基于小波神经网络方法的空调系统传感器漂移故障诊断

WNN Based Drift Bias Fault Detection Method for Sensors in Air-Conditioning Systems

对空调系统中的温度、压力、流量传感器的漂移故障,提出了一种基于小波神经网络的传感器故障诊断方法。该方法首先采用小波分析方法对历史故障数据和正常数据进行分析,从而提取数据的频带特征,通过神经网络对这些特征进行学习,使神经网络分析能够对待诊断数据的进行故障诊断。仿真实验的结果表明,该方法对传感器的漂移故障能够实现有效地诊断。

A fault detection method based on wavelet neural network （ WNN）is presented for detecting sensors with drift bias of temperature, pressure and flow rate. This method uses wavelet packet to analyze the history normal data and fault data, then the neural network can study feature of frequency realm obtained hy wavelet packet analysis , so it eventually can be used to detect the fault information of the new data. The simulation results show that this method can realize effective fault detection and diagnosis for the drift bias of the sensors.