IWGFR benchmark test on signal processing for boiling noise detection, stage 2: Analysis of data from BOR-60

Data from boiling experiments in the BOR 60 reactor in USSR has been supplied by IAEA to enable analysis techniques to be compared. The signals have been analysed at RNL using two basic techniques, High Frequency RMS analysis and Pulse Counting analysis and two more sophisticated methods, Pattern Recognition and Pulse Timing Analysis. All methods indicated boiling successfully, pulse counting proved more sensitive than RMS for the detection of the onset of boiling. Pattern Recognition shows promise of a very reliable detector provided the background can be defined. Data from an Ionisation chamber was also supplied and there was good correlation between the neutronic and acoustic signals. (author). 25 figs, 4 tabs