Weibull & predicatable variation

[JeffMathena]

Hello,

I have a naive question about Weibull analysis concerning small sample sizes and predictable variation in parts and assembly processes that effect the result you're looking for, like reliability / life.

From my perspective if you only put somewhere on the order of 5 - 10 units in test, you are not going to have a representative sample of the product due to all the predictable variation that will naturally occur to the parts and assembly process over time. How do you take this into account using Weibull as a predictive process?

Thanks for any and all help,

Jeff

 

[BigInch]

You must first determine the minimum number of test samples you have to select from a group in order to assure your desired confidence level will be meet in the predictions that you will make using Weibull analysis methods. After you have done that, then apply your Weibull analysis procedures to determine probability of service to design life.

We will design everything from now on using only S.I. units ... except for the pipe diameter. Unk. British engineer