STUDI PERAMALAN BEBAN PUNCAK WILAYAH MAKASSAR MENGGUNAKAN ARTIFICIAL NEURAL NETWORK

Urban load monitoring is useful for integrating environmentally friendly distributed energy sources with power plants. The information will help treat district power plants like virtual power stations for effective coordination with high capacity power plants. As a reference in planning, both in planning in operations and in planning for the development of electricity, one of the most important things is to carry out proper monitoring (forecasting) in order to find out the electricity needs in a certain period of time (Supranto. J, 1981). Monitoring is an activity or effort to estimate how the system will be in the future based on the previous time parameters. The purpose of this study is to determine the peak load in the next 4 months, obtained from the load monitoring simulation using the Artificial Neural Network method and also to determine the accuracy of the monitoring results using the Artificial Neural Network method using MSE calculations and the error value. ANN calculation method is applied to predict the magnitude the peak of load in the Makassar area. Data used is the peak load data for the Makassar region from January to December 2016. This calculation method produces an MSE Training of 0.049, using the TRAINGD calculation method with 5 neuron LEARNGD training adaptations. Find the results of the peak electricity load predictions from January to April 2017 of 976,92MW; 951,92MW;949,96MW; dan 1011,57MW