Enhancement of Accuracy in the Estimated Power-Angle Curve Considering Current Immediately after Fault Clearance

A fault that occurs near a power station may cause a loss of synchronism of generators in the power station against the remaining generators and result in a widespread blackout. When swing of generators in the power station with respect to the remaining generators is similar to that of a generator in one machine and an infinite bus system, a power-angle curve (P-δ curve) is often used for studying the stability of the system and deciding the number of generators to be shed for stabilization. In operation, power system stabilizers predict loss of synchronism, determine the number of generators to be shed, and then shed them in a very short time period. Such stabilizers use data measured during a period of about 200 ms after fault occurence. The voltage and current measured on the bus and line connected to a power station are used to compute the P-δ curve. When a line is removed to clear a fault, the current in the damper circuit of a generator decays significantly in a very short period. This causes significant changes in the line current. However, when the P-δ curve is estimated using these data, and is applied to a stability study, the number of generators to be shed sometimes exceeds the optimum number. This is caused by the practice of measuring the current immediately after a fault is cleaed. Therefore, the authors propose a current detection method in which the portion of current that decays immediately after a fault is cleared. Such current has three distinct terms. Two are exponential functions in which one has extremely short time constant, while the other has a longer time constant. By using these estimated functions, the fast decaying component is removed from the measured current, which can then be used to estimate the P-δ curve more precisely. The method was verified via simulations using an infinite bus system model and the EAST10 model of IEEJ.