Neutronics Modeling and Simulation of Sharp for Fast Reactor Analysis

Reactor design is an iterative process among various disciplines such as reactor physics, thermal fluid dynamics, mechanics, reactor and plant dynamics, fuel behavior, and instrumentation and control, and requires a wide range of modeling and simulation tools that can accurately predict key system performance and safety characteristics. Most code systems for fast reactor design were initiated more than twenty years ago and designed to accommodate the computing resources, tools and methods available at that time, although considerable relevant advances of methods and codes have been made more recently. Improvements to the available tools are thus needed to design future reactor systems that must meet performance, safety and economic goals higher than the current ones. The uncertainties and biases in the various areas of reactor design activities need to be reduced by enhanced prediction capabilities. These improvements are important to minimize the costly and lengthy procedures of building multiple representative mockup experiments to confirm the predictions. The challenge in neutronics analysis is to efficiently generate solutions to the Boltzmann equation by taking into account the geometric complexity and complicated energy dependence of nuclear data. The Monte Carlo method can represent these details, but needs sufficiently low statistical uncertainty, reliable variance estimates and uncertainty propagation. Computing resource requirements still remain unmanageable for many types of routine design analyses, including accurate estimations of detailed pinby-pin power distributions, effects of small perturbations, thermal feedback, error propagation via fuel depletion, and transient analysis. Thus, its use is typically limited to steady-state reference solutions without thermal feedback. As a result, the current design tools rely heavily upon deterministic methods based on various approximations and sophisticated multistep procedures [1-3]. Detailed energy variable treatments are done only at the pin cell level (or using a homogeneous mixture of fuel, coolant, and structure materials) with approximate boundary conditions. Then, by performing a series of subdomain calculations with a larger problem domain but fewer modeling details, space and energy condensed parameters are defined and tabulated for global core calculations. Global analyses are performed for three-dimensional models composed of homogenized regions with low-order approximations of the Boltzmann equation (e.g., diffusion approximation). Detailed pin-by-pin information is then recovered by reconstruction (de-homogenization) methods. In order to explore a broad range of design space and incorporate innovative design features, it is necessary to minimize approximations and increase the modeling capability from first principles. Significantly improved capabilities to simulate multiphysics phenomena are also This paper presents the neutronics modeling capabilities of the fast reactor simulation system SHARP, which ANL is developing as part of the U.S. DOE’s NEAMS program. We discuss the three transport solvers (PN2ND, SN2ND, and MOCFE) implemented in the UNIC code along with the multigroup cross section generation code MC-3. We describe the solution methods and modeling capabilities, and discuss the improvement needs for each solver, focusing on massively parallel computation. We present the performance test results against various benchmark problems and ZPR-6 and ZPPR critical experiments. We also discuss weak and strong scalability results for the SN2ND solver on the ZPR-6 critical assembly benchmarks.