Search

The errors arising from the computational representation of a helical vortex by straight segments, particularly at the small values of pitch that are typical of the wakes of hovering rotors and wind turbines, are considered. For a helix of constant pitch and radius, as in the far wake, the main error in calculating the velocity at a control point of the same radius arises from the pairs of 'aligned segments,' whose junction is in line with the control point. The aligned segments are those closest to the control point; for each M segments per revolution of the helix, there is one pair of aligned segments. They are particularly important at small pitch because of their proximity to the control point. An approximate analytic expression is derived for the velocity due to the section of the helix represented by the aligned segments. Its use is shown to improve significantly the accuracy of the induced velocity for the moderate values of M that are used in many calculations. For a vortex of changing radius, typical of the near wake, the aligned segments are also crucial. Replacing them by an extension of the analytic approximation also leads to a significant improvement in the accuracy of the induced velocity at moderate values of M. Calculations of the expanding wake of a three-bladed wind turbine are used to compare the improvement in accuracy at moderate M to that obtained by increasing M without replacement. Replacement caused a much smaller increase in execution time per iteration.

The accuracy of representing a helical vortex, as found in the wakes of helicopters, wind turbines, and propellers, by a sequence of straight segments is considered. The accuracy is assessed by comparison with recent results for the induced velocity of a helix of constant pitch and radius. This comparison is motivated by the small values of the vortex pitch behind wind turbines and hovering rotors; small pitch leads to errors associated with the proximity of subsequent turns of the helix to the control point at which the velocity is required. Three cases are considered. The first, the velocity on the helix axis, has an analytic solution, which is used to demonstrate that the general accuracy of the straight segment approximation is second order, as has been found in previous comparisons with the velocity field of a vortex ring. For the second case, where the control point has the same radius as the vortex, the segments aligned with the control point are mainly responsible for the error. The error varies from first to third order as the number of segments per revolution of the helix is increased. Third, the self-induced velocity is determined to within an accuracy comparable with the effects of the vortex structure, of which little is presently known in general. The effects of vortex curvature are not significant and easily dealt with.

The aerodynamic flow about a wing-body junction is studied through experimental procedures. Tests are carried out in a wind tunnel wherein an unswept aircraft section containing the wing-body junction is subjected to simulated turbulent flow. The flow's mean velocity are measured by means of a five-hole pressure probe. The results are in very good agreement with the data obtained using an eddy viscosity turbulence model.