Application of zero-sequence filter on transformer differential protection

Delta-Wye transformer connections create discontinuities in the zero-sequence network as the zero-sequence current can flow at one side of the transformer without flowing at the other side. This effect generates a zero-sequence differential current that can make the differential unit trip. Traditional solutions applied to remove the zero sequence differential current where based on delta connected CTs. Zero-sequence filters in digital relays are software implemented. In many digital relays the zero sequence filter can be enabled or disabled. On the other hand, some relays can remove the zero-sequence current calculated from the phase currents or from the ground currents (currents measured in the neutral grounding). This paper reviews the transformer configurations that require the enabling of the zero-sequence filter by taking into account not only the connection group but also the construction of the magnetic core (this aspect is not always considered), explaining in detail the phantom or virtual tertiary effect of three-legged wyre-wye transformers. Real false trips due to this effect are included. The paper also explains the differences between both methods used for the zero-sequence current calculation (the one based on the phase currents and the one based on the ground current). The influence on the differential unit, harmonic restraint and common external fault detectors is analyzed. The first method can lead to a reduction of the differential current and to an erroneous phase selection during an internal fault. However, "2 out of 3" logics both for harmonic blocking and for a phase directional comparison unit can be implemented increasing the stability The second method provides very good sensibility and phase selection but does not allow the implementation of the "2 out of 3" logics reducing the stability. Cases based on real events and RTDS simulations are reviewed.