978-1-5386-3917-7/17/$31.00 ©2017 IEEE Considering variations of network topology in optimal relay coordination using Time-Current- Voltage characteristic Navid Bayati, Akbar Dadkhah, S. H. H. Sadeghi, Behrooz Vahidi Department of Electrical Engineering Amirkabir University of Technology Tehran, Iran N-bayati@aut.ac.ir Ali Eftekhari Milani Department of Electrical Engineering Politecnico di Milano Milan, Italy ali.eftekhari@mail.polimi.it Abstract— Power systems are usually subjected to some changes in connections during operation, which leads to varied short-circuit levels in different parts of them and Directional Overcurrent Relays (DOCRs) may be subjected to miscoordination by changing short circuit levels. Therefore, all network topologies should be considered for coordination among overcurrent relays in a power system. Reduction of operating time of relays is highly important in a protection system. This paper uses the overcurrent relays with a time-current-voltage curve for coordination among overcurrent relays with respect to all the different topologies. This improves coordination results of the relays due to increasing number of variables and considering voltage of the relays. This paper uses a combination of Genetic algorithm (GA) and Linear Programming (LP) for improving operating time and setting current due to the increased number of constraints for coordinating the overcurrent relays. The proposed method was implemented for an 8-bus system. Optimization results show a considerable reduction in the operating time of overcurrent relays and number of miscoordination, compared to conventional relays and the conventional optimization methods. In addition, coordination among DOCRs is maintained by changing topology of network. Keywords— overcurrent relay, hybrid GA, linear programming, relay coordination, non-standard curve component I. INTRODUCTION Overcurrent relays are responsible for determining and excluding the sections with short circuit and error in a power system within the shortest time. One or more backup relays are considered in the coordination of overcurrent relays for improving reliability and providing options for each relay. Therefore, one or more variables are for any overcurrent relay based on an overcurrent relay curve. Optimization methods are usually used for determining values of variables. Optimization mainly aims at minimizing operating time of the overcurrent relays through maintaining coordination among them.in [1-5] Optimization methods such as GA, Particle Swarm Optimization (PSO) and Hybrid Particle Swarm Optimization– Gravitational Search Algorithm. Have been employed for coordination among the overcurrent relays. Linear programming (LP) was used in [6] for coordinating DOCRs in which only a linear variable is considered. Non-linear programming (NLP) was used in [7] for coordinating DOCRs. A power system is usually considered fixed for coordinating among DOCRs; however, coordination among DOCRs may subject to changes due to power line outage under real conditions. The coordination of overcurrent relays in [8] was performed with respect to a change in network topology. In [9] DOCR coordination was performed with considering short circuit level changes. Power systems may also have distribution Generation resources (DGs). Short circuit levels in a system change with the DGs’ performance level changing. Relay coordination considering changes and elimination of the DGs in the power system were considered in [10]. Various characteristic curves have been used for reducing operating time of relays and the miscoordination among relays. Five different variables were used in [11] for coordinating relays. Consequently, fitness function will be reduced due to increasing number of variables, but optimization time and divergence increase. Therefore, combined optimization methods, such as GA and LP, were used [12]. Usually, time-current curves are considered for DOCRs curves in relay coordination studies. Currently, a relay curve can be changed because the relays are microprocessor-based. A logarithmic curve was used in [13], which reduced operating time of relays. The effect of voltage of buses was also considered in DOCR curve. In other words, the time-current-voltage curve was introduced for the relays. Three variables were considered due to number of variables in coordination of DOCRs and this method reduced operating time of relays. The curve of overcurrent relays in this paper was considered as time-current-voltage with three settable variables in it. GA- LP method was used for coordinating DOCRs. The topology of a variable network was also considered to ensure coordination of relays in all topologies. The proposed method reduces optimization time and operating time of the relays. A comparison is also made between the advantages of the proposed method with the conventional relays using the GA-LP and GA methods. The comparison expresses that the proposed