Support to the identification of anomalies in an external neutron source using Hurst Exponents.

Abstract

A new methodology is proposed here to identify anomalies in the neutron flux as a result neutron production trips in an Accelerator Driven System - ADS type reactor. This methodology is based on the calculation of Hurst exponents, where the neutron flux as monitored in the reactor core is treated as a temporal series. In several recent articles, related to fractional diffusion, the Hurst exponent is used as an estimate for the order of the fractional derivative. Our object of study considered a reactor based on the Myrrha simulated with the Serpent Monte Carlo code for two kinds of trips occurred in the production of neutrons, as follows: a Peak of Production (PP), the Unprotected* Accelerator Beam Overpower (UABO), and the spurious Beam Trip (BT). In order to estimate the Hurst exponent it was used two different methods, namely the Rescaled Range Analysis (R/S) and the Detrended Fluctuation Analysis Method (DFA). The results obtained showed that the R/S methodology had some advantages in a comparison with the DFA in indicating the occurrence of those anomalies. Also being able to provide a scale for the assessment of the intensity of the trip occurred, showing to be an useful tool to support anomaly identification in the neutron flux of ADS reactors.