Design and Optimization of UHF Partial Discharge Sensors using FDTD Modeling

Abstract

Partial discharges may occur in defective insulation systems of high voltage equipment, such as gas insulated substations and power transformers. These discharges generate electromagnetic waves that can be detected using ultrahigh frequency (UHF) sensors. UHF sensors must meet certain sensitivity criteria over a wide frequency range so as to be capable of detecting small discharges. Sensor frequency response is measured using a gigahertz transverse electromagnetic calibration system. Previous research has shown that finite-difference time-domain (FDTD) simulation of the calibration process can accurately model the response of the existing sensors. The work reported here demonstrates how a new sensor can be designed entirely using FDTD modeling. The proposed new partial discharge sensor has a physical construction that was selected to make it more robust, simple to manufacture, and convenient to install on metal-clad high-voltage apparatus. The internal structure of the UHF sensor was developed and optimized entirely within the FDTD software domain before the physical device was manufactured and tested. Simulated and experimental calibration results were found to be in agreement to within 10%. This finding validates the design methodology and optimization process. The approach described in this paper will help to streamline the design of UHF partial discharge sensors for specific applications in future.