Since 1946 I have been convinced, from the viewpoints of safety, economy and simplicity (facilitating calculations and development) that the most rational solution to nearly all aircraft propulsion problems is the Fixed Shaft Constant Speed Variable Torque Turbine.

But as the engine manufacturer is not the only judge of the propulsion systems adopted by the airframe manufacturers, I have been and still am therefore obliged to manufacture engines, the technical and commercial success of which I do not deny, while being convinced that more rational solutions exist. Because of this Turbomeca has manufactured, and continues to manufacture—alone or in co-operation with other engine manufacturers—free turbines, pure jets and twin shaft by-pass jets, engines which I do not propose to discuss today.

Many of the problems we have experienced in Europe in this field so far are attributable to the lateness and difficulties of establishing with users, what systems they would like to see in an operational role. These remarks, of course, relate particularly to the application satellites, and not to the scientific sector of the work, which has enjoyed a great degree of interest by the scientific community in calling for various experiments to be incorporated in spacecraft. As a result of this in Europe, we have seen through ESRO and through national programmes, a considerable degree of achievement in spacecraft which have been orbited, and which in nearly all cases have been very successful and have met or exceeded their design life. Through this work, therefore, there exists an extremely valuable baseline of experience, both in the establishments and the institutions concerned with specifying, planning, and managing the programmes on behalf of the scientists, and also in industry that has developed the equipment for these programmes.

The specialists called upon after the failure of a structural component to establish the true cause of the failure find themselves often in the fortunate position that they are able to tell from the fracture what went wrong. In such cases they can readily make suggestions either to improve a bad construction or to replace an unsuitable material. The consequences regarding fulure non-destructive inspection of similar components are often less readily appreciated.

The present paper deals with a number of aircraft component service failures. First, the paper discusses how inspection procedures can be improved as a consequence of service failure analysis. Secondly the paper considers the possibility that changes in existing inspection procedures may lead to unexpected defect discoveries. The third topic shows that the service failure analyst can employ standard non-destructive techniques in unusual ways with increased effectiveness. Finally it is shown that the application of inspection techniques using fluids can influence the subsequent analysis of a failure.

The reinforcement for holes in tensile stress fields sometimes takes the form of an extensive thin plate attached to the sheet, especially if the reinforced region is of integrally machined or bonded construction. Such reinforcements are often placed eccentrically, that is on one side of the sheet only. It has been shown both experimentally and analytically that such an eccentrically placed reinforcement can promote considerable bending stresses.

Wittrick gives curves showing the effect of eccentricity in the case of the circular hole with an eccentrically located annular plate reinforcement in a plane, axially symmetric stress field. High, but only local, bending stresses exist which simple theory will not predict. The stresses are markedly higher than those given with the corresponding symmetric reinforcement. For example, under axisymmetric loading it is possible to design an optimum plate-reinforced neutral hole (i.e. the stress concentration factor is 1·0 in both sheet and reinforcement) if the reinforcement is deposed symmetrically about the plane of the sheet. If however it is placed totally eccentrically then for realistic sheet stresses bending raises the sheet concentration factor 2·0.