ABSTRACT
Maintenance concept, especially when related to the context of power grid, is of great importance. As time is passing by the reliability requirements for systems are increasing, while the maintenance cost are not getting any cheaper (IEEE/PES Task Force, 1999). However, the common practice is to use time independent mathematical model for maintenance and parameter values are just point estimates obtained from the historical data-and it doesn’t really matter what kind problem we might consider. Whether it is a power grid maintenance/inspection optimization task, or reliability predictions under the influence of maintenance/ inspection rates, the “tradition” is the same-obtain point estimates and don’t bother oneself by the data any more. In addition, uncertainty of data is almost never taken into consideration, or event when it is-not as the central part of the whole picture. We demonstrate in this paper that Bayesian framework is applicable to propagate data uncertainty through all the maintenance analysis aspects.
