Bayes analysis obedience class reliability failure data distribution
|School||Central South University|
|Course||Probability Theory and Mathematical Statistics|
|Keywords||Bayes analysis zero-failure data Doof data the rate of failure reliability|
Bayes reliability analysis is an analysis method which applies bayes method to all kinds of reliability models. When the latest technical widely used and advanced technologic constantly development, the data which we got in time curtailed test happened to be "the zero-failure data" or "data only one failure". That is to say in time-setting test, there is no product failure or only one product failure. It is a difficult problem to the reliability theory which based on failure data .So now looking for a way which can analysis data scientifically and effectively under the condition of "zero-failure data" and "data only one failure" has became a new and a very important field in reliability investigation.Weibull distribution, logarithmic normal distribution, and two-parameter exponential distribution are the distributions frequently used in reliability analysis. The thesis has used Bayes reliability method analyzing the zero-failure data which got by these reliability distributions.To zero-failure data, which obey logarithmic normal distribution, the thesis has made a discussion on E-bayes estimation on rate of failure, and on base the develops estimation of reliability parameter has also been given. To Weibull distribution, the reliability demonstration testing procedure has been discussed, that is when the life of set products obey Weibull distribution, if the reliability parameter of product is request attain a certain value, then how much test time it should have in total. When in testing the reliability of product, there will be often appear a situation of a single product has failed. It is called DOOF data. Comparing with the situation of zero-failure data, this kind of situation has its own particularity, so the thesis has taking the DOOF data into account according to the rule of life obey two-parameter exponential distribution.