基于机器学习的铅铋堆中棒束振动预测  

Prediction of Rod Bundle Vibration in Lead Bismuth Reactor Based on Machine Learning

作　　者：王东东 王端 周志伟[1] 张泂 杨红义[1] 杨宏伟[1] 

机构地区：[1]中国原子能科学研究院,北京 [2]中国核工业学院,天津

出　　处：《核科学与技术》2023年第2期133-140,共8页Nuclear Science and Technology

摘　　要：流致振动是一种复杂的非线性过程。在反应堆运行时,冷却剂的流动会引起结构的振动,长期的振动可能使结构发生疲劳损伤或联接件松动磨损,迫使核电厂停堆检修,造成较大的经济损失。本研究基于机器学习算法建立了流致振动参数在线实时监测模型,提前预知掌握流致振动参数在未来一段时间内的变化规律,为实现提前采取有效措施及时避免流致振动诱导事故发生的能力提供技术参考,保障反应堆系统的运行安全。Flow induced vibration is a complex nonlinear process. During the operation of the reactor, the flow of coolant will cause the vibration of the structure. Long term vibration may cause fatigue damage to the structure or loose wear of the connecting parts, forcing the nuclear power plant to shut down for maintenance, resulting in large economic losses. Based on the machine learning algorithm, this research has established an online real-time monitoring model for flow induced vibration parameters to predict and master the change law of flow induced vibration parameters in the future in advance, which provides technical reference for the ability to take effective measures in advance to avoid flow induced vibration induced accidents in time, and ensures the operation safety of the reactor system.

关 键 词：流致振动 机器学习 在线监测 

分 类 号：TM6[电气工程—电力系统及自动化]