Search

Power transformer plays a very important role in power system, its long-term operation will cause various kinds of faults. Accurate identification and timely elimination of transformer faults are the basis of safe operation of power grid. As one of the most commonly used fault diagnosis methods, dissolved gas analysis (DGA) technology is used to identify fault types through dissolved gas in transformer oil, and its reliability has been proved. In order to analyze these gases and diagnose transformer fault types with the results, many methods have been developed, such as Key Gas Method, Method of Duval, IEC 60599 Method, Method of Dornenburg and Method of Rogers, etc. In some cases, the accuracy of these traditional methods is reduced and cannot be applied for diagnosis, since they have fixed input features and is not flexible for input combination. In order to achieve the propose of solving this defect, this paper introduces a deep belief network-based DGA method to diagnose the faults and states of power transformers with customized input features. For this work, six fault classifications were considered based on the nine characteristics extracted from the gases precipitated from the insulating oil of power transformers. The deep belief network was tested using oil samples collected from power transformers. Experiments have shown that the performance of the network has obtained relatively good accuracy results.

The missing and abnormal data in power transformer operation and monitoring greatly affect the accuracy of fault diagnosis and thus threaten the stable operation of power systems. To conduct outlier detection and improve data quality for safety warning, this paper proposes a transformer operation data preprocessing method based on KNN (K-nearest neighbor) and LOF (local outlier factor) for power transformer operation data classification. Firstly, this paper analyzes the characteristics of transformer operation data. Secondly, the local reachable density of the input data is calculated by LOF algorithm. The local outlier factor score of the data is derived according to the local reachable density, and the abnormal data is output according to the abnormal score. Then, KNN algorithm is utilized to classify the relevant data around the abnormal value and missing value of the transformer. The data are filled or corrected according to the classification results. Thirdly, the elbow method is used to determine the optimal K value and cluster operation data by K-Means algorithm. Finally, the proposed method is applied and verified with real transformer operation data in case study. The results show the method can effectively detect and correct the abnormal and missing data, conduct transformer data cleaning and preprocessing and provide accurate and effective data samples for transformer fault diagnosis.