Unravelling sediment transport driven by a multimodal wind-wave spectrum

Unravelling sediment transport driven by a multimodal wind-wave spectrum

Hoogervorst, Chris (TU Delft Civil Engineering and Geosciences)

Antolínez, Jose A.A. (mentor)
Aarninkhof, S.G.J. (mentor)
Tissier, M.F.S. (mentor)
Portilla‐Yandún, J. (graduation committee)

Delft University of Technology

Civil Engineering

2022-06-24

This thesis analyses the importance and application of considering multiple coexisting wave trains in sediment transport predictions.

The wave trains at offshore and nearshore locations are analyzed by wave spectral partitions (Portilla at el., 2009). The temporal variability of consistently occurring wave trains can be analyzed through so-called wave families (Portilla et al., 2015). The wave partitions show that a unimodal wave spectrum occurs 35% of the time in front of the coast of the province of North Holland in the Netherlands. This means an error is introduced 65% of the time as unimodal spectra are assumed in traditional methods that use generalized wave parameters to estimate sediment transport. When the wave field is not unimodal, the angle between wave trains is considerable. Only 39% of the time is the angle between a wave train and the most energetic wave train smaller than 60°. For 43% of the time, the angle is between 60° and 120°. The remaining 17% are wave trains propagating in opposite directions. The difference in wave angle is half of the time larger than 50° when a threshold of one meter of deepwater significant wave height over the whole spectrum, defined by Holthuijsen (2007), is used.

Unsupervised machine learning techniques Principal Component Analysis, Maximum Dissimilarity Analysis, and the kmean clustering algorithm are used to group similar wave spectrums over time, spatially group wave families, analyze the wave climate's spatial-temporal variability, relate wave - and weather conditions, and assess their impact on wave-driven potential sediment transport. The Principal Component Analysis is used as a dimensionality reduction technique. This study shows the physical meaning of the Principal Components and their temporal variability in the wave frequency-directional spectrums, where Principal Components describe wave energy in each frequency-directional bin. The kmean clustering algorithm groups similar wave spectrums over time and gives insight into wave conditions, like calm and stormy conditions, called Sea States. The Maximum Dissimilarity Technique changes the focus between common conditions or extreme events. The wave partitions, wave families, and Principal Components give additional insight into the Sea States.

Potential sediment transport estimations are made by traditional methods that use a representative wave frequency and direction and through vectorizing and superpositioning the estimations by wave partitions and wave families. The outcomes show that it is challenging to predict sediment transport for multimodal wave conditions with state-of-the-art sediment transport formulations and models. The results show that the wave period significantly impacts sediment transport. The sediment transport for coexisting wave trains causes about 58% of the time more sediment transport than traditional methods. Here the not considered interaction between wave trains on bedload and sediment suspension will enhance sediment transport. Furthermore, the sediment transport estimations give insight into the wavefield and weather conditions that contribute most to the yearly wind-wave induced potential sediment transport caused by high energetic or consistently occurring conditions.

Sediment transport
Principal Component Analysis
multimodal wave conditions
Maximum Dissimilarity Analysis
kmean clustering
wave partitions
wave families
ESTELA
synoptic atmospheric pattern
frequency-directional spectrum
wave train

http://resolver.tudelft.nl/uuid:fa4d3279-3443-41bc-a995-6d81f34d6936

2023-06-30

53.196594, 4.478243

Student theses

master thesis

© 2022 Chris Hoogervorst