Numerical simulation of scour below pipelines using flexible mesh methods

Juan Núñez, James Percival, Benjamin Yeager, S.j. Neethling, Matthew D. Piggott2016 · DOI

TL;DR

This paper presents a novel 2D numerical model using flexible, adaptive meshes to simulate seabed scour beneath pipelines. By coupling fluid dynamics with sediment transport and employing dynamic mesh refinement, the model accurately captures the complex evolution of scour, offering a more efficient and flexible approach for analyzing marine structure stability.

What Problem Does This Paper Address?

The stability of marine structures, particularly pipelines, is critically dependent on understanding and accurately predicting seabed scour. Scour, the erosion of sediment around structures, can compromise their integrity. Traditional numerical models often struggle with the complex interplay between fluid flow and dynamic seabed changes, especially in scenarios involving intricate geometries or the need to couple different scales of motion. Existing methods can be computationally expensive, require structured grids that limit flexibility, or lack the ability to adapt the mesh resolution dynamically to capture localized erosion and deposition phenomena effectively. This paper aims to address these limitations by developing and demonstrating a flexible, adaptive mesh approach for simulating scour.

Methods

The core of this work lies in the implementation of a numerical model within the open-source Fluidity computational fluid dynamics (CFD) code. Key methodological components include:

  • Discontinuous Galerkin Finite Element Method (DG-FEM): This approach is used for discretizing the hydrodynamics (Navier-Stokes equations) and the sediment transport equations. DG-FEM offers advantages in handling complex geometries and sharp gradients, which are crucial for capturing turbulent flow and localized scour.
  • Unstructured Meshes and Dynamic Mesh Adaptivity (r-adaptivity): The model utilizes unstructured tetrahedral meshes, providing flexibility in representing complex geometries. Crucially, it employs dynamic mesh adaptivity, specifically Laplacian smoothing, to reposition mesh nodes in response to bed deformation. This ensures that mesh quality is maintained and computational efficiency is optimized by concentrating resolution where it's most needed (e.g., in the scour hole).
  • Coupled Hydrodynamics and Sediment Transport: The model couples the fluid flow calculations with sediment transport and bed evolution models. This includes using empirical formulas for bed-load and suspended sediment transport, such as the Engelund & Fredsøe (1976) equations for bed-load and a modified Exner equation for bed change.
  • Control Volume (CV) Discretization for Sediment Concentration: The sediment concentration field is handled using a control volume discretization, ensuring conservative properties are maintained.

Datasets

The numerical simulations were motivated by and compared against experimental data from Mao's experiments (1986). This experiment involved a hydraulically smooth pipe with a diameter of 0.1 m placed over a sand bed with a median particle diameter of 0.36 mm, subjected to a mean flow velocity of 0.5 m/s. The study uses this setup to validate the model's ability to reproduce observed scour profiles and evolution.

Key Insights and Ideas

  • Adaptive Meshing is Crucial for Scour Simulation: The dynamic adjustment of the mesh (r-adaptivity) is shown to be essential for maintaining mesh quality and computational efficiency when simulating evolving seabed topography. Without it, excessive element stretching can lead to inaccurate results.
  • Tunnel Erosion Dynamics Captured: The model successfully captures the initial "tunnel erosion" phase, where water is diverted through a small gap beneath the pipe, leading to high velocities and significant localized scour.
  • Transition to Lee-Wake Erosion: The model demonstrates the transition from tunnel erosion to "lee-wake erosion," where vortex shedding downstream of the pipe becomes the dominant scour mechanism as the scour hole deepens.
  • Potential for Improved Accuracy: While the model shows good qualitative agreement with experimental data, the authors identify areas where further improvements in turbulence modeling and sediment transport formulations could lead to more accurate quantitative predictions, particularly in the early stages of scour.

Key Results

The study presents a preliminary investigation demonstrating the capabilities of the developed model. Key qualitative results include:

  • Visualizations of Bed Evolution: Figure 2 provides visualizations of the bed profile and velocity magnitudes at different time steps (0.1, 1, 1.5, and 2 minutes), illustrating the development of the scour hole.
  • Comparison with Experimental Data: Figure 4 shows the progression of bed profiles predicted by the model, which qualitatively match experimental data for a specific Shields parameter (θ = 0.098). The model predicts concave scour profiles consistent with observations.
  • Mesh Quality Preservation: Figure 3 demonstrates the effectiveness of Laplacian smoothing in maintaining mesh quality around the pipeline, preventing excessive element distortion compared to a mesh without smoothing.

Quantitative results are not extensively detailed in this preliminary study, but the authors note that the model results are stable for the experimental conditions considered, though some differences in scour depth were observed compared to previous simulations and experiments.

Limitations and Future Work

The authors acknowledge several limitations and outline promising avenues for future research:

  • Turbulence Modeling: The current model does not explicitly include a turbulence closure scheme, which is identified as a critical aspect for accurately capturing the interaction between vortex shedding and the bed, especially in the early stages of scour. Future work will explore incorporating turbulence models like k-ε or Large Eddy Simulations (LES).
  • Sediment Transport Formulas: While empirical formulas are used, there's potential for improvement, particularly in accounting for the gravitational influence on large bed slopes and implementing sediment slide models to prevent unrealistic slopes.
  • Seepage Flow: The study omits the physics of seepage flow beneath the pipeline, which is a significant factor in initiating tunnel erosion. Future work will aim to incorporate this.
  • 3D Simulations: The current study is 2D. Future work will extend the model to three dimensions for more realistic scenarios, such as impinging jets and industrial/environmental scour studies.
  • Morphological Acceleration: The paper briefly mentions morphological acceleration factors but notes potential conservation errors. The current approach avoids this by directly evolving the mesh.
  • Bed Slope Instabilities: The authors acknowledge that bed slope instabilities can lead to unrealistic profiles and are exploring methods like low-pass filtering for smoothing, though this may introduce some inaccuracy. Future work will investigate sediment slide techniques.
  • Multiphase Models: Exploration of multiphase models for flow-sediment interactions using meshless methods is also planned.

Summary updated

Abstract

Evaluating bed morphological structure and evolution (specifically the scoured bed level) accurately using numerical models is critical for analyses of the stability of many marine structures. This paper discusses the performance of an implementation within Fluidity, an open source, general purpose, computational fluid dynamics (CFD) code, capable of handling arbitrary multi-scale unstructured tetrahedral meshes and including algorithms to perform dynamic anisotropic mesh adaptivity. The flexibility over mesh structure and resolution that these capabilities provide makes it potentially highly suitable for coupling the structural scale with larger scale ocean dynamics. In this very preliminary study the solver approach is demonstrated for an idealised scenario. Discontinuous Galerkin finite-element (DG-FEM) based discretisation methods have been used for the hydrodynamics and morphological calculations, and automatic mesh deformation has been utilised to account for bed evolution changes while preserving the validity and quality of the mesh. In future work, the solver will be used in three-dimensional impinging jet and other industrial and environmental scour studies.
BibTeX
@misc{nunez2016numerical,
  author = {Juan Núñez and James Percival and Benjamin Yeager and S.J. Neethling and Matthew D. Piggott},
  title = {{Numerical simulation of scour below pipelines using flexible mesh methods}},
  year = {2016},
  doi = {10.1201/9781315375045-13},
}
Figure