Alarm Method for Loose Part of Nuclear Reactor Based on Fractal Theory

To improve the alarm veracity rate for Loose Part Monitoring System（LPMS） of nuclear power stations,a new alarm method based on the nonlinear fractal theory is presented in this paper.Correlation dimensions for background noise,loosen part impact signal and loosen part impact signal under background noise are studied respectively.It is found that the change of correlation dimension reflects the situation of loosen part well during nuclear reactor operation.Correlation dimensions for steel balls with different mass have approximate values,and the difference between the correlation dimensions of steel ball impact signal and background noise is significant,even if the steel ball impact signal is under background noise.It has been proven by experiments that the impact signal can be detected efficiently even if signal-to-noise ratio is poor.