Detection and location of PD activities using an array of fiber laser sensors

Abstract

Partial discharge (PD) measurement has been deployed as a well established test method to monitoring the health of the insulation inside generators, transformers, gas insulated switch gear, and bus bars. There are different technologies to monitor the health of insulation by PD measurement, e.g., UHF high frequency capacitance measurement and acoustic emission using piezoelectric sensors. UHF technique has proven to be very sensitive but had troubles locating the sources of such PD activities. Piezoelectric acoustic sensors seem to be able to locate PDs, but their signal is weak and locating the sources using time of arrival method proves to be complicated: acoustics signal can travel through the insulation media such as oil or SF6, but they could also hit the metallic container and then travel to where the sensor is mounted. This paper will discuss a break through. First, a fiber laser can be formed by including a twin grating structure over an active fiber segment having Erbium and Ytterbium doping and pumped with 980 nm laser light. A DBR laser capable to follow very high frequency response is formed. Furthermore, this cavity can be made to be wavelength specific. A string of these can be installed over the walls of the transformer or a multi-segment GIS or Bus Bar. The signal from these laser sensors will become self identifying. These reflected signals can be demultiplexed with an array waveguide grating (AWG), which places the signal into buckets of various center wavelengths; therefore, each bucket can be related to where the sensor has been installed. The fiber laser sensor gives a higher signal to noise ratio than piezoelectric sensors, and these DBR lasers can be serially connected to diagnose the location of the PD events. There is no need for complicated software and algorithm to calculate and interpret the difference in time of arrival. Furthermore other fiber optics sensors such as temperature, moisture and loading can be installed to help to correlate PD with these operational and environment condition.