PERAMALAN BEBAN LISTRIK JANGKA PENDEK DENGAN METODE ADAPTIVE NEURO FUZZY INFERENCE SYSTEM (ANFIS) di PT. PLN (PERSERO) WILAYAH SURABAYA UTARA

An important process in the operational management of an electric power system is short-term electric charge prediction, primarily to ensure adequate power availability and optimize resource use. The Adaptive Neuro Fuzzy Inference System (ANFIS) method in P.T. PLN (Persero) North Surabaya Region was used in this study to develop a short-term electric charge prediction model. In the process of developing an ANFIS model, historical electrical load data is used as an input. Training and model adjustment processes are carried out to improve adaptability to complex patterns found in electrical load data. Testing and validation is carried out with independent data to evaluate the performance of the model in predicting future electrical loads. The results show that the ANFIS model can predict short-term electric loads with a high degree of accuracy and a low error rate. For 2019–2023, the predictions generated using MAPE and RMSE metrics are below 1%, indicating that the ANFIS model has an excellent ability to predict test data because it provides prediction accuracy below 10%. The model accurate rate can be seen on the graph on the curve. In addition, this research contributes to the development of better predictive techniques that will help decision-making processes in the power industry