jack Edwards and his students and have been embedded into several production-level Navier-Stokes solvers for computing general reactive and multi-phase flows.

k-z transition and turbulence models.

The k-z two-equation model is a novel formulation developed by Dr. Hassan Hassan and his students. It is coordinate-invariant and is capable of modeling wall-bounded and free-shear flows with high levels of accuracy. This model (as well as the Spalart-Allmaras model) has been combined with a transition prediction framework to yield a unified approach capable of predicting transitional and turbulent flows. The transition model itself accounts for Tollmein-Schlicting, second-mode, and high disturbance environment mechanisms and predicts both the onset and extent of transition.

Dynamic solution-adaptive gridding algorithms (DSAGA).

Several approaches for dynamic adaptive mesh movement have been developed by Dr. D. Scott McRae and his students. Extensions suitable for time-accurate flows, multi-block grids with non-contiguous interfaces, and unstructured grids have been developed. The techniques have been applied to a variety of problems, including high-speed inlet unstart simulations and pollutant source tracking in advanced air quality prediction models.

Hybrid large-eddy / Reynolds-averaged (LES/RANS) simulation strategies.

Several approaches for blending RANS modeling techniques near solid surfaces with LES methods away from solid surfaces and in free-shear flows have been developed and applied to internal flows characteristic of high-speed propulsion devices by Drs. Jack Edwards, Hassan Hassan, and their students. These techniques are zonal in scope and depend on various flow-dependent blending functions that distinguish whether RANS or LES techniques are appropriate for a particular region in the flow.

Immersed Boundary Methods for high Reynolds number flows

Immersed boundary methods are a means of virtually replicating the effects of an object on the surrounding flow without resorting to a body-fitted grid. Drs. Edwards and Choi within the AECFDL have developed an immersed-boundary method that uses power-law interpolations for the velocity field to remove the conventional restriction to low Reynolds number flows. The basic technique has been validated extensively for incompressible flows and is currently being extended to compressible, turbulent, and multi-phase flows.

[Top]

The AECFDL’s in-house computing facility contains several PCs and workstations. TECPLOT (Amtec Engineering) is used for flow visualization, while GridGen (Pointwise, Inc.) and GridPro (Program Development Company) are used for mesh generation. CAD-to-grid capability is provided by licenses for Solidworks and CadFix. While some calculations are run on local workstations, most utilize NC State’s 1500 + processor Linux Cluster (‘Henry 2’), operated by the High Perfomance Computing component of NC State’s Information Technologies Division. Through a Partnership program , the AECFDL has built a sub-cluster containing 150 processing cores that is integrated within the main Henry 2 cluster.

[Top]

A good description of the past history of the AECFDL and a description of the educational philosophy of its members was presented as “A Perspective of CFD Education,” AIAA Paper 99-0911, by Drs. D. Scott McRae and Jack R. Edwards at the 37th Aerospace Sciences Meeting, Reno, NV, January 1999 Our course sequence may be found in the Courses page.

[Top]