Availability demonstration with confidence level based on reliability and maintainability

Abstract

After a first failure, with appropriate measures, such as repair or maintenance, a technical product or system can be put into the function state again. Thus, usually the service life of the system is not completed. For the evaluation of such a repairable system, the availability is an important parameter. Typically, if e.g. the reliability of a system or component is demonstrated, a confidence level is considered, especially if the demonstration is based on limited information. Consequently, the availability demonstration, based on the same information needs to be expressed including a confidence level, too. In this paper a new procedure for the availability demonstration with confidence level is presented. The procedure is based on the pure samples of failure and repair times. It allows individual samples sizes, i.e. they do not have to be equal. The procedure does not require special distribution types of failure and repair behavior. It enables the demonstration of the time-dependent and average availability. Furthermore, the procedure is illustrated by a case study of a repairable system with known samples of failure and repair times. The influence of the samples sizes is investigated. Finally, the potential of the new approach to be applied to more general scenarios is shown based on an example, too.