I must start by thanking the Royal Aeronautical Society for the invitation to present this 27th Lanchester Memorial Lecture. It is an honour and a privilege to follow in the footsteps of the distinguished scientists and engineers who have given the first 26 lectures in this series. These lectures have included many outstanding reviews of a wide range of different topics and I am very conscious that they have set a standard that I, for my part, will find difficult to match. I hope, however, that by choosing a topic that has only been mentioned in passing in just a few of the previous lectures, I will be able to make my own distinct, individual contribution to this tribute to the memory of a man who was not only a great scientist and engineer and talented musician but who, by his writings as long ago as 1907, still carries a message for us today in 1987.

Three low-order panel methods developed for the analysis of supersonic flows have been used to make calculations for a number of test configurations, with the aim of establishing the range of problems for which this class of computational method yields adequate solutions. The programs investigated were the Woodward USSAERO/C program, the related NLRAERO program, and the British Aerospace Warton supersonic panel program. Results obtained using these programs have been evaluated against theoretical and experimental data for a number of test cases covering wing-alone, body-alone and wing-body geometries. It is concluded that low-order panel methods can provide adequate solutions for supersonic flows about wings, bodies and wing-body combinations, provided the assumptions implicit in the linearised potential flow model are not violated. Examples show that the prediction of lift and pitching moment curve slopes for quite complex configurations may be acceptable, but that the detailed pressure distributions are not always predicted satisfactorily. In particular, serious problems are encountered in calculating the flow about wings with rounded supersonic leading edges due to the linearised flow model which is used.

In the Queen Mary College gust-tunnels, unsteady flows are generated by oscillating flaps attached to the downstream upper and lower surfaces of the contraction nozzle of a semi-open test section, open-return, low-speed wind-tunnel. The flow perturbations produced on the mainstream of velocity U∞, are of the travelling-wave type, with wave-velocity Q. Attention is drawn to the contradictory early measurements of Q/U∞. New data are presented which show Q/U∞ apparently diminishing along the tunnel axis from a high value near the nozzle exit to an asymptotic value of about 0·6 far downstream. Using a digital phase meter especially developed for the purpose it is shown that the explanation for this behaviour lies in the two-dimensional nature of the wave process in the region of the flaps.

A planar vortex sheet model for analysis of thin wings in inviscid incompressible flow is presented. In this model a network of spanwise quadratically varying semi-infinite doublet sheets is introduced which produces a continuous trailing vortex wake. The present method has been applied to wings fitted with partial span trailing edge flaps after appropriate modification to account for the flap juncture. Also, the problem of wings in sideslip is attempted by incorporating the zero load condition at the down stream wing tip. The comparison of results shows that the proposed model retains the simplicity of the vortex lattice model to a large extent while overcoming the limitations of the standard vortex lattice model regarding lattice arrangement and number.