A Wavelet Decomposition Approach to Improve Modelling Performance of Artificial Neural Networks in Solar Energy Research
Abstract
In solar energy research, estimating solar potential over a location of interest is an important step towards the successful planning of renewable energy projects. However, solar data are not available for every point of interest due to the absence of meteorological stations and sophisticated solar sensors. Therefore, the solar radiation has to be estimated with accurate solar radiation models. This paper presents a hybrid technique to increase the modeling performance of artificial neural networks (ANN), commonly used in solar radiation applications. A wavelet multiresolution is performed to decompose the meteorological data into different time and frequency scales before it is presented to the ANN models. As a result, improving the learning process of the ANN models. The proposed approach is compared to traditional ANN model and validated using well-known statistical validation metrics including coefficient of determination (R2) and root mean square error (RMSE). In addition, wavelet cross spectrum (WCS) is used as a visual indicator of the model performance in time, frequency, and phase domains. The results show that the modeling performance of the ANNs is considerably improved from R2 of 89.21% to 96.34%.