| Abstract [eng] |
In a relevant field of risk evaluation a probability of an extreme event is often estimated using a certain probabilistic model, based on some laws of general knowledge (i.e. physics). These models contain sets of variables and initial conditions. Due to lack of information about an extreme event some variables need to be estimated while taking data uncertainty into consideration. The classical statistics methods and Bayesian approach is mainly investigated and applied in order to obtain distributions of uncertain variables. The main goal of this work is to propose a methodology for estimation of extreme event probability considering data uncertainty. The integration of classical statistics to Bayesian approach is used for obtaining and updating variables estimates, when statistical data, analyst judgements, expert elicitations and all other sources of information are available for modelling the probability of an extreme event. The probability of plane crash on Ignalina Nuclear Power Plant is taken for experimental calculations. Plane crash for one flight kilometre frequency is the characteristic which has to be estimated taking data uncertainty into account. |
