Full-Amplitude Probability Distribution Analysis-Based Information Entropy Signal Processing Method for Sodium Bubble Detector

Under conditions of high temperature and high pressure, the sodium-cooled fast reactor’s steam generator (SG) will leak and cause a sodium-water reaction, which may lead to an explosion in severe cases. Therefore, timely detection and warning are required at the leakage early stage. An electromagnetic vortex flowmeter (EVFM) can be used for leakage detection of SGs in sodium-cooled fast reactors. To achieve reliable leak detection, EVFM needs to have a fast response time and small leakage detection capabilities. This article analyzes the full-amplitude probability distribution of EVFM signals with or without leakage and finds the distribution differences between single peak and double peak. Considering the randomness of bubble noise, an information entropy-based signal processing method for EVFM is proposed, and information entropy is used to measure the uncertainty of signal amplitude waveform. By comparing the information entropy difference at different amplitude segment numbers, the optimal segment number is determined. Finally, a DSP-based sodium bubble detection system was developed, and offline and online verification tests were carried out. Results show that the developed system can detect SG leakage as low as 0.1 g/s with a response time of less than 1 s.