UHF sensors positioning on the power transformer tank to enhance the partial discharge localization accuracy

Partial discharge (PD) localization in power transformers by ultra-high frequency (UHF) technique based on the time difference of arrival equations is state of the art. In this way, after receiving the PD electromagnetic waves (EMWs) by the installed sensors on the tank and processing them, the PD location is estimated. The PD that occurs in the power transformer insulation is not visible to the eyes and other advanced equipment. Therefore, its location should be accurately determined using available methods for remedial actions. The localization accu-racy by the UHF method suffers from negative effects, including non-line of sight propagation of EMWs, reflection and refraction of EMWs and arrival time detection error. The impact of these negative effects on the localization accuracy can be mitigated by optimal placement of the sensors on the tank. To install the sensors, transformer manufacturers drill holes in the transformer tank. Because this work is done only once during the construction stage, transformer designers must consider places on the transformer tank, such that the acceptable localization accuracy is obtained for all potential PD sources. This paper proposes a new simple and straight-forward method based on the distance between the vertexes of the hyperbolic surfaces and their intersection point to find the best positions on the tank to install the sensors. Simulation results performed by Monte Carlo method and CST microwave studio and comparing the results with the arbitrary sensors positioning results validate the efficiency of this method.

Automated summary

Partial discharge (PD) localization in power transformers using the ultra-high frequency (UHF) technique based on time difference of arrival equations is advanced. By receiving and processing PD electromagnetic waves (EMWs) through installed sensors on the tank, the PD location can be estimated. The UHF method for localization accuracy is affected by non-line of sight propagation, EMW reflection and refraction, and arrival time detection error. This study proposes a new method using the distance between hyperbolic surfaces to determine optimal sensor installation positions on the tank.