Conference Agenda

System-scale TH (STH) analysis codes have extensively been used in WCNR safety evaluation along with quantifying the prediction uncertainties in close conjunction with adopting the best estimate (BE) safety analysis. There still exist some deficiencies in the BE safety evaluation, however, originating mainly from our limited knowledge or poor understanding of underlying fundamental physics on key TH phenomena associated with two-phase flow hydrodynamics and heat transfer, which are broadly relevant to WCNR safety concerns.

TH experiments and analyses for WCNR performance analysis and safety evaluation, in general, need to be carefully checked in terms of their scalability to assure whether they are realistically representative of prototypic situations. The basic concern of ‘scalability’ originates from the differences or gap existing between the prototypic and down-scaled systems due to their idealization and/or simplification.

The scalability of experimental data used for validating TH analysis codes will be discussed, focusing on STH codes with their application to WCNR safety evaluation accompanied by the uncertainty quantification.

Discussion will be focused mainly on our unsatisfactory understanding of fundamental physics associated with the constitutive relations adopted in STH codes, many of which were developed based on unrealistic observation under non-prototypic geometric and TH conditions, and partly on the limited numerical capabilities of STH codes in describing multi-dimensional features of dominant phenomena. Then the perspectives of advanced TH safety evaluation are introduced aiming at improving the modelling and simulation (M&S) capabilities.

Results of a Best-Estimate Plus Uncertainty analysis of the LOFT L2-5 experiment performed with millions of cases is presented. The results are used to examine how the techniques traditionally used in analyses are equipped to handle and address likelihoods significantly less probable than at the 95%/95% level. The paper describes the required changes to the underlying probability distribution functions that were required to ensure physical results. The paper presents changes to the model required to achieve sufficient robustness for the process and changes to the typically used uncertainty distributions.
The analyses show that the simulation results do not form a single population of data. The input distributions that lead to the various clouds of data are presented. Physical explanations are provided that link the specific changes in uncertainty parameters to the various outcomes.

Best Estimate Plus Uncertainties (BEPU) approaches are often perceived as complex in licensing processes by licensees. The complexity of the BEPU approach is generally considered justified, as it a priori offers the potential for a more accurate estimation of safety margins. Since it tends to be less conservative than deterministic methods, it raises legitimate questions about its maturity from a regulatory standpoint, particularly given the challenges of nuclear safety assessments.

A relevant example is the case of Loss of Coolant Accidents (LOCA), where BEPU methodology can play a crucial role in assessing fuel behavior (e.g., peak cladding temperature, rupture…). Analyzing these transients involves multiscale (from sub-channel to reactor level) and multiphysics phenomena (multiphase thermohydraulics, fuel and cladding thermomechanics, neutronics, etc.).

Recently, BEPU methodologies have been proposed by licensees in France for the safety demonstration of operating and newly designed reactors. IRSN’s analyses have led to two key questions:

• Can numerical modeling be considered "best estimate" given the diversity of phenomena involved?
• Can we fully rely on the estimation of uncertainties?

This paper highlights some of IRSN’s concerns regarding these aspects, drawing on expert judgments or explicit CATHARE modeling. IRSN believes that BEPU methodologies could play a role in safety demonstration due to their ability to naturally incorporate different combinations of multiscale and multiphysics phenomena. Nevertheless, using BEPU does not exclude some penalties to be required for covering certain limitations.

Since the 1980s, the Best Estimate Plus Uncertainty (BEPU) methodology has been a cornerstone for deterministic safety analysis of design basis accidents in nuclear power plants. Despite endorsements from the International Atomic Energy Agency (IAEA) and various national regulatory bodies, its industrial application remains limited. At the BEPU 2024 conference in Lucca, Italy, a panel of global experts convened to discuss strategies for expanding BEPU’s industrial use. The panel session focused on:

• Enhancing understanding of uncertainty analysis and Input Uncertainty Quantification (IUQ).
• Addressing challenges and exploring future advancements in BEPU methodology.

• Sharing insights and experiences in BEPU application.

• Fostering dialogue on extending BEPU to broader industrial applications.

Key outcomes of the discussions included:

• Advancements in UQ/IUQ methods by NEA, highlighting the need for international cooperation.
• The role of IUQ in boosting confidence in BEPU.
• Evolution of BEPU methodologies for LBLOCA, addressing high burnup fuels and ECCS rule revisions.
• Expansion of BEPU applications from LOCA safety analyses to non-LOCA scenarios, severe accidents, containment integrity, and dose consequences.
• The necessity of a graded approach for broader BEPU application.
• Limited BEPU use in some countries due to regulatory conservatism and high development costs.
• Potential benefits for non-LWR applications, necessitating new experimental data.

This paper summarizes the main contents and outcomes of the panel discussions.

In order to solve the problem that the planned life extension of Qinshan No.3 Nuclear Power Co., Ltd. in China (hereinafter referred to as Qinshan No.3 Nuclear Power Plant) leads to the increase of spent fuel, and the capacity of existing spent fuel dry storage modules is insufficient, based on the original 1~6 (QM-400) spent fuel storage modules, the intensive spent fuel dry storage facilities (M1 and M2 spent fuel storage modules) have been developed. Compared with QM-400 spent fuel storage module, M1 and M2 modules have larger storage capacity and higher energy density. In order to demonstrate the thermal safety of M1 and M2 modules, a thermalhydraulic program is used to establish the thermal analysis model of M1 and M2 modules based on conservative initial assumptions, and calculate the temperature of each region under normal operation and accident analysis of the module under extreme weather conditions. At the same time, the three-dimensional fluid CFD program is used to verify the calculation results of the thermalhydraulic program, and the calculation results of thermalhydraulic program and CFD program are integrated, The thermal safety of M1 and M2 modules is demonstrated. Finally, from the perspective of engineering feasibility and the condition of nuclear power plant site, M1 module is adopted as the implementation plan of intensive spent fuel dry storage facility.

There is currently a growing interest in analysing the behaviour of Advanced Technology Fuels (ATF) under development. Among the new evolutionary ATF designs, the FeCrAl and Cr coated claddings are the most promising. On the other hand, the LBLOCA sequences in Pressurized Water Reactors (PWR) are among the most demanding for safety systems and have a small safety margin. To perform this analysis, NFQ and UPM developed an in-house version of the TRACE5P6 system code for FeCrAl cladding, as TRACE5P6 is not designed to simulate this cladding material. Then, a BEPU analysis of LBLOCA sequences in a PWR was performed for Zry, FeCrAl and Cr-coated Zry cladding and the safety margins were obtained for each case. The results show that the safety margins for ATF materials are greater than those for the Zry case.

The steam system downstream of the main steam isolation valves (MSIV) is normally not structurally verified for the hydraulic loads that can occur following a steam line break (SLB). Also, the turbine trip is normally not classified according to nuclear safety grade standards. Mechanical failure of the steam system, or a failure of the turbine trip system, can therefore not be excluded following a SLB. This could lead to additional steam outflow in addition to the break flow. As a consequence, blowdown of two steam generators could occur if a single failure is assumed on one MSIV. Also, considering extreme external events such as an earthquake or antagonistic actions, the integrity of the turbine building itself, along with the whole steam system outside containment, could be questioned, potentially leading to blowdown of all three steam generators.

In NUREG-0138 it is expected that the assumption of a stuck control rod would compensate for any penalties associated with the blowdown of two steam generators. In the present study, this statement has been investigated using the system thermal hydraulics code TRACE with its built in neutronic point kinetics model. SLBs with blowdown of one, two and three steam generators are analyzed. The effect of the stuck control rod assumption is also studied. An increase in maximum power is observed when blowdown of several steam generators is assumed. However, as expected in NUREG-0138, a large decrease in maximum power is seen if no stuck rod is postulated. Finally, the impact on DNBR is discussed.