D. Mortin, J. Krolewski, and M. Cushing
The assumption of the constant failure rate can lead to excessive costs
and non-optimum design decisions. As this paper shows, simply summing
constant failure rates can produce results which are highly inaccurate.
Highly inaccurate results can introduce significant error in decisions
made for everything from product design to logistics support requirements
such as spares and maintainers. If we shift from reliability engineering
analysis, the ability to address hazard rates versus time based on root-cause
failure mechanisms will become cost-effective and can become an integral
part of the concurrent engineering approach to product development.
The notion of the constant failure rate should no longer be accepted as
a rule. In stead, statistical distributions and assumptions must
be shown to be appropriate every time they are used. Simplicity alone
is not a sufficient reason to use any given methodology or approach.
article in .pdf format (Adobe Acrobat required) is available to CALCE