基于BP神经网络的核电厂主动容错控制方法研究  

Active Fault Tolerant Control Research for Nuclear Power Plant Based on BP Neural Network

作　　者：李金阳 夏虹 刘永阔 成守宇 巩诚 

机构地区：[1]哈尔滨工程大学核安全与仿真技术国防重点学科实验室,黑龙江哈尔滨150001

出　　处：《原子能科学技术》2012年第7期827-830,共4页Atomic Energy Science and Technology

摘　　要：针对核电厂中的传感器故障,采用改进的BP神经网络算法对传感器进行神经网络训练,建立各种运行状态下的动态模型库,并应用BP神经网络对系统进行实时检测。当传感器发生故障时,采用控制率重构的方法进行容错控制。在核动力装置模拟器上以稳压器为对象进行了仿真实验验证,结果表明该方法对于核电厂中的传感器故障进行容错控制是有效的。In view of the sensor fault of nuclear power plant, the sensor was trained by adopting improved back propagation （BP） neural network method, and the dynamic model bank in different states was set up. The system was detected by using BP neural network in real time. When the sensor goes wrong, it will be controlled by reconstruction. Taking pressurizer as the case, a simulation experiment was performed on the nuclear power plant simulator. The results show that the proposed method is valid for the fault tolerant control of sensor faults in nuclear power plant.

关 键 词：BP神经网络 容错控制 核电厂 传感器 

分 类 号：TL36[核科学技术—核技术及应用]