Certain investigations on location and detection of partial discharge in transformer oil by direction of arrival estimation using circular array of ultrasonic sensors

Abstract

In a transformer, Partial Discharge (PD) is the small electrical discharges which take place across a localized area of the insulation system. The early identification of PD in the transformer is the most significant challenging task in order to prevent the equipment from failure. By the detection of PD at early, the preservation cycle and the efficiency of the equipment get increased. The proposed research work is developed, in order to address the issues related to the PD detection as less accuracy rate, maximum time complexity and error rate, the proposed research is developed. In this proposed research, the single partial discharge and multiple partial discharges in the transformer oil is located by using Cirular Array of Ultrasonic Sensor (CAUS). In first work, the proposed method involves four significant stages such as Fast Independent Component Analysis (Fast ICA) algorithm, Total Least Square (TLS) algorithm, Sparse Representation of Array Covariance Vector (SRACV) technique and Particle Swarm Optimization (PSO) algorithm. At first, the PD signal is de-noised from the CAUS by the performance of Fast ICA. Secondly, the TLS algorithm is executed for changing the wide band signal from CAUS into a narrow band signal. After that, the SRACV technique is implemented in order to separate the Direction of Arrival (DoA) in three directions from PD to CAUS. Followed by, the PD sources are located in an appropriate location through the utilization of pitch and azimuth angles of three DoAs. Then, the PSO algorithm is executed in order to estimate the exact coordination of the three planes. Thereby, the single PD in transformer oil is successfully identified with the help of the proposed PSO method. newline