RG for tests with multi system multiconfiguration and mode classification

[CE]

In my prototype (developmental) stage, I have 3 different configurations (builds). At each new configuration, 3-4 new systems (which would have the fixes from previous configuration, everything in the system is brand new again) are put to test. During the test, I record the time of the observed failure modes, and fix them if possible or assign an effectiveness factors, if the fixes are to be implemented later.

I don't seem to find a data type that incorporates all these issues:

failure times (could be interval data)
multiple units
multiple failure modes
different configurations

I tried the times-to-failure data sheet, but that one doesn’t allow for different configurations! The success/failure data sheet allows for different configuration, but this is not the type of data I have!

Thank you

[adam]

First, you would need to use the Multiple Systems with Unknown Equivalent Times data type. Then select the Crow-Extended model. This will allow you to enter data from multiple systems with multiple failure modes, for a Single configuration. Remember that the Crow models typically apply to a single configuration (or single phase/stage).

For different configurations you would have to use Reliability data. After you analyze each configuration separately, you can then calculate a reliability achieved at the end of each configuration (assuming exponential distribution, the reliability is exp(-lambda*t)). This will then give in your case 3 reliability values. You can then take these 3 values and enter them into the Reliability data type in RG, and use one of the available models to track the reliability improvement over each configuration.