Abstract
One of the technical aspects that support the optimal operation planning of a power plant when viewed in terms of system reliability and economic is about short-term load forecasting. The objective of this research is to forcasting hourly short-term electric load peak (17:30 to 22:30 GMT) at loading area of Transmission Distribution Banda Aceh Unit of PT. PLN P3B Aceh 150-20 kV by using Adaptive Neuro Fuzzy Inference System (ANFIS) method. The toolbox used to predict short-term electric load in this research is by using MATLAB software R2007b and Microsoft Excel 2007. ANFIS structure is trained using ANFIS Sugeno models, three types of membership functions with three and four fuzzy sets for each type of membership function. ANFIS structure is trained using a hybrid algorithm. From the simulation results obtained that the structure of the input membership functions of ANFIS 3 gbell with three fuzzy sets as the ideal structure. Further results of ANFIS estimation compared with the moving average method. From the simulation results is shown that ANFIS models generate MAPE 3.42%, while the forecasts using the moving average method generate MAPE 6.58%.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
