Abstract :
However careful the initial design, advanced electronic equipment used in rigorous environments usually needs a reliability growth programme (R.G.P.), consisting of prolonged functional operation during development, to enable it to reach its potential reliability. Such a programme highlights those types of unforeseen design failures which will not show up except by working the equipment over a period. Their appearance is further accelerated by environmental stress, but this should not exceed design limits. Provided that there is rapid and efficient reporting of defects followed by modification, the result will be a general reliability growth, reflected in increased reliability of production models in the field. To assess progress in reliability achievement during the test programme, mathematical models are employed. In this paper two such models are described and compared, first the well-known Duane model but with some recent developments in estimation, and secondly a useful alternative model newly developed within the UK. Other models are referenced. An alternative simplified approach to modelling is given and the limitations of these techniques are pointed out.
