Sammendrag
This project memo describes and illustrates methods for handling uncertainty related to extraordinary or high-impact low-probability (HILP) events in power systems. It documents some of the work carried out in the project “Analysis of extraordinary events in power systems” (the HILP project), within the work package on the handling of reliability data and uncertainties (WP3). The work is aligned with a case study in collaboration with Statnett related to their system development plans, and a simple motivating example is introduced and used for illustration throughout the project memo: The estimation of the risk associated with a hypothetical HILP event in terms of the associated expected annual interruption cost. This quantity is associated with several and severe uncertainties. Some uncertainties depend on the uncertainty in input parameters, especially in the estimated frequency of occurrence of the interruption event. Contributions to input parameter uncertainty can be characterized as aleatory (i.e. stochastic) or epistemic (i.e. due to a lack of knowledge), and these uncertainties can be represented e.g. by probability distributions. The project memo illustrates alternative methods to provide additional insight compared to analysing the HILP event without considering uncertainties explicitly or only considering expected values. The methods include sensitivity analysis methods (standard “local” methods and more sophisticated “global” methods), and uncertainty analysis methods (including probabilistic and “possibilistic”) for representing uncertainty in input parameters and quantifying the resulting uncertainty in output parameters.