Multifractal detrended fluctuation analysis of boiling water reactors

Abstract

The aim of this work is to determine a relationship between multifractal parameters and Boiling Water Reactor (BWR) stability; therefore, a Multifractal Detrended Fluctuation Analysis (MF-DFA) of the data series obtained from the Average Power Range Monitor (APRM) and the Local Power Range Monitors (LPRM) during stable operation and during an instability event of a BWR-5, which concludes with the SCRAM, was performed. First, a much larger number of data than recommended was considered to calculate the multifractality for the instability event (4,200 data points) and for stable operation (5,700 data points) and determine if there was any difference among these analyses. Subsequently, the data were divided into periods of 50 s (500 data points) to detect the moment when the reactor begins to operate in an unstable manner, here it was not necessary to have enough information to calculate the multifractal parameters because it was only of interest to detect any trend or difference between the multifractal parameters that indicated instability in the operation. The results show that the data from the stable operation present a behavior with higher multifractality compared to the data from the unstable operation. In fact, the results indicate that the data from the unstable operation of the reactor tend to monofractality. This can be seen more clearly through the width of the multifractal spectrum, it is larger for the data corresponding to the stable operation. The proposed methodology was successfully applied to a challenging case study (Forsmark stability benchmark) as it is difficult to detect unstable reactor operation.