The keynote talk will cover different aspects of scour and rock scour protection encountered in offshore wind engineering. Scour in the marine environment has been studied since the 1980's. Physical model tests of pipeline scour and scour around vertical piles are reviewed extensively in Sumer and Fredsøe (2002). Computational Fluid Dynamic (CFD) morphological modelling of flow and scour processes took off in the 1990's, with the first comprehensive CFD results of pile scour reported in Roulund et al (2005) for steady current. CFD modelling has progressed significantly since but remains challenged by long computational times. The recent application of 3D shallow water non-hydrostatics solvers used in Margalit (2017), may bridge the gap between field scale conditions and acceptable computational time. Scour scales with structural dimensions. With the ever-increasing size of turbines and foundations, the magnitude of scour holes has in more and more cases required prevention in form of scour protection. Crushed rock is by far the most used type of scour protection in offshore wind. Rock is a construction material to be specified, produced and tested to ensure a quality product. Standard gradings and test methods are given in the Rock Manual (C683 – CIRIA et al, 2007) or the European Standards EN-13883:2002 part 1&2. The hydraulic stability of rock structures is a function of wave and current exposure. For scour protections the wave loading is at the seabed, far away from the surface waves. The so-called near bed wave orbital velocity, Um , replaces significant wave height, Hs as the governing seastate parameter. Um basically combines wave height, – period and water depth of irregular waves to a single seastate parameter acting at the seabed. The so-called Shields parameter describes the balance between hydraulic loading and rock weight. For irregular waves in combination with current, Roulund et al (2016) suggest a parametric approach for calculation of the Shields parameter of a given seastate. Rock berms are typically applied in offshore wind for separation and coverage at cable crossings and shallow buried cables. In Roulund et al (2017), it is demonstrated that the Shields parameter can be applied to describe the hydraulic stability and gradual damage development for rock berms, allowing use of relatively small rock gradings for berm construction material. Figure 1 shows a pre-laid rock protection on the seabed before the foundation pile is driven. Installation of scour protections and rock berms is typically carried out by so-called fall pipe vessels. The circular rock pads of scour protection may cause edge scour (Petersen et al, 2015). Rock berms installed as cable crossing protection have in areas of strong currents been observed to cause severe scour (Roulund et al, 2018). The berm alignment relative to current direction was found to be a key parameter for scour development.

Rock berms or concrete mattresses have traditionally been applied for protection of shallow buried, surface laid or crossing pipelines and cables. In most cases the installed protection has worked satisfactorily, but under certain adverse conditions, field evidence has shown examples of severe scour both at mattresses and rock berms. The paper presents and discusses physical model tests of rock berm scour in combined waves and current. Both single and double berm configurations were studied. The conclusions from the tests were two-fold addressing: 1) Scour development caused by rock berm(s), and 2) Structural robustness of the rock berm when subject to wave action and seabed model scale effects. The model tests support field evidence that scour is prone to develop in strong current environments and when the protection is aligned oblique or perpendicular to the prevailing currents. The tests also demonstrate larger structural integrity of wide crested berms as opposed to sharp crested berms.

A fundamental requirement for any scour assessment and scour protection design is the ability to determine the Shields parameter for combined wave and current conditions.The Shields parameter can be calculated for current combined with monochromatic waves using the approach of Soulsby (1997) in combination with the wave friction factor concept. For current in combination with irregular waves, the same approach is suggested using a wave orbital velocity, U-m, for representation of the irregular sea state. U-m is defined as 1.41 times the standard deviation of the near bed wave orbital velocity.The Soulsby (2006) expression for U-m is compared with a hyperbolic expression and validated using numerical methods and laboratory measurements.A large number of expressions exist for the wave friction factor as a function of relative bed roughness. From a literature study, the paper proposes a combination of existing expressions to cover relative bed roughnesses from sand over gravel to coarse armour rock.

The Race Bank Offshore Windfarm is part of the UK Round 2 extension. The windfarm comprises 91 6.3 MW turbines and two substations with total installed power of 573 MW. Construction work is scheduled for 2016/17. The windfarm is located in UK waters in the southern part of the North Sea about 30 km north of the Norfolk coastline and 30 km east of Lincolnshire. The Race Bank site has a diverse morphology located on large sandbanks with mobile sandwaves and megaripples and intersected by deeper featureless swales where clay soil formations are found close to or at the sea floor. The morphological complexity of the site was recognised early in the project development and numerous bathymetric surveys have been undertaken to help quantify sandwave migration rates and expected seabed variability. Morphological seabed assessments have been conducted by ABPmer throughout the project development. In the detailed design phase the ABPmer assessments have been combined with in-house DONG Energy analysis to provide design seabed levels and anticipated ranges of natural seabed changes within the lifetime of the windfarm. The paper presents an overall morphological characterisation of the windfarm site and provides examples of the very diverse seabed morphology encountered. When quantified, the morphological diversity can be absorbed in the monopile foundation design and cable installation contractors can select appropriate installation tools. The paper thus comes with a recommendation to ensure that bathymetric surveys in morphological active sites are carried out early and repeatedly during the project development stages.

The flow around a vertical circular pile exposed to a steady current is studied numerically and experimentally. The numerical model is a three-dimensional model. The model validation was achieved against new experimental data (which include two-component laser-Doppler anemometry (LDA) flow measurements and the hot-film bed shear stress measurements, and reported in the present paper) and the data of others, and a k-omega turbulence model was used for closure. The model does not have a free-surface facility and therefore is applicable only to cases where the Froude number F is small (Fr < O(0.2)). The flow model was used to study the horseshoe vortex and lee- wake vortex flow processes around the pile. The influence on the horseshoe vortex of three parameters, namely the boundary-layer thickness, the Reynolds number and the bed roughness, was investigated. In the latter investigation, the steady solution of the model was chosen. A study of the influence of the unsteady solution on the previously mentioned flow processes was also carried out. The ranges of the parameters covered in the numerical simulations are: The boundary-layer-thickness-to-pile-diameter ratio is varied from 2 x 10(2) to 10(2), the pile Reynolds number from 10(2) to 2 x 10(6), and the pile diameter-to-roughness ratio from 2 to about 10(3). The amplification of the bed shear stress around the pile (including the areas under the horseshoe vortex and the lee-wake region) was obtained for various values of the previously mentioned parameters. The steady-state flow model was coupled with a morphologic model to calculate scour around a vertical circular pile exposed to a steady current in the case of non-cohesive sediment. The morphologic model includes (1) a two-dimensional bed load sediment-transport description, and (11) a description of surface-layer sand slides for bed slopes exceeding the angle of repose. The results show that the present numerical simulation captures all the main features of the scour process. The equilibrium scour depth obtained from the simulation agrees well with the experiments for the upstream scour hole. Some discrepancy (up to 30 %) was observed, however, for the downstream scour hole. The calculations show that the amplification of the bed shear stress around the pile in the equilibrium state of the scour process is reduced considerably with respect to that experienced at the initial stage where the bed is plane.

A 3-D flow code, EllipSys3D, tested and validated, has been implemented along with a morphologic model to simulate the scour process around a vertical circular pile in a steady current in the case of non-cohesive sediment. The k-omega turbulence model has been used for closure. The morphologic model includes (1) a two-dimensional bed load sediment transport description, and (2) a description of surface-layer sand slides for bed slopes exceeding the angle of repose. The simulation captured all the bed features, i.e., the scour hole and the formation of a downstream dune at the initial stage, and the truncated cone-shaped scour hole in the equilibrium stage. The maximum equilibrium scour depth obtained from the simulation compares fairly well with the measurements.

This paper deals with scour around a circular pile exposed to a steady current. A 3D numerical model incorporated with the k-omega,SST closure coupled with the sediment-continuity equation and a bedload sediment transport formula has been used to predict the scour. 3D calculations have also been carried out for a plane rigid bottom for reference purpose. The predicted flow features are apparently in fairly good agreement with the experimental data. Early calculations indicate that the model is able to predict the scour properties satisfactorily in the initial stages of the scour process, up to scour depth of 0.6-0.7 times the pile diameter. Calculations that describe the entire scour process (including the equilibrium stage) are underway.