Uncertainty in suction bucket installation can increase the risk of localized soil erosion and foundation failure, with potential consequences for installation time and offshore wind project costs. Using HPC-powered, grain-scale fluid–soil simulations, CEEC improved the accuracy of engineering models for predicting erosion and suction behaviour, helping reduce installation uncertainty and support more cost-efficient foundation installation.
| Industry | Scientific Domain | Code |
| Wind Energy | Fluid–Particle Interaction | waLBerla |
Description of the Challenge
Suction bucket foundations are increasingly used as support structures for offshore wind turbines, particularly for jacket-supported turbines and emerging floating wind anchoring systems. They are installed by applying an internal under-pressure that drives the bucket into the seabed while reducing soil resistance through seepage-induced weakening of the soil skeleton.
Compared to monopile installation, suction buckets offer several advantages such as lower underwater noise emissions, reduced installation time, and simpler decommissioning. However, suction installation introduces specific risks related to soil fluidization and piping, which remain difficult to predict using conventional geotechnical design tools.
Typical installation sites for offshore wind farms are located in water depths of 20–60 m for bottom-fixed turbines, where the seabed often consists of medium-dense sands, silty sands, or layered sediment deposits, although clay and mixed soil profiles also occur. During installation, excessive suction pressure may generate strong hydraulic gradients within the seabed. If the hydraulic gradient exceeds a critical threshold, piping erosion can occur, forming localized channels of fluidized soil that break the hydraulic seal beneath the bucket skirt.
Such piping events can halt the installation process entirely. In practice, this may lead to:
- Installation failure, requiring reinstallation or relocation of the foundation
- Significant vessel downtime
- Additional geotechnical investigations or redesign of the installation procedure
- In extreme cases, replacement of the entire foundation element
Regulatory frameworks such as DNV standards for offshore foundations require robust assessment of foundation integrity and installation feasibility. Uncertainties in predicting suction-induced failure mechanisms can therefore lead to conservative safety margins, increasing foundation cost and limiting the applicability of suction buckets in more challenging soil conditions.
However, predicting piping failure during suction bucket installation remains a challenging problem due to the strong coupling between fluid flow and granular soil mechanics. Conventional geotechnical engineering models typically treat the soil as a continuum and rely on empirical correlations derived from laboratory or field tests. But, piping and fluidization are inherently grain-scale processes, involving localized rearrangement of soil particles and pore-scale fluid flow, which are difficult to capture with traditional continuum-based models.
Why this Matters for Industry
Given that installation vessel time and foundation fabrication represent major contributors to offshore wind project cost, improving the prediction of installation behaviour directly contributes to reducing the Levelized Cost of Energy (LCOE) for offshore wind.
More specifically, installation failures of offshore wind turbine foundations have direct implications for project CAPEX and construction schedules. Offshore installation campaigns rely on specialized heavy-lift vessels operating under limited weather windows. A failed suction installation may therefore result in multi-day delays, causing significant cost overruns and potentially delaying turbine commissioning.
Improved predictive capabilities for suction bucket installation therefore provide several benefits for industry:
- Reduced installation risk through improved prediction of critical suction thresholds
- Improved pre-installation geotechnical assessment, enabling better interpretation of CPT data and soil surveys
- Reduction of conservative safety margins, allowing more efficient foundation designs
- Improved installation procedure planning, including optimized suction ramp profiles
- Extension of suction bucket applicability to deeper waters and larger turbines
The rapid scaling of offshore wind turbines toward 15–20 MW capacity requires larger and deeper foundations, increasing both loads and installation complexity. Reliable prediction tools are therefore also essential for enabling the next generation of offshore wind structures.
Furthermore, suction bucket technology is also being considered for floating wind mooring anchors, where similar installation mechanisms occur. Improved modelling of piping and soil fluidization thus has additional relevance beyond bottom-fixed wind turbines.
Finally, suction bucket foundations provide environmental advantages compared to monopiles, including lower underwater noise during installation and simpler removal at the end of the turbine lifetime. Improving installation reliability could therefore multiply the environmental benefits of suction bucket foundation technology by accelerating its adoption.
▶ Scientific Background
As previously stated, predicting piping failure during suction bucket installation remains challenging because of the strong coupling between fluid flow and granular soil mechanics, which is not well captured by Conventional geotechnical engineering models. These models typically treat the soil as a continuum and rely on empirical correlations derived from laboratory or field tests, and do not capture the inherently grain-scale processes of piping and fluidization, involving localized rearrangement of soil particles and pore-scale fluid flow.
In contrast, CEEC’s work employs fully-resolved coupled Lattice Boltzmann Method (LBM) – Discrete Element Method (DEM) simulations that enable direct simulation of the interaction between individual soil particles and the surrounding fluid flow. This approach resolves the hydrodynamic forces acting on each particle and provides detailed insight into the micromechanical processes governing soil fluidization.
The simulations reproduce the installation behaviour of a suction bucket segment and allow direct comparison with Conventional geotechnical engineering models commonly used in engineering practice. In a studied case, the conventional analytical model typically predicts soil resistance forces that differ by up to a factor of two compared to the high-fidelity simulations. However, after calibration using high-fidelity simulation data, the analytical model predictions improved significantly, with deviations reduced to less than approximately 20%.
This demonstrates how grain-scale simulations can bridge the gap between micro-scale physical mechanisms and macro-scale engineering models, enabling improved prediction of installation behaviour while retaining computationally efficient design tools.
▶ Technical Details: HPC Computing Approach
The simulations rely on a fully resolved particle–fluid coupling approach, where the fluid phase is simulated using the Lattice Boltzmann Method (LBM) and the soil grains are modelled using the Discrete Element Method (DEM).
Such grain-resolved simulations are computationally extremely demanding due to the need to resolve fluid flow around every individual particle. High-performance computing resources are therefore essential for performing these simulations.
The simulations were implemented within the waLBerla multiphysics framework, which provides scalable GPU-enabled implementations of the Lattice Boltzmann method and particle coupling algorithms. This allows efficient execution on modern HPC systems.
Compared to traditional continuum-based geotechnical models, the computational cost of particle-resolved simulations is significantly higher. However, they provide detailed insight into the underlying physical mechanisms and enable the calibration of simplified engineering models, which can then be applied in routine engineering design.
Without access to HPC resources, such grain-scale simulations for realistic geotechnical systems would be computationally infeasible.
Results and New Insights
The simulations provide detailed insight into the micromechanical processes occurring during suction bucket installation.
Key findings include:
- Identification of critical suction thresholds required to initiate penetration
- Visualization of localized fluidization zones and soil particle rearrangements near the bucket skirt
- Observation that outer skirt friction increases during suction installation due to suction-induced soil inflow toward the outside of the caisson
- Demonstration that inner soil resistance remains negligible during the studied installation scenario
The simulations also enable quantitative comparison with analytical installation models.
These results demonstrate that high-fidelity simulations can be used to derive improved engineering correlations and provide guidance for optimizing suction installation procedures.
Industrial Takeaways
The presented modelling framework offers several practical benefits for offshore wind developers and geotechnical engineers:
- Improved prediction of installation feasibility for suction bucket foundations
- Enhanced interpretation of CPT data through physics-based calibration of analytical models
- Reduced uncertainty in suction pressure limits during installation
- Lower risk of installation failure due to piping erosion
- Potential reduction in conservative design margins, enabling more cost-efficient foundations
In addition, the approach enables the development of calibrated surrogate models that combine the accuracy of high-fidelity simulations with the computational efficiency required for routine engineering design.
Ultimately, improving the reliability of suction bucket installation contributes to:
- Reduced installation risk and vessel downtime
- Lower foundation CAPEX
- Reduction of offshore wind Levelized Cost of Energy (LCOE)
The modelling framework could also be integrated into future geotechnical engineering software tools and used in collaboration with offshore wind developers, foundation manufacturers, and certification bodies.
▶ Codes and Software Stack
waLBerla (Widely Applicable Lattice Boltzmann from Erlangen)
The simulations were performed using the open-source waLBerla multiphysics simulation framework, which provides scalable implementations of the Lattice Boltzmann Method and particle-resolved fluid–particle coupling.
Key characteristics include:
- Implementation of the Lattice Boltzmann Method (LBM) for simulating fluid flow
- Coupling with the Discrete Element Method (DEM) for modelling granular soil particles
- Particle-resolved CFD, enabling accurate simulation of fluid–particle interactions
- High scalability on modern HPC systems, including GPU-accelerated architectures
- Performance portability across supercomputing platforms via code generation
The framework enables detailed simulations of particle-laden flows, making it suitable not only for suction bucket installation but also for related applications such as:
- seabed scour around offshore structures
- sediment transport processes
- dredging and erosion modelling
The open-source nature of the code allows further development and potential adoption by both academic and industrial users.