Estimation of sub-criticality using extended Kalman filtering technique

Reactivity is widely used as the paramount means for defining nuclear reactor status. The measurement of reactivity can be only made in an indirect way. Traditionally, reactivity is estimated by the Inverse Point Kinetics (IPK) method. However, this technique suffers from some serious drawbacks like high sensitivity to reactor parameters and less immunity to noise content in the input signals, hence effective only during power range operation. In this paper, the extended Kalman filter (EKF) technique, which is based on stochastic model of reactor kinetics is proposed for subcriticality estimation in nuclear reactor. The proposed technique can work in noisy environment and modeling errors and uncertainties in parameters do not affect the estimation severely as the feedback gain is continuously adjusted during the estimation process. The performance of proposed technique for the reactivity estimation has been evaluated using power variation data sets collected from a PHWR (Pressurized Heavy Water Reactor) and a research reactor. It has been found that with the application of EKF technique, reactivity in a highly subcritical core can be estimated with reasonable accuracy. The EKF based approach has been found to yield higher accuracy, noise suppression and robustness than done by IPK based approach.