Development of methodology for efficient fuel design evaluation of the Advanced High Temperature Reactor (AHTR)

Abstract

This paper presents a methodology for developing a neutronics model to be used for a fluoride-salt high temperature reactor (FHR) fuel design optimization study. The reactor design is based on the Advanced High-Temperature Reactor (AHTR) with hexagonal fuel element with fuel “planks” (plates) developed at the Oak Ridge National Laboratory. The overall objective of this research was to develop a methodology accurate enough, yet practical and fast enough for a fuel plank design optimum search that minimizes the fuel cycle cost (FCC) of FHR. Resonance self-shielding calculations for double heterogeneous fuel planks of AHTR design is currently unavailable in most codes. Therefore, automated modeling of the double heterogeneity was implemented to allow accurate use of multi-group simulations. The methodology generates MCDancoff Factors using SCALE6.2; its adequate accuracy was verified. A MCDancoff Factor surrogate model was created so on-the-fly MCDancoff Factors could be provided for depletion calculations over the design space for parametric studies. A simple non-linear reactivity model with assembly-to-whole-core reactivity correction was applied to analyze performance of multi-batch refueling. Results indicate that cycle length is maximized at 40% packing fraction with a carbon-to-heavy metal ratio (CHM) of approximately 250, which minimizes the cost of outages. In contrast, discharge burnup and fuel utilization is maximized at around 20% packing fraction and a CHM of approximately 600, which minimizes fuel cost, but not necessarily the FCC cost, which also accounts for outage penalty. Void and power reactivity coefficients were determined over the design phase space to establish the envelope for stable design. The minimum FCC is the optimum trade-off between the two, within the limits of the stability envelope. In combination with the cost model the objective function is non-linear and requires global optimization methods. This paper presents the methodology needed for global optimization, while the optimization itself will be presented in a follow up companion paper.