Pass-Fail Series-System Reliability Proximate Lower Confidence Limits

It is common of Pass-fail component test data in the projects. Reliability of pass-fail series systems has reference to such systems as exponential, normal, weibull. This paper estimates to some of common methods obtaining approximate lower confidence for reliability of series systems from independently binomial (pass-fail) component test data.To such methods as ML, MML, LM, SR, MMLSR ,we converte binomial (pass-fail)components test data (n_j, s_j)j = 1,m into one component pseudo binomial(pass-fail) data (n, s), and convert it into pseudo exponential tested data with type I censoring, (n, $) The approximate lower confidence for reliability ofseries systems is derived. Approximate lower confidence for reliability of series systems is obtained by using the second general Bayes formula derived by Zhou Yuanquan. In the last,, a large sample is abstracted to simulate the process with Monte-Carlo technique. Every series is tried for 10000 times, the programmes are compiled with Visual C++ 6. 0. It is presented that the MML method is bold by analysising the result simulated when data meet the specification. And the MMLSR metheod is bold also.The SR mothed is verified to be the least conservative among the conservatively approximate methods after analysising. So the SR method is recommended.