Analysis and indicial modelling of helicopter tail rotor orthogonal blade vortex interaction

This study builds on previous experimental investigations of orthogonal tail rotor blade vortex interaction, and investigates semi-empirical modelling of the phenomenon. In the early stages of the work, weaknesses were identified in the published experimental data that limited their use as a modelling correlation source. For this reason, a detailed analysis of data from the experimental study of Wang et al. (2002) was conducted to allow the validation of semi-empirical modelling strategies applied to orthogonal blade vortex interaction. Indicial modelling was identified as a suitable modelling strategy due to its computational efficiency and also its current use in the helicopter industry. The long-term intention is that the subsequent integration of the orthogonal blade vortex interaction model into a full helicopter aerodynamic model will enable a more complete simulation that considers orthogonal blade vortex interaction during the design stages of a helicopter's development. A number of modelling strategies were considered during this study. Initial models were based on the Kussner function for an aerofoil encountering an upgust. The orthogonal interaction was captured by representing the axial core flow of the tip vortex as an upgust in the shape of a Lamb vortex that engulfed the entire vortex. This resulted in a markedly greater lift response compared to experimental data because the chordwise distribution of axial velocities due to the interacting tip vortex were not properly represented. The modelling approach was then improved to account for this distribution. This produced good agreement with the experimental data, where the vortex centre interacted with the blade. The predicted response was found to be symmetric about the vortex centre, which was in contrast to the asymmetry found in the experimental data. The asymmetry was investigated using a two-dimensional panel method simulation of the orthogonal interaction. It was hoped that the asymmetry could be accounted for by the rotational flow of the tip vortex, however, the panel method demonstrated that there was insufficient rotational flow to account for the magnitude of the asymmetry found in the experimental data. To investigate this asymmetry further a numerical simulation of the wind tunnel experiment was used. This simulation was inviscid and featured a three-dimensional source panel method to represent the wind tunnel walls, a lifting line calculation for the blade of the vortex generator, and a free-wake solution for the wake of the vortex generator. The simulation had previously been found to simulate the experimental wake shape well. The axial velocities predicted by this model at the location of the installed interacting blade were extracted and used as an input into the indicial model. The indicial model then reproduced the asymmetric lift response found in the experimental data; however, the magnitude of the measured lift response was not well represented. This difference may be associated with the inviscid nature of the numerical simulation or flapping of the vortex generator blade observed during the experiment. As a first step towards understanding the difference, the angle of incidence of the vortex generator was reduced in the numerical simulation until the circulation over the interacting blade matched the experimentally measured value. This resulted in a closer agreement in the magnitude of the blade vortex interaction response for ail spanwise locations. The differences between the prediction based on the prescribed Lamb type axial flow distribution and the indicial prediction based on the horizontal cross flow velocities extracted from the numerical simulation, indicate that the shape of the wake and its corresponding induced flow influence the interaction response. The prescribed indicial prediction features a sharper drop off in lift response compared to the indicial prediction forced by the simulated velocities. This can be attributed to the curved wake shape and the trailing vorticity sheet in the numerical simulation, which represent real features of the experiment. (Abstract shortened by ProQuest.)