Progress in international best estimate plus uncertainty analysis methodologies

Abstract

In 1988, following decades of nuclear safety research (USNRC, 1988a), the United States Nuclear Regulatory Commission (USNRC) in 1988 amended its regulations (USNRC, 1988b) to allow the use of best-estimate evaluation models for the analyses of the loss-of-coolant accident (LOCA) in a light water reactors (LWRs). This was followed by similar action by other national nuclear safety authorities. Until this time, LOCA evaluation models applied a strict set of deterministic rules (see (USNRC, 2000)) that specified conservative models and assumptions to be applied simultaneously. This placed implicit restrictions on plant operations, limiting the nuclear industry’s ability to innovate and apply experience gained for the benefit of their customers. The rule change presented the opportunity for a variety of benefits, from simple procedural and operational flexibility to more significant power uprates.In response to this regulatory paradigm shift, the nuclear power industry has since advanced the nuclear-thermal-hydraulic systems analysis capability for a broad spectrum of technical issue resolution applications. Today these industry stakeholders include the major nuclear fuel and plant suppliers (e.g., Westinghouse, GEH, Framatome, MHI, etc.), their utility customers, the advanced and small modular reactor (SMR) design organizations, safety and regulatory authorities, industry advocate organizations such the Electrical Power Research Institute (EPRI) and the Nuclear Energy Agency (NEA) of the European Organization for Economic Co-operation and Development (OECD), and dedicated research institutions such as the U.S. DOE and the academic community. To preserve their position on the leading edge, these stakeholders have maintained continuous modeling and simulation development efforts to serve their customers and sponsors.Of particular note is the ongoing activity in new plant design reviews: both plant suppliers and national safety and regulatory authorities are challenging their analysts to demonstrate greater fidelity and reliability in code prediction of realistic design and safety margins. Research pertaining to this objective has focused on the expanded use of best-estimate-plus-uncertainty (BEPU) analysis methodologies (i.e., beyond LOCA), modern numerical methods, multi-physics/multi-scale component modeling (including computational fluid dynamics), and physics-based risk models. This paper reviews the emergence of BEPU analysis methods, their applications, their adaptations among nuclear states, and the insights revealed from the past two decades of use in the nuclear industry.