Abstract:

Standby redundancy is a method for increasing the reliability of a system through the use of additional "backup" units. Assuming perfect failuresensing and switchover, we consider such a system when the units are identical, nonrepairable, and their lifetimes follow a twoparameter exponential distrubution. Based on unit Type II censored data, interval estimates of the system threshold parameter, mean time to failure, and reliability at time [see pdf for notation], are made through structural inference (a group theoretic approach to fiducial theory). A discussion of ncontent structural tolerance intervals for the distribution of the lifetime of the system is also given. The design of a redundant system is considered by developing an algorithm which will give the minimum number of standby units required so that, with probability at least [see pdf for notation], the system reliability at time [see pdf for notation] is at least a [see pdf for notation]. Numerical examples are given to illustrate the theory. 