An artificial neural network model of coastal erosion mitigation through wave farms.

In this work, a novel approach based on artificial intelligence (AI) to assess the efficiency of wave energy converter (WEC) farms in coastal protection isdeveloped. We consider as a case study a beach subjected to severe erosion: Playa Granada (S Spain). More specifically, we analyse the changes in the dry beach area (quantified through the Pelnard-Considère equation) with and without wave farm protection by means of an Artificial Neural Network (ANN) model. The model is selected after a thorough comparative study involving forty ANN architectures, with one and two hidden layers, and two training algorithms (Levenberg-Marquadt and Bayesian regression). The best results are obtained with a [5-10-1] architecture trained with the Bayesian regression algorithm. Once validated, this ANN model is applied to optimize the design and position of the wave farm. The results confirm that ANN models are a useful design tool for hybrid wave farms. • Development of artificial neural network (ANN) model to design wave farms. • Dual function of wave farms: carbon-free energy production and coastal protection. • ANN with Bayesian algorithm and a [5-10-1] architecture successfully validated. • ANN model applied to find the optimum wave farm design for an eroding beach. [ABSTRACT FROM AUTHOR]