Coordenação de relés de sobrecorrente em sistemas elétricos de potência utilizando metaheurísticas bio-inspiradas: uma abordagem considerando o algoritmo de busca dos cucos via voos de lévy

Abstract

The integrity and robustness of the Electric Power System (EPS) are guaranteed by relaycontrolled protection schemes. The function of these relays is to identify anomalies, locate them and alert the operation of the system, triggering alarms, signals and commands. The primary protection relay acts as the first assistance mechanism against system faults, the back-up protection should only act in the event of irregularity and non-operation of the primary system. One of the problems arising from the use of overcurrent relays (OCRs) relates to their coordination, wich the function to obtain the optimum configuration so that faults are quickly extinguished wherever they occur, even when the relay closest to the fault fails. Therefore, the optimization of OCRs coordination is defined as a problem of minimizing the operating time of the relays under selective conditions. The necessary motivation for this research was found, which aims to approach, evaluate and validate the applicability of bio-inspired metaheuristics to solve the overcurrent relay coordination problem and compare it with other existing methodologies reported in the specific literature. This work was developed at UNIPAMPA facilities, being the material used made available by the Gesep, the proposal contemplates the mathematical formulation of the overcurrent relays coordination problem in four different test situations: a radial distribution system, a second system with 3 power generation fonts, an approach involving a realistic circuit, and finally, the complex 9-bus system. The OCRs used follow the IEC Standard of normal time curve, and the coordination solution is obtained through the use of mathematical optimization techniques known as metaheuristics, inspired by the behavior of elements of nature: Firefly Algorithm, Particle Swarm Optimization and Cuckoo Search Algorithm. The implementation and simulation of the were is done in Matlab. The results found with the tested methods are satisfactory and show the possibility of reaching solutions of equal or superior quality when compared to studies of other methodologies available in the specific literature. Among the three implemented methodologies, the adjustments obtained via Cuckoo Search reported the best quality solutions for the parametrization in the protection relays, of the simulated cases. As the complexity of the problem increases, with the addition of restrictions and decision variables, and, especially, for non-linear programming problems, heuristics emerge as the most viable alternative in the search for the best solution.