SRC Forum  Message Replies
Forum: Reliability & Maintainability Questions and AnswersTopic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
(src_forum@alionscience.com
)
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Original Message:
Posted by: Eric Andersen
(eric.andersen@smithsmedical.com
)
Date posted: Mon Mar 5 15:35:35 US/Eastern 2007
Subject: 95% Acceptance Rate
Message: I need to come up with pass fail criteria for system reliability. I would like to use an acceptance rate of 95%. What rationale would you use that 95% is adequate and correct?
Reply:
Subject: Reliability Acceptance
Reply Posted by: J. L. Romeu
Organization: System Reliability Center
Date Posted: Fri Mar 9 11:13:11 US/Eastern 2007
Message: We will start with some definitions:
Reliability (x) = P { X > x } where “X” is the Life of the system and “x” a design value.
This means that, given the distribution of Life X (and of course, its parameters), then the device or system whose Life is X will have a probability R(x) of surviving “mission time x”, under normal operating and environmental conditions. This implicitly implies having a generally accepted, quantitative definition of “mission time” and “normal” conditions.
Hence, the first justification that “95% of the times (meaning nineteen out of twenty times on the long run average) the device survives its mission time, or fulfills its intended use” constitutes an “adequate and correct reliability criterion” is that such a quantitative assessment is really an adequate number. For example, assume that we are talking about children vaccines, intended to immunize kids for a full year, 95% of the times they are applied. This implies that, at most 5% of the times, they may leave a child exposed to illness for less than a year. If such a vaccine is going to be applied to the United States infant population, then 95% is an inadequately low percentage. For, not too many users (i.e. mothers) will agree that one out of twenty children vaccinated, may acquire the disease it is being vaccinated against, and possibly suffer serious consequences.
Another important justification for the establishment of reliability of 95% may be that the parameters of the Life distribution yielding this value are achieved within cost and other problem constraints. For example, assume a device Life is distributed Exponential, with Mean Time to Failure (MTTF) of 10,000 hours. Also assume that such MTTF can only be extended under difficult and costly efforts, and that the device must fulfill a mission time of 500 hours. The probability of survival of such a device, for the mission time of 500 hours, is a barely higher than 95%. If the device user is completely satisfied with such a percent survival, then this criterion can be considered “adequate and correct” for the establishment of device reliability, given cost and other peripheral circumstances.
Summarizing: two of the most important criteria for establishing a 95% (or any other) reliability level are: (1) that such a level satisfies the expectations and needs of the device user, and (2) that the parameters required from the device Life to achieve such a level can be reasonably attained, within cost and other problem considerations.
For further information, see System Reliability Center (SRC) START Sheet ‘Reliability Estimations for Exponential Life” //src.alionscience.com/pdf/R_EXP.pdf
