基于主成分分析的建筑空调系统传感器故障检测

在已经建立的建筑空调系统仿真器的基础上,针对系统的温度、湿度、流量传感器提出了一种基于主成分分析的传感器故障诊断方法.该方法根据系统正常的历史运行数据建立数理统计模型,通过传感器实际测量数据与正常数据阵在故障子空间投影的比较,对传感器的故障进行检测.仿真试验表明,该方法能够诊断出固定偏差和漂移故障,为进一步研究传感器的故障诊断提供了必要的基础.

PCA-Based Fault Detection Method for Sensors in Building Air Conditioning Systems

A fault detection method using principal component analysis (PCA) is presented for detecting sensors with fixed and drift bias of temperature, humidity and flow rate on the base of HVAC simulation developed. According to the normal history data of the system, PCA model is to be used to detect the sensor faults by comparing the projection onto the residual subspace with which the real measurement vector and normal vector is projected. The simulation tests showed that the method could be useful for detecting the sensor faults of fixed bias and drift bias of temperature, humidity and flow rate.