Mohamed, Hamd (2018) Partial Discharge Detection and localization Using Software Defined Radio in the future smart grid. Doctoral thesis, University of Huddersfield.
Abstract

Partial discharge (PD) occurs if a high voltage is applied to insulation that contains voids. PD is one of the predominant factors to be controlled to ensure reliability and undisrupted functions of power generators, motors, Gas Insulated Switchgear (GIS) and grid connected power distribution equipment. PD can degrade insulation and if left untreated can cause catastrophic insulation failure. However, PD pulse monitoring and detection can save cost and life prior to plant failure. PD is detected using traditional methods such as galvanic contact methods or UHF PD detection methods.

Recently, an alternative method for PD detection and monitoring using wireless technology has become possible. Software Defined Radio has opened new opportunities to detect and monitor PD activity. This research makes use of SDR technology for PD detection and monitoring. The main advantages of SDR technology are that it is cost-effective and it is relatively immune against environmental noise. This is because the noise at electrical power stations is from around a few KHz to a few MHz and this is well below the SDR frequency range and PD frequency band (50-800 MHz). However, noise or interference also exists in the PD frequency band. These interferences are narrow band and mainly from FM, TV broadcasting and mobile telephony signals whose frequencies are well known, thus these interferences can be possibly processed and removed.

In this research two SDR products (Realtek software defined radio RTL-SDR/Universal software radio peripheral USRP N200) are used to detect PD signals emitted by a PD source that was located at a distance of 1 m in case of RTL-SDR device while in case of USRP N200 the PD source was located at a distance of 3 m. These PD signals once received by an SDR device are recorded and processed offline in order to localize the PD source. The detected PD signal was around 20 dB above background noise in case of the RTL-SDR device and 25 dB above background noise in case of using the USRP N200. Selecting the appropriate SDR device depends on factors such as high sensitivity and selectivity. Furthermore, although USRP N200 is more expensive than RTL-SDR dongles, USRP N200 was preferred over RTL-SDR as it demonstrates higher sensitivity and overall better results. PD detection using SDR devices was conducted in the frequency domain. These result were validated using a high-end costly device, i.e. spectrum analyzer. Generally, SDR devices demonstrate satisfactory results when compared to spectrum analyzers. Considering that spectrum analyzers cost around £10,000, while a USRP N200 SRD device costs less than £1000, SDR technology seems to be cost-effective.

Following PD detection, PD localization was performed using USRP N200 results, and a localization algorithm based on Received Signal Strength (RSS) was adopted. The localization result was within a 1.3-meter accuracy and this can be considered as a relatively good result. In addition, and for the purpose of evaluating the proposed scheme, more experiments were conducted using another system that is based on radiometric sensors which is WSN PD system. The estimated error was 1m in case of using the SDR-USRP N200 system and 0.8 m in case of using the WSN PD system. Results of both systems were very satisfactory, although some results at the corners of the detection grid were not good and the error was higher than 3 meters due to the fact that the RSS algorithm performs poorly at corners. These experiments were used to validate both systems for PD detection and localization in industrial environments.

Information
Library
Documents
[img]
Preview
Mohamed THESIS.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (7MB) | Preview
Statistics

Downloads

Downloads per month over past year

Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email