Group Captain de la Plain has said that certain defects are reported formally to MoA as the Supply Department. This paper outlines briefly some of the procedures involved in dealing with these reports, particularly those aimed at improving Reliability and mentions some forthcoming changes which have the same aim in view.

In the course of each year some 25 000 of these reports are handled by RDA (Defects) and in spite of “weapon concept” philosophies, some 200 different investigating authorities are involved. Since the capacity for investigation and re-design is limited, and because the traditional system for reporting does not discriminate sufficiently between important and less important defects, this mass of reports may prevent some defects from receiving the urgent attention they deserve. The Services have agreed therefore, to be more selective in their reporting since it is not possible to remedy every defect in existing aircraft and available effort must be directed into areas where the maximum improvement in aircraft safety, availability and reliability will result.

To Enable a Design Authority to act upon User Experience, full and accurate information is required about the defect as soon as possible after it has occurred.

A skin crack in an aircraft structure, the loss of a turbine blade in a jet engine or excessively high brush wear in a generator, are all defects. The reasons for them being defects can be quite different in each case. One can therefore readily see that to solve these defects the designer needs to know not only everything about the failure, but also the environment (in the broadest possible sense) that led to the failure

First let us look at the history of our engines’ reliability. Fig. 1 shows the reliability in terms of unscheduled removals per 1000 hours and in-flight shut-downs for four Rolls-Royce civil engines since the time of their introduction into service.

It is significant that the general trend is similar and particularly, that after the initial years settling down period, the level of reliability defined in this way has stayed almost constant, despite considerable increases in overhaul life and, for the Dart, great increases in power output as later types of engine were produced.

Special reliability clauses have been introduced into a small minority of our contracts for military aircraft. This does not imply that reliability is otherwise ignored. Where operational requirements call for certain standards of reliability and where these are translated into equipment specifications or test schedules, the contract covers them in the same way as any other performance requirements. This has been particularly the case in contracts for aircraft engines.

Three of the tenets of a Contracts Manager are (a) that he should not be seen, (b) he only speaks if spoken to, and (c) he ensures that every answer he gives is capable of being interpreted in three different ways. You can imagine, then, how embarrassing it is for me to make what is supposed to be a forthright statement to an audience which I imagine contains a number of customers and potential customers.

I suppose one of the most ulcer-making things that can happen to an airline management is to see a piece of brand new and expensive aircraft machinery sitting on the tarmac unserviceable. With the increasing cost of civil aircraft, profits can only be made by keeping them in the air at a high degree of utilisation.

When I was first asked to take part in this Colloquium I declined on the grounds that there had been too many reliability conferences already, and I feared that another would only take the time of good engineers off their day to day tasks of improving the reliability of British products. Having heard some of the contributions, I think I should revise that opinion. We have heard some good material. I was prevailed upon to say a few words on the problems of military aircraft and this I will do. I shall omit almost entirely references to equipment because, as I see it, the problems on the equipment side are fairly similar whether the aircraft is a military or a civil one.

I Would like to emphasise at the outset that the title of this short paper is “Some thoughts on contracting for reliability”, and that what I say does not necessarily represent the official policy of my Company, if indeed a company can have any very firm official policy on a subject which is amenable to so many and varied methods of treatment. For this same reason I am not attempting to recommend any particular treatment of the problem, but rather to talk around some aspects of the matter, and to try to consider how some equitable incentive method of contracting to produce a defined standard of reliability might be applied.

The operational sortie of some military aircraft involves a patrol phase. Typical of these aircraft are Airborne Early Warning (AEW) aircraft, or interceptors operated in a “combat air patrol” mode. For this type of operation, the available endurance on patrol is obviously an important attribute, and in presenting the performance of an aircraft the maximum patrol time as limited by the fuel capacity of the aircraft is usually computed and quoted.

However, the overall reliability of a complex modern weapon system is such that the “mean time between failures” (MTBF) is usually not large in comparison with the fuel limited flight time. A significant proportion of flights will therefore be curtailed by failures, which necessitate a premature return to base, rather than the flight continuing to the maximum flight time.

The aeroelastic properties of modern aircraft are such that necessary calculations, of critical flutter speeds for instance, can be remarkably complicated. A flutter analyst requires an inordinate degree of skill and experience if he is to acquire a “sense of feel” in calculations. The difficulty of deciding how to adjust a critical flutter speed predictably by some structural modification is a typical outcome of this complication.

The pressure distribution on the surface of an aerofoil section at near sonic speeds is not sensitive to small changes in free-stream Mach number, and the distribution at M0=1.0 is often referred to as the sonic-range pressure distribution. Attempts to predict this form of pressure distribution have met with some success in the past, the work of Spreiter and Alksne, using an approximate solution of the non-linear equations of transonic small-disturbance theory, being of particular note. Reference 1 demonstrated the application of this theory to a family of non-lifting, sharp-nosed aerofoils and found good agreement with experiment at free-stream Mach numbers close to unity.

The purpose of this note is to present some recently acquired data showing the effects of variations in the open area of slotted liners on the pressures on a twodimensional aerofoil at sonic speed in the 36 inch X 14 inch transonic wind tunnel at the NPL. The basic slotted liners for this tunnel have eleven longitudinal slots, giving an open-area ratio (σ) of 0.091. Other slot openings are obtained by closing appropriate slots with “Sellotape.” Nine values of σ ranging from 0 to 0.091, were used in this investigation. A wall static-pressure orifice, sufficiently far upstream of the model to be unaffected by its presence (3-6 chords), was used as a reference pressure for tunnel speed. The relationship between the Mach number given by this reference pressure and the true tunnel Mach number (M0), at the model location, was obtained by careful calibration.

Engineering structures are usually designed to minimise weight or some other cost criterion, subject to specified loading and functional conditions being met. The techniques of decision theory may be applied if the cost can be described mathematically in terms of the design variables. In particular, dynamic programming may be applicable to designs for which the parameter values can be derived sequentially. Dynamic programming is here introduced to obtain minimum weight layouts for two types of cantilever structure and comparisons are made with known Michell solutions. The results suggest that other near optimum configurations could be systematically derived under conditions less restrictive than those assumed by the Michell theory.