Decision Tree Assisted Controlled Islanding

• Published in: IEEE transactions on power systems Vol. 21; no. 4; pp. 1790 - 1797
• Main Authors: Senroy, N., Heydt, G.T., Vittal, V.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2006; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptive control systems; Control systems; Controlled islanding; decision tree; Decision trees; Electric power distribution; Electric power generation; Electric power systems; Islands; Machine learning; Power system dynamics; Power system interconnection; Power system protection; Power system relaying; Power system reliability; Power system stability; Protective relaying; R-Rdot relay; self healing; special protection systems; Strategy
• ISSN: 0885-8950; 1558-0679
• DOI: 10.1109/TPWRS.2006.882470

Controlled islanding refers to the controlled separation of an interconnected power system into electrically isolated regions. The objective of this paper is to develop adaptive controlled islanding as a component of an emergency power system control strategy. There are two primary aspects of controlled islanding: 1) where to island and 2) when to island? Assisted by a decision tree (DT) approach, this paper seeks to address the "when to island" aspect. A decision tree based tool is proposed to recognize conditions existing in the system that warrant controlled islanding. A 29-generator, 179-bus system is employed to demonstrate the tool. Simulation data are used to train DTs, and the online performance of DTs is then evaluated as part of a controlled islanding strategy