Investigation of undertow in reflective beaches using a GMDH-type neural network

Undertow is considered to be one of the dominant mechanisms in the erosion of beaches. In this paper, the distribution of undertow velocity is represented based on experimental evidence. A new polynomial model is suggested to calculate undertow velocity, based on experimental data from the Coastal Laboratory of Kagoshima University in Japan involving regular waves approaching natural and reflective beaches. This study addressed the question of whether GMDH-type neural networks could be used to estimate the undertow velocity based on specified variables. Results indicate that GMDH-type neural networks, in validation with the data obtained from the irregular wave experiments performed at the Hydraulic Laboratory of Imperial College (London, UK), provide an effective means of efficiently (R2 = 78%) recognizing the patterns in data and accurately predicting a performance index based on investigated inputs.