Transformer anomaly recognition method based on abnormal characteristics of DGA data
Peng Neo Zhang, Bo Qi, Pei Yang, Ruoyu Zhang, Chengrong Li · 2018
Power transformer is one of the key electrical apparatus in power system and its reliability has directly correlation on the safety of power system. This paper presents anomaly recognition method of oil chromatographic data based on distributed computing, aiming at the existing problems that the study fails to use abnormal features of oil chromatographic data effectively to identify the abnormal and lack of the problem of identifying method for massive data anomaly. First, obtain the characteristics of abnormal data through the analysis of a large number of oil spectrum anomaly data sequence. Then, Form anomaly recognition method. The method uses abnormal features of oil chromatographic data as judgment index of abnormal mode. If any abnormal pattern is met, the abnormal data can be identified. Abnormal recognition of transformer oil chromatographic data is realized on distributed computing platform. Accuracy rate of anomaly recognition is up to 89.3%. Experimental results show that this method is accurate and can be applied to distributed computing platform.