基于主成份分析的介损在线监测数据处理

Data Processing of On-line Monitoring tan δ Based on Principle Component Analysis

电容型设备绝缘在线监测数据量大 ,应予有效处理以提高分析与诊断的效率。利用主成份分析对数据进行正交变换可提取其主要特征成份以实现数据的压缩与去噪。实例分析证明该方法压缩效果非常好 ,同时进一步降低压缩率 。

Data amount of on line monitoring capacitive equipment is very large. How to process the data is very important for analysis and diagnosis. In this paper, principle component analysis is applied to process on line tan δ data. Through orthogonal transformation, principle components of the data can be extracted, then data compression and denoising can be realized. The practical results show this data processing approach is very effective. Even in lower compression rate, the regularity component and noise can be distinguished better.