Moch. Rizal Ramadhan, Sujito, Rajib Muhammad Basthomy, M. Fazlur Rahman Addakhil, Tito Al Afrin Uwais, Moh. Zainul Falah
Electricity, as a primary energy need, experiences increasing demand each year. Load forecasting becomes crucial to maintain a balance between electricity demand and supply, especially in areas with heterogeneous customer characteristics such as the operational area of ULP Lawang. The heterogeneous customers in this context refer to both residential and large industrial sectors. This study implements Double Exponential Smoothing (DES) on historical feeder load data, with the smoothing parameters α and β optimized using Golden Section Search (GSS), based on minimizing error values according to the Mean Absolute Percentage Error (MAPE) standard. The results indicate that Golden Section Search (GSS) produces more optimal parameters compared to conventional approaches, achieving higher forecasting accuracy. Out of eight feeders analyzed, six showed load growth, while two exhibited a decrease. The Kebun Teh feeder recorded the most significant load increase. Overall, the forecasting results indicate that peak loads remain within the safe limits according to cable capacity standards set by the International Electrotechnical Commission (IEC). © 2025 IEEE.
Universitas Negeri Malang, Dept. Electrical Engineering and Informatics, Malang, Indonesia; Chaoyang University of Technology, Dept. Information and Communication Engineering, Taichung, Taiwan
