A new fast-learning algorithm for predicting power system stability

Ahmed A. Daoud, George G. Karady, Ibrahim A. Amin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

13 Scopus citations


This paper presents a new fast learning, on-line method for the prediction of power system transient instability and an example of its application to a single machine and infinite bus. The proposed algorithm is adapted from a proven robotic ball-catching algorithm, which includes fast learning. For instability prediction, the ball location is replaced by measured relative generator rotor angle. Using the measured relative rotor angle, the control algorithm predicts the rotor angle at a future time. The relative rotor angle is sampled at a rate of 600 times per second. This new fast learning algorithm predicts the rotor angle 500 milliseconds into the future. The increase of the generator relative rotor angle beyond a predetermined threshold is a prediction that loss of synchronism will occur. When loss of synchronism is predicted a protection scheme can initiate a stability aid such as generator tripping, braking resistor and/or fast valving.

Original languageEnglish (US)
Title of host publicationProceedings of the IEEE Power Engineering Society Transmission and Distribution Conference
Number of pages5
StatePublished - 2001
Event2001 IEEE Power Engineering Society Winter Meeting - Columbus, OH, United States
Duration: Jan 28 2001Feb 1 2001


Other2001 IEEE Power Engineering Society Winter Meeting
Country/TerritoryUnited States
CityColumbus, OH

ASJC Scopus subject areas

  • General Engineering
  • General Energy


Dive into the research topics of 'A new fast-learning algorithm for predicting power system stability'. Together they form a unique fingerprint.

Cite this