A Multivariable Autoregressive Model of the Dynamics of a Boiling Water Reactor

Abstract

An autoregressive (AR) model with pseudo-random binary sequence (PRBS) test signals was applied to the dynamics of the Japan Power Demonstration Reactor, a boiling water reactor (BWR). The decision of the order of the AR model was based on the Akaike criterion. Multi-input test signals of the PRBS were applied to the steam-flow control valve and the forced circulation pump speed control terminal. Seventeen variables including the instrumented fuel assemblies were observed. The AR model identification facilitated building the BWR dynamics model as a multivariable system. The experiment indicated that the BWR dynamics with rather intensive nonwhite noise interference was effectively represented by the AR model, which was compared with a linear theoretical dynamics model. The results suggested that the identified AR model plays an important role in verifying, modifying, and improving the theeoretical dynamics model.