Asset management in smart grids using improved Dissolved Gas Analysis

Abstract

Asset Management Systems (AMS) are pivotal to build reliable and safe smart grids. An important function of AMS is the monitoring and diagnosis of the power transformers. Various tests are performed on power transformers to detect incipient faults. Among the available methods, Dissolved Gas Analysis (DGA) has been widely used and shown promise. However, interpreting the results of the DGA is challenging due to the availability of wide variety of methods such as Rogers ratio, Doernenburg ratio, key gas procedure of IEEE, Basic gas ratio and Duval triangle methods of IEC. The accuracy of the interpretation methods influences AMS performance leading to reliability issues in the grid. This investigation compares the accuracy of Duval method and basic gas ratio method to detect transformer faults from real-time fault data obtained from power transformers. Our results on data obtained from Electrical Research and Development Association for seven transformer incipient faults shows that the Duval method is accurate than the basic gas ratio method for identifying incipient transformer fault based on DGA results. Further, the basic gas ratio was not able to detect two of the seven faults. These results illustrate the need to integrate Duval method to detect power transformer faults within AMS.