Reactor core flowing monitoring device

In a BWR type reactor, a regression model is applied to signals of two local neutron detectors on the same string. Estimated values of signals of a detector disposed at an upper position (at downstream) obtained by inputting the signals of the detector on the lower position (at upstream) to the model do not contain a fluctuation ingredient caused by bubbles generated between the detectors. Then, if the difference of the observed values of the detector at the upper position and estimated values of the detector at the upper position obtained by the regression model is determined, time series data of the fluctuation ingredient of the bubbles generated between the detectors can be obtained. The regression model is a mathematical model estimating regression of the output by a linear linkage of the values of input signals in the past. Since the fluctuation ingredient caused by the bubbles contained in each of the detectors can be separated irrespective of frequency, the transferring time and the speed of the bubbles between the detectors can be calculated accurately. (N.H.)