Learning from field data - improving reliability in a targeted manner
The statistical analysis of field failures allows the description of the real failure behavior of products in use. It is a central component for determining the field-test correlation and thus forms the basis for well-founded statements on product reliability.
The aim of the analysis of field failures is the mathematical description of the actual failure behavior in the field. The information available is often incomplete and in censored form.
Meaning of censored data
When evaluating censored data sets, it is crucial to take the unknown service life characteristics into account. If these are ignored, the real product behavior can only be inadequately described - with the result that successor products are incorrectly designed or inappropriately tested.
Companies that have access to components that have failed in the field or can track failures using order data benefit in particular from systematic field data analysis. It provides valuable insights for optimizing the reliability assurance of future products.
Methods for evaluating censored field data
Various statistical methods are available for the evaluation, which differ depending on the data situation. Most field events can be classified into the following three categories:
- Censored data with known lifetime characteristics of defective and intact units
- Censored data with known lifetime characteristics of defective units and known run distance distribution of intact units
- Censored data with known lifetime characteristics during the warranty period and unknown lifetime characteristics after the end of the warranty period
Strategic benefit
Field data analysis is becoming increasingly important. More and more companies are beginning to establish targeted strategies for using field data. A central concept is to collect reliability-relevant operating data in order to create real load spectra. This makes it possible to set up service life models or improve existing models.
Our services
- Statistical analysis of field failure data from service and complaint data
- Derivation of mileage distributions from sales and service data
- Development of concepts for determining field data beyond the warranty period
In combination with other methods
Field data analysis is closely linked to warranty management and the creation of service life models. Together, these methods make it possible to precisely evaluate field results and derive targeted measures for development and validation.
