基于孤立森林算法的空调系统运行异常检测研究

Research on Abnormal Operation Detection of Air Conditioning Systems Based on Isolated Forest Algorithm

常规的空调系统运行异常检测以故障样本特征检测为主,但存在错检、漏检的问题,会影响空调系统的正常运行,为此,设计了基于孤立森林算法的空调系统运行异常检测方法。首先提取空调系统运行异常特征,并根据异常特征向量确定空调系统异常位置。然后基于孤立森林算法设定空调系统运行异常检测阈值,获取空调系统运行异常的正态分布数据,将异常特征数据点孤立到同一个平面上,通过阈值区分特征点是否异常,完成异常检测。采用对比实验,验证了该方法的检测准确性更高,能够应用于实际生活。

Conventional abnormal operation detection of a ir conditioning systems mainly focuses on fault sample feature detection,but there are issues such as false detections and missed detections that affect the normal operation of the air conditioning system.Therefore,a method for detecting abnormal operation of air conditioning systems based on the isolated forest algorithm was designed.Firstly,extract abnormal features of the air conditioning system operation,and determine the abnormal location of the air conditioning system based on the abnormal feature vector.Then,based on the isolated forest algorithm,the threshold for anomaly detection in the operation of the air conditioning system is set,and the normal distribution data of the anomaly in the operation of the air conditioning system is obtained.The anomaly feature data points are isolated to the same plane,and the threshold is used to distinguish whether the feature points are abnormal,thus completing anomaly detection.Through comparative experiments,it was verified that the detection accuracy of this method is higher and can be applied in practical life.