Search

This article is devoted to simulating the generation and propagation of EM waves caused by the partial discharge (PD) inside a void defect. Although many researchers have focused on the micro discharge process, few researchers are studying its impact on ultrahigh-frequency (UHF) signal generation. For simplicity, several studies replace PD with a predetermined Gaussian source. This article proposes an accurate pulsed current model for UHF signal generation which further links the waveform of the source with the discharge condition. At the micro level, the PD is caused by the penetration of void defects, whose steep rising edge further generates electromagnetic waves. Based on that, this article simulates the PD UHF signal by combining the micro discharge process and the macro transmission process. For the micro discharge process, a positive streamer caused by dielectric barrier discharge (DBD) is studied, and the discharge current can be obtained. For the macro transmission process, the finite-difference time-domain (FDTD) method is used to calculate the propagation of UHF signals. The proposed method builds a more direct relationship between observed UHF signals with the discharge process; thus it is helpful to establish an inverse model from UHF values obtained by sensors outside to the insulation defect inside of the equipment.

This article proposes to use a time reversal (TR) algorithm for locating the partial discharge (PD) source through ultrahigh frequency (UHF) EM waves. The TR algorithm is operated with the high-fidelity digital twin model of the physical equipment, which is conducive to replicating the process of EM transmission in the virtual space, and the geometric influence of the power equipment is factored into the theoretical calculation at the time of PD location. However, the common TR algorithm is operated under closed conditions without any power loss of signals, which limits its application to common equipment. In this article, the TR algorithm is adapted to achieve PD localization under semi-enclosed conditions, which meets the localization requirements of more electric equipment. Besides, a 3-D gas insulation switchgear (GIS) model is constructed to verify the proposed method, and the finite difference time domain (FDTD) method is used to calculate the process of EM wave transmission. The TR simulation is performed using three detected sensors inside the GIS. Ultimately, the results of numerical simulation are obtained to demonstrate that the proposed model is efficient and adaptive for PD localization, especially under semi-enclosed conditions. This provides an idea of inversion analysis from outside UHF signals to the inside insulation defect position through a digital twin model.