Confidence intervals for steady state availability of a system with exponential operating time and lognormal repair time

Long-run availability of a system assuming lognormal repair times and exponential failure times is considered. With many repairable systems, the time to repair is well characterized by a lognormal distribution. However, due to the difficulties associated with the lognormal distribution, repair times are often modeled using the lognormal distribution assuming that the variance of the lognormal distribution is known or by using some other distribution such as the exponential distribution instead of the lognormal distribution. In this paper using the generalized p-value approach, we propose confidence intervals and exact tests for steady state availability of systems using the two-parameter lognormal distribution for repair time and exponential distribution for operating time. Here, both parameters of the lognormal distribution are assumed to be unknown. A couple of examples are given to illustrate the proposed procedures. A simulation study is given to demonstrate the performance of the proposed procedure. Results are extended for the long-run availability of a system consisting of several independent parallel sub-systems. (C) 2002 Published by Elsevier Science Inc.