The evaluation of uncertainty constitutes the necessary supplement of best-estimate calculations performed to understand accident scenarios in water-cooled nuclear reactors. The needs come from the imperfection of computational tools, on the one side, and the interest in using such a tool to get more precise evaluation of safety margins. The paper reviews the salient features of three independent approaches for estimating uncertainties associated with predictions of complex system codes. Namely, the propagations of code input error and calculation output error constitute the keywords for identifying the methods of current interest for industrial applications, while the adjoint sensitivity-analysis procedure and the global adjoint sensitivity-analysis procedure, extended to performing uncertainty evaluation in conjunction with concepts from data adjustment and assimilation, constitute the innovative approach. Throughout the developed methods, uncertainty bands can be derived (both upper and lower) for any desired quantity of the transient of interest. For one case, the uncertainty method is coupled with the thermal-hydraulic code to get the code with capability of internal assessment of uncertainty, whose features are discussed in more detail.

The present paper deals with the proposal of an additional safety barrier for the class of large (1000 MWe or more) Light Water Reactors (LWR) now in operation, in construction, or under design. Emphasis is given to the motivations or the needs for the barrier. Two main parts of the paper can be distinguished. The following topics are discussed in the former part (section 2): (a) the weakness of the barrier constituted by the current design of nuclear fuel; (b) the continuously increasing complexity of the system, with main reference to the Instrumentation and Control (I&C); (c) the role that the Large Break Loss of Coolant Accident (LBLOCA) had for arriving at the current layout of the Reactor Coolant System (RCS). Furthermore avoiding the severe accidents in 1979, 1987 and 2011, is at the basis of the proposal. In the latter part (sections 3 and 4), the elements of the proposed technological safety barrier are discussed: the As-Low-As-Reasonably-Achievable (ALARA) principle, the Best Estimate Plus Uncertainty (BEPU) approach, the Extended Safety Margin Detection (E-SMD) hardware, the Emergency Rescue Team (ERT) strategy (or a virtual entity for the reactor) and the Independent Assessment (IA) concept. The additional safety barrier, although not demonstrated in the paper, is expected to reduce for a factor in the range 10–1000 the probability of core melt and to have a cost in the order of 1% the cost of a nuclear reactor unit.

The competitiveness of small modular reactors (SMRs) has been planned based on design simplification, short construction time, passive safety systems, and enabling self-financing by ramp-up construction. Due to the global energy challenges, SMRs have received pervasive attention from a wide range of researchers, designers, developers, stakeholders, and customers. Besides the many advantages related to the design of SMRs, there are challenges ahead of these reactors. SMR licensing is one of the most critical challenges in the front deployment of these reactors. This challenge stems from innovations in SMR designs and systems, such as modularity or deployment for desalination, energy storage, hydrogen production, process heat, and district heating. Due to the lack of experimental data and technical knowledge, the licensing challenges for non-water coolant SMRs are more complicated. Nearly all previous generation reactor licenses were based on conservative analysis while the decision-making methods based on best-estimate and realistic approaches have received more attention in recent years. Thus, the method known as the best estimate plus uncertainty (BEPU) approach is selected for licensing in some cases. At this time, using the BEPU approach in licensing for conventional NPPs is a mature technology and ready for industrial application. Nevertheless, because most previous reactors were licensed based on conservative methods, developers and even regulatory bodies resist re-assessments based on the BEPU approach, while using the choice of conventional conservative methods is a type of roll-back for next-generation SMRs. Thus, this work reviews the BEPU approach and clarifies the possibility of using this approach in the licensing process of SMRs. The lack of experimental data and tight coupling of phenomena along with uncertainty quantification are the main challenges ahead of using BEPU in the licensing process of SMRs.

RELAP5 is a system code developed at the Idaho National Environmental and Engineering Laboratory for thermal hydraulic analysis of nuclear reactors. The code RELAP5 is widely used for safety analysis studies of commercial nuclear power plants. However, recent released version of RELAP5/3.2 and over present significant capabilities for analysis of nuclear reactor research systems. As a contribution to the assessment of RELAP5/3.3 for research reactor safety analysis, experimental data from the University of Massachusetts Lowell Research Reactor - UMLRR are used. The UMLRR is a 1 MW light water moderated and cooled, graphite-reflected, open-pool type research reactor. This paper presents the development and the validation of a UMLRR-RELAP model using experimental data. For this purpose, a series of experiments were performed for benchmarking RELAP5 calculations for research reactor systems. As a result of this study, the UMLRR nodalization is shown to be representative of the experimental data reactor behavior.

In the domain of reactor transient simulation, the identification of thermal-hydraulic phenomena (THPs) plays a major role. The system codes should model all influential THPs and should be validated against integral effect tests and separate effect tests which cover all influential THPs. The validation and the uncertainty quantification should cover every model related to an influential THP. A list of 116 THPs, recently established, covers all water-cooled reactors and design basis accident (DBA) analyses. It synthesizes more than 30 years of Organisation for Economic Co-operation and Development and International Atomic Energy Agency activities conducted by several safety specialists. A new tentative method to identify THPs was proposed based on two sources of information, the parameter evolutions in transients (depressurization, voiding, refill, heating, ...) and the set of balance equations with source and sink terms for convection, diffusion, interfacial transfers, and wall transfers. This method is just based on Gen-2 pressurized water reactor (PWR) transient analysis at the system, component, and basic process levels. The comparison of the 116 THP list with the list of the tentative methods is made in this paper for the particular case of GEN-2 PWRs. No major contradiction was found. Both methods identified phenomena at the system, component, and process level. The 116 list better identified special components that require “special models” in system codes. The use of equations identified many more local process THPs, which may help ranking phenomena in a scaling analysis and identifying validation needs. The comparison confirms a potential synergy and complementarity between approach 1, which is based on the 116 THP list, and approach 2, which is based on the tentative method; the outcome of this comparison suggests further efforts to combine them and complement them in a new international collaborative context. This analysis reports ongoing discussions between members of the Forum and Network of System Thermal Hydraulics Codes in Nuclear Reactor Thermal-Hydraulics network of system code developers.

After three quarters of a century using nuclear fission to produce energy, Nuclear Reactor Safety and Risk constitutes an established technological sector. A key feature is continuous updating following new discoveries and progress in knowledge, resulting in extensive and elaborate safety methodologies, which are still not internationally accepted, generally applicable or technically consistent. Each country developed its own methods, guides, traditions and requirements to deal with evolving design, safety, siting and licensing issues. There is a clear parallel in societal risk perception between nuclear radiation exposure in accidents and viral infection in pandemics and the fear of the “unknown”. Unfortunately, over the last 20–30 years the declining introduction of electricity by nuclear fission in the countries that contributed most to its earliest development also has broken the bond between new scientific advancements and improvements of existing safety methodologies. By looking at the origins and fundaments of nuclear technology, we consider the following topics of both deterministic and probabilistic interest: a) Loss of Coolant analysis; b) nuclear fuel accident performance weaknesses; c) role of containment and ultimate heat sinks; d) residual risk and emergency system deployment, and e) independent and risk informed decision making assessment. As a key outcome, we propose modifying the traditional licensing methodology, and the use of active and/or passive systems by being subsumed into a broader Engineered Safety Features Management process. Furthermore, we emphasize the need of connecting the As Low As Reasonably Achievable principle with the analyses to demonstrate the safety of nuclear installations minimizing the need for excessive “paper” safety analyses and licensing efforts.