Detection of local sodium boiling in a simulated LMFBR fuel subassembly

Out-of-pile experiments were carried out to study the feasibility of the local sodium boiling detection by use of temperature, flow and acoustic noises. The studies showed that the temperature fluctuation transferred the information of local sodium boiling to the end of the bundle, but little to the outlet. A new anomaly detection method, which used the algorithm of whiteness test of the residual time series data of autoregressive model, especially using flow fluctuation, was effective detecting the anomaly such as local sodium boiling. On the other hand, acoustic noises with local sodium boiling became much higher than the background such as the mechanical and flow noises. A waveguide method, which was utilized for the present experiments, was useful for the boiling detection.