Weibull analysis should be applied by organization when probability of failure changes over time. In practice it’s applied in the case of issues which don’t appear directly in the production plants (e.g. during PPAP tests), but start to occur in the warranty field.
Special attention when organization should investigate of Weibull analysis preparing is information about increased problem occurrence in the warranty field
Depending on the parameters, it can assume a normal distribution as well as exponential distribution (k = 1).
The k parameter distribution determines the behavior of the probability of failure over time and it’s distinguished on three category:
– for k <1 the probability of failure decreases over time.
– for k = 1 (exponential distribution) the probability is constant. This suggests that failures are external random events.
– for k> 1 the probability increases with time – parts functionality is degradation over time
The reason why it should be used is to provide to customers risk analysis of the potential number of defective components that may still occur in the warranty field.
In practice, two factors are taken into account:
- mileage after which the final customer noticed the unproper functioning of the system
- time after which the problem was noticed by the final customer
It should be also remembered, that each analysis must be approached individually. The following information is examples of the most commonly used assumptions:
- European market daily mileage: 50 km
- Japanese market daily mileage: 30 km
It’s also possible to estimate the average daily mileage if the organization has a significant population of defective components. Thanks to this assumption, it’s possible to estimate the number of defective parts in relation to the duration of the warranty time and mileage covered by the warranty period.
The output from Weibull analysis should be evaluation of range with minimum and maximum quantity of defective units which are still founding in the field basing on which both client and organization can determine next steps. The above information is possible to identify based on the below chart.
More information with practical examples are provided on dedicated trainings.