Neural Network Modeling of Breaking Wave

Wave braking is one of the most complex and important phenomena in the coastal engineering. For this reason, various empirical formulas based on the linear statistical approach have been developed to estimate breaking height and water depth, called to breaker index. This study presents an artificial neural network (ANN) model as an alternative method for predicting breaker index on a gravel and sandy beach. The published available experimental data for sandy beach is used as input system. Further, hydraulic model experiments are performed to obtain the experimental data for gravel beach. A fundamental three-layered feed forward type of network trained using the usual back-propagation training is developed to obtain breaker index from the input of the deep water height, wave period and sea bed conditions. The predicted breaking height and water depth confirmed usefulness of the proposed ANN model for the application of breaking wave.