Need help on Reliability Growth analysis

[123help]

Hello,

I am working on Reliability growth analysis on a product. We have one unit that is running under continuous tests with different test/operating modes. For eg. the initial test was run from 0 to 200 hrs and failures were recorded, then the test was continued for another 100 hrs. Total hours during this mode = 300 hrs. Then the test mode was changed and the test was run from 300 hrs to 600 hrs and failures logged, so the hours on the unit by end of test mode 2 will be 300 + 300 = 600 hrs.

I am using the Crow AMSAA model to calculate the MTBF and failure intensity. How should I identify BC and BD modes, since the Corrective actions were done on some failures but some fails remained in the next test mode. And also according to Crow AMSAA the Total test time is from 0 to T. But how will I do this if the unit is operating continuously and the test modes are changed without stopping the operation of the unit.
Please help.
Thanks,

[Tarik El-Azzouzi]

When you say different test modes, do you mean that the initial test phase’s conditions were different then conditions of the second phase? If the test conditions are different, are they intended to cause different parts of the system to fail?

‘BC’ modes are failures for which you had corrections (improvements) during the test.

For the failures you didn’t have fixes for during the test, you can treat them as ‘BD’ (delayed) modes in case you intend to implement fixes at the end of the test (if you want to consider the improvements in the system due to these delayed fixes, then you have to estimate an effectiveness factor). On the other hand if you don’t get a chance, and do not intend, to implement any delayed fixes for these failure modes, in that case you treat them as ‘A’ (ignored) modes.

For more details about the Crow-AMSAA model, please visit:
http://www.weibull.com/RelGrowthWeb/Crow-AMSAA_(N.H.P.P.).htm
http://www.weibull.com/RelGrowthWeb/Crow_Extended.htm