ABSTRACT
To build new systems fulfilling increasing requirements on reliability and performance, reliability information is needed. Sources of information are specially designed, usually accelerated, reliability tests or field data from similar systems and components in use. In this chapter we address the latter situation.
From operational use we obtain failure data. However, these data have to be transformed to reliability information to be useful to the designer. Any interpretation of data must take into account some model of reality and some assumption concerning the relevance of the observed situation to a future similar situation of a new system. These two questions are to some extent addressed in this chapter.
Most systems in operation are repairable. However, most reliability analysis methods are addressed to nonrepairable systems. Simple graphical techniques to find the relevance of different modelling approaches are discussed in a forthcoming dissertation by the first author. Some results are given in this chapter.
Usually we have many sources of potential reliability information concerning a component under study. All these sources should be utilized as far 94 as possible. In a recent dissertation the third author of this chapter studied problems of taking failure data from several sources into account. Some of the ideas are presented.
