Host: The Japan Society of Mechanical Engineers
Name : The 30th International Conference on Nuclear Engineering (ICONE30)
Date : May 21, 2023 - May 26, 2023
As a kind of clean energy, the efficient and sustainable development of nuclear energy can reduce carbon emissions and greatly reduce the burden on the environment. However, due to the particularity of fuel type, once a nuclear accident occurs, it will cause serious consequence, especially the three major accidents in the history of nuclear power, which have increased the public’s concern about nuclear safety. How to timely and effectively judge the occurrence of nuclear accidents has become an important part of nuclear power plants (NPPs) safety analysis.
This paper proposes a method for NPP accident diagnosis by establishing a Bayesian inference model. During the operation of a NPP, a large amount of monitoring data will be generated. The operator can determine the occurrence of an accident based on experience through abnormal changes in monitoring signals. However, relying solely on human judgment cannot more systematically maximize the use of all available monitoring information. In this study, according to the results of NPP’s safety analysis, the key variables that change when the accident occurs are extracted to establish a Bayesian network model, and the quantitative causal relationship between each node is represented by the conditional probability table (CPT), so as to achieve both qualitative and quantitative accident reasoning. This method can be applied to the emergency decision support system of NPPs, and the detection signals are imported into the established Bayesian network model as evidence information to give the possibility of a specific type of accident. Taking two typical accidents of HTR-10 unit as an example, this paper expounds modeling process of the Bayesian network and verification results, which shows that the Bayesian inference model can better infer the occurrence of an accident. It not only makes full use of the existing knowledge of safety analysis, but also combines the current operating state of the NPP.
