Adaptive Coordination of Time Overcurrent Relays (OCR) in a Radial Distribution Network with Distributed Generation using a Modified LINKNET Pairing Algorithm

With the increasing demand of the grid as of today, power system structures continuously complexing that's why protection and coordination is a seemingly tedious task for relay engineers. This research aims to design an adaptive coordination that when a change in network topology occur, relay engineers can learn the primary and backup relay pairings of the power system and the relay settings fit for these pairing. The proposed design modified the LINKNET algorithm to adapt in a change in network topology that is originally designed for a fixed topology. The program is coded using Python programming language which is best for adapting Artificial Neural Network on the system, especially for the relay setting optimization. Simulated over 32 test cases of n - 1 contingency, the program was able to provide the primary and backup relay pairings and the optimal relay settings. The tests backed-up by statistical analysis using Mann - Whitney U test provided results of the time of operation for some cases are lower than 0.05 and some cases are larger than 0.05 with the reference case without the n - 1 contingency or no change in network topology. The program created was able to provide the desired relay pairings with optimal settings fir for coordination. This program helps relay engineers to ease their tasks as complex power systems increasingly prevalent. The program provided results that are validated in ETAP illustrating Time - Current Characteristic Curve that shows that when primary protection fails to trip, backup relays were the next to execute.