DOMAIN MAP 5: Offshore Wind Turbine Dynamics and Scour Mechanics¶
Synthesised from 45 batch-agent summaries (batch01_agent1 through batch09_agent5), covering approximately 900 papers from 1981 to 2026. Extraction filtered for offshore wind turbine dynamics, scour mechanics, foundation comparison, aero-hydro-servo-elastic simulation, Korean offshore wind programme, design standards, and environmental loading.
Domain at a glance¶
mindmap
root((D5 · Offshore wind<br/>scour mechanics<br/>900 papers))
OWT dynamics
1P 3P resonance avoidance
Campbell diagram
Aero-hydro-servo-elastic
OpenFAST TurbSim
10-min DLC framework
Foundation comparison
Monopile dominant
Jacket multi-leg
Tripod emerging
Suction bucket novel
Gravity base niche
Scour mechanics
Local vs global
HEC-18 / HEC-23
Sumer-Fredsoe
Melville-Coleman
Time-dependent scour
Metocean loading
Wind wave current
Joint probability
Hindcast databases
Extreme value analysis
Korean Yellow Sea
DNV + IEC standards
DNV-OS-J101 retired
DNV-ST-0126 monopile
IEC 61400-3 offshore
DNV-RP-C212 soil
ISO 19901-4 geotech
Site evidence
Robin Rigg scour loss
Borkum Riffgrund 1
Anholt Scroby Sands
East Anglia One
Korean demo sites
Scour countermeasures
Rock dumping
Frond mats
Geotextile bags
Solidified soil
Monitoring-based
Relevance to PhD
J1 J2 J3 centre of gravity
J5 probabilistic
V1 V2 field + SHM
Op3 pipelineScour-to-frequency causal chain¶
The central chain that every scour paper rests on:
%%{init: {"theme":"base","themeVariables":{"fontSize":"16px","fontFamily":"Inter, system-ui, -apple-system, sans-serif","primaryTextColor":"#1a1a1a","lineColor":"#666"},"flowchart":{"nodeSpacing":25,"rankSpacing":40,"padding":16,"useMaxWidth":true}}}%%
flowchart TB
Flow["<b>1 · Metocean flow</b><br/><span style='font-size:14px'>wind · wave · current</span>"]:::env
Scour["<b>2 · Scour hole forms</b><br/><span style='font-size:14px'>local or general<br/>depth grows with time</span>"]:::sc
Stress["<b>3 · Overburden stress</b><br/><span style='font-size:14px'>lost over footprint</span>"]:::s
Stiff["<b>4 · Foundation stiffness drops</b>"]:::st
Freq["<b>5 · Natural frequency shifts down</b>"]:::f
Res["<b>6a · 1P / 3P resonance risk</b>"]:::r
Fatigue["<b>6b · Accelerated fatigue damage</b>"]:::fa
Flow ==> Scour ==> Stress ==> Stiff ==> Freq
Freq ==> Res
Freq ==> Fatigue
classDef env fill:#e3f2fd,stroke:#1565c0,stroke-width:2px
classDef sc fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef s fill:#fff8e1,stroke:#f57f17,stroke-width:2px
classDef st fill:#e8f5e9,stroke:#2e7d32,stroke-width:2px
classDef f fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef r fill:#ffebee,stroke:#c62828,stroke-width:3px
classDef fa fill:#f3e5f5,stroke:#7b1fa2,stroke-width:3pxFoundation-type scour sensitivity ranking¶
flowchart TB
Mono[MONOPILE<br/>Δf 5-15% per 1D<br/>Most sensitive]:::hi
Jack[JACKET<br/>Δf 3-8% per 1D<br/>Moderate]:::me
Trip[TRIPOD BUCKET<br/>Δf max 5.3% at 0.6D<br/>Low sensitivity]:::lo
Grav[GRAVITY BASE<br/>Small Δf<br/>Low · rarely deployed]:::lo
Mono -.-> Reason1[Single column<br/>all capacity in one footprint]:::r
Jack -.-> Reason2[Load redistribution<br/>through bracing]:::r
Trip -.-> Reason3[Three footings share load<br/>+ lid bearing]:::r
Grav -.-> Reason4[Dead weight + large base<br/>stiffness-dominant]:::r
classDef hi fill:#ffebee,stroke:#c62828,color:#b71c1c
classDef me fill:#fff3e0,stroke:#ef6c00,color:#e65100
classDef lo fill:#e8f5e9,stroke:#2e7d32,color:#1b5e20
classDef r fill:#fafafa,stroke:#888,color:#555,font-size:11px1. ESTABLISHED KNOWLEDGE¶
1.1 Foundation response under combined loading¶
Combined V-H-M loading governs offshore foundation design. Every major experimental programme from the Oxford-Cambridge-UWA lineage (Bell 1991, Martin 1994, Cassidy 1999, Byrne 2000, Villalobos 2006, Bienen 2007, Fu 2017, Cheng & Cassidy 2016) confirms that vertical, horizontal, and moment loads are coupled and must be treated through work-hardening plasticity frameworks. Deterministic single-load analysis is unconservative. Failure envelopes for skirted, bucket, and shallow foundations are now well characterised for both drained and undrained conditions in sand and clay (Butterfield & Gottardi 1994, Bransby & Randolph 1998, Gourvenec 2007, Vulpe et al. 2013-2014).
1.2 Soil-structure interaction controls OWT natural frequency¶
Across more than 30 papers spanning 2010-2025, there is unanimous agreement that SSI fundamentally alters the dynamic response of offshore wind turbines relative to fixed-base assumptions. Ignoring SSI leads to errors of 5-30% in first natural frequency prediction (Alamo 2016, Zaaijer 2006, Bhattacharya & Adhikari 2011, Fallais 2022, Stuyts 2022). Foundation model choice causes up to 22% variation in accumulated fatigue damage at mudline (Aasen 2017) and up to 180% fatigue deviation in misaligned wind-wave conditions (Katsikogiannis 2019). The monitored first natural frequency of operational monopiles is systematically 5-15% higher than design predictions (Stuyts 2022, McAdam 2023), primarily because legacy API/DNV p-y methods underestimate foundation stiffness for large-diameter monopiles (Haiderali 2016/2023, Velarde 2017, Choo & Kim 2015, DW Kim 2016).
1.3 Scour is a primary threat to foundation integrity¶
Scour is the leading cause of bridge failure in the US (~22 collapses/year, Yao 2013) and a critical concern for offshore wind monopiles. Scour reduces lateral stiffness, shifts natural frequency downward, and accelerates fatigue damage. Quantified impacts include approximately 24% fatigue life reduction at 1.3D scour depth (Cao 2024), up to 14% natural frequency reduction in horizontal bending modes (Tseng 2018), and 38-48% reduction in ultimate moment capacity depending on local vs. general scour morphology (Ciancimino 2022). For tripod suction bucket foundations, a maximum 5.3% frequency reduction at scour depth of 0.6D has been measured via centrifuge testing (Kim et al. 2025). Vibration-based methods (natural frequency tracking) are the dominant indirect monitoring paradigm (Prendergast 2013-2017, Kariyawasam 2020, Kawabe 2023, Kazemian 2023).
1.4 Cyclic loading changes foundation response¶
Pile stiffness generally increases with moderate cycling in sand (LeBlanc 2010; lateral capacity doubled after 4800 cycles in Almeida 2024), contradicting older degradation assumptions. However, cyclic loading below the yield moment produces negligible plastic deformation (Kim et al. 2014, Byrne 2000), while exceeding it leads to cumulative permanent rotation. Partially drained behaviour during storm loading lies between drained and undrained extremes and cannot be predicted by either limit (Hsu 1998, Villalobos 2006, Stapelfeldt 2020). Two-way loading in saturated dense sand produces larger rotations than one-way loading under partly drained conditions (Nielsen 2017).
1.5 Suction caissons are viable for offshore wind¶
Suction installation drastically reduces penetration force compared to pushed installation (Villalobos 2006). Field trials at Bothkennar (clay) and Luce Bay (sand) confirm high initial stiffness, hysteretic cyclic response, and degradation only at high loads (Houlsby et al. 2005, 2006). Tripod configurations offer distinct advantages in rotation control: tripod yield moment is half that of monopod, but rotation at yield is only 20% of monopod values (Kim et al. 2014). Industry design guidelines exist (OWA 2019), and penetration resistance formulae from API/DNV/Houlsby remain the standard, though discrepancies with field data persist (Zhang X 2024, Suryasentana 2025).
1.6 Coupled aero-hydro-servo-elastic simulation is mandatory¶
For both fixed and floating OWT, coupled multi-physics simulation is now the standard expectation. Multiple code-to-code benchmarks (OC5 Phase III at Alpha Ventus, Popko 2019; DNV GL JIP, Jonkman 2019) have validated simulation tools. Advanced AHSE models affect generator-converter dynamics primarily above rated wind speed (Carmona-Sanchez 2019). For floating platforms, rigid-floater assumptions cause up to 37% eigenfrequency error (Aguilera 2025); platform motions introduce nonlinear wake phenomena not captured by fixed-turbine models (Messmer 2024, Schulz 2024).
2. ACTIVE FRONTIERS (2023-2025)¶
2.1 PISA-based foundation modelling replacing legacy p-y¶
The PISA (Pile-Soil Analysis) methodology and its extensions are becoming the accepted standard for monopile lateral response. PISA predicts foundation stiffness much closer to field-monitored values than legacy API/DNVGL p-y curves (McAdam 2023, Sastre Jurado 2022, Kheffache 2024). Macro-element models calibrated against 3D FE (REDWIN project: Skau 2018; OxCaisson: Suryasentana 2020) offer computational efficiency suitable for time-domain fatigue simulations. However, the transition from p-y to macro-element models in commercial design codes is still contested.
2.2 Vibration-based scour monitoring for OWT¶
While vibration-based scour monitoring is well validated for bridges (Prendergast series, Kariyawasam 2020, Kawabe 2023), translation to offshore wind monopiles remains laboratory-only or numerical. Unsupervised novelty detection using healthy-condition data shows promise (Abdelhak 2024, Smith 2023). Hierarchical Bayesian models can infer population-level soil stiffness distributions and detect scour as anomaly across a wind farm (Smith 2023). Combined scour-seismic analysis for suction bucket OWTs has just begun (Jia 2024, Ngo 2022).
2.3 Physics-informed ML and digital twins¶
Physics-enhanced ML consistently outperforms pure data-driven or pure physics models (Karniadakis 2021, Shen 2023, Haywood-Alexander 2024). Differentiable modelling embeds physical equations within neural network training loops (Shen 2023). Full-scale digital twin validation has been achieved for the TetraSpar FOWT at 10-15% error on fatigue DELs (Branlard 2024). Probabilistic digital twins using ensemble approaches improve risk-based inspection planning (Bull 2025). However, predictive/prescriptive DT levels remain prototype-stage only.
2.4 Low-frequency fatigue dynamics¶
Up to 65% of fatigue damage relates to low-frequency dynamics (periods greater than 1 day) that are invisible in standard 10-minute analysis windows, directly challenging the DNV DLC framework (Sadeghi 2023). A short-to-long-term correction factor has been proposed but not yet adopted by standards.
2.5 Distributed acoustic sensing for OWT SHM¶
Phi-OTDR DAS captures both local damage (loose bolts) and global tower strain over km-scale distances. First full-scale OWT fibre optic validation achieved by Xu & Soga (2024). Virtual sensing below mudline via Gaussian Process Latent Force Models has been validated in situ at Westermeerwind Park (Zou et al. 2022).
3. CONTESTED CLAIMS¶
3.1 Scour geometry treatment in design¶
DNV standards prescribe 1.3D uniform scour depth for monopiles under current-only conditions. Multiple papers show this is context-dependent: local scour (leaving overconsolidated soil) produces different failure mechanisms from general scour (simple mudline lowering) (Qi 2016, Ciancimino 2022, Tseng 2017, Vicente 2023). Whether design codes should distinguish these two cases remains under active discussion.
3.2 Dynamic vs. static scour prediction¶
Cyclic pile movement creates scour holes significantly deeper and wider than static predictions via multi-stage storm-backfill mechanisms (Al-Hammadi 2019). Field data at Robin Rigg show scour correlates more with tidal current direction than water depth, and storm events can increase or backfill scour depending on wave-current angle (Garcia 2022). Current design codes treat monopiles as static.
3.3 IEC design load return periods¶
IEC standards use a 50-year maximum return period, far below the 10,000-year MRP for oil and gas (Wilkie 2019). Whether offshore wind structures in hurricane/typhoon-prone regions need higher MRP design loads is contentious. The IEC extreme turbulence model may also underpredict variance from ramp-like wind events (Hannesdottir 2019).
3.4 SSI model fidelity for fatigue assessment¶
Macro-element models and p-y models yield materially different fatigue damage estimates (Sorum 2022). The community has not converged on which model is correct for design. Macro-element models capture history-dependent stiffness and damping that p-y cannot reproduce (Page 2017, Skau 2018), but p-y curves remain entrenched in DNV standards due to simplicity and regulatory acceptance.
3.5 BEM vs. higher-fidelity aerodynamics for large/floating turbines¶
BEM fails under yaw misalignment for flexible rotors (Shaler 2023), and BEMT reliability for FOWT simulation is questioned (Schulz 2023). Free-wake vortex methods on GPUs are becoming viable (Blondel 2024), but the community has not reached consensus on which unsteady aerodynamic phenomena are relevant for FOWT rotor design (Schulz 2024).
3.6 Constitutive model for cyclic monopile ratcheting¶
SANISAND-MS and PM4SAND yield different accuracy levels for cyclic strain accumulation in monopile sands (Orakci 2024). No single constitutive model satisfies all cyclic loading requirements, and validation remains at the element-test level; boundary-value-problem-level validation against field data is absent.
4. VERIFIED GAPS¶
4.1 Combined multi-hazard loading on OWT foundations¶
No paper addresses the full coupled problem of scour + liquefaction + fatigue + corrosion under long-term operation. Each hazard is studied in isolation. Seismic design of OWT foundations is particularly sparse as offshore wind expands to seismically active regions (Korea, Japan, Taiwan, US west coast). Combined earthquake + storm fragility functions are absent.
4.2 Field-scale validation of vibration-based scour monitoring for monopiles¶
All scour-vibration work targets bridge piers. No field-validated vibration-based scour monitoring system exists for offshore wind monopiles (Hachem 2023, Kazemian 2023, Kawabe 2023 -- all bridge studies). Kim et al. (SNU, 2023) propose a 3-year research plan for centrifuge + integrated numerical + monitoring for OWT scour.
4.3 Long-term cyclic degradation under realistic loading¶
Most experimental campaigns apply hundreds to thousands of cycles. Multi-decadal field monitoring data for OWT foundations do not exist. Variable-amplitude, multi-directional, stochastic loading histories over a 25-year lifetime remain poorly validated experimentally (Zorzi 2020).
4.4 Silty sand characterisation for Yellow Sea conditions¶
Multiple Korean centrifuge programmes (Kim DJ 2014, Choo et al. 2021, Jeong et al. 2019/2021) test in silty sand representative of Yellow Sea conditions, but a validated constitutive model specific to this soil type for use in numerical OWT foundation analysis is absent. CPTU interpretation for transitional/silty soils lacks a robust methodology (Fioravante 2022).
4.5 Integrated digital twin spanning aero-hydro-geotech chain¶
Superstructure digital twins exist (Branlard 2024), and foundation monitoring databases exist (Stuyts 2022), but no operational digital twin integrates real-time scour evolution, soil-foundation state, and structural response in a closed loop. Integration of scour monitoring into OWT digital twin frameworks remains an open problem.
4.6 Screw pile and hybrid foundation design standards¶
Screw piles offer quieter installation and enhanced tensile capacity (Davidson 2020, Cerfontaine 2021/2023), but lack long-term field data, cyclic behaviour models, and installation certification for large diameters. Hybrid monopiles reduce displacement by 30-50% (Trojnar 2023), but no design codes or widely accepted p-y formulations exist for these configurations.
4.7 Population-level SHM for offshore wind farms¶
Only Smith (2023) addresses population-based SHM across a wind farm using hierarchical Bayesian models. Scaling these methods to full-size farms with hundreds of turbines and diverse soil conditions remains unexplored. Fleet-wide fatigue extrapolation from instrumented leaders to entire wind farms is still under development (Weijtens 2016, Ziegler 2019).
5. QUANTITATIVE BENCHMARKS¶
5.1 Foundation geometry and capacity¶
| Parameter | Value | Source |
|---|---|---|
| Tripod bucket dimensions (prototype) | D=6.5 m, L=8.0 m, c-c=26.9 m | Kim et al. 2014 |
| Tripod yield rotation / monopod yield rotation | 20% | Kim et al. 2014 |
| Monopile market share for OWT | 77% | Abdelhak 2024 |
| Foundation cost share of OWT capital | 20-33% | EWEA, cited in Aasen 2017, Kim 2014 |
| Suction caisson critical uplift displacement | ~0.02D (mono), ~0.01D (group) | Zhu et al. 2019 |
| Lateral capacity increase after 4800 cycles | ~2x | Almeida 2024 |
5.2 Scour effects¶
| Parameter | Value | Source |
|---|---|---|
| DNV design scour depth (current-only) | 1.0-1.3D | DNV standard; Tseng 2017, Vicente 2023 |
| Fatigue life reduction at 1.3D scour | ~24% | Cao 2024 |
| Fatigue life increase from TMD (1% mass) | ~65% | Cao 2024 |
| Max frequency reduction (tripod SB, 0.6D scour) | 5.3% | Kim et al. 2025 |
| Natural frequency reduction from scour (monopile, Taiwan) | up to 14% | Tseng 2018 |
| Moment capacity reduction: local scour | 38% | Ciancimino 2022 |
| Moment capacity reduction: general scour | 48% | Ciancimino 2022 |
| Bridge collapses/year due to scour (US) | 22 | Yao 2013 |
5.3 Dynamic response and fatigue¶
| Parameter | Value | Source |
|---|---|---|
| Fatigue damage variation from foundation model choice | up to 22% (fixed), 180% (cyclic) | Aasen 2017, Katsikogiannis 2019 |
| Monitored fn vs. design fn (monopiles) | 5-15% higher measured | Stuyts 2022, McAdam 2023 |
| Low-frequency fatigue contribution (>1 day period) | up to 65% | Sadeghi 2023 |
| AEP loss from blade roughness | 2.9-8.6% | Kelly 2022 |
| Wind farm CF decline rate | 0.26-0.72 pp/yr | Kelly 2022 (UK, US, AU) |
| Nonlinear wave loading fraction | up to 40% of total | Ding 2024 |
| Mudline moment difference (fixed vs. flexible base, seismic) | 34% | Yang 2019 |
5.4 Simulation and monitoring accuracy¶
| Parameter | Value | Source |
|---|---|---|
| OC5 Alpha Ventus multi-code validation | Benchmark | Popko 2019 |
| Digital twin fatigue DEL accuracy (field) | 10-15% error | Branlard 2024 |
| Monopile DEL extrapolation from strain gauges | <4% monthly error | Ziegler 2019 |
| Jacket model updating eigenfrequency error | 30% to 1% via SSI updating | Augustyn 2020 |
| GPLFM virtual sensing below mudline | First in-situ validation | Zou et al. 2022 |
| DAS full-scale OWT validation | First achieved | Xu & Soga 2024 |
| Scour depth ML prediction (1D CNN) | R2 = 0.85 | Huynh 2025 |
| Scour depth ML (ANN-PSO, tripod) | R2 = 0.99 | Jatoliya 2024 |
5.5 Reference turbines and platforms¶
| Reference | Specification | Primary use |
|---|---|---|
| NREL 5 MW | 126 m rotor, geared | Legacy benchmark |
| DTU 10 MW | 178.3 m rotor | Emerging benchmark |
| IEA 10 MW offshore | 198 m rotor, direct drive | Bortolotti 2019 |
| IEA 15 MW + VolturnUS-S | Semi-submersible | Allen 2020 |
| OC4-DeepCwind semi-sub | 5 MW platform | Floating validation |
| OC3-Hywind spar | 5 MW platform | Floating validation |
5.6 Korean offshore wind programme¶
| Item | Detail | Source |
|---|---|---|
| Korean Ministry R&D (2017) | Design standards for OWT support structures and concrete substructures | Korean Ministry report |
| KEPCO/SNU scour project (2024) | Year-1 review: vibration-based scour monitoring, CFD/DEM planned for Year 2 | KEPCO project Q&A |
| Korea MOF regulation (2024) | Floating OWT now under Ship Safety Act; stability inspection mandated | Regulatory amendment |
| Korean 3 MW tripod suction pile | Field-measured fn at each installation stage, 1.5% agreement with numerical | Ryu 2019 |
| KAIST centrifuge (Kim DJ 2014) | Tripod bucket in silty sand at 100g; reference for Korean foundation research | KAIST centrifuge programme |
| Yellow Sea silty sand tests (Choo 2021) | Monopod bucket H-M envelope at 70g; 1.5x capacity increase at 20x loading rate | KAIST centrifuge at 70g |
| KEI guidelines (2021) | Environmental impact assessment guidelines for offshore wind farm construction | Korean government report |
| KOCED standards | Centrifuge preparation, miniature CPT, bender elements, dynamic testing | Korean standards series |
| CGO 15 MW jacket project (2024) | Cost reduction for T&I of 15 MW jacket substructures using suction pile technology | Korean R&D |
| Yoon et al. (2023) | Underwater noise baseline from OWT off southwest Korea | Field measurements |
Cross-cutting observations¶
Methods convergence. Centrifuge testing (appearing in >60 papers) remains the gold standard for geotechnical validation. FEM dominates numerical analysis but is being supplemented by large-deformation methods (CEL, MPM, SPH, PFEM) and ML surrogates. Physics-informed ML is displacing pure data-driven approaches in structural and geotechnical monitoring.
Standards lag. API/DNV p-y curves, developed from small-diameter piles in the 1970s-80s, remain codified despite documented inadequacy for large-diameter OWT monopiles. PISA-based methods are gaining acceptance but are not yet in standards. The 10-minute DLC framework may miss up to 65% of fatigue damage from low-frequency dynamics. No international code provides unified multi-hazard (scour + seismic + wind + wave) design guidance for OWT foundations.
Korea-specific context. The Korean offshore wind programme is active but early-stage relative to North Sea deployments. KAIST centrifuge programmes provide the primary experimental database for bucket foundations in Yellow Sea silty sand. KEPCO-SNU collaboration is pursuing vibration-based scour monitoring as a priority. Korean standards (KOCED series) codify centrifuge specimen preparation and testing but do not yet address offshore wind foundation design specifically. The regulatory framework is evolving, with floating OWT recently brought under the Ship Safety Act (2024).