{AR} models and Time-Frequency analysis for characterization and acoustic detection of water leak in sodium

In this paper, we show the advantage of using the acoustic signals detection approach for detecting pressurized water leakage in the liquid sodium. On the one hand, contribution of the time-frequency analysis for the characterization of the process has been illustrated on data from experimental tests. On the other hand, the study of the performances of autoregressive models has demonstrated a significant reduction in detection delay when using acoustic approach versus commonly used detections techniques such as hydrogen detection