The Proceedings of the International Conference on Nuclear Engineering (ICONE)
Online ISSN : 2424-2934
2023.30
Session ID : 1409
Conference information

RESEARCH ON RELIABILITY EVALUATION METHOD OF ARTIFICIAL INTELLIGENCE SOFTWARE IN NUCLEAR POWER PLANTS
SuMei FengYue LiuShaoHua WangXinYu WangZhiLiang Zhang
Author information
CONFERENCE PROCEEDINGS RESTRICTED ACCESS

Details
Abstract

In recent years, people from academia and engineering applications domain have gradually tried to introduce artificial intelligence software to solve some control and maintainability problems of nuclear power plants to ensure the stable operation of the plants. The introduction of artificial intelligence(AI) software can also solve economic problems further and even explore the possibility of application in safety. However, the current intelligent decision-making has defects such as unexplainable black box problem, output uncertainty problem, and imperfect accuracy, which make the application of intelligent software in nuclear power plants lack confidence. Therefore, it is necessary to properly verify whether the reliability of artificial intelligence software meets expectations. However, the method of artificial intelligence reliability assessment has not been established. How to evaluate the reliability of intelligent software to determine whether it can be used in the nuclear power plant is the main content of this paper.

Content from these authors
© 2023 The Japan Society of Mechanical Engineers
Previous article Next article
feedback
Top