A Method Using Strong Tracking Filter for Compensating the Response Delay of Self-Powered Neutron Detectors

Abstract Self-Powered Neutron Detector (SPND) has been widely used for monitoring neutron flux in reactors. There are two typical types of SPNDs, including prompt type and delayed type. The prompt type SPNDs’ output current are mainly contributed by gammas generated from the neutron capture process. Rhodium SPND is a typical delayed type SPND and has been widely used for neutron flux monitoring in reactors. Compared with other prompt SPNDs such as Cobalt SPND, Rh-SPND contains a higher neutron sensitivity. However, long response time and high noise gain are the main problems that restrict the application of SPNDs with delayed current components. In order to solve the above two problems, a method using the Strong Tracking Filter has been proposed in this paper. Compared to the traditional Kalman Filter, Strong Tracking Filter will offer robustness to the modeling errors and ensure the whiteness property of the filter innovation sequence. This gives Strong Tracking Filtering a high adaptivity to sudden changes (e.g. reactor load rejection) in the useful signal, even in the presence of noise. Following the derivation of the ideal iterative inversion relation from current to neutron flux, a filtering matrix containing noise filtering and neutron flux inversion was established using Strong Tracking Filter. Then it’s shown that this method can give a shorter response time than the Kalman Filter method for the same noise gain, and its strong robustness to initial values and sampling frequency was also demonstrated by the test. Finally, this method was used to obtain the real-time neutron flux from the SPND current in the other reference, demonstrating its effectiveness and excellence.