Analysis of the turbine steady-state data by means of Generalized Pareto Distribution

The steam turbine operational data is analyzed with a goal to derive dependencies between possible defects of the turbine and the statistical characteristics of the steady state data. In modeling of exceedances over high thresholds extreme value theory is applied and particularly Generalized Pareto Distribution (GPD) properties and peaks over threshold (POT) model. As a final result it is concluded that there is a correlation between qualitative and quantitative features in GPD and occurrences of the turbine defects.