MTBF distribution over repeated lifetimes

[Vlad]

Hello

Is it correct to say that the reliability function is a pdf built over an infinite number of samples (population), i.e. over infinitely repeating trials, each of system lifetime length? If we want to determine the probability of achieving a given MTBF within a particular lifetime, do the lifetimes' MTBFs follow a normal distribution (for a sufficiently large sample size, i.e. larger lifetime), and can we use the Central Limit Theorem to determine the level of confidence in achieving a desired MTBF?

thanks,
Vlad

[Pantelis]

Vlad,

The reliability function is (1-cdf), computed from a finite number of samples and then utilized to make inferences on the population.

Now, and assuming a sufficiently large sample size, then using the CLT, one can say that the distribution of the means can be approximated with a normal distribution, and use that to construct confidence interval around the mean. Now this does not mean that the population itself is normally distributed. In other words the means have to be computed using the assumed underlying distribution.

[Pantelis]

Just as a follow up I just noticed that you said - "... large sample sze, i.e. larger lifetime),". I fail to see what the two have in common.

[Manjukiran]

Can you please provide me with link where I can get online free information on Introduction to Reliability and Maintainability

[RS Support]

Check out weibull.com