SISTEM INFERENSI FUZZY MAMDANI DALAM PREDIKSI PRODUKSI AIR DEMIN MENGGUNAKAN ATURAN BERDASARKAN DECISION TREE

West Java Steam Power Plant (PLTU) Pelabuhan Ratu PGU (Power Generation Unit) PT. PLN Indonesia Power is a plant that has reliable and efficient performance, to get the number one best performance assessment in 2022 with a value of 102.02 To support production, PLTU requires the main fuel, namely demin water to produce pressurized dry steam in boilers. It takes an average of 200 to 800 Tons of demin water to produce electricity on average of 4000 to 7800 MWh (Netto) in one shift for 8 hours. In order to support the reliability of PLTU electricity production, demin water production at the Water Treatment Plant must be maintained according to the request for the use of the generating unit and the level of demin water permitted when operating (at least 10 meters or 2327.4 tons) so that the demin water transfer system does not require a transfer pump so the system runs optimally and efficiently. In this study, computerized calculations were carried out using the Mamdani Fuzzy logic method to obtain predictions of the amount of demineralized water production needed by the generating unit. This method uses the concept of using the concept of a Decision tree decision tree in forming Rules. The rules are made based on criteria in determining the amount of demin water production, including the generation load, demin water demand and demin water stock. The rules formed has an accuracy of 95%. From the results of the analysis of the prediction of demin water production compared to the actual production data has an AFER (Average Forecasting Error Rate) value of 8.80%, then the accuracy of the resulting truth is 91.2%. It can be concluded that this method can be used to predict the amount of demin water needed by a power plant unit.

Keywords - Decision tree, Demin Water, Fuzzy Logic, Mamdani Fuzzy Inference System, Prediction.