21 / 2024-05-30 18:28:35

Preliminary Research on Integrating AI Technology to PSA Modeling Process

AI,PSA,FT

Currently, the Probabilistic Safety Assessment (PSA) modeling process for nuclear power plants takes quite a lot of time and energy, and usually the human-made logical models are accompanied with mistakes due to human error. There is an urgent need for using AI technology to help analyst analyze design documents and build PSA models. At present, AI technologies such as image recognition and target detection in computer vision, and machine reading comprehension in natural language processing have become mature, and which facilitates the use of AI technology in PSA modeling process.

This paper gives a preliminary three-step roadmap for achieving the goal of a more intelligent PSA modeling process. Firstly, building a standardized and comprehensive database to define key information for system and equipment in nuclear power plant. Secondly, digitizing or constructing visual chart to reflect the logical process of a system. Thirdly, transferring the system digital chart to system fault trees by following certain PSA assumptions. This paper believes that AI technology could potentially make a difference in improving the quality and efficiency of PSA modeling process.