Analyzing failure data with maintenance amp restoration factors

[Melissa]

I am trying to use Weibull++ to analyze failure data that was collected on in-house instruments. Unfortunately, this data is rather sparse and corrective maintenance was often required on the instruments. I've tried modeling the failures assuming that the corrective maintenance restored the components to a "like new" condition (by treating each interval between failures as a separate life observation), but I don't believe that this is accurate. Can you suggest a better way to treat this data to get a more accurate representation of the true life? I also have BlockSim which I know has the capability of putting in "restoration factors" once I know the underlying life distribution of the component. The second part of the question is how do I determine what the "restoration factor" is?

[Adam]

Another approach to your problem might be using the NHPP with a Power Law intensity function. This model assumes that the system is As-Bad-As-Old after the repair, which might be more representative in your situation. This model is available in RGA 6.

Regarding the estimation of the restoration factor, we at ReliaSoft, are currently working on a model where the restoration factor can be determined from the failure data.

Please let me know if you have any further questions.