上海交通大学学报 ›› 2022, Vol. 56 ›› Issue (3): 267-278.doi: 10.16183/j.cnki.jsjtu.2021.502

所属专题: 《上海交通大学学报》“新型电力系统与综合能源”专题(2022年1~6月)

• 新型电力系统与综合能源 • 上一篇    下一篇

基于数据驱动的核电设备状态评估研究综述

许勇1,2, 蔡云泽1,3(), 宋林2   

  1. 1.上海交通大学 自动化系,上海 200240
    2.福建福清核电有限公司,福建 福清 350318
    3.上海交通大学 系统控制与信息处理教育部重点实验室; 上海工业智能管控工程技术研究中心,上海 200240
  • 收稿日期:2021-12-08 出版日期:2022-03-28 发布日期:2022-04-01
  • 通讯作者: 蔡云泽 E-mail:yzcai@sjtu.edu.cn
  • 作者简介:许 勇(1983-),男,福建省莆田市人,博士生,现从事核电设备故障诊断研究.
  • 基金资助:
    国家自然科学基金(61627810);国家科技重大专项(2018YFB1305003)

Review of Research on Condition Assessment of Nuclear Power Plant Equipment Based on Data-Driven

XU Yong1,2, CAI Yunze1,3(), SONG Lin2   

  1. 1. Department of Automation, Shanghai Jiao Tong University, Shanghai 200240, China
    2. Fujian Fuqing Nuclear Power Co., Ltd., Fuqing 350318, Fujian, China
    3. Key Laboratory of System Control and Information Processing of the Ministry of Education; Shanghai Engineering Research Center of Intelligent Control and Management, Shanghai Jiao Tong University, Shanghai 200240, China
  • Received:2021-12-08 Online:2022-03-28 Published:2022-04-01
  • Contact: CAI Yunze E-mail:yzcai@sjtu.edu.cn

摘要:

核电设备全生命周期的状态评估对提高核电厂安全性、经济性影响重大.以往国内核电厂对系统、设备、构筑物的运维评估手段多依赖于设备自身报警机制、简单阈值判断或者现场工程师经验.随着在线监测系统在核电厂的应用实施和海量设备运行数据的积累,利用数据驱动技术进行设备健康状态评估已经成为行业关注重点.对此,介绍核电在线监测系统现状,分析主要核电设备存在的常见故障,并将核电设备的状态评估归纳为异常检测、寿命预测和故障诊断共3大问题,分别综述其研究和应用现状,重点阐述深度学习技术在该领域的应用潜力.在此基础上,进一步分析核电厂设备状态评估面临的挑战和可能的解决方案.

关键词: 核电设备, 状态评估, 异常检测, 寿命预测, 故障诊断

Abstract:

The condition assessment of the entire life cycle of nuclear power equipment has a significant impact on improving the safety and economy of nuclear power plants. In the past, operation and maintenance of systems, equipment, and structures of domestic nuclear power plants, mostly relied on the alarm mechanism of equipments, the simple threshold judgments of parameters, or the empirical judgments of engineers. With the implementation of online monitoring system in nuclear power plants, a large number of equipment operation data have been accumulated, and the use of data-driven technology to assess the health of equipment has become the focus of attention in the industry. In this paper, the current situation of the online monitoring system of nuclear power equipment was introduced and the common malfunction of nuclear power equipment was analyzed. The condition assessment of nuclear power equipment were categorized into three major problems (i.e., anomaly detection, life prediction, and fault diagnosis), the situation of research and application were summarized respectively, and the application potential of deep learning technology in this field was emphasized. Based on this, the challenges and possible solutions to the condition assessment of nuclear power plant equipment were further analyzed.

Key words: nuclear power equipment, condition assessment, anomaly detection, life prediction, fault diagnosis

中图分类号: