Search Results (767)

Crack detection in mobile inspection scenarios is constrained by both the extremely slender geometry of crack targets and the real-time inference requirements on edge devices, which expose systematic limitations of general-purpose object detectors. This paper proposes YOLO-Crack, a closed-loop solution that couples geometry-statistics-driven module design with end-to-end edge deployment validation. On the algorithmic side, we first quantify crack geometric properties and then introduce (i) a crack-aware cross-dimensional fusion attention (CFCA) module to strengthen feature representations, (ii) a dual-path feature enhancement module (DFEM) to preserve fine details during upsampling, and (iii) an empirical smooth quality window adjustment with shape consistency regularization to stabilize bounding-box regression for slender cracks. Experiments on the Crack500 dataset show that YOLO-Crack achieves 78.8% precision, 51.4% recall, and 65.7% mAP@0.5, improving over the YOLOv11n baseline by 4.2, 1.7, and 2.9 percentage points, respectively. On the engineering side, we deploy YOLO-Crack on a Jetson Orin NX mobile robot platform and evaluate it in a real ROS pipeline; the measured end-to-end throughput reaches 25.5 FPS, meeting real-time video processing requirements. The proposed framework provides a practical reference workflow for edge vision tasks, from geometry analysis to engineering verification. Full article

To address challenges such as complex background interference, limited long-range modeling capabilities of CNNs, and poor generalization in steel-concrete cross-material scenarios, this study proposes an enhanced detection framework. This framework integrates a TAPFormer with morphological reasoning rules. The method utilizes TAPFormer as the backbone network. It captures global topological features of cracks through a Task-Aware Query mechanism. This approach compensates for the deficiencies of traditional convolutional operators in modeling the continuity of thin and long cracks. Furthermore, a mask reasoning module based on geometric priors is developed to handle unstructured interferences, such as marker pen marks, welds, and concrete holes. This module defines logical criteria, including edge curvature consistency, axial aspect ratios, and endpoint extension directions. These criteria are used to perform topological repair and filter false positives in the initial segmentation masks. A hybrid dataset containing 4500 cross-material damage images was used for validation. The results show that the proposed method achieves a mean IoU of 86.72% and an F1-score of 90.36%. Notably, the method filters over 91.0% of false positives caused by manual marker pen marks in interference-rich scenarios. Compared to mainstream state-of-the-art models, the IoU improves by at least 5.48%. The results show that the proposed framework improves the robustness and logical self-consistency of crack identification in complex engineering environments. Full article

An ordinary state-based peridynamic (OSPD) approach combined with an interaction integral method is proposed to calculate dynamic stress intensity factors (DSIFs) and simulate crack propagation in two-dimensional cracked brittle solids. Numerical investigations are carried out for mode I and mixed-mode cracked plates under static, quasi-static, and dynamic loading conditions. A local damping scheme is incorporated into the peridynamic equations of motion to achieve convergence in static and quasi-static analyses. The influence of circular holes on DSIFs and crack propagation paths is systematically examined. Quantitative analyses of elastic deformation and quasi-static fracture behavior for mode I and mixed-mode cracks are verified through the uniaxial tension of a slab. The peak values of DSIFs exceed their static counterparts under dynamic loading. Complex dynamic fracture phenomena, including crack branching in both straight and inclined edge cracks, are successfully captured. The results obtained by the OSPD approach are validated through comparisons with theoretical benchmarks and finite element results, demonstrating high accuracy and effectiveness in calculating elastic deformation and stress intensity factors (SIFs), as well as accurately predicting crack propagation paths in quasi-static and dynamic fracture problems in brittle solids. Beyond the benchmark problems, the proposed OSPD approach is particularly well-suited for investigating more complex fracture systems. Full article

Aluminum is undoubtedly a key material in modern industry. Flat plates made of aluminum alloys are widely used in construction, aeronautics, automotive, and others. The current paper presents an analysis of the behavior of a thin plate made of Al7075-T651 aluminum alloy, subjected to a uniaxial stress, and clamped at one end. The results of the numerical simulation with FRANC2D software have been used for accurate determination of the stress intensity factors (K_I, K_II) and being validated for the simple cases using analytical calculations. The Failure Assessment Diagram (FAD) based on the toughness ratio Kr and the load ratio Lr has been used to evaluate the structural integrity of cracked components based on the load, its position, crack size, and the fracture properties of the material. The FAD analysis results highlight the significant influence of crack position on the values of the K factor. The edge and inclined cracks lead to increases in stress intensity factors and to the occurrence of mixed-mode loading conditions. The study demonstrates the effectiveness and usefulness of the proposed methodology in the analysis of structures with discontinuities and emphasizes the importance of crack positioning in assessing the safety of engineering components. Full article

Surface defect detection is an important task for quality assurance in steel manufacturing. Although YOLO-style detectors are widely used due to their strong performance, they often struggle to accurately localize edge-dominant defects such as crazing and fine cracks. This limitation arises because such defects exhibit weak feature representations. In addition, their high-frequency structural details are progressively degraded during repeated downsampling. To address this issue, a YOLO-based detection framework named EDEN-YOLO is proposed. It incorporates an in-place Edge-Enhance module into the YOLOv8 baseline to improve structural sensitivity. Specifically, a Local Feature Enhancement (LFE) module is designed to capture edge-sensitive patterns. A Gated Module is further introduced to perform spatially selective recalibration of backbone features. This design enhances edge responses while suppressing noise. Experiments on the NEU-DET benchmark demonstrate the effectiveness of the proposed method. EDEN-YOLO achieves 80.5% mAP@0.5 on NEU-DET, showing an improvement over the reproduced YOLOv8 baseline while introducing a moderate increase in model complexity by 0.52M parameters and 1.3 GFLOPs. A supplementary evaluation on the GC10-DET dataset shows that EDEN-YOLO achieves 65.2% mAP@0.5, compared with 61.0% for the reproduced YOLOv8 baseline. The qualitative results show that the proposed module produces more compact feature responses. Full article

Accurate surface defect classification is a critical requirement in structural health monitoring and infrastructure inspection, where defects, including cracks, spalling, delamination and noncrack regions, often appear with low-contrast and complex background textures. Motivated by the need for a robust and discriminative framework that can enhance defect visibility and focus learning on damage-critical regions, this research proposes a novel damage-aware DenseNet-201 (DA-DenseNet-201) model for surface defect classification. As a critical novelty, a damage-aware adaptive contrast-limited adaptive histogram equalisation (DAC) filtering strategy is introduced as a preprocessing stage. The proposed DAC filter dynamically adjusts contrast enhancement parameters based on damage indicators, selectively amplifying crack edges and defect textures while preserving healthy surface regions and suppressing noise. Building on this method, enhanced images are processed using a pretrained DenseNet-201 backbone, retaining the benefits of dense feature propagation and efficient gradient flow. To strengthen the discriminative learning of DA-DenseNet-201 further, an attention refinement block is integrated into the network, combining channel attention to emphasise defect-relevant feature responses and spatial attention to localise damage regions accurately. In addition, a multiscale feature fusion mechanism aggregates feature maps from multiple dense blocks to capture fine-grained crack patterns, texture-level degradation and high-level semantic damage information. Extensive experiments conducted on surface defect datasets demonstrate its effectiveness, achieving a superior classification accuracy of 98.93%, along with notable improvements in sensitivity, specificity and the intersection over union compared with state-of-the-art models. These results confirm that the proposed DA-DenseNet-201 provides a reliable and high-performance solution for automated surface defect classification. Full article

Automated visual inspection of photovoltaic (PV) cells is an important component of solar-module quality assurance. However, adapting modern pre-trained vision backbones to PV defect classification remains challenging because full fine-tuning requires substantial memory, naturally imbalanced datasets can reduce sensitivity to rare defect classes, and edge-oriented inspection workflows impose computational constraints. Parameter-efficient fine-tuning (PEFT) methods, such as Low-Rank Adaptation (LoRA) and Quantized LoRA (QLoRA), have been widely studied in natural language processing, but their use for PV defect classification remains underexplored. This study presents a controlled benchmark of LoRA and QLoRA against full fine-tuning for PV cell defect classification. Four adaptation strategies—full fine-tuning, LoRA with rank 8, LoRA with rank 16, and 4-bit QLoRA with rank 16—are evaluated using a ConvNeXt-Tiny backbone on a 17,377-image polycrystalline PV cell electroluminescence dataset referred to as POLY, covering five classes: intact, cracked, broken, surface-diffuse, and surface-point. The natural 6.7× class imbalance is preserved without synthetic resampling, and a group-aware StratifiedGroupKFold protocol based on available cell or panel-image identifiers is used to reduce identifiable leakage across folds. All PEFT variants slightly outperform full fine-tuning in macro-F1 while training 26–52× fewer parameters. QLoRA_r16 achieves the highest macro-F1 score of 79.92 ± 0.75%, compared with 78.26 ± 0.94% for full fine-tuning, while training the same number of parameters as LoRA_r16 (1.060 M; 3.67% of the adapted model). QLoRA_r16 also improves F1 on the intact (+4.75 points) and surface-diffuse (+2.62 points) classes relative to full fine-tuning. This class-wise pattern suggests that quantized low-rank adaptation may influence minority and visually ambiguous categories; however, the present experiments do not isolate the independent effect of NF4 quantization from adapter rank, batch size, or optimization dynamics. Under the training configuration used, QLoRA_r16 records the lowest observed peak training GPU memory, approximately 30% below full fine-tuning (1727 MB versus 2478 MB). Because QLoRA_r16 was trained with batch size 16 whereas the other methods used batch size 32, this reduction should be interpreted as an end-to-end configuration effect rather than as the isolated effect of 4-bit quantization. Overall, the results indicate that PEFT is a promising and resource-efficient alternative to full fine-tuning for PV defect classification, although batch-matched memory experiments, direct embedded-device profiling, and cross-dataset validation remain necessary before making deployment-level claims. Full article

The use of the stress intensity factor K to characterize the severity of crack tip stress fields is widespread throughout engineering. The relationship between K and the crack growth rate is then usually represented empirically by a straight line Paris law relationship on logarithmic axes. This study develops an analytical relationship between the two by linking crack growth to the accumulation of fatigue damage ahead of the moving crack tip. A stress-based fatigue model was used, with inputs from plastic 2D plane stress FE analyses representing an edge crack by a sharp semi-circular notch. Stress–distance profiles ahead of the crack tip were extracted at the maximum and minimum points of a range of fatigue loading cycles. These were then used with data from smooth specimen LCF tests to predict the build-up of fatigue damage at regularly spaced locations ahead of the crack tip and hence crack growth rates. Full da/dN–ΔK curves were generated for the nickel-based superalloy RR1000 at 20 °C with loading R-ratios of 0, −1 and 0.5. The R = 0 and R = −1 crack growth rate predictions agreed well with experimental data, as did the steeper growth rate slope calculated at R = 0.5. The method was then extended to predict overload behaviour. Full article

Slabbing failure of roadway-side backfill bodies critically threatens gob-side entry retaining stability. This study establishes an elastic thin-plate model with edge cracks, employing an innovative load transformation to reduce the three-dimensional in situ stress state to the combined action of roof–floor uniform load and equivalent axial bending moment. Based on fracture mechanics and elastic-plastic theory, the stress intensity factor $K_1$ and crack initiation load q are derived in closed form. Results show that q is positively correlated with plate thickness t and bending moment M and negatively with crack length a in the dominant range. Applying the nonlinear Hoek–Brown criterion, the failure zone width ${\mathrm{r}}_{\mathrm{p}}$ at the crack tip is shown to exhibit an approximately exponential relationship with $K_1$ for unbolted backfill. Introduction of tensioned bolts via a stress concentration factor $\mathsf{\eta }$ transforms the failure zone growth from exponential to asymptotic saturation, quantitatively confirming the crack-arresting effect. A sensitivity analysis identifies plate thickness as the dominant parameter. The model bridges the gap between initial slabbing and progressive V-shaped notch formation. Full article

Continuous manufacturing has emerged as the prevailing paradigm in the modern pharmaceutical industry, imposing stringent demands for efficient, real-time inspection methods. Furthermore, deploying high-performance deep learning models on industrial edge devices remains challenging due to computational constraints and the difficulty of detecting micro-defects (e.g., micro-cracks and spots). This paper proposes EMN-net, a lightweight defect detection model built upon the YOLOv8n architecture. The proposed algorithm integrates a MobileNetV3 backbone, the Efficient Local Attention (ELA) mechanism and the Normalized Wasserstein Distance (NWD) loss function to balance computational efficiency with sensitivity toward micro-defects. Evaluated on a self-built industrial tablet dataset expanded to 3086 images, EMN-net achieves an mAP50 of 97.8%, representing a 2.5% improvement over the baseline YOLOv8n. the computational complexity is reduced to 4.4 GFLOPs, while the inference throughput reaches 118 FPS, satisfying the real-time requirements of high-speed production lines. Additionally, the model exhibits improved robustness under simulated motion blur and sensor noise. EMN-net presents a balanced automated visual inspection solution for edge devices in continuous pharmaceutical manufacturing. Full article

To address the durability limitations of conventional crack sealants under coupled extreme temperatures and traffic loads in long-life pavements, a bio-oil composite-modified patching material was developed using 90# base asphalt as the matrix, synergistically modified with crumb rubber (CR) and epoxidized soybean oil (ESO). To resolve the contradictory requirements for high elasticity and thermal expansion/contraction coordination in sealants, ESO was introduced; its polar epoxy groups optimize phase compatibility and promote low-temperature stress relaxation without restricting thermal deformability. Rheological evaluations revealed that the optimal system (OPT) successfully extended the service temperature window from PG 76–−24 °C (baseline) to PG 82–−24 °C, significantly enhancing its adaptability to extreme climatic fluctuations. At −24 °C, OPT exhibited a reduced creep stiffness (S) of 164 MPa and an increased creep rate (m) of 0.312, with a cracking resistance ratio (k) as low as 525.6; the quantitative significance of these metrics lies in granting the sealant superior stress relaxation capacity, enabling it to accommodate dynamic crack widening without interfacial debonding or brittle fracture. Fatigue testing via time sweeps demonstrated that N_f50 reached 2890 cycles, highlighting robust long-term resistance against high-frequency shear strains induced by tire edges. Micro-mechanistic analyses (FTIR, TG/DTG, and DSC) confirmed that the modification is primarily driven by physical blending. The elevation of the thermal decomposition threshold (T5%) to 302.4 °C and the residue at 600 °C to 44.8% provide a critical safety margin for high-temperature construction heating, preventing thermal degradation. Furthermore, the glass transition temperature (T_g) decreased to approximately −35.2 °C. These findings establish a rigorous quantitative and mechanistic framework for designing sustainable, high-performance patching materials for resilient pavement maintenance. Full article

The Vertical Shaft Machine (VSM) method is increasingly used in ultra-deep prefabricated shafts. However, as its application extends into hard ground, existing segment designs still largely follow soft soil experiences, resulting in insufficient material utilization and poor economic efficiency. Based on the first VSM shaft in South China, this study establishes a refined finite element model validated by field monitoring and subsequently constructs a structural response database. A GA-XGBoost surrogate model combined with the SHAP method quantifies the contributions of key parameters—concrete strength, rebar diameter, and steel plate thickness—to shaft structural stress. Following the optimization objective of reducing material consumption while maintaining the overall structural performance of the original design, an optimization scheme for Ring 0 reinforcement is proposed. Results show that SHAP analysis effectively identifies the contribution ranking of each parameter to the structural response: for Ring 0, concrete strength contributes the most while rebar diameter shows low sensitivity; for the cutting edge ring, steel plate thickness and concrete strength contribute significantly, whereas tie bars show the lowest sensitivity. After optimization of Ring 0, reinforcement consumption per linear meter of segment is reduced by 43.43 kg, and steel content decreases by 57.91 kg/m³. Verification confirms that the stress distribution remains largely unchanged and crack width meets specification limits. Tie bars in the cutting edge ring play an irreplaceable structural role during concrete pouring and should not be directly optimized. The proposed scheme reduces material consumption while ensuring structural safety, offering a reference for optimizing VSM shaft segment structures in hard ground conditions. Full article

To improve temperature uniformity and reduce thermal stress-induced cracking during laser directed energy deposition (laser DED) repair of TiAl blades, this study proposes a refined induction heating coil design based on coupled electromagnetic-thermal finite element simulation. A temperature-dependent model of the induction heating process for a cast 45XD TiAl blade was established and used to compare circular and elliptical coil cross-sectional shapes. The elliptical coil reduced the magnetic field concentration at the leading and trailing edges and decreased the maximum temperature difference across the blade cross-section to below 100 K, thereby improving transverse temperature uniformity. To further improve the temperature distribution along the blade length, a variable-pitch solenoid coil with sparse turns in the middle and dense turns near both ends was designed. This arrangement improved the balance between local heat generation and heat dissipation and reduced the temperature variation within the central 10 cm region of the blade to about 10 K. Experimental validation showed engineering-level agreement with the simulation results, and the blade body was stably maintained at 1020–1030 K, satisfying the preheating requirement for laser DED repair of TiAl blades within the tested design set. Full article

An explicit dynamic framework based on the Nonlocal Macro-Meso-scale Consistent Damage (NMMD) model is proposed to simulate complex multiple-crack evolution in quasi-brittle materials subjected to blast loading. Three numerical examples—a single-edge-notched half-plate, a thick ring, and a hollow mortar cylinder containing a small borehole—are analyzed. The results show that crack initiation, propagation, branching, and coalescence can be naturally captured by the proposed framework without remeshing. Reliable predictions are obtained only when sufficient mesh resolution is used to resolve nonlocal interactions and the time step satisfies the explicit stability criterion. Comparisons indicate that fewer but more dominant crack paths are predicted by the model, suggesting a conservative tendency in estimating the number of fragments. Crack-path selection is significantly influenced by material heterogeneity, which enables secondary cracks to evolve into dominant crack paths. Crack multiplication and network connectivity are promoted by increased blast pressure, whereas crack complexity and spatial extent are reduced by higher damping coefficients. Full article

To investigate the effect of scribing wheel angle on the scribing behavior of LCD glass, an SPH-based numerical model was established in LS-DYNA and validated against experimental results for reaction force and median crack depth. The results show that the model can accurately capture the mechanical response and crack propagation during the scribing process. At a scribing depth of 10 μm, the maximum relative errors between simulation and experiment were 5.17% for reaction force and 2.36% for median crack depth. The results for the 110° scribing wheel indicate that median cracks mainly initiate and propagate rapidly during the penetration stage, while the median crack depth becomes nearly stable after the preset depth is reached, and the subsequent rolling stage has little influence on further crack growth. As the wheel angle increases from 90° to 140°, the experimental mean peak reaction force increases from 2.66 N to 9.97 N, the maximum effective stress increases from 374.4 MPa to 732.8 MPa, and the median crack depth increases from 68 μm to 97 μm. Experimental observations further show that small wheel angles tend to cause debris accumulation and edge chipping, whereas excessively large wheel angles are likely to induce lateral cracks. Overall, a wheel angle of about 110° provides better cross-sectional quality, surface quality, and crack controllability for 0.2 mm-thick LCD glass. Full article