基于BP神经网络的凝汽器故障诊断系统研究

The Research on the Method of Using BP Neural Network to Fault Diagnosis of the Condenser

摘要: 凝汽器作为核动力装置二回路热力循环的冷源，其功能的完整性与优劣性直接影响着汽轮机能否安全、可靠地运行。本文对船用核动力凝汽器的故障及其相关征兆进行了分析，建立了基于专家经验与理论知识的凝汽器故障特征知识库，并结合BP神经网络设计了凝汽器诊断专家系统，为验证系统的有效性，对凝汽器典型故障进行了仿真实验,从实验结果来看，系统完全能够正确识别凝汽器故障。

Abstract: As the condenser is the cold source of nuclear power plant’s second loop thermodynamic cycle, its function of integrity and merits affect the security and reliability of turbine directly. In this paper, the faults with related symptoms of marine nuclear power condenser were analyzed to build the characteristics base of condenser faults based on expert experience and theoretical knowledge, combined with BP neural network to design a condenser expert diagnosis system. For verifying the effectiveness of the system, we played simulation experiment on typical faults of condenser. The experimental results proved that the system can identify the faults of condenser absolutely.