Premium
Reliability acceptance testing for new series systems
Author(s) -
Blumenthal Saul
Publication year - 1992
Publication title -
naval research logistics (nrl)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.665
H-Index - 68
eISSN - 1520-6750
pISSN - 0894-069X
DOI - 10.1002/1520-6750(199206)39:4<579::aid-nav3220390410>3.0.co;2-k
Subject(s) - mean time between failures , reliability (semiconductor) , reliability engineering , series (stratigraphy) , exponential distribution , acceptance testing , poisson distribution , computer science , statistics , test (biology) , exponential function , test plan , failure rate , mathematics , weibull distribution , engineering , paleontology , mathematical analysis , power (physics) , physics , software engineering , quantum mechanics , biology
The mean time between failures (MTBF) of a newly minted series system, all of whose components exhibit wear, will tend to be much larger than the MTBF of the same system after it has become fully aged. When fully aged systems are used for the testing, acceptance tests with a criterion regarding the MTBF of a well‐aged system can be based on the assumption that times between system failures are independent, with identical exponential distributions. However, these tests are shown to offer essentially no consumer protection when applied to new systems. Tests are derived which are correct when new systems are under test but the acceptance criterion refers to the MTBF of a well‐aged system. The derivation uses an approximate Poisson distribution which is valid if the total number of systems on test is sufficiently large.