Fault section estimation in power systems using an Adaptive Genetic Algorithm

Abstract

This paper proposes a methodology based on the unconstrained binary programming (UBP) model and an Adaptive Genetic Algorithm (AGA) to solve the fault section estimation problem in power systems. The UBP model is formulated using the parsimonious set covering theory for associating the alarms of the protective relay functions informed by the SCADA (supervisory control and data acquisition) system and the expected states of the protective relay functions. The proposed AGA uses only two control parameters and it has automatic and dynamically calibrated recombination and mutation rates based on the saturation of the current population, having an immediate response to possible premature convergence to local optima. Test results for a part of South-Brazilian electric power system have shown that AGA presents robustness, efficiency and less processing time compared with others methods previously published.