Decision Tree Algorithm for Real-Time Identification of Critical Voltage Control Areas

Abstract

Voltage stability, is driven by the balance of reactive power in a transmission power grid, and depends on the reactive power produced by the power stations and the value generated by the capacitive power lines and static compensators. Of particular interest is to identify those Critical Voltage Control Areas (CVCAs) in a transmission power grid, that may suffer reactive power deficiencies. Since speed of analysis is critical for on-line applications the approach will address the development of a scheme whereby CVCAs can be identified using data mining techniques from on-line power system snapshot (PMU). The database for storing/retrieving result of Modal Analysis can be used to construct decision trees (DTs) for each of the identified CVCAs using key power system attributes. The objective of this paper is to propose an new real-time methodology for identification of CVCAs, which is key function for preventive and remedial actions against instability of the Electrical Power System (EPS). This is carried out under various system operation and contingency conditions by using off-line trained decision trees generated and on-line PMU measurements. Numerical results on the 12-bus test system, shows the suitability and effectiveness of the proposed method.