Single-ended traveling wave fault location on radial distribution lines

In this thesis a single-ended traveling wave fault location algorithm is developed for autonomously locating short circuit faults on a radial distribution line using the high frequency traveling wave transients. The traveling wave pattern observed at the sub-station is correlated with the traveling wave pattern predicted using time tree analysis for different fault locations and fault resistance. Genetic search techniques are used to evolve an initial population of possible fault locations to determine the most likely fault location. It is shown through extensive EMTP simulations that the scheme is capable of finding the location of three phase faults, inter-phase faults and single-phase faults for fault resistances ranging from 0 - 1000 ohm on a radial distribution line with five sub-feeders. A new high speed FPGA based data acquisition system is developed suitable for capturing traveling wave fault data from a radial distribution line with the necessary fidelity for the proposed fault location algorithm. The data acquisition system is deployed on a Medium Voltage distribution line in the Santa Caterina region of Brazil. A branched communication network is constructed out of RG-58 coaxial cable and a Time Domain Reflectometry device is used to capture the reflection pattern under different fault conditions. The fault location algorithm is adapted to work with TDR a opposed to fault generated traveling waves. The location algorithm is capable of locating faults with resistance between 0 and 75 ohm up to three zones away from the injection point