Koordinasi Adaptif DOCR Pada Sistem Transmisi Loop Multi Generator Menggunakan Modified Firefly Algorithm-Artificial Neural Network

Abstract

Protection system is a very important component in the electric power distribution system. With good reliability, the distribution of electrical power will be better, one of which is by minimizing disruption to the system quickly, precisely and accurately. Components used in protection systems are directional overcurrent relays (DOCR). In this study proposed optimal and adaptive protection coordination using the Modified Firefly Algorithm-Artificial Neural Network (MFA-ANN) tested on the IEEE 9 real bus loop system with 4 generation combinations. Optimization using MFA will get the Time Delay Setting (TDS) and Ipickup (lowset) values. The ANN used is Backpropagation Lavenberg Marquardt (BPLM) After the BPLM learning process, we will get the weight and bias values of the training results and later as a reference value to get the value of TDS and Ipickup relay. The process will produce relay settings automatically based on the results of optimization and MFA-ANN training which are then simulated on the IEEE 9 bus loop system. The results achieved by MFA-ANN are suitable methods for modeling optimal and adaptive relay coordination systems. ANN training with the BPLM algorithm produces the smallest MSE value of 5.9308xE-10.