Medium Voltage Network Residual Earth Fault Current Estimation Methods

Extensive cabling during 2010s has drastically changed the earth fault behaviour of the rural area distribution network. Against the assumptions of traditional earth fault analysis, cable net-work zero sequence series impedance is nonnegligible, thus zero sequence voltage applied over the zero sequence impedance during an earth fault generates a resistive component to the earth fault current in addition to the capacitive component. In the resonant earthed neutral sys-tem, capacitive earth fault current can be compensated with inductive Petersen coils, but the resistive current component cannot be compensated with Petersen coils. Increase of resistive earth fault current will increase the absolute value of the residual earth fault current flowing to ground during the earth fault and consequently cause dangerous touch voltages. The reactive component of the residual earth fault current is mostly known but the resistive component is associated with multiple uncertainties. The harmonic component is out of the scope of this thesis and thus omitted. Due to the uncertainties, calculation of the resistive earth fault current has proven to be complicated, but if residual earth fault current is to be calculated accurately, the resistive component must be calculated or estimated first. The SFS 6001: 2018 standard states that if the residual earth fault current in resonant earthed neutral system is unknown the value can be assumed 10% of the network capacitive earth fault current. However, as extensive cabling increases resistive earth fault current production of the network, the validity of this assumption has caused concern. Therefore, the aim of this thesis was to develop a practically oriented model for estimating residual earth fault current that can easily be applied to multiple locations in the network. Secondly, the validity of the 10% assumption specified by the standard was studied in Elenia’s network. The network information system used in Elenia is currently unable to take into account the cable network zero sequence impedance, thus a statistical examination was performed based on network data from 45 primary transformer areas. The measurements from centralized Petersen coil regulators were utilized in the examination, since the regulators provide real-time measurement of the network resistive earth fault current production. In the statistical examination the dependency of resistive earth fault current from other network parameters was studied. The objective was to identify variables that correlate with resistive earth fault current, so that they could be used to estimate the resistive earth fault current. After the correlation analysis, correction factors were assigned to the variables and the results were compared to the measurements from the regulators. The conclusion was that the resistive earth fault current can be estimated to be 5% of the total capacitive earth fault current. This result was applied to residual earth fault current calculation and the obtained values were again compared to the values calculated from the measurements. There was only a minor difference, which implies that the developed model yields accurate results. More importantly, the developed model proved to provide more accurate results than the estimation method specified in SFS 6001, that acted as a reference. In addition, there are two alternative interpretations of the method specified in the standard, so depending on the interpretation, the results were either too high or too low when applied to Elenia’s network. However, the results of this thesis are heavily dependent on the properties of the network, thus results should only be applied to networks with similar configuration