Weibull influence of different failure modes

In a weibull analysis all the units in the analysis must have the same failure mode. If different failure modes are not suspended the weibull line fit will be poor. When there is a big difference in Beta there will be a bend viable in the graph. This bend in the curve is visible in the blue dots of the graph below. The blue dots are the combination of units of the data-set A and B. This is also visible in the correlation value r² witch is lower when the two failure modes are combined.

Therefor it is important to check the actual failure mode of each failed unit. And never combine 2 failure modes in one wiebull graph.

Different failure modes must be included in the data-set but must be suspended. More information about suspensions see here


Data

Legend

η (eta) = Characteristic life (63.2% of the samples will have failed)
β (beta) = slope of the weibull plot
n = amount of samples the units in test