Nuclear reactor period calculation algorithm based on extended Kalman filter

Abstract

Reactor period is an important parameter in nuclear power plant which indicates the operational status of nuclear reactor, and it is usually obtained by a nuclear instrumentation system (NIS). In the NIS, current signal from neutron detector is pre-amplified and conditioned into voltage signal which is proportional to the power level of nuclear reactor, and then is acquired and processed by a digital computer in the NIS. The acquired voltage signal always contain measurement noise which has a great influence on the calculation result of reactor period and causes a lot of uncertainty in the result. Reactor period calculation is still a difficult issue so far which has not been satisfactorily solved. In this paper a calculation algorithm of reactor period for a digital NIS is studied based on the advanced signal processing technology: Extend Kalman Filter (EKF). A discrete stochastic state space model for EKF design was first derived, and then the calculation algorithm of reactor period was given. Two calculation results based on the simulation and experimental data were also given to proof the correctness and effectiveness of the proposed calculation algorithm. The results show that the extended Kalman filter-based calculation algorithm of reactor period is feasible and effective, and it can suppress the influence of measurement noise on calculation result to recover the signal's original features.