Overtopping Neural Network

Overtopping Neural Network is a prediction tool for the estimation of mean overtopping discharges at various types of coastal structures. The output of this program tool includes the Neural Network prediction of the mean overtopping discharge at a coastal structure and several other parameters indicating the uncertainty of the prediction.

Open Overtopping Neural Network prediction tool. disclaimer NN Web Application

The Neural Network is also implemented in the Deltares software tool BREAKWAT. This tool, which also includes wave overtopping prediction based on TAW (2002) can be used for preliminary design of coastal structures such as breakwaters. Deltares has broad expertise with physical and numerical modelling of wave overtopping and design of coastal structures. More information can be found below:

Deltares expertise: Coastal Structures
Preliminary design breakwaters: BREAKWAT
Physical Modelling: Deltares facilities
Numerical modelling: ComFLOW

The predictions based on the Neural Network can be used for the conceptual design of coastal structures; they may not be used in the final design stage, since the results should be verified based on dedicated physical model tests for the particular wave conditions and structure geometry of the structure to be built. The predictions are based on a data-set based on small-scale physical model tests; the predictions are to some extent affected by model effects, scale effects (see deliverable D40 from the CLASH project), limited accuracy of measurement equipment, limited accuracy of wave generation techniques (compared to nowadays state-of-the-art techniques), inconsistencies in the data-set, and lack of data in certain fields of application. Although reliability levels are given in addition to the predictions, these reliability levels do not account for most of these influences. Therefore, the Neural Network predictions may only be used as first estimates of mean overtopping discharges.

Background

For the prediction of the mean overtopping discharge a Neural Network model is used. This model was derived by WL | Delft Hydraulics (now Deltares) from 8372 input-output combinations obtained from measurements performed in hydraulic scale models at several institutes (Aalborg University, Denmark; Danish Hydraulic Institute, Denmark; WL | Delft Hydraulics, The Netherlands; Hydraulic Research Wallingford, UK; Leichtweiss Institute für Wasserbau, wks+gwk,Germany; Modimar, Italy; University of Edinburgh, United Kingdom; Universidad Politécnica de Valencia, Spain; and others in Iceland, Japan, Norway and U.S.A).

This Neural Network for Wave Overtopping Predictions has been developed by Deltares | Delft Hydraulics and has been co-sponsored by the Commission of the European Communities within the framework of the CLASH project.