ABSTRACT
A new efficient uncertainty importance ranking method is proposed for comprehensive TH code uncertainty assessment. The proposed methodology is a hybrid two stage qualitative/quantita-tive method. The details for the qualitative uncertainty importance assessment in TH applications (so called modified PIRT), is well described in the previous papers by the authors [1–5]. This paper’s focus is to present a quantitative methodology which utilizes the results of the qualitative assessment for more precise evaluation of the important phenomena. Given the computational complexities of the TH codes, an efficient uncertainty importance measure is introduced that is defined in multiples of coefficient of variation (×ρ) changes in a given input parameter or variable over the resulting changes in the standard deviation of the figure of merit. Different levels of input change (multiples of coefficient of variation) are devised here for accurate ranking of uncertainty contributors. Comparing the output change as a fraction of the overall uncertainty range will result in a ranking index to show the contribution of each uncertainty source. In this paper a brief overview of the importance analysis is first given. Current methodologies for uncertainty importance are discussed and their applicability to TH analysis is elaborated. A description of the proposed methodology, along with an example of its application to LOFT-LB1 uncertain parameters will be discussed.
