Development of Probabilistic Risk Assessment Methodology Using Artificial Intelligence Technology: 1. Automatic Fault Tree Creation

Futagami, Satoshi; Yamano, Hidemasa; Kurisaka, Kenichi; Ujita, Hiroshi

Permalink

https://hdl.handle.net/2142/121832 Copy

Description

Title

Development of Probabilistic Risk Assessment Methodology Using Artificial Intelligence Technology: 1. Automatic Fault Tree Creation

Author(s)

• Futagami, Satoshi
• Yamano, Hidemasa
• Kurisaka, Kenichi
• Ujita, Hiroshi

Issue Date

2023

Keyword(s)

• Fault tree
• Reliability database
• Probabilistic risk assessment
• Artificial intelligence (AI)

Abstract

Probabilistic risk assessment (PRA) of nuclear power plants is a laborious task to perform since step-by-step, systematic, and comprehensive processes are required to input necessary data. It needs highly skilled PRA modelling technicians who can fully understand a vast amount of the design documents and reliability of the systems, structures, and components. One of promising break-through technologies for the issue of practical use of PRA is Artificial Intelligence (AI) technology. Therefore, this study is intended to develop PRA methodology using the AI technology. For this purpose, as a first step, the authors have been conducting a three-year program including the development of AI tools for automatic fault tree (FT) creation and automatic fault detection methodology for building reliability database. These AI tools are intended to enable any users to easily perform PRA with the same quality without user effect. This paper describes overall development plan of PRA methodology using the AI technology and the progress of automatic FT creation tools development. For the automatic FT creation, The AI tool is developed for creating FT from the design documents such as paper source-based piping & instrumentation diagram, single-line diagram, etc. The automatic tools are implemented for FT creation process using AI technology and rule based automatic programs with annotation graphical user interface tool. For example, AI technologies such as image recognition and optical character recognition are adopted to identify component and piping from piping & instrumentation diagram image. The authors checked and demonstrated Prototype Automatic Tools for FT Creation with comparing PRA analysis data and results of Sodium Fast Reactor. For the automatic fault detection method, The AI tool is developed for extracting failure occurrence locations (system/equipment), failure modes, and causes from Japanese reliability databases of NUCIA(for light water reactors) and CORDS(for sodium-cooled fast reactors), and transforming them into a database using AI technologies.

Type of Resource

text

Language

eng

Handle URL

https://hdl.handle.net/2142/121832

Sponsor(s)/Grant Number(s)

MEXT Inoovative Nuclear Research and Development Program Grant Number JPMXD0222682583

Owning Collections