基于BP神经网络的凝汽器故障诊断研究

通过对现有凝汽器运行中常见的典型故障及征兆集的分析,进一步完善了凝汽器的典型故障知识库,对故障征兆的具体表达方法进行了分析。运用 Matlab 神经网络工具箱和隶属度函数两种方法对凝汽器的运行状态进行故障监测和诊断,通过实例验证表明应用 Matlab 神经网络工具箱方法不仅计算简便,而且诊断结果具有较高的可靠性。

Study on Fault Diagnosis of Condenser Based on BP Neural Network

According to the analysis of typical fault and symptom set that commonly found in condenser, the typical fault knowledge base of condenser is further established, besides, the concrete expression of fault symptom is discussed. Two methods are employed to monitor and diagnose the condenser operating status, which are utilizing Matlab neural network toolbox and membership grade function. Example verification indicates that it is not only simple to calculate applying Matlab neural toolbox, but also has higher credibility to the diagnosis results.