SRC Forum - Message Replies
Forum: Reliability & Maintainability Questions and Answers
Topic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Posted by: Wayne
Date posted: Thu Sep 5 22:25:54 US/Eastern 2002
Subject: AMSAA valid?
experience a problem when using the AMSAA model that is my Instantaneous MTBF is higher than my test duration. For e.g i test for a duration of 1000hrs with only 2 failures that occur in the very beginning of the test. Based on this set of data i use the AMSAA Model and the Instantaneous MTBF m(t) of about 1500, which is HIGHER than my test duration. Why is it so? I thought that MTBF cannot exceed the number of test hours. Can enlighten? Is the m(t) a accurate representation of the system at the end of RDGT?
Subject: Reliability Growth
Reply Posted by: Patrick Hetherington
Date Posted: Fri Sep 6 6:34:19 US/Eastern 2002
I will assume you are applying the model correctly and it is possible that it tells you the instantaneous MTBF is greater than test time. The model reflects what is input and if you had two failures at the beginning of the test and none for a long time the model is reflecting tremendous growth, given those failures were corrected. This is most likely not the case. The AMSAA model is fitted with maximum likelihood estimators and should generally be used when 20 or more data points exist. Otherwise a modification to the shape parameter must be used. Using only two data points to make projections is unwise. Mil-Hdbk-781 will provide formulas for goodness of fit and confidence intervals. I would apply those before presenting your claims.