Developing a neutronic core model of the NuScale Power Module using Serpent

Abstract

In this study, the reactor core of the NuScale power module was simulated using the Serpent Monte Carlo code and KENO-VI from the SCALE code system, with the objective of analysing key core parameters (such as k-effective and neutron flux) as well as safety parameters (fuel temperature coefficient and moderator temperature coefficient). Neutronic calculations were performed for the first cycle of the reactor core (i.e., the fresh core). Fuel assembly models and a full-core model were developed, with detailed geometry and material specifications. NuScale Chapter 4 of the Safety Analysis Report served as the primary source for the specifications and key parameters used in this study. However, it is important to note that NuScale Chapter 4 did not provide all the necessary parameters for model development. Fuel enrichment and gadolinium content of the Fuel Assemblies (FAs) for the first cycle, the inner and outer radii of the pressure vessel, the nuclide composition of the spacer grids, and the chemical composition of the stainless steel for the Reactor Pressure Vessel (RPV) and the reflector were obtained from other sources, as acknowledged in this dissertation. The results presented in this document indicate that the k-effective is higher in models that do not contain gadolinium or contain smaller amounts of it (2%,). Gadolinium acts as a burnable absorber and is commonly used for reactivity control in most light water reactors, due to the high neutron absorption cross-section of several gadolinium isotopes and its good solid solubility in UO₂. The calculation results show that the criticality values for the full-core and fuel assembly are 1.17481 ± 0.00015 and 1.13991 ± 0.00014, respectively, based on the core layout described in Section 3.2 and the layout of the fuel assembly in Section 3.2.1 The k-effective value for the full-core was found to be 3 490 pcm higher than that of the fuel assembly. This difference is attributed to the higher uranium enrichment in the full-core model. Additionally, the k-effective results for the full-core in this study deviated by 3 698 pcm from the first comparison model developed by other researchers. This discrepancy may be due to the inclusion of additional components in the full-core simulation, such as the radial stainless-steel reflector, bypass channels for coolant flow, the barrel, the thermal and biological shield, and the Reactor Pressure Vessel (RPV). Furthermore, the first comparison model did not specify the actual dimensions and positions of the bypass channels, which could have contributed to the discrepancy. Further verification indicated that the introduction of coolant channels decreased the k-effective results. Additional verification was conducted by modifying the base full-core model to align with the second comparison model developed by other researchers and is different from the first comparison model. The first and second comparison models are discussed in detail in the dissertation. The difference between the base model and the second comparison model was found to be 0.00088 (88 pcm). The work proceeded chronologically, beginning with the first comparison model, followed by the second comparison model. However, the results of this study show better agreement with the second comparison model. Furthermore, the k-effective results for the full-core base model, calculated using the SCALE code, align with those obtained using Serpent, with a difference of 0.00075 (75 pcm). Thus, a successful code-to-code comparison has been achieved. The distribution of the neutron flux in the full-core base model did not align with the expected cosine-like flux shapes, which are commonly observed in bare reactor cores. This deviation is due to the presence of different types of fuel assemblies in the reactor core, which introduce variations in the material composition of the fuel. The moderator temperature coefficient falls within the typical values for a pressurized water reactor. The moderator temperature coefficient was found to be negative; more negative than the fuel temperature coefficient. Outside the scope of this study, it is worth noting that Boron concentration can influence the MTC. At the beginning of the cycle (BOC), higher Boron concentrations are typically used to control reactivity, and if not properly managed, could lead to a positive MTC. Burnable absorbers are commonly employed to mitigate this and ensure a stable, negative MTC throughout the reactor's operation."