Application of reactor noise methods in safety and reliability of nuclear power plants

摘要

For monitoring temperature coefficient of reactivity of pressurized water reactors a method based on the correlation of fluctuations in signals of in-core neutron detectors and core-exit termocouples and neural network paradigm is used. It is shown that the moderator temperature coefficient of reactivity can be predicted with the aid of the backpropagation neural network technique by measuring the frequency response function between the in-core neutron flux and the core-exit coolant temperature.