Fuel lattice design in a boiling water reactor using an ant-colony-based system

► We present an ant-colony-based system for BWR fuel lattice design and optimization. ► Assessment of candidate solutions at 0.0 MWd/kg 235U seems to have a limited scope. ► Suitable heuristic rules enable more realistic fuel lattice designs. ► The election of the objective has a large impact in CPU time. ► ACS enables an important decrease of the initial average U-235 enrichment. This paper presents a new approach to deal with the boiling water reactor radial fuel lattice design. The goal is to optimize the distribution of both, the fissionable material, and the reactivity control poison material inside the fuel lattice at the beginning of its life. An ant-colony-based system was used to search for either: the optimum location of the poisoned pin inside the lattice, or the U 235 enrichment and Gd 2O 3 concentrations. In the optimization process, in order to know the parameters of the candidate solutions, the neutronic simulator CASMO-4 transport code was used. A typical 10 × 10 BWR fuel lattice with an initial average U 235 enrichment of 4.1%, used in the current operation of Laguna Verde Nuclear Power Plant was taken as a reference. With respect to that reference lattice, it was possible to decrease the average U 235 enrichment up to 3.949%, this obtained value represents a decrease of 3.84% with respect to the reference U 235 enrichment; whereas, the k-infinity was inside the ±100 pcm’s range, and there was a difference of 0.94% between the local power peaking factor and the lattice reference value. Particular emphasis was made on defining the objective function which is used for making the assessment of candidate solutions. In a typical desktop personal computer, about four hours of CPU time were necessary for the algorithm to fulfill the goals of the optimization process. The results obtained with the application of the implemented system showed that the proposed approach represents a powerful tool to tackle this step of the fuel design.