Unit Three Kartini, M. Nur Adiwana
This paper proposes a novel hybrid methodology for short-term forecasting of hourly. The F-Multi Criteria Decision Making-Neural Network (F-MCDM-NN) hybrid model is designed to forecast Global Horizontal Irradiance (GHI) 1 day ahead based on Meteorological data for the target photovoltaic generation. The proposed hybrid modeling is based on meteorology data i.e. wind speed, wind direction, pressure, rainfall, temperature, humidity, gust, and visibility, data especially for optimizing the operation of power generating electricity from photovoltaic (PV) generation system. This modeling is a hybrid of Fuzzy-Multi Criteria Decision Making (F-MCDM) algorithm modeling and Neural Network (NN) model. The novelty of this combination model is taking into account the meteorology data. A set of data on GHI was from the PV station in Indonesia, which is used as research test data. The first hybrid model implements F-MCDM as an input data preprocessing technique prior to the NN method. The error statistical indicators of the F-MCDM-NN hybrid model using the Mean Absolute Percentage Error (MAPE) is 7.5% and the Root-Mean-Square Error (RMSE) is 73.7 W/m2. The hybrid model prediction is then compared to measured data and simulation results indicate that the F-MCDM-NN-based hybrid model presented in this research can calculate hourly GHI with satisfactory accuracy with better result compared to the Neural Network method. © 2024 Praise Worthy Prize S.r.l.-All rights reserved.
Department of Electrical Engineering, Universitas Negeri Surabaya, Indonesia
