Success-Run

[Raphael]

I'm a brazilian student. I want to know, what's "Success-run"?

[David]

Hi Raphael,

I am not sure what you are referrring to. Could you please provide some additional information, such as the context in which you might use that terminology? The one thing that comes to mind is one-shot items. In this case you test an item and it either passes or fails. An example of this is a missile. It either hits the target or it does not.

I hope this helps.

[Raphael]

I work in an automobilistic industry and I want to evaluate the reliability of production system. For this I need to use the success-run test. Sorry, because my English isn’t good.

Tank you!

[David]

OK, I am still not sure, but let me take another shot. Are you possibly referring to a zero failure test? A test where you are trying to prove a certain goal (i.e. reliability or MTBF)? The setup for this type of test can be done within the Design of Reliability Tests (DRT) in Weibull++ 7 (http://weibull.reliasoft.com). In general, the DRT uses the cumulative binomial distribution to estimate the required test units or test time to meet a specified requirement. Additional information on the DRT can be found at http://www.weibull.com/hotwire/issue24/relbasics24.htm. You could also use simulation to design the test as presented in http://www.weibull.com/hotwire/issue60/hottopics60.htm.

If this is not what you are looking for, then I would suggest that you talk to the person who is asking you to do this. They should be able to explain to you what they mean by "success-run."

[Raphael]

David, tank you! I'll talk with the person who asked me about "sucess-run".

[joshi]

Even when one considers single shot systems then reliability estimation based on success-failures can be used, however, it does not take into account ther time of operation/mission time; system/sub-system operation times including continuous systems like electronics which may be either active/standby, etc. If goes by this methodology and few development samples are availalble for testing then the question is how does one estimate reliability after say 5 tests ?

What will one do for electro-mechanical systems ? Ten tests with zero failures during development with design reviews, environmental tests, life tests etc can be used for R estimation based on probability of success. Do we then demonstrate reliability based on designer's confidence, QA/QC philosophy especially for non-repairable systems.