Skip to content

Batch 03 Agent 5 -- Literature Synthesis (positions 561-600)

Generated: 2026-04-17 | Files scanned: 40 (2 garbled/unreadable)


Individual Paper Summaries

# Author(s) Year Title Core Finding Method Tags
1 Xie et al. 2022 Gaussian Process Regression for Missing Geotechnical Data (Garbled text -- unreadable) GPR, kernel selection geotechnics, ML, missing-data
2 Xu et al. 2022 3D Numerical Analysis of Stilling Basin, Kala Hydropower 3D FE evaluation of stilling basin and bedrock stability at Kala hydropower station 3D FEM (FLAC3D) hydraulic-structures, FEM, rock-mechanics
3 Yeter et al. 2022 AI-aided Life Extension of Offshore Wind Support Structures Review synthesizing AI/ML for SHM, fatigue, corrosion, and risk-based life extension of OWT support structures Literature review OWT, AI, SHM, fatigue, life-extension, review
4 Yu 2022 AIME: Autoencoder-based Integrative Multi-omics Embedding Autoencoder integrates multi-omics data with confounder adjustment, outperforming PCA/other methods on CCLE and TCGA datasets Autoencoder, multi-omics integration autoencoder, multi-omics, confounder, bioinformatics
5 Zhang et al. 2022 Spatial Variability of Potato Yield on Sand Soil, NW China Soil sand content correlates with >66% of tuber variables; K-means clustering identifies critical factors for site-specific management Geostatistics, K-means clustering spatial-variability, geostatistics, agriculture
6 Zeng et al. 2022 Undrained Stability of Dual Tunnels in Layered Soils Inclination-fixed worst-band exists for placing bottom tunnel; three collapse patterns identified FELA (finite element limit analysis) tunnelling, stability, layered-soil, FELA
7 Zhao et al. 2022 PDF and Credible Region for Multivariate Geotechnical Data (Garbled text -- unreadable) Probabilistic, multivariate geotechnics, uncertainty, statistics
8 Zhou & Yang 2022 Water Droplet High-Velocity Impact on WT Blade Surface Contact force has two peaks; forces and strains decrease with lower impact velocity and angle SPH-FEM coupling blade-erosion, impact-dynamics, SPH-FEM
9 Zhu et al. 2022 Data-driven Li-ion Battery Capacity from Voltage Relaxation ML on relaxation voltage features achieves RMSE 1.1%; transfer learning extends to other chemistries at <1.7% RMSE ML, transfer learning battery, data-driven, transfer-learning
10 Zou, Niu et al. 2022 Initial Imperfections on Fatigue of Floating WT Tower Flanges Review: initial imperfections significantly affect fatigue assessment of tower flange connections in FWTs Review FWT, fatigue, flange, imperfections, review
11 Zou, Lourens & Cicirello 2022 Virtual Sensing of Subsoil Strain via GPLFM for OWT Monopiles First in-situ validation of GPLFM for below-mudline strain estimation; robust under high soil-foundation model uncertainty GPLFM, Kalman filter, Bayesian OWT, virtual-sensing, SHM, monopile, Gaussian-process
12 Zuo et al. 2022 Deep Learning in Optical Metrology: A Review DL outperforms traditional methods in fringe denoising, phase retrieval, phase unwrapping, and error compensation Review (DL for optical measurement) deep-learning, metrology, review
13 Kelly 2022 Lifetime Performance of Wind Turbines and Wind Farms (PhD) Blade roughness reduces AEP by 2.9-8.6%; UK/US/AU farm performance declines at 0.26/0.72/0.43 CF pp/yr; ML prediction improves MAE by 56-90% BEM, long-term correction, ML OWT, performance-degradation, roughness, ML
14 Abdullah et al. 2023 Geotechnical Simulation of Historical Buildings, Najran Dissolution of clay minerals and salt crystallization are key degradation drivers; PLAXIS 3D confirms field observations FEM (PLAXIS 3D), SEM/EDX/XRD geotechnics, FEM, heritage, weathering
15 El Absawy et al. 2023 Variable Cross-Section Tower-Soil-Structure Interaction for OWT Variable cross-section tower model improves dynamic response accuracy by 15% over constant cross-section models Euler-Lagrange, BEM, Morison, JONSWAP OWT, SSI, tower-dynamics, numerical-model
16 Achouri et al. 2023 SHM via PSO-BOA-ANN Using FRF Hybrid PSO-BOA-ANN outperforms individual algorithms for damage severity prediction from natural frequencies PSO-BOA-ANN, modal analysis, FRF SHM, swarm-optimization, ANN, damage-detection
17 Ahmed & Paskyabi 2023 Negative Shear Effects on 15 MW OWT in Low-Level Jets Significant structural load variations during low-level jet events with negative shear above hub height Aero-hydro-servo-elastic simulation OWT, low-level-jet, structural-loads, 15MW
18 Alajmei 2023 Proppant Distribution via Buckingham Pi Correlation Novel correlation predicts proppant distribution across perforation clusters with <20% error using Buckingham Pi dimensional analysis Buckingham Pi, experimental correlation hydraulic-fracturing, dimensional-analysis, proppant
19 Amjadian et al. 2023 Seabed Sediment Characterization at Offshore RE Sites Characterizes seabed sediments at contrasting offshore renewable energy sites to inform foundation design Field characterization seabed, sediment, offshore-RE, site-characterization
20 Arafa et al. 2023 RN-Autoencoder for Imbalanced Cancer Genomic Data Autoencoder for dimensionality reduction + RN-SMOTE for class imbalance yields up to 18% accuracy improvement Autoencoder, RN-SMOTE, classification autoencoder, imbalanced-data, genomics, cancer
21 Bahdasariants et al. 2023 Joint Impedance in Dynamic Leg Simulations Evaluates whether joint impedance improves dynamic leg simulations with explicit and implicit solvers Biomechanical simulation biomechanics, impedance, simulation
22 Bapir et al. 2023 SSI Review: Modeling Techniques for Seismic Response State-of-the-art review of SSI modeling techniques and their impact on seismic response of building structures Review SSI, seismic, review, buildings
23 Barreto & Ortega 2023 OWT Dynamic Response Under Peruvian Conditions Dynamic response of 5MW monopile OWT under Peruvian stochastic wind-wave environment using FAST Aero-hydro-servo-elastic (FAST) OWT, dynamic-response, monopile, FAST
24 Bateman et al. 2023 p-y Curves from ROBOCONE Tests via Similarity Approach Similarity-based linear transformation converts in-situ ROBOCONE lateral load curves to p-y curves for pile design in clay ROBOCONE, similarity approach, CPT p-y-curves, lateral-loading, piles, clay, in-situ
25 Bernardi et al. 2023 Tower Modeling Effect on NREL 5MW Aero-elastic Response (File unreadable) Aero-elastic modeling OWT, tower-modeling, aero-elastic
26 Boschi et al. 2023 Design of Anchored Wire Meshes for Slope Stabilization Interaction model based on 3D FE replaces need for ad hoc FE analyses; validated against punching tests 3D FEM, macroelement model slope-stability, wire-mesh, FEM, design
27 Bouneguet et al. 2023 Dynamic Response of Suction Caisson Foundations Dynamic impedance of suction caissons highly affected by diameter, soil stiffness, and caisson length 3D FEM, substructuring suction-caisson, dynamic-impedance, SSI, OWT
28 Carvalho et al. 2023 Systematic Review of Pile Capacity Prediction ANN gaining popularity over traditional semi-empirical methods; gaps in helical piles and instrumented tests Systematic literature review pile-capacity, ANN, review, foundations
29 Cerfontaine et al. 2023 Silent Piling for Offshore Jackets: DEM + Centrifuge Screw piles installed by rotary jacking at low reaction force; DEM matches centrifuge data; tensile head force at low advancement ratios DEM, centrifuge modelling screw-pile, silent-piling, offshore, DEM, centrifuge
30 Chen, Cao & Zhu 2023 Data-Driven SHM and Predictive Maintenance: Survey Most data-driven SHM remains in simulation; real-world implementation faces digressive performance and inefficient computing Survey (IEEE IoT Journal) SHM, data-driven, predictive-maintenance, review
31 Chen & Fan 2023 Factor Selection for Landslide Susceptibility with ML/DL Factor selection improves ML models but can reduce DL model accuracy; autoencoder architecture improves DL performance ML (RF, SVM, XGBoost), DL (CNN, LSTM), autoencoder landslide, ML, DL, feature-selection, autoencoder
32 Cho et al. 2023 Sediment Connectivity Linked to Hydrologic Pathways Conceptual model linking sediment sources/connectivity to water cycle, geomorphic processes across scales Conceptual model sediment, hydrology, geomorphology, connectivity
33 Christiansen et al. 2023 Anthropogenic Mixing by OWT Foundations in North Sea OWT foundation drag causes large-scale mixing impacts in shallow stratified North Sea waters Numerical ocean modelling OWT, environmental-impact, mixing, North-Sea
34 Chrysanthopoulos et al. 2023 IoT TDR/FDR Sensors for Soil Water Content IoT-based TDR+FDR system enables real-time soil moisture monitoring through full unsaturated zone depth TDR, FDR, IoT soil-moisture, IoT, monitoring, agriculture
35 Climaco et al. 2023 mrDMD for WT Gearbox Damage Detection Multi-resolution DMD extracts damage features from stochastic vibration data better than Fourier/TSA/EMD mrDMD, vibration analysis condition-monitoring, gearbox, OWT, mrDMD
36 Cui et al. 2023 Stress Effects on Soil Thermal Conductivity and Energy Piles Stress state affects thermal conductivity and heat transfer efficiency differently in saturated vs unsaturated soils Numerical (thermo-hydro-mechanical) energy-piles, thermal-conductivity, unsaturated-soil
37 Cunha et al. 2023 ML for Structural Dynamics and Vibroacoustic: Review ML applied to SHM, active noise/vibration control, and vibroacoustic surrogate design; physics-guided ML is key future direction Review ML, SHM, vibration-control, surrogate, review
38 Davey & Ochoa-Cabrero 2023 Extended Finite Similitude and Dimensional Analysis Finite similitude theory extends beyond physics to any quantitative theory; countably infinite similitude rules overcome dimensional analysis limitations Finite similitude, Lie group scaling, similitude, dimensional-analysis, Buckingham-Pi
39 Diachenko-Bohun & Sosnova 2023 Environmental Monitoring in Ecology Education History and pedagogy of environmental monitoring for ecology students (Ukrainian language) Pedagogical review environmental-monitoring, education, ecology
40 Exner et al. 2023 Dynamic vs Static Impedance in Electroconvulsive Therapy Dynamic and static impedance are strongly correlated in bifrontal ECT; low static impedance correlates with better seizure quality Retrospective clinical analysis, mixed-effects regression ECT, impedance, clinical, neuroscience

SYNTHESIS

CONSENSUS

  1. Data-driven methods dominate structural and geotechnical monitoring. Papers 3, 11, 13, 16, 30, 35, 37 all converge on the conclusion that ML and data-driven approaches (ANN, GP, transfer learning, mrDMD) substantially improve damage detection, capacity estimation, and virtual sensing compared to physics-only baselines. The trend holds across OWT SHM, battery health, gearbox monitoring, and pile capacity prediction.

  2. Soil-structure interaction fidelity matters. Papers 15, 22, 23, 27 agree that simplified SSI models (fixed base, constant cross-section towers) introduce significant errors in dynamic response. Variable cross-section and frequency-dependent impedance representations consistently yield 10-15% accuracy improvements.

  3. Autoencoders are versatile across domains. Papers 4, 20, 31 show autoencoders effective for dimensionality reduction and data integration in genomics, cancer classification, and landslide susceptibility. Their ability to learn compressed, noise-reduced representations transfers well regardless of application domain.

  4. Fatigue and degradation govern OWT life-cycle costs. Papers 3, 10, 13, 35 agree that aerodynamic performance loss (2.9-8.6% AEP from roughness), flange imperfections, and gearbox damage are the primary life-limiting mechanisms for offshore wind assets.

DEBATES

  1. Feature selection for DL models. Paper 31 (Chen & Fan) finds that traditional feature selection methods (IGR, RFE) actually reduce DL model accuracy for landslide prediction, contradicting the common practice of pre-filtering inputs for all model types. Autoencoder-based selection is recommended instead for DL, creating a methodological split between ML and DL pipelines.

  2. Physics-based vs data-driven SHM. Paper 30 (Chen, Cao & Zhu) highlights that most data-driven SHM works remain in simulation with "digressive performance" in real-world deployment, while paper 11 (Zou et al.) demonstrates successful in-situ GPLFM validation. The gap between lab/simulation results and field deployment remains a live debate.

  3. Dimensional analysis limitations. Paper 38 (Davey) argues that classical Buckingham Pi analysis offers no solution when scale effects are present, while paper 18 (Alajmei) successfully applies Buckingham Pi for proppant distribution prediction (<20% error). The disagreement reflects domain-specific applicability rather than fundamental invalidity.

GAPS

  1. In-situ validation of virtual sensing for OWT. Paper 11 is explicitly one of the "first cases of in-situ validation" of GPLFM below mudline. The field needs far more operational validation campaigns to build confidence in virtual sensing replacing physical sub-mudline instrumentation.

  2. Real-world deployment of data-driven SHM. Paper 30 identifies the transition from simulation to field as a major unsolved challenge. Computing efficiency for real-time analytics and performance degradation under environmental variability remain open problems.

  3. Helical and screw pile design methods. Paper 28 identifies a lack of studies on helical piles and instrumented pile load tests. Paper 29 begins addressing screw pile installation but the design framework is nascent.

  4. Silent piling acceptance. Paper 29 demonstrates feasibility of noise-free rotary jacking installation for offshore jackets, but scaling to full production and regulatory acceptance are unaddressed.

  5. Environmental impacts of large-scale OWT deployment. Paper 33 shows foundation-induced mixing alters North Sea stratification, but long-term ecological consequences are not quantified.

METHODS INVENTORY

Method Papers Domain
FEM (PLAXIS, FLAC3D, general) 2, 6, 14, 26, 27 Geotechnics, tunnelling, slope stability
Aero-hydro-servo-elastic (FAST, OpenFAST) 17, 23 OWT dynamics
ML classifiers (RF, SVM, XGBoost, ANN) 16, 28, 31 SHM, pile capacity, landslide
Autoencoders 4, 20, 31 Multi-omics, cancer, landslide
Gaussian process (GPR, GPLFM) 1, 11 Geotechnics, OWT virtual sensing
DEM + centrifuge 29 Screw pile installation
SPH-FEM coupling 8 Blade erosion
mrDMD (multi-resolution dynamic mode decomposition) 35 Gearbox condition monitoring
Transfer learning 9 Battery capacity estimation
Buckingham Pi / dimensional analysis 18, 38 Proppant distribution, scaling theory
Geostatistics + K-means 5 Agricultural spatial variability
FELA (finite element limit analysis) 6 Tunnel stability
Swarm optimization (PSO, BOA) 16 SHM damage detection
TDR/FDR IoT sensors 34 Soil moisture monitoring

BENCHMARKS

Paper Metric Value Context
Zou et al. 2022 (#11) In-situ strain validation First GPLFM field validation Westermeerwind Park, Netherlands
Zhu et al. 2022 (#9) RMSE capacity estimation 1.1% (base), <1.7% (transfer) 130 commercial Li-ion cells
Kelly 2022 (#13) AEP loss from roughness 2.9-8.6% Exemplar rotors via BEM
Kelly 2022 (#13) CF decline rate 0.26-0.72 pp/yr UK, US, AU wind farms
Kelly 2022 (#13) ML prediction improvement 56-90% MAE reduction vs power-curve method
El Absawy 2023 (#15) Variable vs constant tower 15% accuracy improvement NREL 5MW monopile
Arafa 2023 (#20) Classification accuracy gain up to 18% Colon cancer genomic data
Alajmei 2023 (#18) Proppant distribution error <20% Buckingham Pi correlation
Achouri 2023 (#16) PSO-BOA-ANN vs standalone Best accuracy for damage severity Cracked aluminum beams

38 papers synthesized (2 garbled). Dominant themes: OWT structural integrity, data-driven SHM, ML/autoencoder applications, soil-structure interaction, and scaling/dimensional analysis.