[en] Researches and studies on nuclear reactor core are usually subdivided into two major fields, named: Thermal-Hydraulic and Neutronic, in which, precise simulation of reactor behaviour in both fields is highly required to ensure the designers that reactor will work in a safe margin. In this study, a thermal-hydraulic analysis of pressurized water reactor core is performed using a porous media approach. Based on this approach, each fuel assembly was modelled and was divided into a network of lumped regions; each of them was characterized by a volume average parameter. In such manner, while complex geometries are easily defined and dealt with, the thermal-hydraulic parameter and phenomena like friction, shear stress, cross-flows, convective heat transfer and etc. are strictly included in simulations. To validate the applied approach, the numerical analysis and COBRA EN code results were compared for a typical PWR core and showed a good agreement. (authors)

[en] A slab model has been widely applied in the thermal-hydraulic safety research fields for a simplified test development. However, the similarity in thermal-hydraulic behaviors is strongly dependent on the flow characteristics having a 2 D or 3 D flow field. The slab model is suitable for 2 D flow characteristics. If the slab model is applied in a strong 3 D flow field, the flow distortion grows highly. In this numerical study, the distortion of the flow field is investigated by changing the aspect ratio of the test section

No abstract available

[en] This paper describes some Uncertainty Quantification (UQ) and propagation performed within the NURESAFE European collaborative project. It especially presents the CEA results in the PREMIUM benchmark, an international benchmark organized by the Committee on the Safety of Nuclear Installations (CSNI, 2013) of the OECD/NEA, devoted to the quantification of the uncertainty of the physical models. It contains a confirmation of the estimation with a propagation of the uncertainties to two experiments including one blindly to check the capacity of the uncertainties to be extrapolated. In addition to the PREMIUM benchmark, other studies are shown, that were carried out to determine the accuracy of different quantifications using different indicators. The difficulties faced during the project are presented and analyzed, and potential high-impact improvements to the method are detected. One of them would consist in developing and assessing an evaluation matrix from SETs and IETs with two parts, one designed for the quantification and the other for the validation. The paper also highlights the need of tools to evaluate the quality of the UQ.

[en] The first time I met George was in March 1998 for dinner. And the last time I had dinner with him was in March of last year, or two decades later exactly. During the first evening, we discussed and confronted our strategies for the short period of time left then until his retirement in 2004; was it sufficient time to contribute to the discipline with something risky and innovative? During the last four-hour dinner, after having evoked life and death rather joyfully and serenely, we looked back and analysed what came out of our two decades of collaboration. This paper, which was prepared with the aim to portray George’s achievements in thermal-hydraulics during the last 20 years of our partnership (1998–2018), is written in the spirit of narrating our epilogue culinary-science-talk, while trying to be faithful to his thoughts and ideas about the developments of the discipline and its perspectives. The paper introduces first the cascade of computational tools and discusses trends related to reactor safety problems and developments needed, as well as the need for new kinds of refined experimental data. Although George was not directly involved in all the examples presented here, he felt so concerned about the success of each project/case presented in this paper that he was virtually part of it; and as he wrote in our last paper (Yadigaroglu and Lakehal, 2016): this is after all his near-home work. Finally, in memory of the two other great scientists who have left us recently, Geoff Hewitt and Sam Martin, the content of the paper includes two cases in which both of them had collaborated directly or indirectly.

[en] SIMULATE-3K is a two group, advanced nodal reactor analysis transient code. It has been used by many utilities for the prediction and analysis of BWR stability events. In recent years, several enhancements of the SIMULATE-3K channel and vessel thermal-hydraulic model have been made. This paper summarizes the status of the SIMULATE-3K core and vessel models. It presents results for the Ringhals-1 Stability Benchmark and for regional instability in a large BWR. Finally, modeling sensitivities to space and time discretization are discussed. (author)

[en] The state feedback optimal control techniques are used in designing the reactor control system. The mathematical plant model with the temperature feedback effects is established from the point kinetics equation and the singly lumped thermal-hydraulic balance equations, and is expressed in terms of state variables. The LQR control system is designed, being followed by the LQG design to determine the optimal conditions both for the rod movements and for the output responses. This approach permits the consideration of the nuclear limitations since the rod movements are included as a system constraint. Also two different servo control schemes are proposed for the purpose of input tracking. The general control characteristics such as stability margins and output responses are investigated

[en] This study has extensively investigated the emerging 3-D printing technologies for use of MIR-based flow field visualization methods such as PIV and LDV. As a result, mixture of Herb essential oil and light mineral oil has been evaluated to be great working fluid due to its adequate properties. Using this combination, the RIs between 1.45 and 1.55 can be accurately matched, and most of the transparent materials are found to be ranged in here. Conclusively, the proposed MIR method are expected to provide large flexibility of model materials and geometries for laser based optical measurements. Particle Image Velocimetry (PIV) and Laser Doppler Velocimetry (LDV) are the two major optical technologies used for flow field visualization in the latest fundamental thermal-hydraulics researches. Those techniques seriously require minimizing optical distortions for enabling high quality data. Therefore, matching index of refraction (MIR) between model materials and working fluids are an essential part of minimizing measurement uncertainty. This paper proposes to use 3-D Printing technology for manufacturing models for the MIR-based optical measurements. Because of the large flexibility in geometries and materials of the 3-D Printing, its application is obviously expected to provide tremendous advantages over the traditional MIR-based optical measurements. This study focuses on the 3-D printing models and investigates their optical properties, transparent printing techniques, and index-matching fluids

[en] New methods of studying reactor thermohydraulic problematic such as CHF, DNBR, boiling and instability etc. were introduced briefly in this paper. Some of these methods have been successfully applied in the reactor thermohydraulic analysis. And the others have not yet been applied in this field. The further study work must be done to show that these new methods can be successfully applied in this field with high accuracy because they are interdisciplinary fields. (authors)

[en] This summary discusses a new methodology to support the construction of surrogate models for computationally expensive models with many input parameters and output responses. In realistic engineering calculations, such as thermal-hydraulics simulation, the cost of the simulation is typically expensive, especially when detailed models are employed to describe system behavior, thereby justifying the need for surrogate models. The cost of surrogate model construction is typically a function of the number of model parameters and the degree of model nonlinearity. For sufficiently complex models, the cost of constructing the surrogate model could be by itself overwhelmingly expensive, thereby diminishing its value. One key requirement of surrogate model construction is the ability to preselect the best parametric form to describe the model behavior over the expected range of parameter variations. This summary presents a new algorithm to help select the best approximation out of a template of functions using recent advances in reduced order modeling (ROM) techniques. The idea of selecting functions out of a template is typically referred to as subset selection in the statistical and applied mathematics community. Different from existing subset selection algorithms, we rely on ROM to help identify the most influential functions for the surrogate model. Previous work has shown that ROM can help reduce the effective dimensionality of the parameter space, which can help reduce the cost of surrogate model construction. In this summary, we adopt a different philosophy, where the fitting functions are treated as pseudo parameters and ROM is applied on the pseudo parameter space, with the result now being a ranked list of the pseudo parameters along with their contribution to the quality of the surrogate model fit. This approach allows the analyst to discard all terms that have negligible impact on the quantity of interest. A relap5 model is employed to demonstrate the application of the proposed methodology. (authors)