Abstract: Analysis of the dissolved gas of power transformers is one of the most widely used online monitoring technologies at present,still it has problems such as poor data quality affected and lack of unified standardization of data collection due to the generally complex scene environment which greatly limit the effectiveness of application of oil chromatographic technology for the equipment condition monitoring,accuracy and timeliness. In this paper,an integrated method of unsupervised anomaly detection for online monitoring data of oil chromatography based on rule extraction is proposed,and the data quality evaluation system for the online DGA monitoring system is established to enhance the availability of oil chromatographic data. Taking a 220 k V power transformer as an example for case study,the paper has verified the effective application of the data quality assessment system has been verified.