Providing continuous and reliable electric power supply is essential, even under crucial operational circumstances. Faults occurring on medium voltage distribution networks have significant affects on power quality. Due to the open electric energy market the importance of meeting demands on maintaining the quality of electric power has risen. Improving the quality indices is not only obligatory for distribution system operators, but also essential for keeping their competitiveness.
Consequently the distribution system operators are motivated to make the restoration of operation after a fault event less time consuming – sufficiently accurate information of a fault’s location provides a faster and more cost efficient restoration, thus resulting in shorter outage durations.
In the past years researches have begun worldwide in automatic disturbance detection, fault location and effective restoration of normal operational conditions. In this thesis, a comprehensive review of the proposed methods regarding fault classification and location is given, moreover several of these methods are introduced in details.
This thesis presents a new phase-to-ground earth fault location method for radial cable distribution networks and various algorithms composed for its examination. This is a modified version of a previously developed method applicable to medium voltage distribution networks containing only overhead lines. Both locating techniques have been developed by the BUTE Department of Electric Power Engineering.
The method uses injections and measurements at harmonic frequencies conducted only at the substation to determine the zero sequence impedance belonging to the earth fault. The calculated impedance correlates to the distance between the substation and the occurred fault. The configuration, topology and preliminary simulation of the observed distribution network is required for the method. The laterals of distribution lines can result in multiple fault locations at the same distance.
The accuracy of the parameters of different cables constituting the simulated distribution network significantly affect the effectiveness of the presented method, hence the possibilities of improving these parameters are investigated. As part of the examination the method’s sensitivity to the changes of cable parameter values, and the minimal fault and open-circuit measurement variation required to sufficiently improve the cable parameter values are determined. The possibilities of using the faults occurred subsequently during the operation of the fault locating device to improve the essential cable parameters and hereby the accuracy of the fault location are also investigated.