A modeling error compensation approach for the feedback control of the nuclear reactor operation

Abstract

Nuclear energy for civil applications has been a matter of renewed interest as it offers clean, safe, and reliable energy sources. A new generation of compact reactors and improved safety strategies have driven refreshed research on the design and control of nuclear plants. New results on the regulation of the power output of nuclear reactors should provide improved operation strategies oriented to reliability and safety. The present work aims to study the stability and performance of a nuclear reactor under nonlinear feedback control actions. The study considered the five states March-Leuba mathematical model of a boiling water reactor. A nonlinear feedback controller that leads to exact cancellation of nonlinearities was obtained by dynamical inversion methods. However, uncertain parameters and unmeasured signals are commonly found in practical conditions. To address this issue, an observed-based estimation scheme was proposed to obtain estimates of the unmeasured signals and the trajectory of the modeling error, which are subsequently used to obtain an approximate version of the exact feedback inversion. Numerical simulations for open-loop stable and unstable (e.g., large-amplitude limit cycle) showed that the proposed control scheme provides closed-loop stability despite large model uncertainties.