Estimation of the Dynamic Reactivity Using Digital Kalman Filtering

The problem of computing the time-dependent reactivity from flux measurements is solved for a stochastic-point-reactor model using linear estimation theory. In order to use this method, an appropriate reactivity model must be formed which can be embedded into the reactor system as one of its state variables. A digital filter, designed using Kalman-Bucy filter theory, operates on the sampled neutron-density output of the stochastic reactor and provides estimates of the state variables of the system, one of which is the reactivity. Digital Computer results are used to show the response of the Kalman filter in computing the reactivity from stochastic-neutron-density measurements. The results show that this analytical filter can be used to determine the reactivity from neutron density which has a certain noise content due to the stochastic processes involved in the reactor system.