基于性能参数对标的加热器故障诊断研究

Study of Fault Diagnosis of Heat Exchanger Based on the Performance Parameters Deviation

为实现对加热器性能的监控和故障诊断,利用神经网络对加热器换热过程进行了建模,可实时计算加热器性能参数的目标值,并利用目标值和运行值的对比,结合加热器的故障诊断逻辑,形成了加热器故障诊断系统,可以对电厂加热器实现在线性能监控和诊断。研究过程中对加热器的机理进行了分析,发现变工况运行时下,加热器的传统计算模型具有一定的局限性;提出了采用神经网络的方法建立加热器模型,在分析确定了输入输出参数后,利用2000多组数据进行训练,形成了加热器换热模型,可以根据实时边界条件计算性能参数的目标值;通过对加热器的故障分析,形成故障诊断逻辑,再结合神经网络模型计算的目标值同运行值对标,实现加热器的故障诊断。

In this study,the heat transfer process of heat exchanger is modeled by using neural network,using this model performance parameter reference value can be predicted,which can be used to calculate deviation to the real-time operating value of the performance parameters.Combined with the fault diagnosis logic,this model can be used for on-line performance monitoring and diagnosis of heat exchangers in the coal-fired power plant.In the process of research,the mechanism of heat transfer is analyzed,and it is found that traditional calculation model of heater has some limitations when it is used under variable operating conditions.The heater model is established by using neural network method,and more than 2000 sets of data are used for training to form the heater heat transfer model.The reference value of performance parameters can be calculated according to the real-time boundary conditions.Through the fault analysis of heat exchanger,the fault diagnosis logic is formed,and with this logic and the deviation of the real-time value to reference values,fault diagnosis of the heat exchanger can be realized.