Application of spectral entropy method in alarm algorithm of loose parts monitoring system of nuclear power plant

<bold>Background</bold>The primary function of the Loose Parts Monitoring System (LPMS) is to provide early detection of loose parts that may exist within the Reactor Coolant System (RCS). The LPMS has a high probability of false alarms due to noise and pulse signal interference outside the typical frequency band of loose component signals.<bold>Purpose</bold>This study aims to mitigate noise-induced high false alarm rates in LPMS of nuclear power plants.<bold>Methods</bold>An alarm method based on signal spectral entropy was proposed to improve the traditional spectral entropy method by enhancing the weighting of the main frequency band in the energy spectrum of loose parts' impact signals. A parametric analysis of the weight coefficients was conducted to determine the optimal value for identifying loose parts' impact signals. Furthermore, the influence of interference signals (background noise and pulse signals) on the alarm accuracy was explored through steel ball impact tests.<bold>Results</bold>Test results show that the modified spectral entropy method can clearly reflect the fluctuation of the spectral entropy curve caused by impact signals with a -5 dB signal-to-noise ratio, while eliminating the interference from background noise and pulse signals.<bold>Conclusion</bold>The proposed method of this study significantly enhances the traditional spectral entropy method's capability to distinguish loose parts' impact signals from interference signals, thereby reducing the false alarms probability of loose parts monitoring system.