I was presented with a problem that I am not sure how to go about solving. If a product is to be used 600 times in a year, what would be the amount of tests I need to do without fail that would suggest with 90% confidence that this product would not fail in that year? Any help would be greatly appreciated.
Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
-
Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Confused about Confidence 1
- Thread starter Aaron_12
- Start date
-
1
- #2
EmmanuelTop
Chemical
You are talking about a product with failure rate (failure on demand) less than 1/600 = less than 0.001667 or < 1.67 E-03
This is equivalent to SIL (Safety Integrity Level) of 2, which is quite common for many industrial devices.
Depending on the product configuration (redundancy, voting system) and hardware fault tolerance you will come up with the required frequency of full proof tests to maintain the required SIL level.
For more information refer to IEC 61508 and IEC 61511. See also
Dejan IVANOVIC
Process Engineer, MSChE
This is equivalent to SIL (Safety Integrity Level) of 2, which is quite common for many industrial devices.
Depending on the product configuration (redundancy, voting system) and hardware fault tolerance you will come up with the required frequency of full proof tests to maintain the required SIL level.
For more information refer to IEC 61508 and IEC 61511. See also
Dejan IVANOVIC
Process Engineer, MSChE
I'm not following how you're able to test for future non-failure, unless there's some mechanical indication that is a prelude to actual failure. Moreover, if you are that concerned about failure, testing the system consumes life, so you need to factor in whether you are forcing the system into earlier failure.
Do you know what the MTBF is supposed to be? Usually, the innate failure rate can be used in the probability integral to predict with some level of confidence what the probability of failure is within a certain period of time, but it's a very coarse estimate. Nevertheless, unless you are talking about wearout or fatigue, classical failures are RANDOM, and are not predictable.
TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list
Do you know what the MTBF is supposed to be? Usually, the innate failure rate can be used in the probability integral to predict with some level of confidence what the probability of failure is within a certain period of time, but it's a very coarse estimate. Nevertheless, unless you are talking about wearout or fatigue, classical failures are RANDOM, and are not predictable.
TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list
- Status
Similar threads
- Locked
- Helpful tip
