Otimizacão da coordenação de relés de sobrecorrente direcionais em sistemas elétricos de potência utilizando a programação inteira binária

Abstract

This work aims to optimize the coordination of microprocessor-based directional overcurrent relays in power systems using Binary Integer Programming (BIP). Two new mathematical models of BIP are presented. The first determines only the Time Multiplier of each relay, while the second determines simultaneously the Time Multiplier and Current Multiplier of each relay. These models have a great advantage over the Linear Programming (LP) and Nonlinear Programming (NLP) models to determine the relay settings directly within the range provided by these instead of the LP and NLP which use continuous variables. Thus, it avoids the rounding of settings to the closest values available in the relays, which can cause failures in coordination. Still, the algorithms used to solve these BIP models do not require an initial guess, unlike what happens in NLP, and the search from getting trapped in local minima. This paper presents the NLP and LP models considered and the necessary changes in order to obtain the two new BIP models. To validate the new mathematical models of coordination of overcurrent relays and compare them with the models which use continuous variables, the proposed methodology is applied in phase and earth protection of two test systems of different sizes considering whether or not the instantaneous unit of each relay. The results are evaluated in terms of the Objective Function, the obtained settings and operating times of relays for faults within the zone of primary protection. Thus, it is shown that the proposed models have the ability to determine the optimal solution of the problem in a reduced computational time and without the need to make any changes to the final solution. These models can also integrate a software aid to decision making by the protection engineer, allowing to interact in the construction of mathematical model to customize the final solution.