Engineering

Abstract: An explanation of an industrial intrusion detection can be trusted only as far as the chain behind it allows, namely the dataset, the detector, and the explanation. Four research questions test this chain on combinations of seven variants of industrial control system and industrial Internet of Things (ICS/IIoT) datasets with three model families (XGBoost, CNN, logistic regression), using two explainers, the KernelSHAP and the LIME. We find that the detector is worth explaining in 19 of 21 combinations. Each explainer mostly repeats its answer, where a new seed keeps about 9 of the top 10 features for KernelSHAP and 8 for LIME. The two explainers agree more than random feature lists would on 18 of 19 combinations. Yet, although both explainers are local, little of this agreement appears to be instance-specific, since lists for two different instances overlap nearly as much as lists for the same instance. In process datasets, where we injected attacks into known variables, their agreement showed no clear link to KernelSHAP being right, so agreement does not imply correctness. We propose four checks per explanation, one per research question. For reproducibility, the settings for each explanation are also reported.

Abstract: Digital payment adoption depends on users’ assessments of functional value, security, trust, and service experience. This study examines intention to use and continue using AI-enabled Visa digital payment services in Bangladesh through an integrated acceptance, security, and continuance framework. A cross-sectional survey dataset yielded 385 usable observations after attention-check and completeness screening. Nine theory-relevant antecedents were examined using simultaneous multiple regression, supplemented by exploratory bootstrap indirect-effect and interaction analyses. Technology readiness was the largest standardized predictor of intention (β = .224, p < .001), followed by satisfaction (β = .143, p = .004), trust (β = .142, p = .004), and performance expectancy (β = .134, p = .005). Perceived security retained a positive association after adjustment for the other constructs (β = .101, p = .032). The model explained 32.4% of the variance in intention. Bootstrap intervals were consistent with small positive indirect associations through trust and satisfaction, whereas the perceived-security × technology-readiness interaction was not significant. Internal consistency was low across the focal three-item scales, restricting the strength of construct-level inference. The findings offer exploratory evidence that security perceptions operate alongside readiness, trust, satisfaction, and expected performance. They support further investigation of an integrated adoption framework with refined measures and longitudinal designs, rather than causal or confirmatory conclusions from the present data.

Abstract: Prostate cancer is the second most commonly diagnosed malignancy in men, yet interpretation of multiparametric MRI (mpMRI) suffers from substantial inter-observer variability and depends on scarce expert radiologists. Expert readers follow a hierarchical, coarse-to-fine visual strategy: they first localize the prostate gland and then direct focused attention to tissue within it. Inspired by this biological selective-attention mechanism, we propose a segmentation-guided framework in which a modality-aware 3D Vision Transformer (ViT) first delineates the prostate gland from T2-weighted, ADC, and DWI volumes, and the resulting anatomical mask then constrains downstream cancer detection. Global self-attention models long-range anatomical context, analogous to the integration of peripheral and foveal information in human vision. Detection was evaluated through two pathways: gray-level co-occurrence matrix (GLCM) texture features with classical classifiers, and a ViT classifier. On 200 patients from the public PROSTATEx cohort, the proposed segmentation model achieved a Dice similarity coefficient of [0.905] versus [0.831] for an optimized U-Net. The segmentation-guided ViT achieved an AUC of [0.987] and an accuracy of [0.998] for clinically significant cancer detection. Ablation experiments confirmed that anatomical guidance improved both pathways. These results show that emulating expert visual search strategies yields accurate and interpretable computer-aided prostate cancer diagnosis.

Abstract: We evaluate whether typicality and uncertainty improve multimodal fusion of winter- wheat leaf area index(LAI)and biomass increments beyond availability-only gating(availability plus agronomic context). Consecutive elementary sampling unit (ESU) visits are kept when their 20m windows overlap (n = 105 LAI, n = 141 biomass); overlap is not a repeated plot. Nested field-grouped cross-validation does not favour the full gate. A nonnegative stacker is more accurate on fully observed ∆LAI (RMSE 0.827 vs 1.069). Availability-only gating has a slightly lower observed RMSE on fully observed biomass increments (101.1 versus 102.4), with no statistically significant difference detected (field Wilcoxon p = 0.23). No consistent field-level benefit of typicality and uncertainty was established. Withholding the current-biophysics expert from the increment predictor measured yt is still used in reconstruction—gives reliability-gated fusion (SARG) the lowest pooled biomass RMSE, but not a paired win versus availability-only (p = 0.42). Concatenated walk-forward does not establish usable next-campaign skill: ridge has a small numerical ∆LAI RMSE gain over the mean-increment baseline (2.190 vs 2.234), and every concatenated R² remains negative. Reconstruction \(\widehat{y}_{t+1} = y_t + \widehat{\Delta y}\) uses the measured current state.

Abstract: This work investigates the link between surface topography at the mesoscopic scale and emissivity at the macroscopic scale within the Morphology-Driven Radiative Analysis and Design framework. The analysis is conducted on periodic lamellar titanium gratings characterized by the period P, groove width W, and groove depth H. Their emissivity is investigated by comparing three modelling approaches: the Agababov model, ray-tracing simulations, and Rigorous Coupled-Wave Analysis (RCWA), used as the electromagnetic reference. A numerical design comprising 8,100 configurations is used to investigate the effects of geometry, wavelength, and incidence angle. The results show that the accuracy of the geometrical-optics models (Agababov and ray tracing) is governed by the relative scales of the surface geometry and wavelength, as well as by the incidence angle. The largest discrepancies with respect to the electromagnetic solution occur at small P/λ, particularly when associated with large H/λ. The resulting validity regions and their dependence on incidence angle are consistent with the trends reported by Tang and Fu. Overall, ray tracing does not systematically outperform the Agababov model, which exhibits errors of the same order over substantial regions of the investigated parameter space, highlighting its ability to provide reliable predictions despite its simpler formulation.

Abstract: Project management research increasingly recognises social and relational processes, yet the field has not developed an integrated project level account of how emotional dynamics emerge, circulate, stabilise and alter collective behaviour within temporary project systems. This paper develops a theory of emotional dynamics as emergent operating conditions: collectively produced emotional patterns that become sufficiently stabilised to shape how formal project mechanisms are interpreted and enacted. The argument is that project actors do not interpret plans, risks, decisions and relationships in emotionally neutral ways. Emotional dynamics arise from project conditions, are activated by affective triggers, propagate through interaction and shared interpretation, stabilise into project emotional climates, and may reach tipping points that redirect collective behaviour. The theory identifies six constructs: project conditions, affective triggers, emotional propagation, project emotional climate, emotional tipping points and project emotional regulation mechanisms. It also distinguishes between amplifying trajectories that increase project fragility and integrative trajectories that strengthen resilience and collective action. The contribution is to shift attention from emotion as an individual trait, interpersonal issue or leadership competency to emotion as a project level process. Existing literatures on emotional intelligence, psychological safety, trust, conflict, leadership and wellbeing provide important building blocks, but they have not been integrated into a project level theory of emotional dynamics under temporary organising conditions.

Abstract: Measurement noise is an important limiting factor in high-accuracy surface texture metrology. Although standardized procedures for noise characterization exist for areal surface topography measurements, no equivalent comprehensive framework is currently available for profile measurements. This review examines measurement noise in contact stylus profile metrology, focusing on its metrological interpretation, characterization methods, and implications for measurement uncertainty. The terminology and relationships between basic, static, dynamic, instrument, and measurement noise are examined within the relevant normative framework. Noise characterization methods are reviewed, including amplitude measures, subtraction and averaging procedures, autocorrelation, and power spectral density analysis. Particular attention is given to the applicability of procedures established within the ISO 25178 framework to profile measurements. The analysis indicates that these principles can be transferred to profile data, providing a basis for comprehensive noise characterization and its consistent inclusion in uncertainty evaluation for high-accuracy contact stylus profile metrology.

Abstract: Agricultural catalogues are awkward to benchmark once the records stay at the centres that produced them. We turn 75,897 bibliographic records from the International System for Agricultural Science and Technology (AGRIS), already held by 7 area-prefix clients, into a federated multi-view benchmark. Each record has four disjoint views (1,096 features), a ten class AGROVOCheuristic, publication-level grouping, a client- and class-aware 70/10/20 cut, and ablinded400-record subject-matter expert (SME) audit. Against the consensus gold the heuristic reached 94.50% accuracy and macro-F1 = 0.945 (sampling-weighted 94.91%). Grouping removes 1,132 duplicate-document groups that would otherwise cross folds. Compute-matched Federated Averaging (FedAvg) is 0.2998 ± 0.0151 macro-F1, against 0.3041 ± 0.0021 when the data are pooled and 0.2884 ± 0.0037 when they are not. Eight-bit uplink quantization keeps 0.3001 ± 0.0147 at a 4× uplink cut (1.6× on total traffic). The full pipeline is archived on Zenodo as release v1.0 (DOI: 10.5281/zenodo.22956341). This is a benchmark and evaluation framework, not a new federated-learning method.

Abstract: Phaseolus vulgaris L. is one of the food legumes of greatest functional relevance in Andean agricultural systems; however, the combined influence of genotype and extraction system on its phenolic profile and antioxidant behaviour remains poorly characterised for Ecuadorian varieties. The present study evaluated four Andean bean varieties (brown, black, panamito and UEB-Bombolín) under extraction with methanol, ethanol and acetonitrile using a 4 × 3 factorial design with three replicates (n = 36). Total polyphenols, flavonoids and anthocyanins were quantified using the Folin-Ciocalteu, Kumazawa and differential pH methods, respectively, while antioxidant capacity was determined using ABTS, DPPH and FRAP. The two-way ANOVA showed highly significant effects of variety, solvent and their interaction on all variables (p < 0.0001), with η²p values of up to 0.994. Principal component analysis explained 78.4% of the total variance and differentiated treatments according to seed-coat pigmentation and extraction efficiency. Brown + ACN showed the highest antioxidant activity by ABTS (73.49 µmol TE/g) and DPPH (37.38 µmol TE/g), whereas Black + ACN recorded the highest anthocyanin content (507.91 mg CY3GE/g) and the highest FRAP response (38.48 µmol TE/g). Multiple regression models showed that polyphenols and anthocyanins accounted for up to 92.1% of the antioxidant variability. The brown variety yielded a protein isolate with 70.81% protein and 76.9% extraction efficiency, demonstrating potential for the development of plant-based functional ingredients.

Abstract: This preliminary proposal suggests an educational approach built on two complementary foundations: deep, intuitive understanding of a selected set of fundamental concepts taught by a human lecturer, and explicit AI-assisted self-learning by students to support lifelong independent learning using the best generally available AI tools as they evolve. These two foundations support a third capability - engineering and scientific judgment. An ‘Analog Electronic Circuits’ course is proposed as an initial test case.

Abstract: Lithological mapping is essential for geological exploration, though fieldwork-based mapping is often limited by time, cost, and accessibility. This study proposes an integrated methodology to support lithological mapping of the Coastal Batholith between San Gregorio and Chepén, using Landsat 8, ASTER, and Sentinel-2 imagery, DEM-derived morphometric variables, and the Random Forest algorithm, applying dimensionality reduction techniques (PCA, MNF, ICA) and field-based validation. Results showed robust performance across all three sensors, with Landsat 8 achieving the highest accuracy (Overall Accuracy = 76.1 %, Kappa = 0.736, F1-score = 76.8 %), followed by Sentinel-2 (OA = 75.3 %) and ASTER (OA = 73.0 %). Morphometric variables, particularly elevation, contributed 44–48 % of the model's total importance, proving key for identifying Quaternary de-posits. Field verification confirmed the reliability of the resulting mapping, demonstrating that integrating multisensor data, morphometric variables, and Random Forest is an effective method-ology for geological exploration.

Abstract: This study investigates how integrated zero-emission mobility systems can support environmental preservation and sustainable transport transitions, in tourism-intensive coastal areas. Using the European LIFE for Silver Coast project as a real-world case study, the research addresses the limited empirical evidence concerning the implementation and governance of a multimodal electric mobility ecosystem in environmentally fragile territories. The study adopts a mixed qualitative–quantitative methodology combining environmental monitoring data, mobility simulations, avoided-emission scenarios, and lifecycle-based environmental assessment to evaluate the impacts of a multimodal and intermodal electric mobility system integrating electric land vehicles, experimental photovoltaic-powered e-boats, and a digital Integrated Mobility Platform (IMP). Results demonstrate that the coordinated deployment of zero-emission transport modes contributes to measurable reductions in pollutant emissions and private vehicle dependency, while improving territorial accessibility and sustainable tourism mobility. The findings further highlight the role of platform-based coordination in facilitating the diffusion and social acceptance of sustainable mobility innovations. From a theoretical perspective, the study contributes to innovation management and sustainability transition research by conceptualizing LFSC place-based systemic eco-innovation model paradigm aligned with the Society 5.0 paradigm, where technological innovation, environmental governance, and user-centered mobility services converge within a collaborative regional ecosystem.

Abstract: Basalt and granite offer complementary opportunities for geological CO₂ storage where conventional sedimentary reservoirs are limited or permanent mineral trapping and geothermal integration are priorities. This review examines how mineralogy, pore and fracture architecture and fluid-rock reactions control storage performance in these crystalline reservoirs. Basalt contains reactive olivine, pyroxene, plagioclase and volcanic glass that release Ca²⁺, Mg²⁺ and Fe²⁺, supporting carbonate precipitation. Field projects including CarbFix and Wallula demonstrate its potential for rapid mineralization. Granite is less reactive and has low primary porosity, making storage dependent largely on natural or engineered fractures. Feldspar alteration and integration with Enhanced Geothermal Systems may nevertheless support combined CO₂ storage and heat extraction. In both lithologies, dissolution and secondary precipitation modify porosity, permeability and injectivity, creating a tradeoff between mineral trapping and sustained fluid flow. Laboratory experiments, field demonstrations and reactive transport models are evaluated to identify the principal controls on mineralization and reservoir performance. Key research needs include improved laboratory-to-field scaling, coupled hydrochemical and geomechanical modeling and dedicated granite field validation. Basalt and granite should therefore be considered complementary components of a diversified geological carbon storage portfolio.

Abstract: Arecanut cultivation is affected by several diseases and disorders that can reduce crop health and productivity. Automated disease identification from field images is challenging because of visually similar symptoms, illumination variations, complex backgrounds, and disease-specific texture patterns. This paper proposes a Multi-Scale Texture-Attention Network (MSTANet) for automated nine-class arecanut health and disease classification. MSTANet integrates multi-scale convolutional feature extraction, a texture-aware branch based on depthwise-separable convolutions, coordinate attention, and feature fusion to capture complementary spatial and textural characteristics. The proposed model contains 1.397 million trainable parameters and processes images resized to 150 × 150 pixels. Experiments were conducted using 5,628 images for model development and an independent test set containing 1,757 images. The model achieved 100.00% validation accuracy with a validation loss of 0.0012. On the independent test set, MSTANet achieved 88.84% accuracy, 90.55% weighted precision, 88.84% weighted recall, and 88.97% weighted F1-score. It further achieved 86.97% balanced accuracy, an MCC of 0.8669, Cohen’s kappa of 0.8643, weighted ROC-AUC of 0.9909, and weighted PR-AUC of 0.9558. The highest class-wise F1-scores were obtained for Healthy Nut (98.72%) and Mahali Koleroga (95.48%). Grad-CAM analysis was employed to provide qualitative visual explanations of model predictions and to examine the disease-relevant regions used by the network. A literature-based comparison with previously reported models addressing the same nine-class arecanut classification problem further demonstrates the competitive performance of MSTANet. The results indicate that MSTANet provides a compact, multi-scale, texture-cale feature learning, texture-aware learning, coordinate attention, explainable artificial intelligence (XAI), Grad-CAM, convolutional neural networks, precision agriculture.

Abstract: Cultural heritage assets are exposed to diverse environmental, structural, geological, biological, and anthropogenic threats, yet IoT monitoring strategies remain largely site-specific. This study presents and validates the ARGUS framework, a threat-oriented methodology that defines monitoring requirements according to deterioration mechanisms rather than monument typology. The framework integrates heritage characterization, risk assessment, threat-to-sensor mapping, and IoT deployment and was evaluated across five contrasting European heritage sites. Priority threats identified through preliminary risk assessment were compared with evidence obtained from environmental, structural, and operational monitoring. The results showed strong correspondence between the anticipated threats and the deterioration processes observed during monitoring, while also enabling their relative importance to be reassessed. Cross-site comparison revealed recurring relationships between environmental forcing and material or structural responses and showed that heritage assets affected by similar deterioration mechanisms required comparable sensor configurations despite differences in typology, materials, and geographical setting. These recurring configurations were consolidated into transferable monitoring packages addressing moisture, climatic exposure, structural stability, pollution, and operational pressures. The ARGUS framework therefore provides a scalable approach for translating conservation risks into targeted monitoring strategies, supporting risk-informed preventive conservation across heterogeneous cultural heritage contexts.

Abstract: Vibration-driven robots represent a lightweight and mechanically simple class of locomo tion systems that achieve propulsion through oscillation-induced frictional interactions with the ground, eliminating the need for wheels, legs, or tracks. Despite their promise for tasks such as in-pipe inspection, debris clearance, and small-scale medical operations, most existing models assume purely elastic coupling or linear viscous damping, thereby 1 2 3 4 5 6 neglecting complex dissipative phenomena. This study introduces and analyzes a two- 7 module vibration-driven robot in which the interconnecting element is a spring–damper 8 system filled with a non-Newtonian fluid. The effective damping in this configuration 9 is nonlinear and explicitly dependent on the angular velocity of the unbalanced rotating 10 mass, resulting in strong coupling between actuator dynamics and energy dissipation. A 11 set of governing dynamic equations is derived to capture these interactions and is validated 12 through numerical simulation. Parametric studies are conducted to quantify the influence 13 of design variables such as mass, stiffness, and fluid rheological properties on locomotion 14 velocity and energetic cost. The results demonstrate that nonlinear damping introduced by 15 non-Newtonian fluids significantly alters locomotion regimes, enabling adaptive trade-offs 16 between stability, efficiency, and speed. These findings highlight the potential of velocity- 17 dependent damping to enhance vibration-driven locomotion and provide new pathways 18 for the design and control of next-generation energy-efficient mobile robots.

Abstract: Small-scale molten salt fast reactor (MSFR) demonstrators are essential for validating reactor physics, fuel cycle integration, and operational characteristics of liquid-fuel fast reactors. However, unlike large commercial MSFR concepts designed for near iso-breeding operation, demonstrators typically exhibit burner behaviour, resulting in a continuous loss of reactivity during operation. This creates a major challenge for concepts aiming to operate without a conventional mechanical control system. To address this issue, we propose and evaluate an innovative stainless-steel sleeve using the HELIOS-based neutronic modelling framework developed for molten salt reactor studies. The sleeve is a stainless steel reflector inserted inside the core vessel, where it simultaneously improves neutron reflection and provides a reserve of fuel volume that can be released gradually by reducing sleeve thickness during operation and increasing the fuel inventory without additional fuel salt. Parametric studies show that an optimised sleeve thickness of approximately 24 cm reduces the critical fuel inventory by about 30%, a massive upfront cost benefit, while maintaining the same initial criticality. Stepwise reductions of sleeve thickness provide up to 3500 pcm of reactivity compensation, exceeding the approximately 1600 pcm reactivity loss expected during the planned 20-year operating period. Beyond reactivity management, the sleeve significantly decreases fast neutron exposure of the reactor vessel. Lifetime neutron fluence is reduced to roughly 25% of the reference case for neutrons above 100 keV and to about 7% for neutrons above 1 MeV. The concept therefore combines passive reactivity compensation, reduced fuel requirements, and enhanced vessel protection. Results indicate that the sleeve offers a promising pathway toward long-term operation of a small-scale molten salt fast reactor demonstrator without reliance on a mechanical control system.

Abstract: Basalt is a promising geological medium for CO₂ storage because of its potential for permanent mineral trapping, although fluid–rock reactions can alter pore structure and hydraulic properties. This study investigates coupled axial deformation, hydraulic and geochemical evolution of basalt during CO₂-saturated brine flow at 88 °C and approximately 210 bar pore pressure. In situ axial deformation, differential pressure and apparent permeability were monitored continuously, while effluent chemistry, mineral saturation states and pore-structure properties were evaluated. Six axial deformation stages were identified, with a pronounced hydraulic transition after approximately 115 pore volumes (PV) injected. Larger deformation magnitudes coincided with decreasing differential pressure and increasing apparent permeability, and absolute axial strain change was positively correlated with stage-averaged apparent permeability (r = 0.919, R² = 0.8437). Absolute permeability increased from 2.41 to 8.55 mD, while measured porosity increased only slightly from 9.15% to 9.51%. Fe and SiO₂ concentrations generally increased during reactive flow, and late-stage saturation indices indicated thermodynamic favorability for selected secondary phases. Nitrogen adsorption measurements showed changes in accessible pore structure. These results demonstrate concurrent deformation, hydraulic and geochemical evolution and indicate that permeability evolution cannot be explained by changes in bulk porosity alone.

Abstract: This study examines how closely the English curriculum for Diploma in Engineering students in Bangladesh aligns with the communication competencies associated with skilled migration and international technical employment. A sequential two-phase framework was developed to distinguish curriculum-level alignment from student-level readiness. Phase I, reported in this paper, integrates systematic curriculum analysis, competency mapping, benchmark comparison, assessment analysis, and sensitivity testing. BTEB English-I (25712) and English-II (25722) were evaluated across ten competency domains informed by English for Occupational Purposes, engineering employability research, International Labour Organization guidance, and relevant national policy. Curriculum emphasis was coded using a transparent 0–5 ordinal rubric and quantified through the Domain Alignment Ratio (DAR), Curriculum Readiness Alignment Index (CRAI), and Readiness Deficit Score (RDS). The CRAI was 44.5% under equal weighting, 38.0% under migration-priority weighting, and 32.3% under a safety-critical scenario. Technical English, safety/procedural communication, workplace communication, and career communication together accounted for 66.7% of the total positive deficit. The assessment structure also showed limited emphasis on oral performance, with listening and speaking representing 10% of total marks compared with 60% allocated to the final examination. Overall, the findings reveal a curriculum that is comparatively strong in foundational English but less explicitly aligned with occupational and workplace communication demands. These results describe curriculum structure rather than individual student proficiency. Phase II is therefore proposed to validate student-level readiness through multi-polytechnic performance assessment, stakeholder evidence, reliability analysis, confirmatory factor analysis, and multilevel modelling. The study offers a transparent framework for linking curriculum evidence with subsequent empirical validation and evidence-informed reform of English communication education for international technical employability.

Abstract: Monitoring the progression of wildfires in near-real-time is essential for active-fire situational awareness and emergency response management. Current satellite-based wildfire monitoring systems face a trade-off between temporal and spatial resolution. Geostationary satellites such as the Geostationary Operational Environmental Satellite (GOES) offer frequent but coarse observations (~5–15 min, 2 km), while polar-orbiting satellites such as those carrying the Visible Infrared Imaging Radiometer Suite (VIIRS) provide fine spatial detail with limited temporal coverage (~12 h, 375 m). To bridge this gap, this study introduces a deep learning (DL) approach for sub-pixel fire segmentation and brightness temperature (BT) estimation from GOES imagery. The proposed approach consists of two steps, a segmentation step to distinguish active fire regions from background and a regression step to estimate the BT of active fire pixels. The model is developed using a dataset of 257 wildfires across the United States, consisting of multi-spectral GOES imagery paired with VIIRS-derived fire observations processed through parallax correction, reprojection, and resampling. The proposed approach processes data within approximately 5 min of observation, indicating near-real-time capability. Compared with a previously developed autoencoder-based model, the proposed method improves fire segmentation performance, increasing the average precision (AP) from 0.105 to 0.368 and reducing the fire-region root mean square error (RMSE) of BT estimation by 72.1% from 81.33 K to 22.69 K. Notably, the proposed approach reduces the omission error rate for medium-sized fires from 9.41% to 1.96% while maintaining zero omission errors for large fires. It also reduces the false positive rate on non-fire images from 66.86% to 8.48%. Ablation studies assess the impact of background values in the VIIRS reference data as well as the impact of loss functions and their weights. Overall, the proposed approach provides a practical solution for near-real-time sub-pixel fire segmentation and BT estimation using existing geostationary satellite systems.