By inviting me to give the 25th Henson and Stringfellow Memorial Lecture the Committee of the Yeovil Branch and the Lectures Committee of the Main Society have done me an honour which I deeply appreciate and I hope that what I have to say may be sufficiently novel and interesting to do justice to this important occasion.

My subject is the story of the mild obsession which first aroused my interest in rotating wings and which has guided my steps at critical times over the past 50 years— a benign obsession in that for most of that time there was nothing that could be done about it so that it has not hampered the engineering work needed to implement the policies of my various employers. It was a matter of great good fortune for me that at just the right time some of the concepts underlying this lecture had hardened sufficiently for Westland to agree to my concentrating on this study during the last few years of my career.

High authority automatic flight control systems are now in wide use in both transport and combat aircraft. These have required the development of failure survival system architectures, and techniques in safety assessment to ensure that the risk of fatal accidents attributable to the automatic control aspect is acceptably remote.

Most of the current systems are analogue in nature, and employ sufficient redundancy to survive at least one failure during operation. Future systems will employ quadruplex or triplex-monitored architectures, capable of surviving at least two failures, and will be mainly digital in concept. The design and safety assessment techniques developed for analogue systems have been adequate for current systems, but the new digital technology and more extensive use of systems in safety critical applications will demand further development of the various aspects involved.

A major contribution to the design and safety assessment of safety critical systems was made during the past two decades in the programmes for the development of automatic landing for civil aircraft, culminating in the system in the Anglo-French Concorde.

Multiple redundant flight control systems have been the subject of continuous development since the 1950s. Early examples include the automatic landing systems of the BAC 1–11 and the VC10; these were followed by the full flight envelope control systems fitted to the Concorde and Tornado. These examples all use analogue computing techniques, with redundancy levels appropriate to the individual fail-passive or fail-operational requirements of each system.

However, recent advances in digital technology have enabled highly sophisticated flight control systems to be implemented which would be wholly impractical using analogue computing methods. Furthermore, digital computing is amenable to self-monitoring techniques which can be used not only to improve system maintainability, but also to augment the cross comparison or majority voting algorithms which have been developed on analogue systems.

The behaviour of a thin tensioned sheet with an opening has been the subject of a number of recent investigations. In this type of problem, as in an aircraft structure, a skin carries all or some of the in-plane load as well as a lateral pressure load. Failure often results from the development of fatigue cracks which propagate from a stress concentration at the cut-out being subjected to structural vibration excited by acoustic pressure fluctuations. The Comet disaster in 1954 was due to the structural failure of the pressure cabin, brought about by the development of fatigue cracks from high stress concentration regions around the corners of the windows. The cases of statically tensioned, uniaxially and biaxi-ally loaded plates having a central crack subjected to acoustic excitation have been studied. These investigations were concerned with the basic features of plate vibration; i.e., mode shapes at a resonance frequency, the resonance frequency versus load behaviour and the crack propagation rates.