Confidence intervals on the reliability of repairable systems

The author discusses confidence intervals procedures on the failure intensity and reliability function appropriate when failure data are generated by multiple systems. He assumes that one is interested in placing confidence intervals on the failure intensity function at a time T and has K systems which have operated from time 0 to at least time T. The failure data for each system over (0,T) is used to calculate the confidence intervals by the techniques presented. The actual operating times for these systems may, however, be greater than T. These confidence interval procedures utilize the methods developed for the author's (AMSAA) reliability growth model (1977). Examples illustrating the application of these methods to burn-in, useful life, and wearout are discussed.<<ETX>>