Skip to content

Literature Synthesis -- Batch 04 Agent 4

Scope: 40 papers (positions 721--760), all published 2024.

Individual Paper Summaries

# Author(s) Year Title Core Finding Method Tags
1 Esposito et al. 2024 Bioinspired Fano-like resonant transmission: frequency selective impedance matching Fano resonance enables frequency-selective impedance matching for acoustic devices Analytical/numerical modeling of acoustic metamaterials acoustics, impedance, metamaterials
2 Gaba et al. 2024 Systematic analysis of enhancing cyber security using deep learning for CPS DL outperforms ML for cyber-attack detection in cyber-physical systems; reviews attack detection methods Systematic review of DL-based CPS security cybersecurity, deep-learning, CPS, review
3 Gaidai et al. 2024 Lifetime assessment of semi-submersible wind turbines by Gaidai risk evaluation method Gaidai reliability method handles multi-DOF extreme load assessment for FOWTs FAST coupled aero-hydro-servo-elastic simulation + Gaidai reliability method FOWT, reliability, lifetime, FAST
4 Gamberini et al. 2024 Validation of aeroelastic dynamic model of active trailing edge flap system on 4.3 MW WT HAWC2 and BHawC flap models validated against field data; reduces prediction uncertainty Field validation of aeroelastic codes (HAWC2, BHawC) aeroelastic, trailing-edge-flap, validation, wind-turbine
5 Gkougkoudi-Papaioannou et al. 2024 Numerical and physical modelling of pore pressure around a monopile foundation Monopile presence increases max pore pressure, especially waveward; pore pressure attenuates with depth 3D coupled hydro-geotech numerical model + COB physical model tests (1:23 scale) monopile, pore-pressure, physical-modelling, wave-seabed
6 Guan et al. 2024 Dynamic response characteristics of offshore floating WT pitch system Independent pitch reduces mean tower oscillation but increases fatigue load in flow-parallel direction Simulation of uniform vs. independent pitch control FOWT, pitch-control, dynamic-response, fatigue
7 Guo et al. 2024 New method for incorporating foundation damping in time-domain analysis of OWT Foundation damping significantly reduces monopile OWT loads; new incorporation method proposed Time-domain integrated analysis with foundation damping model monopile, foundation-damping, OWT, time-domain
8 Haghi & Crawford 2024 Data-driven surrogate model for wind turbine damage equivalent load Sequential ML maps wind speed time series to DEL, replacing costly aeroelastic simulations; transfer learning improves wake conditions ML surrogate (ANN) trained on aeroelastic outputs + transfer learning surrogate-model, DEL, fatigue, ML, wind-turbine
9 Hameed et al. 2024 Hybrid ELM-DOA for modeling liquefaction triggering in sand-silt mixtures ELM-DOA achieves R2=0.935; nonlinear normalization boosts all models by ~25% Extreme Learning Machine + Dingo Optimization Algorithm; GUI developed liquefaction, ML, optimization, geotechnical
10 Haywood-Alexander et al. 2024 Spectrum of physics-enhanced ML: survey on structural mechanics Categorizes PEML along physics-data axes; demonstrates on Duffing oscillator Survey + working examples with open code PEML, physics-informed, structural-mechanics, survey
11 Hietanen et al. 2024 Novel techno-economical layout optimization for floating wind farm design Multi-parametric optimization (AEP + mooring + cables) increased profit by EUR 34.5M vs. grid layout Layout optimization with ScotWind site 10; mooring and cable cost models FOWF, layout-optimization, techno-economic, LCOE
12 Hsiao et al. 2024 Explainable AI models for predicting liquefaction-induced lateral spreading XGBoost + SHAP reveals key drivers of lateral spreading prediction from 2011 Christchurch data XGBoost with SHAP explainability liquefaction, XAI, lateral-spreading, geotechnical
13 Hubert et al. 2024 Dynamic response of WT wake to surge and heave step motions Floating platform surge/heave motions introduce complex wake dynamics affecting downstream turbines Wind tunnel experiment with scaled FOWT FOWT, wake, surge, heave, wind-tunnel
14 Jasim & Ahmed 2024 Bearing capacity of organic soil reinforced by sand dune and sodium silicate columns End-bearing columns at L/D=6 with 8 columns improved bearing capacity by 666% Plaxis 3D numerical analysis soil-improvement, organic-soil, bearing-capacity, FEM
15 Jiang et al. 2024 Causes of red tide in offshore Rongcheng based on coupled physical-biological model Coupled model identifies physical/biological causes of kelp production loss from red tide Coupled physical-biological ocean model oceanography, red-tide, coastal, not-directly-relevant
16 Jilo et al. 2024 Numerical analysis of underground tunnel deformation at Lega-Dembi gold mine Rock bolts + shotcrete effective; GSI and UCS most influential on tunnel deformation Combined continuum-discontinuum numerical method (FEM, FDM, DEM) tunnel, mining, numerical, geotechnical
17 Kam et al. 2024 Cost-benefit analysis of cyber risk mitigation in offshore wind: a survey Lack of CBA methods specific to offshore wind cybersecurity; only 6 of 18 articles met criteria Systematic literature review (Scopus, WoS) cybersecurity, offshore-wind, CBA, survey
18 Kim et al. 2024 Foundation types of fixed offshore wind turbine Reviews gravity, monopile, jacket, tripod, suction bucket; hybrid foundations increase lateral resistance Review + FEM + centrifuge tests + fatigue analysis OWT-foundations, monopile, jacket, suction-bucket, review
19 Lamei et al. 2024 Hydro- and aero-elastic response of FOWTs in frequency domain Frequency-domain coupled aero-hydro-elastic model captures full 6-DOF motion + structural elasticity Analytical: linear diffraction + BEM + FEM in frequency domain FOWT, hydroelasticity, aeroelasticity, frequency-domain
20 Larionov et al. 2024 Method of successive approximation in geotechnical mechanics modelling Successive approximation as alternative to FEM for circular roadway under hydrostatic pressure Mathematical modelling, successive approximation method geotechnical, analytical, tunnel, modelling
21 Lee et al. 2024 Monitoring and simulation of bridge pier scour and deposition in floods Real-time MEMS scour monitoring + 1D/2D hydrodynamic simulation validated with field data; bridge safety curve developed MEMS sensors + numerical mobile-bed model scour, bridge, monitoring, SHM, flood
22 Liu et al. 2024 State-of-the-art review: AI-enhanced computational mechanics in geotechnical engineering ANN (35%), RF (19%), SVM (17%) dominate; mechanical properties account for 59% of AI applications Comprehensive review of AI in geotechnical computational mechanics AI, geotechnical, review, ANN, computational-mechanics
23 Memari et al. 2024 Review: drone + deep learning for wind turbine blade inspection Integrates aerial inspection, DL image processing, and structural integrity assessment for WTB Review of drone-based DL defect detection methods WTB, drone, deep-learning, inspection, SHM
24 Messmer et al. 2024 Enhanced wake recovery from nonlinear dynamics of FOWT Both side-to-side and fore-aft motions accelerate wake recovery via nonlinear spatiotemporal dynamics; lock-in at St in [0.2, 0.55] Wind tunnel experiments, laminar inflow, Strouhal number analysis FOWT, wake, nonlinear-dynamics, wind-tunnel
25 Millar et al. 2024 ArchIMEDES: computer vision tracking of masonry arch bridge scour stability Computer vision (ArchIMEDES) links structural response to scour conditions on FlexiArch scale model Scale experiments + computer vision algorithm scour, bridge, computer-vision, SHM, masonry
26 Moore et al. 2024 Insolation cycles control resonance frequency drifts at a natural rock tower, Utah Insolation (not air temperature) is primary driver of annual resonance frequency drifts 2-year ambient vibration monitoring + numerical insolation modelling resonance-frequency, SHM, rock-slope, environmental
27 Morley et al. 2024 Soil ratcheting behind integral bridges using centrifuge modelling Centrifuge modelling successfully simulates backfill strain ratcheting; global rotations and base sliding are significant Centrifuge testing with high-accuracy actuation centrifuge, integral-bridge, ratcheting, soil-structure
28 Onyenanu et al. 2024 Mathematical model for palm fruit digester via dimensional analysis Buckingham Pi theorem yields R2=0.9956 for digester throughput and efficiency Dimensional analysis, Buckingham Pi theorem dimensional-analysis, Buckingham-Pi, process-engineering
29 Orakci et al. 2024 Soil constitutive models for OWT under lateral cyclic loads Compares SANISAND-MS and PM4SAND for cyclic strain accumulation (ratcheting) in monopile sands Triaxial test simulations in OpenSees and PLAXIS monopile, cyclic-loading, constitutive-model, ratcheting
30 Pham et al. 2024 Critical review of physical-mechanical principles in geostructure-soil interface mechanics Comprehensive review of THM behaviour at soil-structure interfaces; identifies gaps in energy geostructure interfaces Critical review of testing, constitutive models, and numerical methods SSI, interface-mechanics, THM, review
31 Rautela & Goyal 2024 Transforming air pollution management in India with AI/ML Convolutional autoencoder achieves SSIM>0.60 for PM2.5 forecasting across India AI/ML (convolutional autoencoder) for PM2.5 prediction air-pollution, AI, autoencoder, not-directly-relevant
32 Ren et al. 2024 Experimental study of tendon failure for TLP FOWT 1:50 scale TLP FOWT tested under tendon failure scenarios with wind and waves Physical model testing at Dalian University TLP, FOWT, tendon-failure, experimental
33 Rodrigues et al. 2024 Speeding up large-wind-farm layout optimization SMAST heuristic + algorithmic differentiation achieves 75x speedup for 500-turbine farms; single SMAST run outperforms 5000 random starts Gradient-based optimization with PyWake/TOPFARM layout-optimization, gradient, large-scale, wind-farm
34 Rodriguez Castillo et al. 2024 Critical review: offshore structures for renewable hydrogen production SWOT analysis of offshore green hydrogen infrastructure; reviews O&G tools adaptable to hydrogen Critical review + SWOT analysis offshore, hydrogen, renewable, review, structural
35 Roy et al. 2024 Latent space correlation-aware autoencoder for anomaly detection in skewed data Robust Mahalanobis distance in latent space detects both near and far anomalies in skewed sensor data Autoencoder with robust MD regularization anomaly-detection, autoencoder, sensor-data, unsupervised
36 Sah et al. 2024 Higher mode implications for dynamic SSI of offshore wind turbine Higher modes must be considered for accurate dynamic SSI response of OWTs Modal analysis of NREL 5MW OWT with various foundation models OWT, SSI, higher-modes, modal-analysis
37 Sathe & Giometto 2024 Impact of numerical domain on turbulent flow statistics for canopy flows Domain aspect ratios and scale separation critically affect LES accuracy; recommends YAR>=3, XAR>=6, SS>=12 LES with Buckingham Pi-based scaling analysis LES, turbulence, canopy, Buckingham-Pi, CFD
38 Savvides & Papadopoulos 2024 Neural network for reliability analysis of shallow foundation failure on cohesive soils FNN achieves relative error <10^-5 for limit pressure and displacement prediction from Monte Carlo data Feed-forward NN trained on MC simulations with Modified Cam Clay neural-network, shallow-foundation, reliability, geotechnical
39 Schmudderich et al. 2024 Two-step dynamic FEM-FELA for seismic slope stability FEM-FELA approach is simpler and more straightforward than LEM for seismic slope FoS with pore pressure effects Coupled dynamic FEM + finite element limit analysis seismic, slope-stability, FEM, FELA, liquefaction
40 Schulz et al. 2024 Wind turbine rotors in surge: unsteady aerodynamics of FOWTs Returning wake effect (from helicopter aerodynamics) identified at typical wave frequencies for IEA 15MW rotor; OpenFAST shows notable differences vs. FVW Wind tunnel + free-vortex-wake panel/lifting-line methods FOWT, surge, unsteady-aero, returning-wake, experimental

Synthesis

CONSENSUS

  1. Foundation damping and SSI matter for OWT design. Multiple papers (Guo, Sah, Orakci, Gkougkoudi-Papaioannou) converge on the conclusion that soil-structure interaction, foundation damping, and higher-mode effects significantly influence the dynamic response of monopile-supported offshore wind turbines. Neglecting these leads to unconservative or inaccurate load predictions.

  2. ML/AI surrogates are viable replacements for expensive physics simulations. Haghi (DEL surrogate), Savvides (shallow foundation reliability), Hameed (liquefaction), and Hsiao (lateral spreading) all demonstrate that data-driven models trained on physics-based simulation outputs achieve high accuracy (R2 > 0.90) at orders-of-magnitude lower computational cost. Transfer learning further extends applicability across conditions.

  3. FOWT platform motions fundamentally alter wake dynamics. Messmer, Hubert, and Schulz independently show that surge, heave, and side-to-side motions of floating platforms introduce nonlinear wake phenomena (lock-in, returning wake effect, accelerated recovery) that fixed-turbine wake models cannot capture.

  4. Scour remains the leading cause of bridge failure. Lee and Millar both confirm this and propose complementary monitoring approaches (MEMS arrays vs. computer vision) to detect scour progression in real time.

  5. Physics-enhanced ML outperforms pure data or pure physics alone. Haywood-Alexander's survey and Liu's review both conclude that hybrid PEML approaches combining domain knowledge with data-driven flexibility produce more generalizable and accurate models in structural/geotechnical mechanics.

DEBATES

  1. Uniform vs. independent pitch control for FOWTs. Guan shows independent pitch reduces tower oscillation but increases fatigue in the flow-parallel direction, presenting an unresolved design trade-off with no clear winner for all conditions.

  2. Which constitutive model for cyclic sand behaviour? Orakci compares SANISAND-MS and PM4SAND and finds different levels of accuracy and calibration complexity. The community has not converged on a standard model for monopile cyclic loading assessment.

  3. Unsteady aerodynamic phenomena relevance. Schulz explicitly states "no consensus has been reached in the research community on which unsteady aerodynamic phenomena are relevant" for FOWT rotors. Their discovery of the returning wake effect adds a new phenomenon to this open debate.

  4. FEM vs. alternative numerical methods in geotechnics. Larionov proposes successive approximation as an alternative to FEM, while Schmudderich combines FEM with FELA. The optimal numerical framework for coupled hydro-mechanical geotechnical problems remains contested.

GAPS

  1. No validated cyclic soil models at the boundary value problem level for OWT monopiles. Orakci's work remains at the material (element test) level; validation against full-scale field data under cyclic lateral loading is absent.

  2. Lack of cybersecurity CBA methods specific to offshore wind. Kam's systematic review found only 6 qualifying papers; detailed cost modelling for offshore wind CPS is virtually nonexistent.

  3. Limited field validation of aeroelastic flap models. Gamberini's work is one of the first field validations of trailing edge flap aeroelastic models, highlighting that most codes use only partially validated flap modules.

  4. Missing connection between pore pressure response and structural response for monopiles. Gkougkoudi-Papaioannou measures pore pressures but does not close the loop to structural load/displacement consequences.

  5. Explainability in geotechnical ML models. While Hsiao applies SHAP to liquefaction prediction, most geotechnical ML papers (Hameed, Savvides, Liu review) treat models as black boxes. The gap between ML performance and engineering trust persists.

  6. Frequency-domain methods for FOWTs are underdeveloped. Lamei notes that most FOWT analyses use time-domain methods; frequency-domain aero-hydro-elastic tools remain rare despite computational advantages.

METHODS

  • Physical modelling: Centrifuge (Morley, Kim), wind tunnel (Messmer, Hubert, Schulz), wave basin (Gkougkoudi-Papaioannou), TLP model testing (Ren) -- scale ranges from 1:23 to 1:50.
  • Numerical: FEM/FDM/DEM (Jilo, Jasim, Schmudderich), coupled aero-hydro-servo-elastic codes (FAST, HAWC2, BHawC, OpenFAST), LES (Sathe), PLAXIS/OpenSees (Orakci).
  • AI/ML: ANN/FNN (Savvides, Liu), XGBoost+SHAP (Hsiao), ELM+DOA (Hameed), convolutional autoencoder (Rautela), ML surrogate for DEL (Haghi), robust autoencoder with Mahalanobis distance (Roy).
  • Dimensional analysis: Buckingham Pi theorem applied in process engineering (Onyenanu) and LES domain scaling (Sathe).
  • Monitoring/SHM: MEMS scour sensors (Lee), computer vision (Millar), ambient vibration resonance tracking (Moore), drone+DL inspection (Memari).

BENCHMARKS

Benchmark Paper Value
DEL surrogate accuracy Haghi High correlation with aeroelastic outputs; transfer learning for wake conditions
Liquefaction ELM-DOA R2 Hameed 0.935
Liquefaction XGBoost (lateral spreading) Hsiao SHAP-interpreted; trained on 2011 Christchurch data
Shallow foundation FNN relative error Savvides < 10^-5
FOWT layout profit gain Hietanen EUR 34.5M over grid layout (ScotWind site 10)
Large-farm optimization speedup Rodrigues 75x with algorithmic differentiation for 500 turbines
Soil improvement (organic soil) Jasim 666% bearing capacity increase (8 end-bearing columns, L/D=6)
PM2.5 forecasting (autoencoder) Rautela SSIM > 0.60, PSNR 28-30 dB
Wake recovery Strouhal range Messmer Lock-in at St in [0.2, 0.55]; enhanced recovery St in [0.2, 0.9]
Returning wake effect Schulz Occurs at ratio of 3P to surge frequency; notable for IEA 15MW at typical wave periods