Computational hydrodynamics applied to an America's Cup class keel - best practice and validation of methods
Doctoral thesis, 2006
Numerical predictions of the flow around yacht keels can be performed with various types of methods. In this work computational fluid dynamic (CFD) computations are validated against wind tunnel studies of an America's Cup Class keel with the purpose of investigating the accuracy that can be achieved with CFD, which methods that are suitable and how they should be used in the best way.
The test case is a quarter scale model of a keel with a fin, a bulb and two winglets, tested in a closed section wind tunnel. Force measurements as well as wake surveys are presented for several winglet pitch angles and two longitudinal positions.
The experimental results are used for validation of one RANS code and one potential flow/boundary layer code. Effects of grid size, stagnation point anomaly and turbulence model are discussed. The errors of the RANS code are slightly larger than the experimental uncertainty, whereas the potential flow/boundary layer results are within the experimental uncertainty, provided that a correct panelisation is used.
The usefulness of wake integration methods are investigated.
Force-estimates based on wake integration applied to the computed wake are proven to be equally or slightly less accurate than integration of surface forces. The measured wake is successfully used to validate the lift distribution along the winglets computed with the RANS code and the potential flow code.
A decomposition of the drag force is made by applying wake integration methods to the computed wake. The difference between vortex drag and induced drag are pointed out.
Apart from the CFD validation work, the thesis contains two short pre-studies. The first concerns methods for verification and uncertainty analysis of CFD results. Three different methods are applied to the results from an analytical test case. The second pre-study describes the first steps towards the implementation of an automatic grid adaption method for a viscous flow solver.