This work evaluates the ability of a hybrid Reynolds-Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) turbulence method to accurately predict the physics of an unsteady separated flow field in an unstructured legacy RANS computational fluid dynamics code. The hybrid method consists of a blending of the k − ω SST RANS model with a one-equation LES model for the subgrid-scale turbulent kinetic energy (k sgs). Unstructured grids provide better resolution of complex geometries which is the motivation for extending this method. Correlations include theoretical data, experimental data and computational results with RANS turbulence models.

The focus of this paper is to discuss the unique challenges introduced through the use of unstructured grids in rotorcraft computational fluid dynamics (CFD)-computational structural dynamics (CSD) coupling. The use of unstructured grid methodology in CFD has been expanding because of the advantages in grid generation and modeling of complex configurations. However, the resulting amorphous distribution of the grid points on the rotor blade surface provides no information with regard to the orientation of the blade, in direct contrast to structured grid methodology that can take advantage of the ordered mapping of points to identify the orientation as well as simplifying airloads integration. A methodology has been developed and is described here, which now permits unstructured methods to be utilized for elastic rotary-wing simulations. This methodology is evaluated through comparison of the UH60A rotor with available flight test data for forward flight.

No abstract

This work compares the aerodynamic and aeroacoustic predictions for flatback airfoil geometries obtained by applying advanced turbulence modeling simulation techniques within Computational Fluid Dynamics (CFD) methods that resolve the Reynolds-Averaged Navier-Stokes (RANS) equations of motion. These flatback airfoil geometries are designed for wind turbine applications. Results from different CFD codes using hybrid RANS-LES and RANS turbulence simulations are correlated and include analysis with experimental data. These data comparisons include aerodynamic and a limited amount of aeroacoustic results. While the mean lift prediction remains relatively insensitive across many simulation techniques and parameters, the mean drag prediction is dependent on both the grid and turbulence simulation method. Aeroacoustic predictions obtained from post-processing of the airfoil surface pressure agree reasonably well with experimental data when consistent boundary layer tripping is used for both the simulation and experimental configuration.

Computational Fluid Dynamic (CFD) methods were used to determine the basic aerodynamic, performance, and stability and control characteristics of the unmanned air vehicle (UAV), Kahu. Accurate and timely prediction of the aerodynamic characteristics of small UAVs is an essential part of military system acquisition and air-worthiness evaluations. The forces and moments of the UAV were predicted using a variety of analytical methods for a range of configurations and conditions. The methods included Navier Stokes (N-S) flow solvers (USM3D, Kestrel and Cobalt) that take days to set up and hours to converge on a single solution; potential flow methods (PMARC, LSAERO, and XFLR5) that take hours to set up and minutes to compute; empirical methods (Datcom) that involve table lookups and produce a solution quickly; and handbook calculations. A preliminary aerodynamic database can be developed very efficiently by using a combination of computational tools. The database can be generated with low-order and empirical methods in linear regions, then replacing or adjusting the data as predictions from higher order methods are obtained. A comparison of results from all the data sources as well as experimental data obtained from a wind-tunnel test will be shown and the methods will be evaluated on their utility during each portion of the flight envelope.