Abstract

This work demonstrates a full scale numerical approach to simulating large aggregate erosion protection (rock armour) on coastlines; specifically, the relative performance of armoured beach faces to dissipate wave energy and limit wave attack overtopping. The flow within the interstices of piled Dn-50 masonry rock armour was simulated using Volume of Fluid (VOF) CFD modelling techniques with a RANS approach. The geometry of the rock armour was constructed using DEM particles to achieve a realistic stacking environment and meet a target void fraction. An adaptive mesh was used to track the movement of the fluid free surface. An irregular JONSWAP wave spectra was imposed at the domain boundaries to simulate sea conditions with the resulting overtopping discharge monitored at the rock armour crest. The simulation was run until the mean discharge rate reached asymptotic convergence. An analytical estimate of the mean overtopping discharge was calculated in parallel using an Artificial Neural Network trained on the EurOtop database of wave overtopping tests, which showed close agreement to the results of the CFD model. The forces on each element of rock armour were tracked and plotted to demonstrate peak loads for different wave energies, with the motion and displacement of top layer rocks modelled using DEM particles. The CFD approach developed in this work demonstrates the potential for full scale numerical modelling to be more frequently used as a tool in coastal engineering. High geometry flexibility, the ability to resolve complex filtration hydrodynamics and the ease of recording forces and moments at arbitrary locations are an advantage in the early design stage, and can be delivered within useful timescales. As a practical design tool, the efficiency of the model is paramount; symmetry and strategic meshing must be exploited to reduce CPU time. The primary benefits come from allowing unusual geometries to be modelled where standard formulae are ineffective, and as a cost-effective prototyping tool to develop designs before physical modelling.