Historical insights in the development of Best Estimate Plus Uncertainty safety analysis

Abstract

Requirements related to licensing and regulation submittals for US commercial nuclear reactors were originally based on what was believed to be the most problematic accident scenarios, conservative safety limits and prescribed analysis methods. However, by the 1980s the large database of experimental and analytical research, and actual reactor operation indicated the initial strongly conservative approach was somewhat misdirected and sometimes even gave non-conservative results. Accordingly, an increasing motivation existed to develop a safety analysis strategy that reflected risk based licensing principals. The first successful advancement of such an approach was embedded in the Code Scaling Applicability and Uncertainty, and Phenomena Identification and Ranking Table methodologies developed by the Research, and accepted by Regulation arms of the USNRC. This strategy is commonly termed the Best Estimate Plus Uncertainty methodology. A historical perspective of the original Code Scaling Applicability and Uncertainty, and Phenomena Identification and Ranking Table processes is provided, as is description of the subsequent improvements through 2010 in Best Estimate Plus Uncertainty methodologies.