times-to-failures

[hustlfm]

I am confused about the term 'times-to-failures' in weibull++7. That is to say, I want to know what's exact mean of this term: is a cumulative time when units are failed or the time between failures?
for example, i have a data set about a machine: F at 200, F at 340, F at 580, F at 600…
i want to know the time of the second failure: is 340 or 200+340=540?

[Pantelis]

The quick answer is that the time is Not Cumulative – if you are in a Weibull++ Life Data Analysis Folio (i.e. Weibull Analysis Folio).
Now in the example that you mention, with a single machine, one needs to be careful because Weibull analysis assumes that each time to failure is independent of the last one, which may or may not be valid for a single machine experiencing sequential failures. As a matter of fact the only time that would be valid is if every repair action restored the machine to as good as new condition. If this is not the case alternative analysis methods should be used. Look at the Recurrent Event Data Analysis module in Weibull++, as this may be more applicable. In this module the times to failure are entered in cumulative fashion.
See also (for recurrent event data analysis)
http://www.weibull.com/LifeDataWeb/recurrent_events_data_analysis.htm (http://www.weibull.com/LifeDataWeb/recurrent_events_data_analysis.htm)
and for an example see
http://www.reliasoft.com/Weibull/examples/w7ex9/index.htm (http://www.reliasoft.com/Weibull/examples/w7ex9/index.htm)

[hustlfm]

Hi Pantelis
Many thanks for your post.
From your post ,it is true when i am in a Weibull++ Life Data Analysis Folio (i.e. Weibull Analysis Folio),so-called times-to-failures means the times between failures?
in addition, RDA analysis is used in single repairable system.i want to know multi-repairable systems can be analyzed by RDA?
finally, there is some method to recognize the exact meaning of "times-to-failures" in weibull++7 and RGA 6:either cumulative time or time between failures?

[Pantelis]

Ok… Just to make sure that we are both on the same page.

When doing Weibull Analysis the times-to-failure are iid (independent and identically distributed). In other words if I am testing 10 identical light bulbs, each one of the times to failure for a light bulb would be iid, and this is the data I would use. Now if I am looking at a single car, and I have times to failure for a component in that car (say the water pump) then each time between the water pump failures would be my times-to-failure assuming that I replaced the old pump with a new one.

Now RDA analysis looks at the whole system without regards to the component being replaced and we are looking at times between events. Note the same holds true for repairable system analysis in RGA 6. In both cases you can analyze multiple systems.

Hope this clarifies and helps.