Optimal Deployment of Controller-Detectors for the HWR System

Abstract

<p>A major concern in the design of control systems for nuclear power reactors is where and how many controllers and detectors are to be deployed in the reactor to satisfy design criteria. In order to answer this concern, we have developed an analytical method in which emphasis is placed on the linear regulator theory and the least square estimation theory.</p> <p>This work has four areas: realization of the measure of the optimality of the controller locations in terms of feedback gain in spatial control; calculation of the static set-points to compensate the excess reactivity of the reference state evaluation of dynamic range of controllers for regulating neutron fluxes; and estimation of additional responses counteracting burnup/fuelling induced random external disturbance.</p> <p>The deviations of neutron, iodine and xenon distributions from the reference states were expanded with the referenced λ-modes. The order of amplitude vector space was reduced by assumption of the dominant mode concept. Performance indices were formed with reduced state vectors and separated control functions. Pontryagin's maximum principle was applied to deterministic components and the square-root filtering to stochastic components.</p> <p>Problems were finally narrowed to solve a series of the algebraic matrix Lyapunov and Riccati equations whose solutions imply a linear transformation of adjoints to state vectors.</p> <p>A computer code ODZCR was developed for designing CANDU zone control systems using the above theories. Analysis of the existing 600 MWe CANDU zone control compartments in terms of their locations and numbers led to the following conclusions. The effective region for both spatial and bulk control was very limited and, hence, the current vertical compartments occupied the region most effectively: also the range of spatial control assigned to individual compartments was coincident with the spatial effective region found in the study.</p> <p>Alternatively we propose a horizontal zone control system that has comparable performance with better predictability.</p>