10-40 L90 LINE CURRENT DIFFERENTIAL SYSTEM – INSTRUCTION MANUALSINGLE-POLE TRIPPING CHAPTER 10: THEORY OF OPERATION1010.5.2 Phase selectionThe L90 uses phase relations between current symmetrical components for phase selection. First, the algorithm validatesif there is enough zero-sequence, positive-sequence, and negative-sequence currents for reliable analysis. Thecomparison is adaptive; that is, the magnitudes of the three symmetrical components used mutually as restraints confirmif a given component is large enough to be used for phase selection. Once the current magnitudes are validated, thealgorithm analyzes phase relations between the negative-sequence and positive-sequence currents and negative-sequence and zero-sequence currents (when applicable), as shown in the figure.Figure 10-15: Phase selection principle (ABC phase rotation)Due to dual comparisons, the algorithm is very secure. For increased accuracy and to facilitate operation in weak systems,the pre-fault components are removed from the analyzed currents. The algorithm is very fast and ensures proper phaseselection before any of the correctly set protection elements operates.Under unusual circumstances, such as weak-infeed conditions with the zero-sequence current dominating during anyground fault, or during cross-country faults, the current-based phase selector may not recognize any of the known faultpattern. If this is the case, voltages are used for phase selection. The voltage algorithm is the same as the current-basedalgorithm. For example, phase angles between the zero-sequence, negative-sequence, and positive-sequence voltagesare used. The pre-fault values are subtracted prior to any calculations.The pre-fault quantities are captured and the calculations start when the disturbance detector (50DD) operates. When anopen pole is declared, the phase selector resets all its outputs operands and ignores any subsequent operations of thedisturbance detector.The phase selector runs continuously. When there is no disturbance in the power system, and the disturbance detector(50DD) from the source assigned as the DISTANCE SOURCE is reset, then the PHASE SELECT VOID FlexLogic operand is set andall other phase selector operands are reset. All current and voltage sequence components are memorized andcontinuously updated.When the disturbance detector (50DD) operates, memory stops being updated and retains two cycles of old current andvoltage values. First, purely fault components of current and voltage are calculated by removing the pre-fault (memory)quantities from presently calculated values. A series of conditions are checked with respect to magnitudes and phaseangles of current and voltage symmetrical components in order to detect the fault type. Secondly, currents are used toidentify the fault type (AG, BG, CG, AB, BC, CA, ABG, BCG, or CAG) according to the Phase Selection Principle figure. If thecurrents fail to identify the fault type, and voltages are available, then voltages are used. If any of the above types isdetermined, then the corresponding FlexLogic operand outlined in the following table is asserted.Table 10-12: FlexLogic operands assertedOperand DescriptionPHASE SELECT AG Asserted when a phase A to ground fault is detectedPHASE SELECT BG Asserted when a phase B to ground fault is detected
