Nuclear power plant accident diagnosis method, device and equipment based on transfer learning, and medium

By using a transfer learning-based approach, a nuclear power plant accident diagnosis model was constructed and updated using a simulation dataset from the early stages of the reactor core's lifespan. This solved the problem of decreased diagnostic accuracy caused by changes in the reactor core's lifespan, ensuring high accuracy and safety of the nuclear power plant throughout its entire lifespan.

CN122113616APending Publication Date: 2026-05-29SHENZHEN INSTITUTE OF INFORMATION TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHENZHEN INSTITUTE OF INFORMATION TECHNOLOGY
Filing Date
2026-02-10
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing nuclear power plant accident diagnosis technologies fail to adequately consider changes in reactor core lifespan, leading to differences in operating parameters at different life stages. This results in a decrease in the accuracy of accident diagnosis models, impacting the safe operation of nuclear power plants.

Method used

Based on the transfer learning method, an initial diagnostic model is constructed by acquiring a simulation dataset of accident conditions at the beginning of the core's lifespan. During the operation of the nuclear power plant, the core status is monitored in real time, the diagnostic accuracy is evaluated periodically, and the model is updated through transfer learning to ensure that the model adapts to changes in the core status and improves diagnostic accuracy.

Benefits of technology

It achieves high accuracy and robustness in accident diagnosis throughout the entire operating cycle of a nuclear power plant, improving the safety of nuclear power plant operation and the efficiency of accident response.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122113616A_ABST
    Figure CN122113616A_ABST
Patent Text Reader

Abstract

The application relates to a nuclear power plant accident diagnosis method and device based on transfer learning, equipment and medium, wherein the method comprises the following steps: obtaining a nuclear power plant accident working condition simulation data set at an initial stage of a core life cycle, training an initial accident diagnosis model according to a constructed training data set; monitoring a real-time core state parameter; during the operation of the nuclear power plant, the real-time core state parameter is monitored, the diagnosis accuracy of the initial accident diagnosis model is evaluated based on current operation parameters according to the core state change, if the diagnosis accuracy decreases by more than a preset threshold, transfer learning is performed on the initial accident diagnosis model, and the accident diagnosis model after the transfer learning is updated as a current effective accident diagnosis model; when an accident transient working condition occurs in the nuclear power plant, a real-time collected operation parameter time sequence is input into the current effective accident diagnosis model, and an accident diagnosis result is output. The application improves the accuracy and reliability of nuclear power plant accident diagnosis.
Need to check novelty before this filing date? Find Prior Art