Search Results (2,414)

In acidic tea soil, boron (B) adsorption and desorption processes are dominated by the complex relationship between soil acidity, mineralogy, and organic matter. This study investigated B adsorption–desorption behavior in five acidic tea soils (pH 3.8–5.6) collected from the Eastern Black Sea region of Türkiye and evaluated the potential of machine learning (ML) algorithms to predict B desorption. Laboratory batch experiments were conducted using five initial B concentrations, and adsorption data were interpreted using the Langmuir isotherm model. Adsorption experiments indicated that B interacted with Fe/Al-oxide-containing clay minerals, which had low but favorable binding affinity, as indicated by Langmuir maximum adsorption capacities (Qmax) ranging from 46.5 to 181.8 mg kg⁻¹. Desorption experiments revealed a high degree of reversibility, particularly in soils with lower adsorption capacities, ensuring potential B leaching. To capture the governing B desorption, six machine learning (ML) algorithms—Extreme Gradient Boosting (XGBoost), Random Forest (RF), Support Vector Regression (SVR), Gaussian Process Regression (GP), Elastic Net Regression (EN), and Multivariate Adaptive Regression Splines (MARS)—were trained on 75 data points. Among the tested models, Elastic Net showed the highest predictive accuracy (R² = 0.735). This model does not replace adsorption experiments. It offers a within-assay determination of desorption given measured adsorption, which may reduce the requirement for separate desorption equilibration and analyses. Permutation importance analysis identified B_ads as the dominant predictor of B desorption, with smaller contributions from pH_ads and EC_ads. The results demonstrate that integrating laboratory experiments with machine learning provides an effective framework for predicting B mobility in acidic tea soils, offering a parameterized experimental framework for describing boron desorption behavior in acidic tea soils. Full article

This work aims to shorten the time of lager beer fermentation through a temperature profile determined by a model-based controller, as an exploratory proposal to reduce fermentation time while maintaining yeast viability and process performance, without compromising the fermentation dynamics or negatively affecting the yeast activity. This study was developed from an engineering perspective focused on the optimization of the beer fermentation process through model-based control, preserving the beer properties of the original process. This exploratory work was carried out in four stages: (1) performance of constant temperature fermentations of a lager-type beer where concentrations of yeast and ethanol were monitored along the process, (2) model parameters adjustment and validation of a beer fermentation mathematical model on the basis of data obtained from experiments, (3) outline of a temperature trajectory, in a simulation framework, from an ethanol controller of movable convergence rate constructed with a nonlinear technique and the mathematical model, (4) experimental implementation of the outlined temperature trajectory in the beer fermentation. Beer batches’ quality-control endpoints suggested by Mexican quality standards frameworks, such as fermentation time, alcoholic and caloric content, and fermentation efficiency, were analyzed. The lag stage was reduced when the temperature profile devised by the controller was employed, resulting in a reduction in the time required to reach the stationary stage. No significant final characteristic variations in bottled beers brewed at constant and variable temperatures were identified. The quality assessment of the analyzed variables was conducted in accordance with the measurement capabilities of the employed equipment and under the applicable Mexican quality standards framework. This proposal presents an alternative systematic strategy to reduce the fermentation time of lager beer, favoring the efficiency and profitability of craft beer production. Full article

Automated waste classification is essential for improving recycling efficiency and supporting sustainable waste management systems. However, conventional convolutional neural network (CNN) approaches primarily focus on localized feature extraction, which may limit their ability to capture complex spatial relationships in heterogeneous waste materials. This study proposes DeepWasteSort-SI-SSO, a Vision Transformer (ViT)-based framework enhanced with a Self-Improved Sparrow Search Optimization (SI-SSO) strategy for hyperparameter tuning. The optimization process focuses on key training parameters, including learning rate, batch size, and dropout rate, to improve convergence stability and reduce the risk of suboptimal local minima. The framework was evaluated on a balanced four-class waste image dataset (paper, wood, food, and leaves; N = 4000) using a five-fold cross-validation protocol. Experimental results achieved an average accuracy of 95.5% (±0.007), a macro-averaged AUC-ROC of 0.975, and a Cohen’s Kappa coefficient of 0.938, indicating strong agreement between predicted and true labels. Comparative experiments against ResNet-50 and a baseline ViT configuration suggest that SI-SSO optimization improves performance stability with only a modest increase in computational cost. These findings highlight the potential of optimized Transformer-based approaches for automated waste image classification under controlled evaluation conditions. Full article

Truck-drone collaborative delivery has attracted increasing attention as an effective means to enhance flexibility and efficiency in complex distribution systems. However, the resulting Vehicle Routing Problem with Drones (VRPD) is NP-hard, and existing heuristics often struggle to balance solution quality and computational efficiency, especially in large-scale and multi-trip settings. To address these challenges, this paper proposes a Structure-Guided Adaptive Large Neighborhood Search (S-ALNS) framework for truck-drone collaborative routing. The proposed approach explicitly exploits problem-specific structural characteristics through a three-phase solution process. First, balanced initial truck routes are constructed using customer clustering. Second, a structured split-based heuristic reallocates suitable customers from truck routes to a drone service. Third, the solution is further refined within an improved ALNS framework, where a structure-guided repair mechanism based on dynamic programming is introduced to efficiently handle batch customer insertions under coupled capacity and feasibility constraints. Extensive computational experiments on instances of varying scales show that S-ALNS consistently produces near-optimal solutions for small-scale instances used for validation. For medium- and large-scale instances, S-ALNS significantly outperforms classical heuristics, including simulated annealing and a standard ALNS baseline, in terms of solution quality while maintaining competitive in computational efficiency. These results demonstrate the effectiveness of incorporating the problem structure into Adaptive Large Neighborhood Searches for complex truck-drone routing problems. Full article

This paper addresses the issue of low detection accuracy in underwater optical images for unmanned underwater vehicles (UUVs) during practical operations, caused by factors such as uneven lighting, blur, complex backgrounds, and target occlusion. To enhance the autonomous perception and control capabilities of UUVs, a high-precision algorithm named MBACA-YOLO is proposed based on the YOLOv13n model. Firstly, the convolutional layers in the backbone network of YOLOv13n are optimized by replacing stride-2 convolutions with stride-1 and embedding SPD layers to enable richer feature extraction. Secondly, the newly proposed MBACA attention mechanism is integrated into the final layer of the backbone network, enhancing effective features and suppressing background noise interference. Thirdly, traditional upsampling in the neck network is replaced with CARAFE upsampling to mitigate noise pollution. Finally, an Alpha-Focal-CIoU loss function is designed to improve the accuracy of bounding box regression for underwater targets. To validate the algorithm’s effectiveness, experiments were conducted on the URPC dataset with the following evaluation protocol: 640 × 640 input resolution, batch size 1, FP32 precision, and standard NMS. All results are from a single random seed with 300 epochs of training. The proposed MBACA-YOLO algorithm outperforms the baseline YOLOv13n model, improving mAP@0.5 and mAP@0.5:0.95 by 3.1% and 2.8% respectively, while adding only 0.49M parameters and 1.0 GFLOPs, with an FPS drop of just 2 frames. This makes it an efficient, deployable perception solution for automated Unmanned Underwater Vehicles (UUVs), significantly advancing intelligent underwater systems. Full article

Continual learning of multivariate time series (MTS) forecasting is critical for process industries where working condition drift is frequent due to the variation in the feed properties and other factors. However, existing continual learning methods struggle with dynamic input dimension changes, and the lack of symmetry-aware feature and dimension regulation further exacerbates the interference of irrelevant variables and dimensional inconsistency. To overcome this problem, G-MLoRA, a continual learning method based on dynamic redundancy clustering and multiple low-rank adapters, is proposed in this paper. This method can effectively enhance the network’s capability for prediction of multivariate time series under dynamic input dimensions. First, it groups MTS via Wasserstein distance-K-means clustering to reduce irrelevant variable interference. Second, each group is assigned to an exclusive LoRA adapter, with pre-trained backbone weights frozen during fine-tuning to lower complexity and mitigate catastrophic forgetting. Third, mini-batch gradient accumulation enables reuse of inconsistent-dimensional historical knowledge. Extensive experiments on two real grinding classification datasets show G-MLoRA outperforms baselines in new/historical knowledge compatibility, especially under dynamic dimensions. Full article

We considerthe single-machine group scheduling problem with controllable processing times (i.e., resource allocation) under a common due-window ($condw$) assignment. The objective is to minimize a total cost composed of earliness, tardiness, due-window-related penalties, and resource consumption. Motivated by realistic production settings such as aerospace component machining and electronics batch assembly, the study addresses the joint optimization of group sequence, job sequence, due-window placement, and resource allocation. For linear and convex resource models, we propose a branch-and-bound ($\widehat{BaB}$) algorithm and efficient heuristics. Numerical experiments show that the $\widehat{BaB}$ algorithm can solve instances with up to 250 jobs and 16 groups. The heuristics ($\widehat{UB}$), including a simulated annealing ($\widehat{SA}$) algorithm, obtain near-optimal solutions with an average error below 0.05% much faster, demonstrating their practical usefulness for real-time scheduling. Full article

Background/Objectives: Motor imagery (MI) EEG-based brain–computer interfaces (BCIs) are promising for neurorehabilitation, but practical use is often hindered by time-consuming per-user calibration and performance instability across sessions/users. Methods: To mitigate this issue, we aim to improve subject-dependent MI classification by leveraging labeled training data from other subjects within the same dataset via transfer learning. We propose Maximizing Single-Feature Separability (MSFS), a lightweight plug-in regularization applied during target–subject fine-tuning. MSFS operates on the network feature layer and constructs batch-wise target positions by maximizing a silhouette-based separability criterion for each feature dimension. The target position computation is implemented in a fully vectorized GPU-friendly manner. Results: We evaluate MSFS on BCI Competition IV-2a and IV-2b datasets using three representative backbone networks (EEGNet, ShallowConvNet, ATCNet). MSFS consistently improves standard transfer learning across both datasets and all backbones. When compared against representative transfer learning algorithms from the literature, MSFS remains competitive against the literature baselines. Ablation analysis confirms the effectiveness of each algorithm component. Few-shot experiments further indicate that MSFS is still beneficial when the target subject provides limited labeled data. Conclusions: MSFS provides a within-dataset transfer learning enhancement for MI EEG decoding, improving target–subject accuracy under limited calibration data without relying on external datasets, and can be readily integrated into common deep MI classification pipelines. Full article

The extreme persistence of per- and polyfluoroalkyl substances (PFAS), exemplified by perfluorooctanoic acid (PFOA), demands remediation technologies that surpass conventional approaches. This study introduces a novel closed-loop adsorption–regeneration–distillation–plasma (ARDP) process designed for high-efficiency PFOA removal with low energy and chemical consumption. Comparative evaluation of anion-exchange resins identified D311 (macroporous methyl polyacrylate) as the optimal adsorbent. In batch experiments with an initial PFOA concentration of 100 mg/L, D311 achieved an adsorption capacity of ~20 mg/g, exhibited rapid kinetics, and achieved high regeneration efficiency (up to 100% under optimized conditions) via a methanol–NaCl solution. Distillation of the spent regenerant recovered approximately 80% of methanol while simultaneously concentrating PFOA for subsequent destruction, accomplished by utilizing waste heat from the plasma system, without the need for additional thermal energy input. Subsequent dielectric barrier discharge (DBD) plasma treatment of the residue achieved 100% PFOA degradation and up to 69% defluorination. The ARDP process proves to be a highly sustainable strategy, characterized by a low specific energy input (4.15 kWh/m³) and minimized secondary waste, making it a promising approach for practical PFAS remediation. Full article

Frother chemistry strongly influences gas dispersion, froth stability, water recovery, and selectivity in flotation circuits; however, conventional frothers may exhibit excessive persistence and partitioning under alkaline conditions, impairing downstream cleaning performance. This study evaluates a novel switchable frother chemistry (Transfoamer^TM) designed to achieve the benefits of strong frothing in the rougher stage while reducing the selectivity losses associated with high frother concentrations in the cleaner stages. Laboratory column tests, batch flotation experiments, and an industrial evaluation at the Caserones concentrator were conducted to characterize frother behavior in terms of gas holdup, foam height, water carrying rate, and persistence. The results showed that the Transfoamer^TM behaved as a strong frother under mildly alkaline conditions, providing gas dispersion comparable to conventional strong frothers. As pH increased, a distinct switching behavior was observed, characterized by reduced gas holdup, foam height, water recovery, and persistence, in contrast to traditional alcohol- and polyglycol-type frothers. Batch flotation tests and plant trials confirmed that combining MIBC with Transformer^TM T-100 improved rougher copper recovery without compromising circuit selectivity. Full article

l-menthol is one of the most popular flavors in the world. The separation of menthol enantiomers is crucial because of the unpleasant taste of d-menthol. This work presents the chiral separation of racemic menthol by simulated moving bed chromatography for the first time. Six preparative columns packed with amylose 3,5-dimethylphenylcarbamate coated on silica gel were used for separation, and a mixture of n-hexane/isopropanol was selected as the mobile phase. The hydrodynamic properties of the SMB columns were studied to minimize the packing asymmetry in the SMB experiment. The binary adsorption isotherm of menthol enantiomers was measured by the adsorption–desorption method. Fixed-bed batch chromatography was carried out to evaluate the adsorption kinetic behavior. Mathematical models, considering the mass transfer resistance and axial dispersion, were applied to describe the dynamics of the chromatographic separation process. The SMB process for chiral separation of racemic menthol was designed by evaluating the separation region using simulations. Reasonable agreements were achieved between the predicted results and the experimental results. Purities for both the extract and raffinate were above 99.0%, and a productivity of 0.267 g_racemate/(L_CSP∙min) and a solvent consumption of 0.431 L/g_racemate were achieved. Full article

Mass M (kg) and peak intrusion L (mm) are jointly minimized for a CFRP-enabled battery pack enclosure under the GB 38031-2025 −40° side-pole extrusion condition. A 50-run explicit FE design of experiments is conducted and deterministically partitioned into 37/5/5/3 for initial training, two sequential enrichment batches, and an independent hold-out test. Bayesian additive regression trees are trained as the primary surrogates for M, L, and Stress, and stress acceptability is enforced through a probability-of-feasibility (PoF) gate anchored to a baseline-scaled cap, σ_lim = 1.2 σ_base = 410.4 MPa. NSGA-II performed on the feasible surrogate landscape yields a bimodal feasible non-dominated set. The two branches correspond to two discrete levels of a key thickness variable x₄: a low-mass regime (n = 106) with M = 100.61–104.81 kg and L = 5.430–5.516 mm at x₄ ≈ 5.60 mm, and a stiffer regime (n = 94) with M = 110.69–115.08 kg and L = 5.362–5.430 mm at x₄ ≈ 8.00 mm. PoF screening eliminates part of the intermediate region where feasibility confidence is insufficient. Independent FE reruns further indicate that the PoF gate reduces deterministic misclassification near the stress boundary (e.g., one near-threshold candidate exceeds σ_lim, whereas others satisfy the cap with margin). Overall, the proposed workflow offers a traceable lightweighting route under extreme-cold uncertainty within a constrained FE budget. Full article

In this study, hydroxyapatite (HAp) was obtained from fishbones of four species: gilt-head bream (Sparus aurata), salmon (Salmo salar), hake (Merluccius merluccius), and megrim (Lepidorhombus boscii). Batch adsorption experiments were performed with Cr³⁺, Ni²⁺, and Zn²⁺ ions under different pH conditions (natural, 3, and 11) and contact times (6 and 72 h), which is innovative in this study and allows a unified comparison across species and thermal treatment (non-calcined vs. calcined). Results indicated that non-calcinated materials were particularly effective for Ni²⁺ and Zn²⁺ removal at natural and acidic pH, whereas calcinated samples were more suitable for Cr³⁺ adsorption under alkaline conditions. Given the precipitation of its insoluble hydroxide under alkaline conditions, zinc removal was limited to natural and acidic pH. Among the tested precursors, megrim and hake-derived (non-calcined) HAp exhibited the highest performance, achieving up to 99.99% removal efficiency at 6 h of contact time and 20 °C. The analysis of the used adsorbents confirmed metal incorporation into the HAp lattice with minimal crystallographic disruption. These findings demonstrate the potential of fishbone-derived HAp as an efficient and low-cost adsorbent for heavy metal removal from aqueous systems, while simultaneously contributing to the valorization of fishery waste. Full article

During the assembly process of the bolter miner cutting drum, the varying installation postures of the cutting picks result in unique and non-repetitive irregular gaps between the tooth seat bottom surface and the cylindrical rotating surface. Such gaps are constrained by dual-surface geometry and lack batch statistical regularity, making traditional methods such as shim filling, selective assembly, or on-site welding inadequate for achieving high-precision fitting and reliable process implementation. To address this challenge, this paper proposes an automatic design method for compensation bodies based on computer-aided design, realizing a shift from experience dependence to algorithm-driven design. This method transforms the complex dual-surface gap filling problem into a serialized geometric modeling process: first, smooth extrapolation of the tooth seat bottom surface is achieved through a point sequence prediction model based on minimum mean square error; second, surface projection is simplified to boundary curve projection, enabling precise mapping onto the cylindrical surface and generating trimming surfaces; finally, a ruled surface is constructed to integrate the extended surface with the trimming surfaces, automatically generating a compensation body fully adapted to the gap morphology. Case verification demonstrates that this method can automatically and accurately generate compensation bodies that meet dual-surface fitting requirements, significantly improving geometric adaptability and weldability. This research not only resolves a critical technical bottleneck in the assembly of bolter miner cutting drums but also provides a universal and scalable computational framework for the intelligent compensation design of non-repetitive dual-surface gaps in complex equipment. Full article

Background/Objectives: The precise classification of non-small-cell lung cancer (NSCLC) into lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) has important role in treatment decisions and in prognosis. Proper subtyping ensures that patients receive the most appropriate therapeutic strategies and allows clinicians to make informed evaluations regarding disease outcomes. This study presents a deep neural-network-based classification approach utilizing genome-wide DNA methylation profiles from the Illumina HumanMethylation450 BeadChip platform. Methods: A total of 5000 of the most discriminative CpG probes are identified through variance-based feature selection in the presented methodology, which are then classified through a five-layer deep neural network with batch normalization and dropout regularization. Training and validation were performed using data from The Cancer Genome Atlas (TCGA), with external validation conducted on two independent Gene Expression Omnibus (GEO) datasets: GSE39279 and GSE56044. Results: The model achieved 96.92% accuracy with an area under the receiver-operating characteristic curve (AUC-ROC) of 0.9981 on the TCGA test set. Robust generalization was obtained in cross-dataset validation experiments, with the GEO-trained model achieving 88.92% accuracy and 0.9724 AUC-ROC when validated on TCGA data. The most influential CpG biomarkers contributing to classification decisions are analysed using SHAP (Shapley Additive Explanations). Conclusions: These findings demonstrate the potential of DNA methylation-based deep learning approaches for reliable NSCLC subtype classification with clinical applicability. Full article