I will take an attempt.
Area of Engineering:
This falls under the category of estimating reliability at a given confidence interval.
Scenario:
In a real life situations if we perform test for a wear out, we have data like:
Mean time to wearout, Variance of mean time to wearout, sample size, and we may want to know the time we can expect 95% of the units still functioning at 90%confidence.
Calculations:
Assuming this data following a normal distibution. (Wear out period typically follow normal).
The lower 90% confidence limit can be calculated by:
Lower Limit of Mean XL= Mean time to wearout (Xbar)-Z (standard error of mean)
Z=1.282 from table.
(standard error of mean= std.dev of mean time to wearout/squareroot of n)
Now the 95% Lower limit value for reliability is:
X at 95% = XL at 90% -(Z*std.dev of mean time to wearout);Z=1.645 from table.
Applications:
These type of estimations are very useful for setting warranty, setting spares inventory and lifecycle cost estimations.
Useful references:
Reliability Engineering Handbook - Kececioglu Vol 1 & 2.
Regards,
Govind.