Previous Page

the attachment line along the leading edge and near the separation line along the trailing edge also off the wing surfaoe. A solution may be regarded as accurate only if it can be shown that, for example, the attachment surface is a continuous streamsurface and intersects the wing along a continuous attach­ment line and, further, that the neighbouring streamsurfaces are also conti­nuous and do not intersect the attachment surface. No method has as yet been tested in this way and, in this sense, no "exact" solution exists for three­dimensional wings, apart from those for non-lifting ellipsoids by H Lamb (1932) and К Maruhn (1941).

Comparing the classical approximations described above with numerically exact solutions, obtained by iteration, we find typically that the approximate loading is slightly too high near the wing tips and that, as a consequence, the overall lift slope may be too high, by nearly 4% in some cases. We may infer that the cause of this error lies mainly in the assumption made in (4.51) and not so much in the many other assumptions or in the interpolation procedure. Thus (4.71) for the chordwise loading can only be approximately true and the aerodynamic centre does not remain at the same chordwise position all along the span, as it would according to (4.72). One could say that there is a threedimensional "small-aspect-ratio effect" in the tip region, moving the loading further towards the leading edge as a consequence of a larger contri­bution to the downwash from the streamwise component of the vorticity distri­bution over the wing. On these grounds, G G Brebner et at. (1965) have pro­posed that the parameter n should depend not only on the aspect ratio but also on the spanwise station y/s. Their modified relation, replacing (4.79), improves the results.