Search Results (886)

Background/Objectives: Schatzker type is assigned early, usually from the anteroposterior (AP) radiograph. A single benchmark accuracy cannot say whether a model read the fracture, the anatomy around it, the annotation, or how the archive was assembled. We ran four inexpensive controls to separate those contributions. Methods: We benchmarked a ResNet-50 on PlaTiF, a 2026 public release built for artificial-intelligence research that pairs 421 AP knee radiographs from 186 patients with expert Schatzker labels and per-image tibial segmentations. Evaluation used stratified group five-fold cross-validation grouped by patient, five seeds and balanced accuracy. Inputs were cropped to the expert tibial segmentation shipped with the dataset, an oracle localisation unavailable at deployment. Four controls ran on identical folds: a regression given no pixel content; ablation of the tibial pixels with its complement; a regression on the expert mask alone; and an augmentation audit for label-erasing invariances. Results: Among the 128 fracture patients the network reached 0.345 ± 0.030 six-class balanced accuracy, +0.168 over a non-anatomical baseline fitted on the same folds and the same labels (95% CI +0.106 to +0.230, p = 0.002). Recall was graded: 0.72 for Schatzker VI, 0.11 for V and 0.04 for IV, the last two below chance (0.167). Erasing the tibial pixels left 0.257 ± 0.025, read on its own as the target bone being unused; its complement, the tibia with everything else removed, reached 0.367 ± 0.012, and the whole radiograph, which carries both, only 0.297 ± 0.032 (+0.071 for the tibia alone, 95% CI +0.031 to +0.111, p = 0.008). A regression on the expert mask alone reached 0.213 ± 0.034 and was not distinguishable from the erased model. On fracture versus no classifiable fracture the network reached 0.833 ± 0.028 against 0.814 ± 0.016 for a model given no pixels (p = 0.264), and a coronal computed tomography section accompanied 126 of 128 fracture patients but 24 of 58 others (p = 2.9 × 10⁻¹⁹). Conclusions: Each headline number admitted an explanation other than the fracture in the target bone. An ablation reported without its complement misstated where the signal lay, and an augmentation audit overturned our own explanation for the failure of type IV. Controls of this kind cost minutes, and this study illustrates why they can be informative when a benchmark is built on a retrospective clinical archive. Full article

Shale gas reservoirs can develop extensive fracture networks via hydraulic fracturing. Nevertheless, gas stored in matrix pores remains hard to produce efficiently. Thermal stimulation helps trigger secondary fracture growth inside the matrix and forms multiscale gas-water flow channels to improve reservoir productivity. In this study, a thermo–hydro–mechanical (THM) model is established using COMSOL Multiphysics. The model accounts for fluid seepage, heat transfer, mechanical deformation, and fracture evolution, and assesses shale gas productivity under combined hydraulic fracturing and thermal injection. The effects of initial permeability, injection pressure, and in situ stress on fracture propagation are investigated. Three heating strategies are considered: preheating, heating during fracturing, and continuous heating throughout production. Simulations reveal that thermal injection improves displacement, permeability, and porosity around injection locations. Continuous heating achieves the most obvious fracture expansion. After 6000 days, the daily gas production of the three schemes reaches 50.6839, 73.2167, and 81.403 million m³/d, corresponding to production increases of 44.5% and 60.6%. Full article

Ultra-low-permeability tight oil reservoirs, including the Chang 6 and Chang 8 formations in the Changqing Oilfield and the Fuyu reservoir in the peripheral Daqing Oilfield, are characterized by poor reservoir properties and limited waterflooding efficiency. Conventional areal waterflooding either fails to establish effective displacement or results in rapid local water breakthrough, leading to rapid production decline and a recovery degree of less than 10%. Large-scale refracturing combined with modification of the water injection strategy has, therefore, become an important approach for improving single-well productivity. However, long-term injection–production imbalance may cause substantial formation-energy depletion and an increase in horizontal stress contrast, which are unfavorable for the development of complex fracture networks during refracturing. To investigate the mechanism and optimize the design of pre-fracturing energy enhancement, rock-mechanics experiments were first conducted on cores from the Fuyu reservoir in the Daqing Oilfield. The resulting pore pressure and stress responses were interpreted based on poroelastic coupling and the effective-stress principle. A three-dimensional coupled reservoir–geomechanical model was subsequently established for a Chang 6 tight oil block in the Changqing Oilfield using formation-specific geological, petrophysical, geomechanical, and production data, and field performance was further used for validation. The laboratory results show that pre-fracturing water injection increases pore pressure, reduces effective confining stress, and decreases the horizontal principal-stress difference, thereby promoting a transition in rock failure from isolated shear fractures toward intersecting fracture patterns. The laboratory-derived mechanical trends were transferred to the Chang 6 model primarily at the mechanistic level, while quantitative parameters were recalibrated using reservoir-specific data. By establishing the relationship between the energy enhancement ratio, defined as the ratio of injected fluid volume to cumulative produced fluid volume, and formation pressure recovery, and further considering sensitivity, economic feasibility, and operational constraints, the optimal energy enhancement ratio was determined to be 0.8–1.0. These results clarify the geomechanical mechanism and key design parameters of pre-fracturing energy enhancement and provide practical guidance for refracturing design in ultra-low-permeability tight oil reservoirs. Full article

Particle-reinforced metal matrix composites, wherein preform pore structure dominates liquid infiltration behavior and final composite quality, are essential for high-performance industries. Conventional empirical models predict pore characteristics for bimodal preforms based on ideal particle stacking assumptions yet ignore real compression-induced microstructural changes including particle contact compaction and particle fracture, yielding systematic deviations from actual pore characteristics. This study adopted high-resolution 3D X-ray micro-computed tomography (μ-CT) to quantify such discrepancies for bimodal SiC_p preforms across six coarse-to-fine particle ratios (0–100%). Three-dimensional pore network models were extracted to quantify key characteristics including areal porosity, surface area, and pore/throat dimensions. The results demonstrated that the average areal porosity fell to a minimum at a 67% coarse fraction then rose, while the pore distribution homogeneity steadily declined. Additionally, μ-CT measurements revealed that particle contact compactness reduced the particle surface area per unit volume at coarse fractions below 25% whereas particle fracture increased it at fractions above 25%, deviating significantly from empirical predictions. Larger coarse particle fractions reduced pore/throat quantities but increased their average size and volume. Beyond using established pore network extraction, this work distinguishes these two competing micro mechanisms and provides reasonable datasets to support bimodal preform optimization for composite manufacturing. Full article

Diabetic refractory wounds are a prevalent and severe complication of diabetes, whose pathological progression is jointly mediated by multiple factors, including oxidative stress imbalance, chronic inflammation, impaired angiogenesis, bacterial infection, and biofilm formation. Current clinical hydrogel dressings generally suffer from drawbacks such as single-function performance, potential toxicity of nano-components, static networks incompatible with dynamic wound conditions, and the absence of bionic repair signals. Therefore, they cannot simultaneously satisfy the dual repair requirements of complex pathological microenvironments and dynamic mechanical properties for diabetic wounds. In this study, a multi-functional dynamically responsive composite hydrogel (MC group) with high-efficiency antioxidant, antibacterial, and pro-angiogenic capacities was fabricated. Using SDS-C₁₈ micelles as hydrophobic units, a rigid–flexible dual-network framework was constructed with polyvinyl alcohol (PVA) and methacrylated hyaluronic acid (HAMA). Manganese dioxide nanozymes were introduced to scavenge reactive oxygen species (ROS) and mitigate oxidative stress. Calcium-ion-mediated dynamic micelle reconstruction was adopted to regulate the hydrophilic–hydrophobic balance, while achieving antibacterial effects and facilitating tissue regeneration. In vitro experiments verified that the MC hydrogel possesses mechanical properties well-matched to human soft tissues (fracture stress: 25 kPa) and excellent biocompatibility (cell viability > 100%, hemolysis rate: only 0.13%). It also exhibits prominent antioxidant activity (DPPH radical-scavenging rate: 36.95%), antibacterial performance (>99.86% bactericidal rate against Staphylococcus aureus, survival rate of Escherichia coli reduced to 15.95%), and cell-migration-promoting activity (endothelial cell migration rate of 83.72% and mouse fibroblast migration rate of 90.88% within 24 h). In the full-thickness skin defect model of diabetic mice, the wound-healing rate reached 99% on day 16. Moreover, it promoted ordered collagen deposition, skin appendage regeneration, and functional microvascular reconstruction, thereby accomplishing high-quality tissue repair. This design synergistically intervenes in multiple pathological links of diabetic wounds, overcomes several key limitations of existing dressings, and provides an innovative strategy for developing smart dressings. Full article

Deep shale gas reservoirs are subjected to anisotropic in situ stresses that break the directional symmetry of hydraulic-fracture growth, promote propagation along the maximum horizontal principal stress, and suppress transverse spreading. This study investigates geomechanical and injection controls on fracture-network evolution in the Yi214 block of the Changning shale gas field using a field-calibrated numerical workflow. The model was calibrated against pre-design microseismic-derived metrics from Well Yi202, with relative differences of 1.8% for fracture length and 3.9% for effective fracture volume; this comparison is treated as calibration rather than independent multi-well validation. A dimensionless directionalization index, I_d = (L/L₀)/(V/V₀), is defined relative to the equal-horizontal-stress case (Δσ_h = 0) as a global proxy for the concentration of longitudinal extension relative to volumetric spreading. One-factor-at-a-time screening evaluated Young’s modulus, Poisson’s ratio, horizontal stress difference, injection rate, fluid-volume intensity, and proppant loading, while final parameter combinations were treated as well-specific engineering design selections rather than mathematical global optima. As Δσ_h increased from 0 to 15 MPa, fracture volume decreased by approximately 44% and average fracture length increased by approximately 12%, raising Id from 1.00 to 1.99; the increase in Id was driven predominantly by reduced volume retention rather than length growth alone. The final combined design cases increased simulated fracture volume by 18–27%, while the reported model-based EUR forecasts increased by 36–40%. The results provide a symmetry-based interpretation of stress-controlled fracture directionalization and a field-calibrated basis for well-specific stimulation design in deep shale reservoirs. Full article

Large hydrocarbon fields are being developed in northern Sakhalin, whereas southern Sakhalin contains prospective resources hosted in unconventional, low-permeability siliceous source rocks. Their development requires stimulation to create conductive fracture networks, whose long-term integrity is critical for production feasibility. This study investigates permeability changes in four artificially fractured specimens subjected to cyclic confining pressure; the specimens are treated as case studies rather than as a statistically representative formation-scale data set. Cylindrical cores, 30 mm in diameter and approximately 30 mm long, containing an induced axial fracture were hydraulically tested under biaxial cyclic confinement. The tests showed irreversible loss of fracture conductivity, with residual permeability after unloading amounting to 7.1–37.4% of the initial value. Direct application of laboratory data to field scale fracture longevity models is inappropriate because cylindrical specimens develop nonuniform circumferential stresses and heterogeneous closure. A procedure is proposed to transfer core scale measurements to a planar fracture subjected to uniform normal stress in the rock mass. Fracture aperture was quantified by X-ray computed tomography (CT) and incorporated into a cell-based contact–hydraulic model accounting for geometric and hydraulic aperture. A nonuniform closure function was used to reconstruct the closure field and correct the laboratory results. At maximum pressure, Kmass/Klab ranged from 0.034 to 0.93 for the three specimens reproduced with acceptable fit; sample 3–4 was retained only as a diagnostic case because of its high root-mean-square error (RMSE). These values are model-based single-fracture scenario estimates pending direct-normal-loading validation. Full article

Conventional Rate Transient Analysis (RTA) models, based on homogeneous fracture assumptions, are inadequate for characterizing flow in complex fracture networks of heterogeneous unconventional reservoirs. This study develops a fractal-based RTA (FD-RTA) workflow integrating lithofacies analysis, microseismic fracture interpretation, and post-fracturing production data from the Yingxiongling shale oil field in the Q’aidam Basin. The workflow is applied to eight horizontal wells completed in layered and laminated dolomites. Results show that the two lithofacies exhibit distinct fractal flow behaviors. Layered dolomite tends to develop preferential flow pathways, characterized by rapid initial depletion followed by declining supply capacity, with the half-flow dimension (δ) decreasing from 0.299 to 0.074 during production. Laminated dolomite displays stronger fracture-matrix interaction and sustained production performance, with δ increasing from 0.469 to 0.678 as multi-scale fractures are progressively activated. The FD-RTA workflow effectively links fracture complexity with production behavior, providing a dynamic characterization tool for evaluating hydraulic fracturing effectiveness in shale oil reservoirs. Full article

Wrist fractures present significant challenges in clinical diagnosis, often leading to treatment delays and compromised recoveries. Manual diagnosis is resource-intensive and error-prone. To address these challenges, we develop DeepWristFNet, a compact convolutional architecture designed for end-to-end wrist fracture classification using a small dataset of 193 wrist X-ray images. The DeepWristFNet architecture integrates multi-scale convolutional operations with Fire and Shuffle modules within a compact network design, followed by fully connected layers for binary classification. We applied data pre-processing techniques such as data augmentation, image enhancement, and image resizing to increase the number of images, improve image quality, and resize images to match the DeepWristFNet input size. The proposed method comprised three phases. In the first phase, we trained, validated, and tested end-to-end and achieved validation and testing accuracies of 99.04% and 87.93%, respectively. Testing was performed on a hold-out subset of image instances that was kept separate from model development. The evaluated hold-out images originated from the same dataset distribution and included the corresponding augmented variants. In the second phase, we further evaluated the learned representation by extracting deep features from the first fully connected layer of DeepWristFNet. ReliefF was then used to select informative features, which were subsequently evaluated using 10 conventional machine learning classifiers. Out of 10 classifiers, 5 classifiers, i.e., Efficient linear SVM, quadratic SVM, Narrow NN, wide NN, and medium NN, achieved 100% testing accuracy on unseen samples. In the third phase, an auxiliary Fuzzy Inference System provides an intensity-based foreground-background representation of the X-ray images. This representation provides complementary visual information for interpretation but is not intended to directly classify or localize fractures. Grad-CAM is additionally used to visualize image regions contributing to the DeepWristFNet predictions, providing a model-specific explanation of the classification decision. Additionally, we evaluated how well the proposed DeepWristFNet approach performed against cutting-edge deep transfer learning models. In the evaluated experiments, DeepWristFNet outperformed the compared pre-trained deep learning architectures on the unseen hold-out subset from the same dataset distribution (test set). This study demonstrates the potential of DeepWristFNet for wrist fracture classification under a small-data setting. However, further evaluation on larger, independently collected clinical datasets is required to establish its robustness, generalizability, and suitability for clinical decision support. Full article

Background/Objectives: To achieve effective clinical decision-making from radiographic images, accurate diagnosis of bone fractures is required, but this is difficult due to anatomical variation, subtle fracture appearance, and variable imaging conditions. Existing deep learning studies for fracture detection are predominantly limited to binary classification or single anatomical regions, limiting their real-world clinical utility. This study demonstrates that a single deep learning model can effectively classify multi-region bone fractures by transforming the task into a 22-class classification problem, utilizing the publicly available MultiBoneX dataset. Methods: The proposed model utilizes an EfficientNet-B3 convolutional neural network integrated with a Convolutional Block Attention Module (CBAM) to enhance feature representation by focusing on diagnostically significant areas. We used regularized preprocessing, data augmentation, and structured training to support strong model learning and generalization. Model predictions were interpreted using Gradient-weighted Class Activation Mapping (Grad-CAM) to highlight the image regions that were most important for class selection. Results: When tested on a held-out test set of 3280 images, the model achieved an overall accuracy of 75.03% (95% CI: 73.57–76.43%), with precision, recall, and F1-score of 77.09% (95% CI: 75.28–78.77%), 72.85% (95% CI: 71.00–74.60%), and 73.45% (95% CI: 71.49–75.15%), respectively. The overall multi-class Matthews Correlation Coefficient (MCC) was 0.7361, providing an additional class-imbalance-aware measure of classification performance. Conclusions: These findings demonstrate the feasibility of a unified, multi-class system for diagnosing bone fractures across diverse anatomical sites, providing a scalable foundation for future AI-assisted radiography. Full article

Deep mining operations in metal mines face escalating challenges from high in situ stress environments, where conventional single-method ground control approaches often prove insufficient. This study proposes an innovative combined protection technology integrating four synergistic components: (i) prestressed rock bolt and plate support for immediate roadway reinforcement, (ii) hydraulic fracturing-based shielding to create fracture zones isolating the orebody from surrounding high-stress regimes, (iii) destress blasting at the orebody crown to form a stress-transfer barrier, and (iv) subsequent cemented paste backfill to provide long-term regional stability. Theoretical formulations are derived for each component: a bolt–rock composite bearing model quantifies the confinement provided by prestressed support; a fracture mechanics-based model characterizes the stress-shielding efficiency of hydraulic fracture networks; controlled blasting theory predicts the stress redistribution achieved through crown destressing; and a backfill arching model evaluates the long-term load-bearing capacity of cemented paste fill. A three-dimensional FLAC3D numerical model was established to simulate six working conditions: no protection, support-only, shield-only, pressure relief-only, backfilling-only, and the integrated four-component system combining support, shield, pressure relief and backfilling. The numerical results reveal that the integrated composite protection system reduces roadway subsidence by 86%. Furthermore, the synergistic effect among all components exceeds the sum of individual contributions, which verifies the necessity of adopting the comprehensive protection strategy under deep high in situ stress conditions. The research findings provide theoretical basis and engineering design references for safe and efficient mining in deep metal mines. Full article

Asphalt ultra-thin overlays have emerged as an efficient and cost-effective pavement preservation technology. However, because of their reduced thickness, their service life is often limited by cracking. Based on semi-circular bending (SCB) test data, this study developed three types of regression models: multiple linear regression (MLR), random forest regression (RFR), and artificial neural network (ANN) regression. The multifactor experimental matrix covered three test temperatures, three nominal maximum aggregate sizes, two binder types, three aging states, and four long-term aging durations. The results show that fracture energy (G_f), flexibility index (FI), and cracking resistance index (CRI) respond differently to individual factors. G_f and FI decreased as temperature increased, whereas CRI increased, indicating that CRI may be misleading at intermediate temperatures and should be interpreted cautiously. The aggregate size exhibits a non-monotonic relationship with the three indicators, reaching its maximum at a size of 8. Short-term aging increased the G_f, and the gain was more pronounced at higher aging temperature, reaching a maximum increase of up to 1.54 times. However, it also accelerated the subsequent G_f reduction during long-term aging, after the decline rate decreased. Higher short-term aging temperatures caused more pronounced deterioration in FI and CRI. Meanwhile, both indices gradually decreased with increasing long-term aging duration, and short-term aging further intensified this reduction. In model comparisons, ANN showed the most stable generalization. The factorwise trends were verified using both the Beta values and SHAP values, demonstrating the feasibility of the models for qualitative discrimination and trend interpretation. The integrated importance ranking identified long-term aging duration and temperature as the dominant factors. External validation further confirmed ANN as the most reliable model for predicting crack resistance evolution. These findings support crack-resistance prediction and provide guidance for optimizing the design of ultra-thin overlay asphalt mixtures. Full article

Human activities such as intensive groundwater abstraction and mine dewatering can profoundly disrupt the natural hydrological functioning of karst aquifers. The Transdanubian karst aquifer in Hungary represents one of Central Europe’s most prominent examples, where decades of coal-mine dewatering lowered groundwater levels by more than 40 m and fundamentally altered the natural recharge–discharge regime. Understanding and forecasting recovery in such complex karst systems remain challenging because of heterogeneous conduit–fracture networks, strong climate sensitivity, incomplete monitoring records, and uncertainty in long-term predictions. This study presents an integrated end-to-end machine learning framework for groundwater characterization and climate-constrained probabilistic forecasting. Monthly groundwater-level records (1970–2026) from five monitoring wells were first reconstructed using a hybrid Moving Average–Random Forest gap-filling approach, achieving high reconstruction accuracy (R² = 0.87–0.98). Self-Organizing Maps subsequently identified four hydrogeological states representing the dewatering, transition, recovery, and near-equilibrium phases, while inter-well weight-plane correlations (>0.95) confirmed strong basin-scale hydraulic connectivity. A Bootstrapped Random Forest model forced by bias-corrected COSMO-CLM precipitation projections under the SSP2-4.5 climate scenario generated probabilistic groundwater forecasts through 2030, achieving high predictive performance (NSE > 0.80; RMSE = 0.10–0.35 m). Forecast results indicate that the basin as a whole is approaching hydraulic equilibrium by 2030, with distinct well-specific trajectories including mild steady decline and near-stable water level. The proposed framework provides a robust and transferable methodology for groundwater characterization and long-term forecasting in complex karst and fractured aquifer systems under changing climatic conditions. Full article

Background: Osteoporosis is a multifactorial skeletal disorder characterized by reduced bone mass, impaired bone remodeling, and an increased risk of fractures, particularly under estrogen-deficient conditions. Artocarpus integer (Thunb.) Merr. contains prenylated flavonoids and chalcone derivatives with diverse biological activities; however, its anti-osteoporotic potential remains largely unexplored. This study investigated the phytochemical composition, molecular mechanisms, and anti-osteoporotic effects of A. integer leaf extract in an ovariectomized (OVX) rat model. Methods: Phytochemical profiling was performed using liquid chromatography–high-resolution mass spectrometry (LC–HRMS). Network pharmacology was employed to identify potential osteoporosis-related targets and signaling pathways. The anti-osteoporotic activity of the extract was evaluated in OVX rats through physiological and biochemical assessments, including body weight gain, uterine weight, serum biomarkers, femoral calcium content, and RT-PCR analysis of genes associated with osteogenesis, osteoclastogenesis, and estrogen signaling. Results: LC–HRMS analysis identified several bioactive compounds, including isobavachalcone, artocarpesin, morachalcone A, genistein, apigenin, luteolin, naringenin, catechin derivatives, and mangiferin. Network pharmacology revealed 96 overlapping targets between A. integer phytochemicals and osteoporosis-related genes, highlighting pathways involved in estrogen signaling, PI3K–Akt signaling, osteoclast differentiation, inflammation, and metabolic regulation. In vivo, OVX rats exhibited increased body weight gain, uterine atrophy, elevated leptin levels, reduced adiponectin concentrations, and decreased femoral calcium content. Treatment with A. integer attenuated OVX-induced metabolic alterations, improved adipokine profiles, and increased femoral calcium content, particularly in the medium-dose group. RT-PCR analysis demonstrated the upregulation of the osteogenic markers Runx2 and Osx, together with the downregulation of the osteoclastogenic markers TRAP. Conclusions: Artocarpus integer leaf extract exhibited promising anti-osteoporotic activity through the coordinated regulation of osteogenesis, osteoclastogenesis, estrogen-related signaling, and bone mineral preservation. These findings support the potential development of A. integer as a nutraceutical candidate for the prevention or management of postmenopausal osteoporosis. Full article

Rock bolt reinforcement governs underground excavation stability through the combined effects of embedment depth, installation pattern, and interaction with the evolving stress and structural environment. Despite substantial advances across mechanistic, empirical, numerical, discontinuum, dynamic, and field-based research traditions, no unified framework currently integrates these approaches across geological and stress regimes. Current practice, therefore, relies on design methods calibrated within specific contexts, producing optimisation outcomes that are model-dependent, metric-sensitive, and not reliably transferable across site conditions. This review critically synthesises evidence from 30 peer-reviewed studies organised into six analytical categories: mechanistic confinement frameworks, empirical classification systems, numerical parametric investigations, discontinuum- and discrete fracture network (DFN)-based optimisation studies, high-stress and dynamic performance analyses, and field-based performance evaluations. The synthesis establishes three principal findings. First, optimal bolt embedment is stress-regime-dependent; plastic-radius-based design logic is appropriate under moderate static conditions but becomes insufficient under high stress or dynamic loading, where energy absorption capacity and controlled yielding govern performance. Second, in discontinuous rock masses, joint geometry and spacing dominate reinforcement effectiveness, shifting optimisation from uniform length selection toward pattern-specific alignment and multi-length configurations that outperform equal-length grids under DFN-controlled conditions. Third, numerical optimisation outcomes are sensitive to the choice of objective metric and modelling paradigm, such that bolt length and spacing recommendations cannot be transferred across analytical frameworks without explicit mechanism comparison. To integrate these findings, a unified conceptual framework is proposed based on regime classification using three dimensionless indicators: the bolt penetration ratio (Π₁ = L/r_p), which relates embedment to plastic zone radius; the structural interception ratio (Π₂ = S/S_j), which relates bolt spacing to dominant joint spacing; and the stress intensity ratio (Π₃ = σ_in situ/σ_cm), which relates in situ stress to rock mass compressive strength. These indicators identify whether confinement-dominated, structure-dominated, or stress-dominated behaviour governs stability, and direct design logic accordingly. The framework does not prescribe universal geometric thresholds; rather, it provides a structured classification pathway that integrates mechanistic and empirical evidence into a coherent and transferable design logic. Probabilistic validation incorporating geological variability, stochastic fracture network modelling, and iterative field calibration is identified as the necessary development path toward a statistically robust optimisation methodology. Full article