Proposta de utilização do expoente de Hurst da detecção de transientes causados pela fonte externa de nêutrons em reatores ADS

Abstract

In this work, a new methodology is proposed for the detection of transients caused by an external neutron source in an ADS reactor. A methodology presented consists of analyzing the neutron flux fluctuations as a time series from the stochastic analysis using the Hurst exponent method, which is a well established tool for the treatment of time series and detectation of intrinsic trends. The Hurst exponent allows to identify events occurring within the time series and characterize them as to the persistence of such events. Events with persistent behavior are the most important case on reactor kinetics. They indicate phenomena that can directly influence any kinetic behavior on the reactor. The study considered the MYRRHA reactor simulated with the Monte Carlo Serpent code for some types of transients, which are: Peak production, Unprotected Accelerator Beam Overpower and Spurious Beam Trip. The results indicate that the R/S methodology was more efficient than DFA methodology to identify transient occurrence, proving to be a useful tool, including the real-time transient detection support for ADStype reactors.