Coastal structures are often designed to a maximum allowable wave overtopping discharge, hence accurate prediction of the amount of wave overtopping is an important issue. Both empirical formulae and neural networks are among the commonly used prediction tools. In this work, a new model for the prediction of mean wave overtopping discharge is presented using the innovative machine learning technique XGBoost. The selection of features to train the model on is carefully substantiated, including the redefinition of existing features to obtain a better model performance. Confidence intervals are derived by tuning hyperparameters and applying bootstrap resampling. The quality of the model is tested against four new physical model data sets, and a thorough quantitative comparison with existing machine learning methods and empirical overtopping formulae is presented. The XGBoost model generally outperforms other methods for the test data sets with normally incident waves. All data-driven methods show less accuracy on oblique wave data, presumably because these conditions are underrepresented in the training data. The performance of the XGBoost model is significantly improved by adding a randomly selected part of the new oblique wave cases to the training data. In the end, this new model is shown to reduce errors on all data used in this work with a factor of up to 5 compared to existing overtopping prediction methods.

Wave overtopping is an important design criterion for coastal structures such as dikes, breakwaters and promenades. Hence, the prediction of the expected wave overtopping discharge is an important research topic. Existing prediction tools consist of empirical overtopping formulae, machine learning techniques like neural networks, and numerical models. In this paper, an innovative machine learning method-gradient boosting decision trees-is applied to the prediction of mean wave overtopping discharges. This new machine learning model is trained using the CLASH wave overtopping database. Optimizations to its performance are realized by using feature engineering and hyperparameter tuning. The model is shown to outperform an existing neural network model by reducing the error on the prediction of the CLASH database by a factor of 2.8. The model predictions follow physically realistic trends for variations of important features, and behave regularly in regions of the input parameter space with little or no data coverage.

A method has been developed to estimate wave overtopping discharges for a wide range of coastal structures. The prediction method is based on Neural Network modelling. For this purpose use is made of a data set obtained from a large number of physical model tests (collected within the framework of the European project CLASH, see e.g. [Steendam, G.J., Van der Meer, J.W., Verhaeghe, H., Besley, P., Franco, L. and Van Gent, M.R.A. (2004). The international database on wave overtopping. World Scientific, Proc. 29th ICCE, vol. 4, pp. 4301-4313, Lisbon, Portugal.]). Moreover, a method was developed to obtain confidence intervals for the overtopping predictions of the neural network.

A method has been developed to estimate wave overtopping discharges for a wide range of coastal structures. The prediction method is based on the technique of Neural Network modelling. For this purpose use is made of a data set obtained from a large number of physical model tests. Moreover, a method was developed to obtain the confidence intervals around these predictions.

For the design of vertical breakwaters, reliable predictions of the horizontal forces are required. Through physical modelling useful predictions can be made but, due to the very complex interaction between waves and structures, the derivation of reliable empirical relations based on such tests can still be rather difficult. Here, based on a large data-set from physical model tests, use is made of Neural Network modelling to predict horizontal forces on vertical structures. In addition, a method is developed to estimate the reliability-intervals around the predictions. The resulting tool is a complementary design-tool for predicting forces on vertical breakwaters and also a suitable tool for application in probabilistic design methods.