Search Results (134)

In the fabrication process of pipelines for petrochemical and other industries, weld reinforcement is often excessive and adversely affects subsequent processes such as anticorrosion treatment and surface coating. Weld reinforcement must be removed through a grinding process. Welding deformation and fit-up errors often lead to highly irregular weld geometries, which makes robotic grinding difficult and causes the task to still heavily rely on manual operation. To address this issue, this study proposes an automatic weld recognition and grinding trajectory planning method based on 3D visualization and deep learning. A weld recognition network, termed WSR-Net, has been developed based on an improved PointNet++ architecture with a cross-attention mechanism, achieving a segmentation accuracy of 98.87% and a mean intersection over union of 90.71% on the test set. An intrinsic shape signature (ISS) key point selection algorithm with orthogonal slicing-based pruning optimization is developed to robustly extract key weld ridge points that characterize the weld trend on rugged weld surfaces. According to the height differences between the weld and the adjacent base metal surfaces, the grinding reference surface is fitted using the weld contour through the moving least-squares method. The ridge line points are projected onto the grinding reference surface along the local normal to generate the expected grinding trajectory points. The grinding trajectory that meets the process constraints is generated through reverse layer slicing. Grinding experiments demonstrate that the proposed WSR-Net achieves robust segmentation performance for both planar and curved surface welds. With the reverse layered trajectory planning method, the proposed method enables high-precision automatic grinding of complex spatially curved surface welds. The results show that the final grinding mean error is 0.316 mm, which satisfies the preprocessing requirements for subsequent processes. The proposed method provides a feasible technical method for the intelligent grinding of spatially curved surface welds. Full article

Sustainable manufacturing performance is increasingly influenced by the quality of decisions embedded in Quality Management System (QMS) activities, particularly those related to problem analysis and preventive action. In industrial environments such as welded bus chassis production, recurring quality defects—although involving small components—can generate sustainability impacts through rework, inspection effort, and energy consumption. Although artificial intelligence (AI) is increasingly adopted to support quality-related tasks, its contribution is often assessed in terms of automation rather than its effect on decision quality. This study presents an AI-supported, prompt-driven decision framework designed to strengthen preventive performance within QMS. The framework is implemented through a deterministic software application that formalizes prompt engineering as a rule-based process, transforming informal human problem descriptions into structured prompts suitable for external AI reasoning tools. The application itself does not embed AI and does not generate decisions; instead, it functions as a transparent decision interface that reduces variability in problem formulation and supports methodological consistency. The framework was validated through an industrial case study conducted in a bus chassis manufacturing plant experiencing recurring defects related to missing or incorrectly positioned welded brackets. Quantitative evaluation using Key Performance Indicators demonstrates reduced analysis cycle time, improved completeness of problem definitions, higher corrective action implementation rates, and lower defect recurrence. Full article

Efficient operation of robotic manipulators in repetitive industrial tasks, such as welding and logistics sorting, requires careful coordination of obstacle representation and motion planning. Traditional methods, such as axis-aligned bounding boxes, generate overly conservative trajectories, while highly detailed models impose excessive computational burden, both increasing cumulative energy consumption in long-duration operations. This paper presents an adaptive sphere-based obstacle modeling framework integrated with energy-aware motion planning for repetitive manipulation tasks. The proposed method employs an improved Whale Optimization Algorithm with nonlinear parameter adjustment and elite guidance mechanisms to generate compact sphere representations through adaptive voxelization. Experimental validation using a 6-DOF UR5 manipulator demonstrates substantial performance improvements over conventional AABB models, achieving 31–66% energy reduction and 12.5–37% shorter configuration-space paths, with competitive modeling efficiency (2.63–3.34 s) compared to 11 metaheuristic algorithms. The framework provides a systematic methodology for integrating obstacle modeling with motion planning, particularly suitable for applications where cumulative energy savings are critical in repetitive operations. Full article

Industrial robots are widely used in critical tasks such as assembly, welding, and material handling as core components of modern manufacturing systems. For the reliable operation of these systems, early and accurate detection of execution failures is crucial. In this study, a comprehensive comparison of machine learning and deep learning methods is conducted for the classification of robot execution failures using data acquired from force–torque sensors. Three different feature engineering approaches are proposed. The first is a Baseline approach that includes 90 raw time-series features. The second is the Domain-6 approach, which consists of 6 basic statistical features per sensor (36 in total). The third is the Domain-12 approach, which comprises 12 comprehensive statistical features per sensor (72 in total). The domain features include the mean, standard deviation, minimum, maximum, range, slope, median, skewness, kurtosis, RMS, energy, and IQR. In total, ten classification algorithms are evaluated, including eight machine learning methods and two deep learning models: Support Vector Machines (SVM), Random Forest (RF), k-Nearest Neighbors (KNN), Artificial Neural Network (ANN), Naive Bayes (NB), Decision Trees (DT), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM-LGBM), as well as a One-Dimensional Convolutional Neural Network (CNN-1D) and Long Short-Term Memory (LSTM). For traditional machine learning algorithms, 5 × 5 nested cross-validation is used, whereas for deep learning models, 5-fold cross-validation with a 20% validation split is employed. To ensure statistical reliability, all experiments are repeated over 30 independent runs. The experimental results demonstrate that feature engineering has a decisive impact on classification performance. In addition, regardless of the feature set, the highest accuracy (93.85% ± 0.90) is achieved by the Naive Bayes classifier using the Baseline features. The Domain-12 feature set provides consistent improvements across many algorithms, with substantial performance gains. The results are reported using accuracy, precision, recall, and F1-score metrics and are supported by confusion matrices. Finally, permutation feature importance analysis indicates that the skewness features of the Fx and Fy sensors are the most critical variables for failure detection. Overall, these findings show that time-domain statistical features offer an effective approach for robot failure classification. Full article

In the context of Industry 4.0, reliable automatic inspection of weld surface defects is critical for structural safety, yet current deep learning-based detectors struggle with the extreme scale variation and anisotropic shapes characteristic of weld flaws such as pores, cracks, and lack of fusion. Existing YOLO-family models, although effective on general-purpose datasets, often fail to robustly localize tiny defects and long, slender discontinuities while remaining lightweight enough for industrial edge deployment. A critical research gap lies in the lack of task-specific optimization for weld defects: standard attention mechanisms are isotropic and cannot capture linear defect continuity, while existing loss functions ignore scale disparity between tiny pores (area < 100 pixels²) and large incomplete fusion defects (area > 5000 pixels²), leading to unstable regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. Here, we propose a dual-optimized lightweight YOLOv11 framework tailored for weld defect detection that addresses both feature representation and bounding-box regression. First, we introduce WeldSimAM, an enhanced attention module that augments parameter-free SimAM with directional (horizontal/vertical) and channel-wise enhancement to better capture the directional texture of linear weld defects. Second, we develop an Enhanced Normalized Wasserstein Distance (EnNWD) loss, which incorporates scale-disparity penalties and relative-area-based weighting to mitigate sample imbalance and improve regression accuracy for tiny and large-aspect-ratio targets. Validated via 10-fold cross-validation on three datasets (self-built + two public), the method achieves 99.48% mAP@0.5 and 73.29% mAP@0.5:0.95, outperforming YOLOv11 by 0.13 and 3.76 percentage points (p < 0.01, two-tailed t-test), with 5.21 MB and 132 FPS on NVIDIA RTX 4090. It also surpasses non-YOLO SOTA methods (e.g., EfficientDet-Lite3) by 3.8–5.5 percentage points in mAP@0.5 (p < 0.05), offering a practical real-time solution for industrial inspection. Full article

Conventional subtractive manufacturing methods, such as cutting, often result in material waste and limitations in geometric complexity. To address these challenges, Wire Arc Additive Manufacturing (WAAM), in which components are built through successive weld bead deposition, has attracted increasing attention across various industrial fields. However, WAAM-fabricated components typically exhibit significant surface irregularities, necessitating additional post-processing that reduces overall productivity. Improving productivity therefore requires effective control and optimization of deposition parameters. This task is particularly challenging in multilayer WAAM processes, as the geometry of previously deposited layers varies with operating conditions. To address this challenge, this study proposes an AI-based framework for controlling surface roughness by rapidly identifying near-optimal process parameters in response to evolving bead geometry. A large-scale simulation dataset was generated by applying a pre-trained deep neural network (DNN) surface roughness predictor to one million bead geometry variations under 72 process parameter combinations. The resulting optimal parameter labels were used to train a classification model that recommends process conditions based on the current bead geometry. Model performance was evaluated using predictor-estimated surface roughness values, achieving Weighted Precision, Recall, and F1-score of 0.98, with an average AUC of 0.977. Five previously generated WAAM specimens were used for comparative analysis between AI-recommended and conventional process conditions using the previously developed and validated surface roughness prediction model, rather than direct physical measurements. This predictor-based feasibility analysis showed that AI-recommended conditions consistently reduced the predicted surface roughness, indicating the potential of AI-driven process optimization to improve surface quality in WAAM and reduce reliance on post-processing. Full article

Laser welding is widely used in complex manufacturing processes and valued for its reliability, flexibility, and high energy density. However, achieving the desired weld quality requires the detection and, ideally, the prevention of defects. Besides other methods, machine learning (ML) has been integrated into laser welding with the primary goal of process optimization and quality improvement, for example, by enabling process adaptation before or during welding to reduce defects. This survey systematically reviews publications from 2015 to 2025 that integrate machine learning and deep learning methods into laser welding optimization or adaptation processes. An extensive analysis identifies which parts of the process and for what purposes ML methods are researched and implemented and how they are evaluated, as well as the sensors, lasers, and materials involved. Furthermore, the findings are analyzed and organized into taxonomies that define overarching meta-categories into which existing approaches can be classified and contextualized. The results reveal that various ML approaches are applied for tasks, such as surrogate modeling, process planning, direct control, and virtual sensing and monitoring. Although many different control parameters and optimization targets are considered, laser power and welding speed dominate as the most frequently adjusted parameters, while penetration depth and weld geometry-related properties are the most common optimization targets. Finally, the survey identifies major challenges, including the lack of benchmarking datasets, standardized evaluation protocols, and interpretable models. Full article

Fine segmentation of point clouds in hull structure junction regions is a key technology for achieving high-precision digital inspection. Conventional hard-segmentation methods frequently yield over- or under-segmentation in junction regions such as welds, compromising the reliability of subsequent inspections. This study presents a computational framework that combines the Gaussian-Uniform Mixture Model (GUMM) with the Hidden Markov Random Field (HMRF) and follows a “coarse segmentation–model construction–fine segmentation” pipeline. The framework jointly optimizes the sampling model, the probabilistic model, and the expectation–maximization (EM) inference procedure. By leveraging model simplification and dimensionality reduction, the algorithm simultaneously addresses initial value estimation, spatial distribution characterization, and continuity constraints. Experiments on representative structures, including wall corner, T-joint weld, groove, and flange, show that the proposed framework outperforms the conventional GMM-EM method by approximately 2.5% in precision and 1.5% in both accuracy and F1 score. In local segmentation tasks of complex hull structures, the method achieves a deviation of less than 0.2 mm relative to manual measurements, validating its practical utility in engineering contexts. Full article

We introduce AHA–SCA, a compact hybrid optimiser that alternates the wave-based exploration of the Sine–Cosine Algorithm (SCA) with the exploitation skills of the Artificial Hummingbird Algorithm (AHA) within a single population. Even iterations perform SCA moves with a linearly decaying sinusoidal amplitude to explore widely around the current best solution, while odd iterations invoke guided and territorial hummingbird flights using axial, diagonal, and omnidirectional patterns to intensify the search in promising regions. This simple interleaving yields an explicit and tunable balance between exploration and exploitation and incurs negligible overhead beyond evaluating candidate solutions. The proposed approach is evaluated on the CEC2014, CEC2017, and CEC2022 benchmark suites and on several constrained engineering design problems, including welded beam, pressure vessel, tension/compression spring, speed reducer, and cantilever beam designs. Across these diverse tasks, AHA–SCA demonstrates competitive or superior performance relative to stand-alone SCA, AHA, and a broad panel of recent metaheuristics, delivering faster early-phase convergence and robust final solutions. Statistical analyses using non-parametric tests confirm that improvements are significant on many functions, and the method respects problem constraints without parameter tuning. The results suggest that alternating wave-driven exploration with hummingbird-inspired refinement is a promising general strategy for continuous engineering optimisation. Full article

This paper proposes JADEGBO, a hybrid gradient-based metaheuristic for solving complex single- and multi-constraint engineering design problems as well as cost-sensitive security optimisation tasks. The method combines Adaptive Differential Evolution with Optional External Archive (JADE), which provides self-adaptive exploration through p-best mutation, an external archive, and success-based parameter learning, with the Gradient-Based Optimiser (GBO), which contributes Newton-inspired gradient search rules and a local escaping operator. In the proposed scheme, JADE is first employed to discover promising regions of the search space, after which GBO performs an intensified local refinement of the best individuals inherited from JADE. The performance of JADEGBO is assessed on the CEC2017 single-objective benchmark suite and compared against a broad set of classical and recent metaheuristics. Statistical indicators, convergence curves, box plots, histograms, sensitivity analyses, and scatter plots show that the hybrid typically attains the best or near-best mean fitness, exhibits low run-to-run variance, and maintains a favourable balance between exploration and exploitation across rotated, shifted, and composite landscapes. To demonstrate practical relevance, JADEGBO is further applied to the following four well-known constrained engineering design problems: welded beam, pressure vessel, speed reducer, and three-bar truss design. The algorithm consistently produces feasible high-quality designs and closely matches or improves upon the best reported results while keeping computation time competitive. Full article

Many engineering design problems are formulated as constrained optimization tasks that are nonlinear and nonconvex, and often treated as black boxes. In such cases, metaheuristic algorithms are attractive because they can search complex design spaces without requiring gradient information. In this work, we propose an Enhanced Henry Gas Solubility Optimization (eHGSO) algorithm, which is an improved version of the physics-inspired HGSO method. The enhanced variant introduces six main contributions: (i) a more diverse, population-wide initialization strategy to cover the design space more thoroughly; (ii) adaptive temperature/pressure control parameters that automatically shift the search from global exploration to local refinement; (iii) an elitist archive with differential perturbation that accelerates exploitation around high-quality candidate designs; (iv) a simple combination of the global HGSO search moves with a lightweight gradient-free local search to refine promising solutions; (v) a constraint-handling mechanism that explicitly prioritizes feasible solutions while still allowing exploration near the constraint boundaries; and (vi) a complexity and ablation analysis that quantifies the impact of each mechanism and confirms that they introduce only modest computational overhead. We evaluate eHGSO on four classical constrained engineering design problems: the stepped cantilever beam, the tension/compression spring, the welded beam, and the three-bar truss. Its performance is compared with seventeen recent metaheuristic optimizers over multiple independent runs. eHGSO achieves the best average objective value on the cantilever, spring, and welded-beam problems and shares the best average result on the three-bar truss. Compared to the second-best method, the mean objective is improved by about $0.84\%$ for the cantilever beam and $0.35\%$ for the welded beam, while the spring and truss results are essentially equivalent at four significant figures. Convergence and robustness analyses show that eHGSO reaches high-quality solutions quickly and consistently. Overall, the proposed eHGSO algorithm appears to be a competitive and practical tool for constrained engineering design problems. Full article

Automated ultrasonic testing (AUT) serves as a vital method for evaluating critical infrastructure in industries such as oil and gas. However, a significant challenge in deploying artificial intelligence (AI)-based interpretation methods for AUT data lies in improving their reliability and effectiveness, particularly due to the inherent scarcity of real-world defective data. This study directly addresses data scarcity in a weld defect classification task, specifically concerning the detection of lack of fusion (LOF) defects in weld inspections using a proprietary industrial ultrasonic B-scan image dataset. This paper leverages state-of-the-art generative models, including Generative Adversarial Networks (GANs) and Denoising Diffusion Probabilistic Models (DDPM) (StyleGAN3, VQGAN with an unconditional transformer, and Stable Diffusion), to produce realistic B-scan images depicting LOF defects. The fine-tuned Transformer-based models, including ViT-Base, Swin-Tiny, and MobileViT-Small classifiers, on the regular B-scan image dataset are then applied to retain only high-confidence positive synthetic samples from each method. The impact of these synthetic images on the classification performance of a ResNet-50 model is evaluated, where it is fine-tuned with cumulative additions of synthetic images, ranging from 10 to 200 images. Its accuracy on the test set increases by 38.9% relative to the baseline with the addition of either 80 synthetic images using VQGAN with an unconditional transformer or 200 synthetic images by StyleGAN3 to the training set, and by 36.8% with the addition of 150 synthetic images by Stable Diffusion. This also outperforms Transformer-based vision models that are trained on regular training data. Concurrently, knowledge distillation experiments involve training ResNet-50 as a student model, leveraging the expertise of ViT-Base and Swin-Tiny as teacher models to demonstrate the effectiveness of adding the synthetic data to the training set, where the greatest enhancement is observed to be 34.7% relative to the baseline. This work contributes to advancing robust, AI-assisted tools for critical infrastructure inspection and offers practical pathways for enhancing available models in resource-constrained industrial environments. Full article

This study applies to a virtual commissioning (VC) workflow with discrete-event simulation in WITNESS Horizon to diagnose and improve an automated brushless stator assembly line. A validated model of the full route—Stator Assembly Machine (SAM), Linear Transport System (LTS), Winding Machine (WM), Terminal Welding Machine (TWM), Inspection Machine (IM) and Electric Tester (ET)—was executed over a one-shift horizon (28,800 s). We compared the baseline configuration with an optimized scenario that retrieved robot tasks and refined LTS routing. Key performance indicators (KPIs) were resource utilization (Busy/Idle/Blocked) and completed operations. The results are quantitative and specific. Blocking at the SAM interface collapsed from 73.32% to 0% at PressPosition and from 80.64% to 0% at Robot2. LTS transitioned from 97.46% Blocked to 0%, with the share of Move/Running increasing to 14.76% (from ~0%). Line output—measured as completed assemblies at SAM—increased from 368 to 425 units per shift (+15.5%). Similar gains were recorded at other stations (e.g., WM1: 351 → 424 operations, +20.8%). These changes reflect the removal of the primary transfer bottleneck and a more balanced utilization across stations. The study demonstrates that VC can deliver actionable commissioning guidance. By quantifying where blocking occurs and testing alternative control strategies in a virtual environment, it is possible to raise throughput while maintaining stable operation. The modeling approach and metrics are reusable for related electromechanical assembly lines. Full article

Welding is one of the most common machining methods in the industrial field, and weld grinding is a key task in the industrial manufacturing process. Although several weld-image datasets exist, most provide only coarse annotations and have limited scale and diversity. To address this gap, we constructed INWELD, a comprehensive multi-category weld dataset captured under real-world production conditions, providing both single-label and multi-label annotations. The dataset covers various types of welds and is evenly divided according to production needs. The proposed multi-category annotation method can predict the weld geometry and welding method without additional calculation and is applied to object detection and instance segmentation tasks. To evaluate the applicability of this dataset, we utilized the mainstream algorithms CenterNet and YOLOv7 for object detection, as well as Mask R-CNN, Deep Snake, and YOLACT for instance segmentation. The experimental results show that in single-category annotation, the AP50 of CenterNet and YOLOv7 is close to 90%, and the AP50 of Mask R-CNN and Deep Snake is greater than 80%. In multi-category annotation, the AP50 of CenterNet and YOLOv7 is greater than 80%, and the AP50 of Deep Snake and YOLACT is nearly 70%. The INWELD dataset constructed in this paper fills the gap in industrial weld surface images, lays the theoretical foundation for the intelligent research of welds, and provides data support and research direction for the development of automatic grinding and polishing of welds. Full article

Accurately predicting welding performance measures like ultimate strength (UTS), weld bead hardness, and HAZ mechanical hardness is crucial for ensuring the structural integrity and performance of welded components. Multitask learning (MTL) refers to a machine learning approach in which one model is designed to handle several interconnected tasks at the same time. Instead of training separate models for each task, MTL shares representations among tasks, allowing them to leverage common patterns while maintaining task-specific distinctions. In this study, we compared two advanced machine learning techniques, namely multitask neural network (MTNN) and stacking ensemble learning, for predicting these parameters based on a shared dataset. A multitask neural network (MTNN) is a specific type of multitask learning (MTL) model that uses a deep neural network architecture to handle multiple related tasks simultaneously. In MTNN, different tasks share some hidden layers while having task-specific output layers. This shared representation allows the model to learn common patterns across tasks while maintaining task-specific outputs. Both methods are evaluated using RMSE and R² to determine their predictive accuracy and overall effectiveness. It showed robust prediction strength, as its RMSE outcomes are 0.1288 for UTS, 0.0886 for weld hardness, and 0.1125 for HAZ hardness, whereas R² values are 0.6724, 0.9215, and 0.8407, respectively. This underlines that it can generalize well in interrelated tasks. Stacking ensemble learning outperformed MTL in the accuracy of individual tasks: the RMSE for UTS is 0.0263 and R² is 0.9863; for weld hardness, it is 0.0467 and 0.9782; and for HAZ hardness, it is 0.1109 and 0.8453. Such results indicate the good ability of ensemble methods to produce highly accurate, task-specific predictions. This comparison reveals the trade-offs between the two approaches. MTL is good in scenarios where the tasks are related and the data are sparse, giving efficient training and good generalization; stacking ensembles work better in the case of accurate, independent predictions. In both cases, they show remarkable potential for improving the predictive power of welding applications, making them suitable precursors to further investigation into hybrid models that bring the best features of both approaches together. Full article