PREDICTING FUTURE BEHAVIOR OF TRANSIENT EVENTS RAPIDLY ENOUGH TO EVALUATE REMEDIAL CONTROL OPTIONS IN REAL-TIME

Electric utilities are becoming increasingly interested in using synch ronized phasor measurements from around the system to enhance their pr otection and remedial action control strategies. Accordingly the task of predicting future behavior of the power system before it actually o ccurs has become an important area of research. This paper presents an d analyses several approaches for solving the real-time prediction pro blem. The first method clusters the initial post-fault swing curves in to coherent groups and fits a low order equivalent model to the specif ic transient event in progress. The model is updated with each new set of phasor measurements and provides a running prediction of future be havior which is valid for approximately 1/2 second into the future. We show how this capability would be useful inside the framework of a pr otection scheme such as the proposed French Defence Plan. If, on the o ther hand, a relatively detailed reduced-order model is available ahea d of time, then it could be used to predict future behavior for severa l different control options. The task in this case is to solve the mod el much faster than real-time using the post-fault phasor measurements as the initial condition. In order to solve systems with detailed loa d model fast enough for real-time prediction, we present a new piecewi se constant current load model approximation technique that can solve a model as complex as the New England 39 bus system with composite vol tage dependent loads much faster than real-time. If the reduced order model is too large for real-time solution, then a pattern recognition tool such as decision trees can be trained off line to associate the p ost-fault phasor measurements with the outcome of future behavior. In this case also, the piecewise constant current technique would be need ed to perform the off-line training see generation with sufficient spe ed and accuracy.