Requirements for Reliability Design
The minimum data requirements for DOE reliability designs are as follows:
- For the Weibull and lognormal distributions, the number of unique failures must be greater than the number of parameters in the model.
- For the exponential distribution, the number of unique failures must be greater than or equal to the number of parameters in the model.
Depending on the type of design you are working with, these requirements affect calculations in different ways.
- For two level factorial and Plackett-Burman, if the
data set does not meet the requirement, the software will
automatically reduce the number of parameters in the model
until it can be calculated. Therefore, the following requirements
apply.
- To use the Weibull or lognormal distribution, you need at least two failures.
- To use the exponential distribution, you need at least one failure.
- In one factor reliability designs, you must have more
than one factor level regardless of the distribution chosen.
Therefore, the following requirements apply.
- To use the Weibull or lognormal distribution, you need at least one failure at each level and at least two different failures for at least one level.
- To use the exponential distribution, you need to have at least one failure at each level.