Analysis and modelling of radio frequency interferences induced by electric arc discharges in high voltage substations

Abstract

In the context of the next generation of electric grids, called smart grids, the use of modern and advanced communication systems in high-voltage substations can significantly improve the efficiency, reliability and safety of the electric power grid. Indeed, the deployment of intelligent sensor networks allows for the development of a more efficient, rapid, and automated remote monitoring, control and diagnosis in major pieces of high-voltage equipment in substations: current, potential, and power transformers, circuit breakers, and high-voltage disconnectors. Wireless sensor networks offer significant benefits in this area. Unfortunately, high-voltage substations are harsh and hostile environments whose wireless communication systems can be interfered to such an extent as to render their performances severely degraded. This work consists of the identification, characterization, and modelling of electromagnetic interférences in substations for the deployment of wireless sensor networks. In the first part, we propose a radio frequency measurement setup for electromagnetic interférences identification and characterization in the frequency range of 800 MHz to 5 GHz in substations. The majority of interference sources comes from radiations caused by partial discharges. The induced signals are characterized by strong transient impulses whose such events follow a cyclostationary process. Next, a characterization process is proposed, by which physical characteristics of partial discharge can be measured in terms of first- and second-order statistics. In the second part, we investigate the modelling of electromagnetic interferences caused by partial discharge sources. First, we propose complete and coherent approach model that links physical characteristics of high-voltage installations to the induced radio-interference spectra of partial discharge sources. The goodness-of-fit of the proposed physical model has been measured based on some interesting statistical metrics. This allows us to assess the effectiveness of our approach in terms of first- and second-order statistics. Finally, a generalized radio-noise model for substations is proposed, in which there are many discharges sources that are randomly distributed over space and time according to the Poisson field of interferers approach. This allows for the identification of some interesting statistical properties of moments, cumulants and probability distributions. These can, in turn, be utilized in signal processing algorithms for rapid partial discharge’s identification, localization, and impulsive noise mitigation techniques in wireless communications in substations.

Date	26 Apr 2016
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	François Gagnon (Supervisor) & Basile L. Agba (Co-supervisor)