B1554
Title: Modeling wind energy with ANFIS with different membership functions
Authors: Gultekin Atalik - Anadolou University (Turkey) [presenting]
Sevil Senturk - Anadolu University (Turkey)
Abstract: The importance of renewable energy sources is increasing day by day due to decreasing of fossil fuels. Wind energy is a kind of green energy sources. Modelling wind energy gains importance because of global warming. There have been done many studies on wind energy. Turkey has a big wind energy potential. So, it is important to model and estimate wind energy in Turkey. Adaptive Neuro Fuzzy Inference System (ANFIS) is an adaptive technique that combines neural network with fuzzy set theory. ANFIS uses a hybrid algorithm. That algorithm combines a backpropagation algorithm and the least squares method to make an estimation. There are many membership functions in ANFIS. These membership functions affect the model performance of ANFIS. We try to model wind energy with different membership functions to determine the best model for wind energy by using ANFIS method. Finally, the best model is determined via goodness of fit criteria.
