A study on practical fault location system for power transmission lines using neural networks

For the efficient operation of power transmission facilities, the authors have developed a new fault location (FL) system and put it to practical use. This system uses neural networks to analyze the distribution pattern of the current induced in overhead ground wires along the poser line. Improved reliability results from the introduction of fuzzy operation of input data, a fault-type decision method and an index expressing the reliability of the fault location result. These FL systems are installed in eight commercial lines, and run normally, one of which experienced a fault due to lightning and successfully located the fault point.