<![CDATA[The Rembang Coal-Fired Power Plant (PLTU Rembang), with a capacity of 2 x 315 MW, is a key power plant in Central Java, where fuel expenses represent the largest cost component. Accurate fuel procurement planning, which relies on projecting electricity sales, is essential to reduce these costs. This study develops and compares four machine learning-based Capacity Factor (CF) prediction models: random forest regression, support vector regression, multiple polynomial regression, and multiple linear regression. The independent variables are selected from internal and external sources using F-tests and t-tests. Among the four models, the multiple linear regression model demonstrated the smallest Mean Absolute Percentage Error (MAPE) of 7.83%. Using this model, the annual CF for PLTU Rembang in 2024-2026 is predicted to be between 82% and 84%, while the CF for February-June 2024 is expected to range from 87% to 91%. With a monthly CF prediction accuracy classified as very good (MAPE of 2.35%), these predictions are valuable for optimizing monthly fuel purchase allocations, considering initial fuel stock and target inventory age (17-30 Days of Plant Operation).]]>
Copyrights © 2025
