SRC Forum - Message Replies
Forum: Reliability & Maintainability Questions and Answers
Topic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Posted by: B. Kelley
Date posted: Tue Mar 23 9:39:40 US/Eastern 2004
Subject: Integrated "Metal Box" MTBF
Have an issue that I need help with.
I need to determine an MTBF on an LRU that is integrated with COTS products. Pretty much a packaged in a box scenario. All I have to work with is:
1. Vendor supplied MTBF numbers.
2. No metering. Failure data, based on returns we have seen, is limited population. (less than 1000 units per LRU.)
Have determined that this will probably not work based
results I was generating. Too many variables and
assumptions involved to determine failure rates.
I came up with a method that I believe is in the ball park,
Inverse Summation factored with parts count, but would be interested in hearing thoughts on this.
My world is electronics, and fixing what is broke. I'm not a
Statistician. Feel free to explain this as if you were talking to the dumbest individual on the planet. Not going to hurt any feelings.
Subject: Field reliability and MTBF estimation
Reply Posted by: Larry George
Organization: Problem Solving Tools
Date Posted: Fri Mar 26 17:15:39 US/Eastern 2004
It is possible to estimate field reliability (what really happens) even if you don't track products. Enter your ships and returns data into http://www.fieldreliability.com/Table.htm, and I will send back nonparametric estimates of reliability, failure rate function, and MTBF (if possible).
Your inverse factored by parts count sounds like what is proposed in "Credible Reliability Prediction," http://www.asq-rd.org/publications.htm. That may be necessary to make a credible MTBF estimate, if there's not enough failure data to estimate it directly as proposed in paragraph 1.