Sistema de detecção e localização de descargas parciais com rede de sensores sem fio.

Abstract

Failures in insulation of power systems equipment can causes irreparable damages and interruption of power supplying. A potential cause for these failures is the occurrence of partial discharges. Partial discharges (PD) are electrical discharges that partially bridges the insulation between conductors. The monitoring of this phenomena is a way to diagnose insulation and take a decision with respect to the equipment. An online system for partial discharge monitoring, with a low cost, can be justified when compared to assets value that operates in a power system. Taking into account this problem, this Master’s Dissertation aims the development of a system based on a wireless sensors network (WSN) integrated to a virtual instrument (VI), for partial discharges detection and localization. The used principle is that transients generated by a PD source irradiates radio frequency (RF) waves. In the system proposed, the sensors nodes are composed by an antenna, a RF low noise amplifier, a RF power detector, a microcontroller and a ZigBee module. The data are sent through wireless network to a central node, in which receive data and send to computer. A virtual instrument being executed in computer, estimates the PD source localization using Received Signal Strength (RSS) Fingerprinting technique. Localization estimation is calculated by K-Nearest Neighbor (KNN) and Weighted K-Nearest Neighbor (WKNN) methods.