An artificial neural network tool to estimate wave overtopping at vertical walls with stilling wave basin structures

Date

2023-12-06

Author

Uzun, Utku

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

70
views

50
downloads

Cite This

Estimating wave overtopping values in coastal defense structures is a very important phenomenon in coastal engineering. Due to the complex behavior of waves, most of the existing wave overtopping estimation methods are empirical and valid for limited conditions. As a new design approach, artificial neural network (ANN) tools are started to be developed. The available ANN tools such as The Overtopping-Neural Network tool can calculate overtopping values at sloped rubble mound structures, and simple vertical walls. However, it cannot calculate wave overtopping values at vertical walls that have bullnose, promenade, and stilling wave basins (SWB). SWB, in shorelines where social and recreational facilities are done regularly, can be an efficient way of designing sea defense structures, since it can result in lower wave overtopping results in shorelines with lower crest level values. In this study, an ANN tool that calculates overtopping discharges at vertical seawalls with SWB is developed. A database for training and testing of the ANN is collected. The most efficient architecture was analyzed and selected with Grid Search method. The new ANN tool is evaluated against wave overtopping discharge empirical formulas in the literature and The Overtopping-Neural Network tool adopted by EurOtop (2018). The new ANN tool shows very good performance with high accuracy metrics on both plain vertical walls and vertical walls with SWB. The new ANN performed better than The Overtopping-Neural Network on plain vertical walls. In addition, the new ANN tool performed better than empirical formulas on plain vertical walls and as well as empirical formulas on vertical walls with SWB.

Subject Keywords

Wave Overtopping, Machine Learning, Artificial Neural Networks, Stilling Wave Basin

URI

https://hdl.handle.net/11511/107769

Collections

Graduate School of Natural and Applied Sciences, Thesis