A unified account is presented of the various inviscid models used to represent three-dimensional vortex flows in aerodynamics; essentially those relying on vortex sheets and linevortices. Recent developments in extending the scope, accuracy, and stability of these models are described. An evaluation of their relative strengths and weaknesses suggests that the different models all have continuing roles to play. It is claimed that vortex modelling has come of age, in the sense that we can now learn about the real world from the behaviour of models, after decades of trying to make the models conform with reality.

Wind tunnel tests have been made on the effect of bottom wall and sidewall fences of various sizes and combinations in reducing the swirl of an S-duct at high angle of attack. It is shown that large reductions can be made: in some cases the swirl is eliminated or even reversed. Fences of modest size can be chosen to combine a large swirl reduction with an improvement of the total pressure distortion at high incidence and only a small effect on mean pressure recovery at low incidence. Some further observations are made on the nature of the general swirl phenomenon.

The ‘inverse method’ proposed in 1936 by Duncan, Collar and Lyon for the calculation of flutter boundaries is re-stated in a form suited to modern methods of dynamical and numerical analysis and programmed for use on a home computer. The method is extended to provide information on sub- and supercritical eigenvalues. For a microcomputer with 32K RAM, flutter problems of orders up to 35 may be accurately solved.

Vorticity panel methods are applied to solve the interacting flow between an outer stream and a two-dimensional jet exhausting into that stream; both are assumed to be at low speed. The cases include a jet emerging from a thick afterbody and an inclined jet. The work is intended as a prelude to obtaining initial values for calculating the mixing of a jet with an outer stream.