Derivation of Wave Overtopping Formulas for Vertical Walls Using GMDH Algorithm

Abstract

To measure the wave overtopping rate over a coastal structure, hydraulic experiments have been conducted, the results of which have also been used to develop empirical formulas. Recently, by collecting a large number of data from many hydraulic tests, the CLASH database was constructed, which was also used to develop machine learning models, e.g. EurOtop-ANN model. In this study, a different machine learning model is developed for wave overtopping over a vertical wall by using the group method of data handling (GMDH) algorithm with the CLASH and new EurOtop datasets. The developed model is presented by a series of formulas, which in turn is programed in a spreadsheet calculator, so that engineers without deep knowledge of machine learning methods can easily use the formulas. The developed formulas are more accurate than other empirical formulas and machine learning models. The estimation errors of 95% confidence interval are also given.