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¶
-
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.
-
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.
-
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.
-
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¶
-
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.
-
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.
-
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¶
-
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.
-
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.
-
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.
-
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.
-
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.