Since the severe accident of Fukushima Dai-ichi Nuclear Power Plants, the enhancement and improvement of the safety regulations of nuclear facilities in Japan from various perspectives have been considered the new regulations are based on defense in depth concept as well as before Fukushima Accident, and have significantly raised design basis assumptions and required countermeasures for natural phenomena to prevent the simultaneous loss of safety functions due to the dependency.
From the above background, according to raised design basis assumptions, Severe Accident (SA) countermeasures against beyond design basis external hazards including earthquakes, tsunami and so on have been considered and the plans of installation of them have been accepted by Nuclear Regulatory Authority. On the other hand, since Probabilistic Risk Assessment (PRA) and Safety Margin Analysis (SMA) are performed in the safety improvement evaluation to ensure the residual risk of severe accident and confirm the actual performance of safety functions. However, since less information could be obtained from the experienced capacity data or operation experiences, the safety insights from PRA and SMA have not been clarified so far.
This paper presents the research and examination results of the reliability of capacity data based on the very few experience data by using Statistical Technique and Bayesian Update Technique to enhance the reliability and validity of PRA and SMA and summarizes kinds of insights obtained by this research and examination.