Performance evaluation of hybrid wind power forecasting models based on the wavelet decomposition techniques

Auto Regressive Moving Average (ARMA) models, which perform a linear mapping between inputs and outputs, and Artificial Neural Networks (ANNs) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS), which perform a non-linear mapping, provide a robust approach to wind power prediction. In this paper a new hybrid method is proposed, based on the application of the six Daubechies wavelet employed to do the 3rd level discrete wavelet decomposition of the original hourly wind power time series, in combination with ANNs, ARMA and ANFIS models, in order to predict the power production of a wind farm located in Southern Italy in different time horizons: 1, 3, 6, 12 and 24 hours. In particular, the results obtained with and without the wavelet decomposition are compared for each of the aforementioned techniques (ANNs, ARMA and ANFIS), by investigating the error of the different prediction systems for various forecasting horizons; the statistical distributions of the error are calculated and presented