Search Results (575)

We propose and empirically evaluate a two-stage approximation to a Generalized Multivariate Functional Additive Mixed Model (GMFAMM) for the joint bias correction of five NASA POWER reanalysis variables: minimum and maximum temperature ($T_{\min }$, $T_{\max }$), relative humidity (RH), solar radiation (Rad), and precipitation occurrence ($P_{\mathrm{bin}}$). Our primary contribution is the first operational-scale evaluation of such a framework (≈200,000 station–day observations, two orders of magnitude beyond previous studies) together with its deployment in an open-access web application. A systematic grid of more than 200 marginal configurations is evaluated on a strict chronological 70/30 hold-out (training 2016–2022; testing 2023–2025) to identify the optimal marginal specification per variable. Against a correctly specified marginal baseline, station-level linear calibration combined with the marginal GAMM removes the bulk of the systematic bias (RMSE reductions of ≈80%, 82% and 30% for $T_{\min }$, $T_{\max }$ and RH). A shared latent step, using the first principal component of the marginal residual matrix as a scalar proxy for ${\Lambda }_0\left(t\right)$, yields additional but conditional out-of-sample reductions (≈17% $T_{\max }$, 10% RH, 9% Rad; negligible for $T_{\min }$, with precipitation occurrence retained in the shared representation but its joint gain treated as exploratory); because it requires co-located donor observations, at ungauged locations the deployed pipeline applies the marginal correction only, whose spatial transfer is confirmed by leave-one-station-out cross-validation. The residual cross-correlation structure is consistent with, though not in itself proof of, Clausius–Clapeyron coupling. The trained artefacts are deployed in ColClim, an open-access R Shiny application that queries the NASA POWER API and the Open-Meteo forecast service for any location in Colombia and delivers historical bias-corrected series and short-range (1–16 day) forecasts. Full article

This paper evaluates the potential for green hydrogen production in Tunisia using nearly 3 GW of renewable electricity distributed across four strategically selected sites: Haouaria, Zriba, Sbikha, and Feriana. These locations were chosen for their proximity to the Trans-Mediterranean (TransMed) natural gas pipeline linking Algeria to Italy, as well as their strong but underexploited solar and wind energy resources. Each site was optimized according to land availability and renewable energy potential: Haouaria is wind-dominant, Zriba employs a hybrid solar-wind configuration, Sbikha focuses on solar, and Feriana integrates both solar and wind over a large area. The analysis reveals a total green hydrogen production capacity supported by approximately 3.1 GW of installed renewable power, with a base-case LCOH ranging from $1.21 to $2.05 per kilogram. El Haouaria emerges as the most cost-effective site due to its highly favorable wind conditions, while the sensitivity analysis shows that LCOH can reach up to approximately $3.8 per kilogram under higher CAPEX assumptions. The findings underscore the viability of a multi-site development strategy and highlight northern Tunisia’s comparative advantage for low-cost green hydrogen production, thanks to its superior resource mix, existing infrastructure, and better water availability relative to Tunisia’s southern regions. Full article

Objectives: To assess the feasibility and reproducibility of predicting antimicrobial resistance (AMR) in Escherichia coli from MALDI-TOF mass spectrometry data using a standardized, open-source machine learning (ML) workflow, we systematically compared four ML algorithms, evaluated the impact of culture conditions, extract storage, and spectral preprocessing on model performance, and validated results through nested cross-validation with statistical significance testing. Methods: A total of 282 clinical E. coli isolates were analyzed. Two MALDI-TOF MS datasets were generated from freshly cultured extracts (T₁) and recultured isolates one year later (T₃), yielding 4468 spectra. A third dataset from the T₁ extracts stored at −20 °C for one year (T₂) was evaluated for spectral stability but excluded from primary modeling likely due to storage-induced degradation. Protein spectra (m/z 2000–15,000) were preprocessed using an in-house developed MALDI-TOF preprocessing pipeline (MTPP) comprising variance stabilization, Savitzky–Golay smoothing, SNIP baseline correction, TIC normalization, LOWESS alignment, and MAD-based peak detection (SNR ≥ 3), yielding 121 m/z features. Four classifiers—Random Forest (RF), Logistic Regression, Support Vector Machine, and Gradient Boosting—were trained to predict resistance to 11 antibiotics using nested cross-validation: outer GroupShuffleSplit (5-fold, isolate-level) for evaluation and inner GroupKFold for recursive feature elimination (RFECV) and hyperparameter tuning (RandomizedSearchCV). Classification thresholds were optimized via the precision–recall curve. Model performance was assessed using AUROC, AUPRC, F1-score, Matthews Correlation Coefficient (MCC), and bootstrap 95% confidence intervals (1000 replicates). Pairwise model comparisons were tested with McNemar’s chi-squared test. Results: Among the 12 antibiotics included in the analysis (meropenem excluded for absence of resistance), resistance prevalence ranged from 1.1% (colistin) to 59.9% (amoxicillin). Colistin was subsequently also excluded from ML modeling due to insufficient resistant isolates (n = 3), leaving 11 antibiotics for prediction. The best predictive performance was observed for ciprofloxacin (AUROC 0.76 [95% CI 0.74–0.77]; F1 0.54; MCC 0.38) and ceftazidime (AUROC 0.68 [0.65–0.71]; F1 0.36; MCC 0.29), using 13 and 37 RFECV-selected features, respectively. Amoxicillin achieved the highest F1-score (0.76), driven by high recall (0.98) but modest AUROC (0.58). No meaningful predictive signal was detected for amikacin, cefepime, or tigecycline (AUROC ≤ 0.57, F1 ≤ 0.17), attributable to extreme class imbalance, and no robust multi-peak resistance signature was detected in this dataset. McNemar’s test confirmed that RF significantly outperformed Logistic Regression for all antibiotics (p < 0.01), while Gradient Boosting performed comparably to RF for ciprofloxacin (p = 0.17) and ceftazidime (p = 0.28). Frozen extracts (T₂) produced lower spectral similarity and were excluded from model training; the aligned T₁₊₃ dataset yielded the most stable performance across metrics. Conclusions: Machine learning analysis of MALDI-TOF spectra enables reproducible AMR prediction for selected antibiotics in E. coli, with ciprofloxacin and ceftazidime showing the strongest signal. Nested isolate-level cross-validation, multi-model comparison with statistical testing, and open-source code provide a transparent, reproducible foundation for integrating ML-assisted MALDI-TOF analysis into diagnostic AMR surveillance. Extract storage at −20 °C degrades spectral quality and should be avoided in ML training workflows. Full article

The implementation of safe and reliable Autonomous Driving Stacks in complex urban environments remains a formidable engineering challenge. While classical modular pipelines provide necessary component-level interpretability, they are inherently rigid, often struggling to adapt to novel environments and failing to provide robust scene interpretation in highly interactive scenarios. In this paper, we present a modular End-to-End ROS-based autonomous driving architecture that upgrades a classical modular baseline by injecting learning-based models into its individual processing layers, integrating GaussianCaR and CLIP for dense semantic BEV perception, expanding the Hierarchical Petri Net state space for safe multi-agent reasoning, refining the planning layer with continuous curve optimization, and replacing the previous reactive controller with an Adaptive Nonlinear Model Predictive Control strategy for superior trajectory tracking. Validated in the CARLA simulator across challenging traffic scenarios and adverse environmental conditions, the proposed architecture raises the Driving Score from 53.81% to 66.46% over the previous baseline, driven by a substantial increase in the Infraction Penalty from 0.59 to 0.79, reflecting a fundamental shift towards safer and more conservative driving behavior at the cost of a moderate reduction in route completion. Against pure End-to-End approaches, our architecture achieves the highest Driving Score at 73.9% and the strongest Infraction Penalty at 0.913, demonstrating that modular interpretability and competitive End-to-End performance are not mutually exclusive. Code will be made publicly available online. Full article

Despite the growing adoption of Augmented Reality (AR) in industrial maintenance, no structured methodology exists to systematically identify which operations are best suited for effective AR deployment. This study addresses this gap by proposing a five-step, Multi-Criteria Decision-Making (MCDM)-based selection framework for determining AR-compatible maintenance operations in high-speed railway systems. The framework—applied under the AFNOR FD X 60-000 standard—integrates maintenance-level compatibility analysis, multi-criteria filtering across five dimensions (operational frequency, execution complexity, safety impact, traceability, and scalability), and expert validation involving 100 railway maintenance professionals. Applied to 12 candidate operations at a high-speed railway maintenance facility in Morocco, the framework identified OP10 (insulating oil level verification of the Main Transformer) as the optimal pilot use case, confirming expert consensus (Kruskal–Wallis: H = 18.479, p < 0.001). The selected operation was subsequently integrated into a hybrid AR–Deep Reinforcement Learning architecture employing a Deep Q-Learning (DQL) agent for adaptive decision support, deployed on a Magic Leap 2 head-mounted device via a Unity-based rendering pipeline with hybrid marker-based and markerless computer vision tracking through Vuforia Engine. Experimental validation conducted over three months under simulated and semi-operational conditions yielded a 34–47% reduction in intervention time, a 55–70% decrease in human error rates, and a 28–42% decline in failure-related costs. While results are currently limited to a single-site context, the proposed methodology is directly transferable to any asset-intensive, regulated maintenance environment beyond the railway sector. Full article

Under the background of the green transformation and energy conservation in the aviation field, hydrogen-fueled aero-engines are the primary direction for achieving sustainable aviation power development. However, the unique thermophysical properties of hydrogen fuel induce extreme thermal load challenges to engine thermal management. Based on the requirements of energy conservation and digital-intelligent technologies, this paper reviewed the recent research progress, important challenges, and future development directions in the thermal management field for hydrogen-fueled aero-engines, and filled the gaps in existing related reviews. (1) As for the liquid hydrogen thermal properties and thermal management requirements, the unique thermal physical properties of liquid hydrogen can easily cause fluctuations in heat load, large temperature differences, and material compatibility issues such as hydrogen embrittlement during storage, transportation, and combustion. The application of thermal barrier coatings, the design of targeted cooling structures, and the regulation of heat loss in the pipeline of the hydrogen supply system require particular attention. (2) As for the technical architecture and optimization of thermal management, the optimization of the high-pressure side manifolds in the cooled cooling air heat exchanger increases the flow uniformity by 18.8% and reduces the weight by 22.5%. The intercooled recuperated engine with the optimum area ratio reduces specific fuel consumption by 5.3% compared to the baseline engine in cruise. However, the system-level optimization research of the above widely recognized solutions is relatively limited in terms of coordinating the energy flow of engines. The baseline engine employed the method of system integration optimization to achieve a 2.99% increase in thrust and a 6.78% reduction in fuel consumption. (3) As for the thermal management modeling and simulation, the intelligent optimization method based on computational fluid dynamics reduces the pressure loss coefficient of the vane-integrated heat exchanger by 36%. Nevertheless, the multiphysics coupling model confronts a contradiction between computational cost and accuracy. (4) As for the comprehensive evaluation method, the advanced configuration of the hydrogen-fueled aero-engine can approximately reduce specific fuel consumption by 68.5% and NOx emission by 12.7% under the same maximum thrust condition. The hydrogen consumption of the proton exchange membrane fuel cells system model compared with the baseline system, optimized by the multi-objective optimization algorithm, has decreased by 15%, while the thermal uniformity has improved by 20–30%. However, the current evaluation system mostly focuses on a single dimension, lacking the analysis of nonlinear coupling among multiple factors and a closed-loop mechanism for evaluation, optimization, and verification. Future research should focus on the matching model of liquid hydrogen’s thermophysical properties and full flight conditions, global multi-energy flows optimization methods, multidimensional collaborative numerical simulation, multiphysics coupling models, and multidimensional comprehensive evaluation systems, to provide closed-loop theoretical support for the efficient, intelligent, and reliable thermal management system for hydrogen-fueled aero-engines. Full article

To address the challenges of large morphological variations of weed targets, background interference in close-range agricultural images, and limited computational resources for visual perception models, this paper proposes a sequential visual perception method for weed recognition and short-range 3D localization in controlled or semi-controlled close-range precision weeding scenarios. The proposed method consists of two main stages: weed detection and 3D localization. In the detection stage, a lightweight MCS-YOLOv8 model is constructed based on YOLOv8n. MobileNetV3 is introduced to reduce the number of parameters and computational complexity, while CBAM and Shape-IoU are adopted to enhance weed-related feature representation and improve bounding-box regression for irregular weed targets. In the localization stage, RAFT-Stereo is used as the initial disparity estimator, and a detection-guided adaptive WLS depth optimization strategy is designed by using the detection bounding boxes and confidence scores. This strategy refines the target-region depth information and supports short-range 3D coordinate estimation. Experimental results show that MCS-YOLOv8 contains 1.6 M parameters and requires 4.3 GFLOPs. Its mAP@0.5 and mAP@0.5:0.95 reached 94.1% and 65.0%, respectively, which were 2.0 and 2.7 percentage points higher than those of the YOLOv8n baseline. Meanwhile, the number of parameters and FLOPs were reduced by approximately 46.7% and 46.9%, respectively. In the 3D localization experiments under controlled conditions, the mean absolute errors in the X, Y, and Z directions were 2.2 mm, 2.6 mm, and 3.2 mm, respectively, with an average 3D Euclidean error of approximately 4.7 mm. Dynamic target experiments further demonstrated that the proposed pipeline could complete indoor dynamic target recognition, 3D coordinate updating, and laser pointing verification under controlled conditions. The results indicate that the proposed method shows effective weed detection and short-range 3D localization performance under controlled agricultural close-range imaging conditions, and can provide a reference for visual perception and end-effector pointing in controlled or semi-controlled close-range precision weeding equipment. Full article

We present an interpretable evolutionary-fuzzy feature extraction framework for high-dimensional electroencephalography (EEG) classification. The proposed method combines an evolution strategy (ES) optimizer with fuzzy membership encoding to automatically discover compact, nonlinear feature representations from raw EEG signals. Applied to a chemosensory experiment distinguishing nasal breathing conditions during taste perception (N = 10 between-subjects participants, 1600 trials, 612 raw features), the framework achieves 89.50% cross-validated accuracy, equivalent to or exceeding all 25-feature baselines, while reducing dimensionality by 95.9% (from 612 to 25 features). The method produces fully interpretable fuzzy rules, enabling neuroscientists to inspect the decision logic rather than relying on nontransparent classifiers. A comprehensive validation including noise robustness analysis (0–30% Gaussian noise) and between-subjects generalization assessment is provided. Due to the between-subjects design, this study focuses on demonstrating the within-dataset discriminative capacity and the interpretability of the feature extraction pipeline, rather than claiming true subject-independent generalization. Full article

Acne vulgaris is a prevalent inflammatory skin condition that requires accurate severity assessment for effective management. As a step toward more objective and reproducible severity assessment, this study developed an optimized transfer learning-based convolutional neural network (CNN) framework for facial acne severity classification using a restricted, imbalanced, non-public facial image dataset. A total of 442 frontal facial images were collected under natural lighting conditions. Acne severity was graded by a board-certified dermatologist using the Investigator’s Global Assessment (IGA) scale and categorized into three levels. The study systematically investigated model architecture selection, hyperparameter tuning, optimizer comparison, cross-validation, and class-imbalance handling through random oversampling, Synthetic Minority Over-sampling Technique (SMOTE), and Generative Adversarial Networks (GANs). The 5-fold cross-validation experiment supported the reliability of model performance across different data partitions, achieving an accuracy of 0.51. Resampling methods showed limited overall benefit, although some methods altered class-wise prediction patterns. To further examine model behavior, Gradient-weighted Class Activation Mapping (Grad-CAM) visualization was used to provide qualitative insight into the facial regions contributing to model predictions. Although the findings are limited by dataset size and diversity, the proposed framework suggests exploratory feasibility for automated acne severity assessment. Rather than serving as an immediately deployable clinical tool, this pipeline provides a preliminary baseline framework that requires further validation using larger, more diverse datasets, particularly to address subtle visual differences between acne severity classes. Full article

Solid oxide fuel cell (SOFC) multi-stack systems require well-balanced reactant distribution to ensure stable operation, high efficiency, and long-term reliability. In this work, a 3D CFD framework was constructed for a vertically arranged 4 × 1 kW SOFC multi-stack system to examine the influence of U-type and Z-type manifold configurations on the distribution of mass flow, overall flow uniformity, and local utilization of the reacting gases. The system consists of four 1 kWe-rated planar solid oxide fuel cell stacks, individually rated at 1 kWe and comprising 40-unit cells. Results show that the preferred connection type differs between the cathode and anode sides. At 30% air utilization, the U-type connection provides better cathode-side air distribution, with mass flow uniformities of 0.9485 among the four stacks and 0.91842 among the 160-unit cells, while its local reaction gas utilization rates remain close to the prescribed value of 0.30. In contrast, the Z-type connection shows superior anode-side fuel distribution under all tested fuel utilization rates, with its mass flow uniformity increasing from 0.9740 to 0.9865 as the fuel utilization rate increases from 30% to 80%. At the representative 50% fuel utilization condition, the local reaction gas utilization rates of the Z-type connection are closer to the target value of 0.50 than those of the U-type connection. These findings highlight the greater suitability of the U-type connection for cathode-side reactant supply, whereas the Z-type connection is more effective for anode-side fuel distribution, providing useful guidance for pipeline connection design and flow-field optimization in vertically stacked multi-stack SOFC systems. Full article

In large-scale fluid transport systems, distributed pump and valve stations must coordinate their operations to prevent overpressure while minimizing energy use and control effort. This paper presents a communication-aware, game-theoretic coordination framework in which stations act as rational agents that iteratively adjust operating setpoints based on locally computed utilities. Existing station-level pressure controllers regulate local pressures and flows, while a slower supervisory negotiation layer governs inter-station coordination using steady-state hydraulic surrogates derived from pump affinity laws and pipeline loss relationships. The proposed framework does not rely on centralized optimization or exhaustive enumeration of strategies. Instead, stations update setpoints sequentially, evaluating incremental changes in utility to determine beneficial adjustments and detect equilibrium conditions. Cooperative behavior emerges naturally when communication is available, enabling stations to internalize the hydraulic impact of their actions on neighboring stations. When communication is lost, the system transitions seamlessly to a non-cooperative mode in which each station optimizes its local objective while maintaining safe operation. Simulation studies conducted on a multi-station pipeline with mixed actuator types demonstrate measurable performance improvements over fixed-setpoint operation. Cooperative coordination reduces total system energy usage from 39.6 MW to 38.8 MW while increasing average control valve openness from 60.4% to 63.7%. Non-cooperative operation converges more rapidly but results in higher energy consumption (39.2 MW) and greater valve throttling. Under partial communication loss, the system preserves near-cooperative energy performance (38.8 MW) with a modest increase in convergence time, demonstrating robustness to degraded communication. Across all simulated scenarios, the iterative game converged to stationary operating points consistent with Nash-equilibrium behavior in non-cooperative settings and Pareto-stationary solutions in cooperative communication settings. Full article

Deep learning models for power quality (PQ) disturbance classification remain critically vulnerable to adversarial perturbations, with classification performance degrading severely under white-box attacks. Existing defenses address individual models in isolation and provide no mechanism for operators to assess whether the system is under attack or which classifier remains trustworthy. This paper proposes a two-stage framework that combines adversarial training with large language model (LLM) reasoning to improve both robustness and interpretability. In the first stage, four architecturally diverse classifiers, including a proposed Multi-Scale Temporal Attention Network (MSTAN), are evaluated under four adversarial attacks (FGSM, PGD, C&W, and UAP), and their failure patterns are recorded as structured vulnerability fingerprints. The ensemble is then retrained via adversarial training on mixed clean and perturbed signals. In the second stage, an LLM analyzes the ensemble predictions alongside the fingerprint knowledge base to perform attack detection, fingerprint-guided meta-classification, and operator-facing threat report generation. On a 17-class, 255,000-signal synthetic benchmark, adversarial training recovers FGSM and PGD accuracy from below 25% to the 53–78% range, with MSTAN achieving the highest post-training robustness (78.26% under FGSM, 65.41% under PGD). The LLM reasoning layer provides an additional 3.5–6.2 percentage point improvement over majority voting by selecting the most reliable ensemble member based on the inferred attack condition, and detects adversarial attacks with 87.6% overall accuracy. To our knowledge, this is the first integration of LLM-based ensemble reasoning into the PQ adversarial robustness pipeline and the first application of the C&W optimization attack to power quality signals. Full article

Ceramic additive manufacturing offers strong potential for fabricating geometrically complex and application-specific components, yet achieving reliable dimensional fidelity remains challenging because dimensional deviation is governed by highly coupled material, process, thermal, and environmental factors. To address this problem, this study proposes an uncertainty-aware computational framework for dimensional error prediction in ceramic 3D printing under variable material and process conditions. The contribution is positioned as a system-level integration of established learning, uncertainty estimation, calibration, and reliability-interpretation components within a ceramic additive manufacturing dimensional-error prediction workflow, rather than as a fundamental methodological breakthrough. The validation is conducted using the publicly available Ceramic 3D Printing Process Control Dataset, a 1000-sample tabular dataset, and the resulting findings are therefore interpreted as dataset-specific computational evidence rather than direct proof of industrial deployment readiness. The methodology begins with a structured data-driven preprocessing pipeline that transforms the Ceramic 3D Printing Process Control Dataset into a multi-condition feature space through data cleaning, one-hot material encoding, min–max normalization, and engineered descriptors capturing extrusion–speed balance, thermal gradients, cooling intensity, deposition density, and material-conditioned interactions. A multi-branch deep computational architecture is then developed to encode material, process, thermal-environmental, and engineered-feature streams separately, followed by adaptive cross-condition fusion to learn nonlinear dependencies across ceramic printing regimes. To improve reliability beyond deterministic regression, the framework jointly models aleatoric and epistemic uncertainty and incorporates calibration refinement to align predictive confidence with observed error behavior, thereby enabling preliminary reliability-oriented interpretation of stable and high-risk operating conditions. Experimental results demonstrate that the full model achieves the best overall within-dataset performance, with a test MAE of 0.0118, RMSE of 0.0172, $R^2=0.999$, MAPE of 1.74%, calibration error of 0.003, PICP of 0.996, reliability score of 0.992, and a stable prediction rate of 98.7%. Although these values indicate strong predictive behavior under the current structured dataset, the exceptionally high $R^2$ should be interpreted cautiously because external experimental validation, larger measured datasets, and cross-machine ceramic printing trials are still required. These findings show that the proposed framework provides an effective system-level computational strategy for dataset-specific reliability-aware dimensional quality prediction in ceramic additive manufacturing and offers a preliminary data-driven foundation for uncertainty-aware intelligent process optimization. Full article

Context: Modern quantum-computing experimentation generates heterogeneous, context-dependent execution data whose scientific value depends on preserving calibration state, compilation decisions, and run outcomes in a traceable and repository-ready form. In the NISQ era, probabilistic outputs, time-varying hardware conditions, and opaque transpilation pipelines create a data-management problem that directly affects reproducibility, traceability, and long-term reuse of experimental records. Goal: This paper aims to address this gap by proposing a specialized metadata and schema model for managing quantum-circuit execution data as governed, machine-interpretable, and evolvable repository artifacts. Proposal: We propose QC-MM, a platform-agnostic metadata model for capturing, validating, and relating contextual evidence of quantum-circuit experiments. The model integrates time-indexed calibration binding, transpilation traceability, lightweight provenance links, validation rules, and controlled schema evolution through a JSON Schema specification. Results: The evaluation follows a multi-scenario protocol and shows that QC-MM captures dynamic calibration context in IBM Quantum Cloud, remains interoperable through a local SpinQ NMR device, and makes transpilation effects traceable through structured records. It also supports repeated-run statistical reporting and links compilation decisions to execution outcomes, including circuit-depth reductions and changes in an estimated fidelity proxy under different optimization settings. Conclusions: QC-MM provides a specialized data-modeling and schema-governance foundation for traceable quantum-experiment repositories. Beyond improving reproducibility-oriented reporting, the proposal contributes to metadata validation, controlled schema evolution, and repository-oriented management of contextual experimental data. Full article

Plunger lift systems are vital for optimizing production in gas wells, but their performance can be compromised by various operational anomalies. Traditional diagnostic methods and conventional convolutional neural network (CNN) approaches often struggle with the complex, transient data from these systems, particularly in capturing long-range temporal dependencies and generalizing from limited, imbalanced datasets. This study presents an enhanced diagnostic framework for plunger lift anomaly detection by leveraging the strengths of a pre-trained Vision Transformer (ViT). The methodology transforms one-dimensional time-series pressure data into two-dimensional image representations using the element-wise summation of Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF), which simultaneously preserves global operational trends and local transient dynamics for vision model analysis. The ViT model, initialized with pre-trained weights, is further optimized using Bayesian optimization (BO) for hyperparameter tuning, and a tailored data augmentation pipeline is employed to improve robustness. Comparative evaluations demonstrate that the proposed ViT-based approach, particularly the ViT + GAF + BO model, significantly outperforms baseline CNN models and their optimized variants, achieving the highest Precision, Recall, and F1-score, with an F1-score of 0.93. Visualizations using t-SNE confirm the ViT’s superior capability in learning discriminative features, showcasing well-separated clusters for different operational conditions compared to CNNs. This research underscores the potential of pre-trained ViTs combined with appropriate data representation and optimization techniques for achieving accurate and reliable anomaly detection in plunger lift systems. Full article