Estimation of Sojourn Time Distribution Parameters Based on Expert Opinion and Condition Monitoring Data

This paper presents an Bayesian approach for the estimation of sojourn time distribution parameters. Expert opinion and condition monitoring data can be used as source of information for the estimation. It is shown how expert opinion and data can be combined and how estimates for the parameters can be calculated by means of numerical methods. It is also discussed how expert uncertainty and systematic expert errors can be handled. Furthermore, procedures are presented for utilizing censored data sets, which are available as discrete time-series from a condition monitoring program. Finally, several examples are presented showing the practical application and the use of the approach.