Design and Optimization of a High Thermal Flux Research Reactor Via Kriging-Based Algorithm

Abstract

In response to increasing demands for the services of research reactors, a 5 MW LEUfueled research reactor core is developed and optimized to provide high thermal flux within specified limits upon thermal hydraulic performance, cycle length, irradiation utilization, and manufacturability. A novel fuel assembly concept which makes use of integral flux traps is postulated to meet these requirements. Each assembly can be rotated into one of three different configurations to produce flux traps of different size, shape, and neutron energy spectrum within the core. A method for predicting and guiding the search for the optimum geometry was sought. Kriging has been chosen to predict the values of eigenvalue and thermal flux at untested geometric parameters. Because kriging treats all measurements as the sum of a global deterministic function and a stochastic departure from that function, predictions come with a measurement of uncertainty. As a result, the analyst can search the design space for likely improvement, or probe areas of high uncertainty for improvements that might have been missed using other methods. The technique is used in an algorithm for constrained optimization of the design, and a set of best practices for use of this are described. The optimized design produces a peak thermal flux of 1.56 x 10[superscript 14] n/cm[superscript 2]s. Safety is demonstrated by presentation of reactivity feedback coefficients and the results of loss of flow and reactivity insertion transient analysis. A single fission target can be used to produce 96 6-day Ci of [superscript 99]Mo per week. When the reactor is oriented to take advantage of high fast flux, steels can be subjected to damage rates of 5.76 dpa per year. Silicon carbide can be damaged at a rate of 2.79 dpa/y. The concept is a safe, versatile, proliferation-resistant means of supplying current and future irradiation needs.