摘要
变风量空调系统的空气处理机组(AHU)出现故障时会使系统舒适性降低,能耗和运维成本增加。本文提出了一种基于改进型主元分析(PCA)和BP神经网络算法,用于AHU的模型建立及故障诊断。结果表明使用改进滤波的PCA检测模型主元数为3个,累计贡献率92.7%。当系统传感器出现5%的故障偏差,模型在送风温度、新风温度、冷冻水流量三种故障中的检测率均大于90%。使用BP神经网络算法对系统故障进行诊断,新风温度传感器故障和冷冻水流量传感器故障诊断率均达到了100%。新风温度传感器由于精度较高检测率为92%。但是三种故障类型检测率都超过了90%。通过分析表明改进型主元分析(PCA)和BP神经网络算法可以有效检测与诊断AHU系统故障。
Failure of Air Handling Unit(AHU)in VAV air conditioning system will reduce system comfort,increase energy consumption and operation and maintenance costs.In this paper,a new algorithm based on improved Principal Component Analysis(PCA)and BP neural network is proposed for model building and fault diagnosis of AHU.The results show that the PCA detection model using the improved filter has three principal components,and the cumulative contribution rate is 92.7%.When the system sensor has a fault deviation of 5%,the detection rate of the model in the three kinds of faults of supply air temperature,fresh air temperature and chilled water flow is greater than 90%;the system fault is diagnosed by BP neural network algorithm,and the fault of fresh air temperature sensor and chilled water flow sensor is diagnosed The detection rate of fresh air temperature sensor is 92%due to its high accuracy,but the detection rate of three fault types is more than 90%.The analysis shows that the improved PCA and BP neural network algorithm can effectively detect and diagnose the fault of AHU system.
作者
庞义旭
鹿世化
陈玮玮
田建新
PANG Yi-xu;LU Shi-hua;CHEN Wei-wei;TIAN Jian-xin(School of Energy and Mechanical Engineering,Nanjing Normal University;Wujiang Yangming Air Conditioning Purification Corporation)
出处
《建筑热能通风空调》
2021年第2期11-14,63,共5页
Building Energy & Environment
基金
江苏省自然科学基金项目(BK20180732)
中国博士后科学基金项目(2018M632332)
江苏省高等学校自然科学研究项目(18KJB470017)。
关键词
主元分析
故障检测
空气处理机组
变风量空调系统
principal component analysis
fault detection
air handling unit
VAV air conditioning system