The goal of this work is to present a synthesis of the improvements and updates developed to deliver the final version of the ANN tool adopted by the second edition of the wave overtopping manual, EurOtop, released on the internet in 2016. This tool consists of three identical but independent ANNs able to predict the main parameters representative of the wave-structure interaction processes, i.e. The mean wave overtopping discharge, the wave transmission and the wave reflection coefficients. The contribution focuses on the modifications of the ANN architecture carried out since the last ICCE conference to achieve an optimized representation of the wave overtopping, especially in case of low and extreme values of the overtopping discharge. The consistency of the ANN predictions is assessed through an artificial dataset including geometrical and climate input parameters that are varied with continuity, while the robustness of the tool is checked by applying the ANN to selected geometries excluded from the training database.@en

Most of the experience collected on wave overtopping is concentrated on emerged or zero-freeboard structures, and only few experiments are available on over-washed dikes. The experience on submerged crest conditions is principally related to rubble mound structures and limited to the wave reflection and transmission processes. The existing tools for the prediction of the average wave overtopping discharge (q) at dikes are therefore targeted to represent overtopped structures, leading to cautious estimates at over-washed dikes and unrealistic overestimations in case of catastrophic scenarios such as breached dikes. This paper proposes a new conceptual and practical method for the prediction of q that is valid for both over-washed and fully breached dikes. It consists of three physically-based formulae whose coefficients are fitted on a new numerical database of smooth structures under variable wave attack. The predictions are compared with the new numerical results and the data and the tools available from the literature, i.e. experimental data, formulae and neural network. The new method gives a more realistic representation of the overtopping and overflow processes, reducing significantly the overestimation in case of breaching and being at least as accurate as the existing tools in case of over-washing.@en