Abstract
The implementation of renewable energies is among the main challenges that we are confronting in the present situation of climate change. In this work, an artificial neural network (ANN) is optimized and used to assess the wave energy resource available to a wave farm over its service life. We select as case study a stretch of coastline in southern Spain. Different ANN architectures and training algorithms are tested for a dataset in deep water composed by: three values of significant wave height, four values of peak period, two values of incoming wave direction, three astronomical tide values, three storm surge values and three values of sea level rise induced by climate change. These deep-water sea states were propagated using a numerical model (Delft3D-Wave) and results were obtained at 176 locations. Thus, more than 114,000 data were used to train and test the ANNs. Once validated, the ANN was used to assess the cumulative wave energy at 704 locations during a 25-year period for three scenarios of rise in sea level according to the Intergovernmental Panel on Climate Change (IPCC) reports: present situation, pessimistic IPCC projection and optimistic IPCC projection. According to the results, the cumulative wave energy in the case study increases with increasing water depths. The greatest values of cumulative wave energy are reached at great depths off a shoreline horn and a port. Importantly, the rise in sea level will induce an increase in the wave energy resource. The ANN developed in this work allows the quantification of wave energy over long-term periods, reducing the computational cost, as well as the choice of the best locations for wave farms considering the effects of climate change.