Enhancing numerical accuracy in the prediction of rotor wake vortex structures

Abstract

In modern high-fidelity computational fluid dynamic simulations, the primary vortex system in hover often breaks down into secondary vortices. The sources of numerical error influencing the prediction of the vortex system were studied by performing high-fidelity simulations of the wake of a two-bladed rotor and comparing the predictions to stereoscopic particle image velocimetry measurements in different measurement planes. Various numerical inputs, including sub-iteration convergence, blade pitch offset, and grid resolution, were varied to resolve discrepancies between the measured and predicted vortex characteristics from a previous study done by the authors. A parametric study on near- and off-body solver sub-iteration convergence demonstrated that although the secondary vortex characteristics converged as the sub-iteration convergence of both solvers increased, a large discrepancy in the number of secondary vortices remained. This discrepancy was investigated by varying the thrust, where it was found that the breakdown of the primary vortex is directly linked to the number of secondary vortices. Dissimilarities in the blade pitch angle, which could not be avoided in the experiment, were modeled by intentionally using an offset in the blade pitch angle of the two blades. It was shown that as blade pitch angle offset increases, vortex pairing becomes more distinct. When vortex pairing occurred in both the experiment and simulation, the decay of secondary vortices in the experiment and simulation agreed best. To better match the experimental resolution, grid resolution was increased and comparing the two simulations, the finer mesh simulation agreed best with the measured primary and secondary vortex characteristics.