Reliability of Large Systems

Abstract

Modern plants and processes can involve the accumulation of unnatural concentrations of hostile substances in high energy situations. The safe storage of the ingredients and products of these processes, the controlled disposal of waste and the successful harnessing of the reaction energies are all problems which tax our technological resources. There is now sufficient evidence from recent years to demonstrate the potential of fault situations to cause vast economic losses in terms of plant damage and loss of production, to cause loss of life and injury to numbers of operators and in some cases to cause the effects of the catastrophe to extend far beyond the bounds of the factory fence. The need to objectively measure and regulate reliability has now long since been clear and has stimulated the growth of a formalised methodology of reliability assessment which finds expression in quantitative terms.

This quantitative component of assessment offers an analogue scale of performance estimation. By this means alternative design strategies can be decisively compared and the reliability spectrum of the component-sections (at any chosen hierarchical level) of a system can be derived. Also the reliability performance of a system may be assessed with respect to some quoted target criterion. Such a target criterion may be derived in reference to plant economics, plant safety or both. Thus a definitive means of gauging the adequacy of the reliability of systems is provided which is increasingly of value as the scales of our technological activities grow. The ability to regulate the reliability of plant to satisfy explicitly derived performance requirements clearly places a valuable and reassuring tool in the hands of the designer. Not only does the tool enable the correct level of reliability to be designed into a system (which is itself cost effective) but also provides a sufficient break-down of performance constituents to enable the most cost effective means of achieving this level to be found.