Useful Life Failure Period
During the useful life failure period, t1 to t2, the failure rate remains substantially constant, and, although some failures may still arise from manufacturing weaknesses or wear-out, the majority of failures are caused by the operating stresses to which the item is subject in its particular application (e.g., temperature, electrical and environmental stresses) and occur randomly (without any time-dependent pattern). During this period, when the failure rate is considered to be constant, the negative exponential distribution describes the times to failure.
The useful life failure period is the interval of most interest from a reliability prediction standpoint because, if a rigorous reliability programme is applied throughout a project lifetime, it is assumed that:
• The majority of early life failures will normally be eliminated before an item enters service.
• An in-service maintenance policy will ensure that items are replaced before wear-out becomes a significant problem.
 |
Note that, because of these assumptions, a prediction based on the exponential distribution will, in general, represent the reliability of a ‘mature’ design whose failure rate comprises mainly stress-related failures. Where the assumptions above are not given proper consideration, predictions will be substantially optimistic.
|