基于PCA和TS在线辨识模型的GIS故障诊断方法

摘要

In this paper,by analysing the characteristics of GIS switch system,the fault diagnosis method is design by using PCA technology and the online identification TS fuzzy model. First,the characteristics of the GIS switch system is studied. The soft measurement system collects information by using Principal Component Analysis technology to get the fault characteristic informa?tion. Using the fault information collected by PCA fault,the information database based on TS fuzzy model is established which could judge the fault of the GIS according to the acquisition of current information. The D-S criteria is introduced to judge the fault. This article provides a method for GIS switch system fault detection for information fusion,and the experimental results show that this method is effective.%论文针对GIS开关系统的特点,利用PCA技术结合在线辨识TS模型,建立起一种新型的故障诊断方法,再利用TS模型计算的模糊权值根据D-S特征判据来判断分析GIS系统发生故障.PCA降维技术将多传感器采集的故障特征信息进行降维,反映故障的特征信息.再利用TS在线辨识模型实时的修正故障模糊规则库,并利用模糊规则库计算出本状态下的故障特征权值.利用统计的观点,统计出系统的故障特征和正常工作的特征,再利用D-S判据的方法判断出95%置信度的故障发生情况.论文所述的方法可以有效地对多故障信息进行融合,在决策层对故障信息进行融合得到有效的故障诊断结果.