[3J01-05] Anomaly Detection Technique and Robot

[3J04] Investigation on anomaly detection technique for cooling system device of sodium fast reactor by acoustic method

(4) Proof of concept of sensing of boiling phenomena using acoustic recognition and deep learning

*Yoshitaka Ueki^1,3, Shunsaku Hashimoto¹, Masahiko Shibahara¹, Kosuke Aizawa² (1. Osaka Univ., 2. JAEA, 3. TUS)

Keywords:Boiling Acoustics, Deep Learning, Machine Learning, Sodium Fast Reactor

Anomaly detection in nuclear power plants enables us to early execute prevention and mitigation measures against accident progression. In sodium-cooled fast reactors, a fuel subassembly local blockage will induce a severe accident associated with coolant boiling. The detection of the fuel subassembly local blockage can help us prevent the start and progress of the severe accident. We employ an acoustic technique to detect anomaly boiling. However, various sorts of noises are expectedly superimposed on accidental boiling sounds in reality. It is inevitable to distinguish the boiling sounds from the superimposing hostile disturbance with high accuracy. To achieve this, we utilize machine learning techniques and assess the feasibility of boiling sensing based on acoustic recognition and deep learning.