Engineering

Abstract: Transport electrification can reduce climate impacts, but its environmental performance depends on vehicle production, energy supply, operation, and end-of-life treatment. This study compares battery electric vehicles (BEVs), non-plug-in gasoline hybrid electric vehicles (HEVs), and gasoline internal combustion engine vehicles (ICEVs) using model parameters relevant to Greece. A cradle-to-grave life cycle assessment was implemented in openLCA using ReCiPe 2016 Endpoint (H), with a functional unit of 150,000 vehicle-kilometres. Both electrified configurations exhibited lower impacts than the ICEV across all 18 indicators with positive results. The BEV achieved reductions of 43.2% in global-warming damage to human health, 44.8% in fine particulate matter formation, and 45.1% in terrestrial acidification relative to the ICEV. The HEV achieved reductions of 51.5% and 55.8% in fossil and mineral resource scarcity, respectively. Each electrified technology recorded the lowest value in nine positive indicators. The BEV–HEV ranking-margin analysis identified equal-impact thresholds of 9.05% for global-warming damage to human health and 4.25% for human carcinogenic toxicity. Negative water-consumption indicators were retained without interpretation as physical water savings. The findings identify category-specific trade-offs and support coordinated vehicle-efficiency and electricity-decarbonization measures without establishing overall environmental superiority.

Abstract: Human-robot collaboration has relied on task models that are specified in advance for each product. Embodied foundation models and agentic artificial intelligence change this premise, because a robot can now interpret an open goal, plan, act, monitor the outcome and recover from failure. This review examines what that change means for collaboration. A critical narrative review was conducted over 231 sources. Embodied intelligence, agentic intelligence, autonomy and collaboration are distinguished, and four collaboration modes are defined by the roles of the human and the robot. Four system architectures are compared: modular agentic, end-to-end embodied, hierarchical agentic embodied, and human-centred shared autonomy. Six groups of enabling technologies and five collaboration mechanisms, from autonomy allocation to failure recovery, are analysed. Applications in six domains, from manufacturing to domestic service, are assessed by the maturity of their evidence. Reported systems are validated mainly on single cells and short tasks, human factors are measured in few studies, and verified safety layers remain separate from learned reasoning. An evaluation framework and research directions for seven open challenges are proposed. The review gives researchers and system builders a basis for bringing embodied and agentic AI into human-robot collaboration and for evaluating the systems that result.

Abstract: This paper presents a traceability-based framework for managing inconsistencies between desired objectives and estimated performance during conceptual design. A seven-step procedure links measurable objective attributes to functions, selected means, physical elements, and performance evidence. Direction-aware, non-negative violation scores prevent compensating overperformance from masking unmet requirements, while coverage, evaluability, and ambiguity criteria identify deficiencies in the assessment record. A retrospective transmission-device case illustrates two baseline concepts and a reconfigured alternative. Under the declared inputs, their weighted inconsistency indices are 0.1542, 0.1079, and 0.0000, respectively. The baseline indices exclude the unevaluated mandatory attribute A51 and use an evaluated-weight denominator of 0.90. The reconfigured alternative has three E3-class calculations, four E4-class specifications, and one E5-class estimate; these classifications do not establish source completeness or physical verification. Simultaneous adverse bounds for A51 and A71 yield an index of 0.0126, so nominal conformance is not robust over the declared ranges. Missing source records, configuration discrepancies, and the absence of prototype tests limit the findings to a methodological illustration. The contribution is an integrated procedure for documenting deficiencies and corrective actions; comparative effectiveness remains to be established.

Abstract: Nondestructive evaluation (NDE) is moving from manually interpreted indications toward connected measurement systems in which sensing, imaging, automated processing, artificial intelligence (AI), and engineering decisions are coupled through a digital thread. This transition increases information availability but also introduces additional points at which measurement meaning, uncertainty, or traceability can be lost. This critical integrative review reframes NDE 4.0 as a metrological-assurance problem and proposes a six-layer architecture spanning the physical measurand, instrument, signal, information, inference, and decision layers. The architecture is combined with an uncertainty map for hybrid physical-digital inspection, a minimum digital evidence record, and a qualification V-model linking intended use to verification, representative validation, bounded autonomy, and post-deployment surveillance. A worked ultrasonic-sizing example illustrates how classical measurement contributions and model-inference uncertainty can be combined without treating AI confidence as a substitute for measurement uncertainty. The synthesis distinguishes metrological traceability from digital provenance, probability of detection from classification metrics, and probabilistic calibration from measurement uncertainty. Cross-sector analysis of aerospace, nuclear, rail, civil infrastructure, and additive manufacturing shows that increasing autonomy raises both information value and evidential burden. The central conclusion is that, for measurement-based engineering decisions, especially in higher-consequence applications, the defensibility of an NDE 4.0 system depends on an adequately traceable, uncertainty-aware, intended-use validation, and auditable sensor-to-decision chain.

Abstract: Natural-language access to structured manufacturing data could let plant personnel query batch records without structured query language (SQL) expertise, yet realistic multi-table industrial schemas remain hard for text-to-SQL systems, and regulated environments demand data sovereignty, traceability, and validatable behavior that cloud-hosted models cannot satisfy. This study presents a fully on-premise large language model (LLM) system that translates questions into SQL over an ISA-88-aligned database built from a pharmaceutical tablet dataset of 1,005 batches. The pipeline combines schema-aware prompting, retrieval-augmented generation (RAG) over validated query examples, execution-based checking, confidence scoring, and human review. Evaluation on 48 questions across three complexity tiers compared a schema-only baseline with five enhancements: controlled vocabularies, retrieved examples, analytical views, execution-guided self-consistency, and schema-specialized fine-tuning. Retrieved examples roughly doubled execution accuracy for two local models, from 23–25% to 46–48%, without reducing accuracy on any question; gains concentrated in two- and three-table joins, where accuracy reached 80%. Fine-tuning a 7-billion-parameter code model on execution-validated synthetic queries was the only intervention that improved the hardest tier, and combined with retrieval it achieved the highest accuracy there. Results show that a lightweight, auditable local pipeline can support regulated manufacturing analytics while flagging outputs that require human validation.

Abstract: This work addresses the limited transparency and real-time adaptability of traditional value stream mapping (VSM) in smart manufacturing. We developed an XAI-enabled dynamic VSM framework that integrates IIoT data, time-stamped process mapping, XGBoost-based waste classification, SHAP explanations, and multi-objective optimization. Using 12 weeks of MES, PLC, sensor, and quality data from a 14-station electromechanical assembly line, the framework identified 38.2% baseline non-value-added (NVA) time, achieved 89.4% classification accuracy, and reduced cycle time by 22.3%, WIP by 31.7%, and increased throughput by 18.9%. Process variability (CV) dropped from 0.34 to 0.19. Before deployment, waste root causes were hidden and decisions reactive; after deployment, real-time heatmaps and SHAP explanations enabled proactive, interpretable interventions. These results demonstrate that combining dynamic VSM with XAI and optimization improves operational performance, operator trust, and continuous-improvement decision-making in Lean 4.0 environments.

Abstract: This study demonstrates a single-step fabrication of blazed gratings on silicon substrates using direct laser greyscale lithography and aluminum coating. Three key parameters were studied for process optimization, i.e., the exposure dose of the photoresist, the focusing method, and the laser scan writing mode. We have shown that a bijective correspondence between dose and thickness ranging from 0 to 127 mJ/cm2, defined across 100 levels, enabled to obtain a linear photoresist thickness profile. Stitching artifacts were then eliminated using fine optical autofocus with overlap writing mode. Finally, the optimized grating (3.3 µm period, 12° blaze angle, 0.5 µm depth) exhibits a well-defined sawtooth profile, confirmed by atomic force and scanning electron microscopy. Optical characterization using Fourier-plane imaging shows a dominant −1st diffraction order, confirming successful blaze behavior. This work establishes direct laser greyscale lithography as a viable, cost-effective, and fast method for fabricating high-quality blazed gratings.

Abstract: Industrial Internet of Things (IIoT) technologies enable interconnected sensing, computation, and actuation across distributed industrial systems, but translating equipment health information into autonomous physical maintenance actions remains challenging. This work proposes and experimentally validates an IIoT-enabled Cyber-Physical Production System architecture that integrates Remaining Useful Life estimation, Digital Twin-based prescriptive logic, distributed control, and on-demand Additive Manufacturing to close the loop between failure prediction and maintenance execution. The architecture distributes sensing, prognostics, decision-making, and manufacturing functions across heterogeneous edge nodes using IEC 61499-compliant function blocks, DINASORE, and MQTT-based communication. A nine-node physical testbed comprising degrading robotic systems and a 3D printer was developed to evaluate individual components and the complete information-to-action workflow. Experimental results demonstrate reliable distributed data acquisition and control, online degradation assessment, and autonomous triggering of replacement-part production. During closed-loop validation, degradation of a previously unseen component caused its predicted health state to cross the prescribed threshold and automatically initiated fabrication, with a mean orchestration latency of approximately 225ms and no human intervention in the decision path. The results demonstrate how IIoT connectivity, distributed intelligence, and cyber-physical actuation can support autonomous maintenance workflows linking equipment health monitoring directly to physical manufacturing resources.

Abstract: Machine-learning predictions are used to assess material properties, monitor manufacturing processes and support inspection and maintenance. Their engineering value depends on whether the information used by a model can be examined and related to the physical task. This narrative review considers interpretable models, feature attribution, response plots, visual explanations and constrained counterfactuals in manufacturing and materials processing. It distinguishes predictive performance from explanation fidelity, physical evidence and decision usefulness. Selected applications illustrate these differences: process–property modelling in additive manufacturing, acoustic monitoring, tool-condition assessment, ultrasonic defect sizing and uncertainty in production processes. A comparison of these examples shows why the prediction target; data representation and validation procedure must be stated before explanation results are interpreted. The review proposes four categories of explanation use—descriptive, diagnostic, prescriptive and governed—and identifies the additional evidence required for each. The main methodological concerns are dependent inputs, unrealistic perturbations, non-independent observations and changes in the model or operating conditions. Explanation methods can help examine a model and guide further investigation, but process changes require separate checks of feasibility and physical consequences. The narrative selection does not establish the prevalence of individual methods across the field.

Abstract: Project-type production in the Heating, Ventilation, and Air Conditioning (HVAC) sector is highly vulnerable to scheduling disruptions caused by high product variety, complex routing structures, and uncertainty in the delivery of critical externally sourced components. In the investigated manufacturing environment, supplier-related delivery variability, rather than internal capacity limitations, constitutes the dominant source of schedule instability. This study develops a data-driven, event-based dynamic rescheduling framework that integrates supplier delay characterization with a Mixed-Integer Linear Programming (MILP) model. Historical procurement records obtained from the enterprise resource planning system were analyzed using K-Means clustering, identifying four representative delay classes of 5, 26, 51, and 127 days. These empirically derived patterns were incorporated into the rescheduling logic. The framework is activated at the disruption information time, t_info, when completed and ongoing operations are fixed to preserve schedule feasibility. A rolling-horizon mechanism freezes the subsequent five-day production window while reoptimizing the remaining horizon. Delays exceeding 40 days trigger an escalation rule that enforces the fastest available logistics mode. The multi-objective formulation simultaneously minimizes service-level agreement penalties, logistics expenditures, and operational energy costs. Implemented in Python 3.9 and solved with Gurobi 11.0, the model attained a 0.00% optimality gap within 1.2-2.4 s across the tested instances. Results show that early disruption information limits logistics costs to approximately 2,397 USD, whereas delayed information increases them to approximately 2,723 USD. The proposed framework provides a computationally efficient and transferable decision-support mechanism for coordinating production rescheduling and logistics expediting under supplier uncertainty, while indicating that expediting is economically justified when its marginal cost per recovered day remains below the corresponding dynamic delay penalty.

Abstract: Metal additive manufacturing has seen rapid growth, yet access remains limited by the high cost of established processes such as Selective Laser Melting (SLM), where entry-level systems typically require investments exceeding $100,000 USD. This paper introduces Powder Arc Fusion (PAF), a powder bed fusion concept that replaces the laser with a controlled electric arc as the energy source for metal powder melting, targeting system costs below $5,000 USD and a desktop-compatible form factor. The PAF concept consolidates powder deposition, arc-based melting, and layer handling into a single integrated machine rather than a laboratory arrangement. The process is designed to operate within a sealed protective atmosphere and to allow layer thickness and melt track width to be varied within a single build, so that fine features and bulk volumes can be produced in one uninterrupted job. An analytical assessment of the melting energy requirements across more than 50 industrially relevant alloys indicates that the required thermal power is expected to remain within the range of a compact, low-cost power supply. The arc stability required for powder bed operation is intended to be provided by a proprietary arc control architecture developed at NeoFuse3D; its technical implementation is the subject of a dedicated forthcoming publication. The work presented here is at concept stage; no prototype has yet been built. The paper positions PAF against established powder bed fusion, wire arc additive manufacturing, and bound metal deposition, and identifies the application areas in which the concept is expected to be viable. PAF targets small and medium-sized enterprises, prosumer users, and research institutions that are currently excluded from metal additive manufacturing by cost and infrastructure barriers.

Abstract: Perovskite solar cells (PSCs) represent a promising alternative to conventional crystal-line silicon photovoltaics (Si PV), but their commercialization requires simultaneous consideration of environmental performance, operational lifetime, end-of-life (EoL) management, and technology-development priorities. This study develops an inte-grated life-cycle and decision-support assessment of PSC and Si PV systems. Life Cycle Assessment (LCA) models were developed in OpenLCA on a functional basis of 1 m² of PV system, considering production, use, and EoL stages and applying ReCiPe 2016 Midpoint (H) and CML-IA Baseline. The environmental results were further integrated into a multi-criteria decision-support framework using cost-criterion normalization, Simple Additive Weighting, and scenario-based sensitivity analysis. PSC exhibited substantially lower climate-related impacts than Si PV, with reductions of 57.04% for global warming–human health (ReCiPe) and 57.35% for GWP100a (CML-IA), together with 48.53–53.41% lower fossil-resource-related impacts. However, pronounced trade-offs were identified, including higher stratospheric ozone depletion (+132.50%), mineral resource scarcity (+22.06%), land use (+754.03%), photochemical oxidation (+5747.52%), and terrestrial ecotoxicity (+731.95%). Under equal weighting, Si PV achieved higher Decision-Support Scores (DSS) for both ReCiPe (78.88 vs. 67.33) and CML-IA (67.05 vs. 62.75). Sensitivity analysis showed stable ordering under all ReCiPe scenarios, whereas the CML-IA ordering reversed under climate- and resource-priority scenarios, demonstrating the influence of environmental priorities on aggregated technology assessment. Although PSCs offer important opportunities associated with lower-energy manufacturing, their sustainable commercialization depends on extend-ing operational lifetime, reducing material- and toxicity-related hotspots, implement-ing scalable and resource-efficient manufacturing, and developing effective recycling and material-recovery systems. The proposed framework demonstrates how LCA combined with multi-criteria decision support can provide a transparent basis for evaluating environmental trade-offs and guiding the sustainable development of emerging photovoltaic technologies.

Abstract: The adoption of conventional effluent treatment plants in small and medium enterprises operating in the textile industry poses serious economic challenges that cause release of untreated wastewater and atmospheric soot emissions. In order to resolve this problem, this research paper proposes a new architectural design called “Integrated Eco-Factory” that not only aims at solving water pollution but also capturing atmospheric carbon soot, harnessing industrial waste energy, and up-cycling the hazardous chemical sludge. Four fundamental subsystems are proposed that work in coordination in the following ways: (1) Cyclonic separation technique as per the Lapple standard for transformation of flue gas soot into inkjet printing ink; (2) Electrocoagulation (EC) reactor made from aluminum that is used to replace biological degradation to achieve 100% process water recycling; (3) Solidification/Stabilization (S/S) method for transforming EC chemical sludge into construction grade Eco-Bricks; and (4) autonomous Tri-Hybrid energy system (Solar Photovoltaic Cells, Thermoelectric Generators, and Micro Hydro Turbines) with ESP32 Internet of Things monitoring system. The experiments carried out for proving the effectiveness of the proposed architecture showed particulate capture efficiency of 92.4%, ink viscosity of 3.18 cP, and surface tension of 38.5 mN/m. The Upcycled Eco-Bricks with (10%) sludge percentage had compressive strength of 14.2MPa and successfully passed TCLP heavy metal leaching safety test. The power generated through the process is about 950Wh/day, providing energy independence from external energy sources. The techno-economic analysis indicates that Capex cost recovery will be completed within 7 months, thereby providing the “Waste to Wealth” model for industries.

Abstract: Large format additive manufacturing using recycled plastic pellets marks a significant evolution in 3D printing, advancing both fabrication methods and circular economy goals. While it remains underutilized due to the inherent variability and unpredictability of the recycled material, difficult to manage in the context of product design and engineering, the experiences of New01Bike, Amarcord and OneLineTwoCoils show that embracing this unpredictability through iterative feedback loops drives manufacturing innovation and creates cult objects. The proposed approach involves fusing PC pellets with a screw industrial extruder, using parametric modeling and part redesign to mitigate warping, and employing Adaptive Layer Height slicing to unlock new geometries and reduce scrap.

Abstract: Wood modification has been developed as a non-biocidal concept to overcome inherent material limitations such as dimensional instability and biological durability. But sustainability assessment must extend beyond initial functional performance to include machinability, interface compatibility, repairability and end-of-life (EoL) cascade potential. This critical narrative review integrates the mechanisms, processing dynamics and circular-economy implications of 4 major modification routes: thermal treatment, acetylation, furfurylation and thermo-hydro-mechanical (THM) densification. We examined a purposive sample of 40 peer-reviewed articles including seminal citations between 1990-2007 and recent publications between 2021-2025. The synthesis shows that thermal modification and acetylation can enhance hygroscopic and di-mensional stability through different mechanisms, but with different trade-offs in mechanical toughness and secondary processing. Furfurylation can improve hardness and decay resistance, but also increase brittleness and recycling difficulties. THM densification improves the local load bearing capacity, but moisture-induced set-recovery is a major limitation. Cross-contamination in heterogeneous waste streams, residual chemical matrices, changed surface-energy interactions with adhesives, and limited standardization of sorting protocols are important barriers to circular material loops across the modification routes. A descriptive particleboard case-study dataset covering 0–70% secondary-wood fractions shows decreasing MOR, MOE and IB and increasing thickness swelling with increasing recycled content. These observations are presented as only study-specific descriptive evidence, and not as evidence of a universal recycling threshold or formal EN 312 P2 conformity, as the current experimental documentation does not contain sufficient information on independent panel replication and inferential statistical analysis. Future work should consider design for disassembly, digital material traceability and multi-criteria sustainability assessment.

Abstract: Abrasive waterjet cutting (AWJC) is widely used for processing Kevlar-reinforced polymer composites; however, surface quality is typically assessed using amplitude-based roughness parameters like Ra and Rz, which offer limited insight into the overall morphology of the machined surface. This study investigates the use of complementary roughness descriptors—focusing on profile height, distribution, spacing, and slope—to characterize the kerf wall morphology generated during the cutting process. Measurements were performed across different zones of the kerf wall under varied cutting conditions, enabling the examination of how specific process parameters influence these descriptors. All investigated parameters significantly influenced the four roughness descriptors, which proved to be strongly inter-correlated, reflecting a unified material removal mechanism. Furthermore, a clear distinction was highlighted between the smooth, uniform top zone and the rougher bottom zone featuring diagonal striations, confirming the sensitivity of these descriptors to local variations in cutting conditions.

Abstract: Laser-Powder Bed Fusion process is one of the most widespread Additive Manufacturing processes for metal and polymeric parts, enabling the realisation of complex, free-form, and low-density structures that are impossible or difficult to realise with traditional manufacturing techniques. However, this process is characterised by defects such as porosity, roughness, layer delamination, and others that limit mechanical behaviour. This work aims to analyse the correlation between mechanical performance and shape complexity in additive manufacturing parts. A complexity index was adopted to characterise the geometry of specimens and lattice structures, and prototyping and testing activities were carried out to evaluate the correlation between the shape complexity and performance. The materials involved in the experimental analysis are Al-Si-Mg, Al-Cu, and PA12, all realised by Laser-Powder Bed Fusion processes. To extend the approach, specimens with a biocompatible resin realised by VAT photopolymerization were also considered. The results show a dependence between shape complexity and mechanical performance, highlighting a non-linear behaviour. Despite the shape complexity offering a significant advantage in achieving tailored component performance in mechanical products, the results showed that designers must use it with particular attention due to fluctuating material performance as shape complexity increases.

Abstract: Continuous thermal monitoring of aluminum electrolysis cells—which operate at ~950 °C under strong magnetic fields in a corrosive fluoride melt—remains an unsolved process monitoring challenge. This paper presents a fiber-optic Raman distributed temperature sensing (DTS) deployment for high-temperature cathode steel bar monitoring in a 380 kA industrial cell supplied by a grid with over 30% wind and solar penetration. A custom fiber ring packaging scheme, pressed against the underside of the cathode bar (temperature >300 °C), achieved >1 m thermal contact length within the confined space beneath the cell. Features encoding supply-side renewable power periodicity, thermal inertia, and local fluctuation intensity were engineered to anchor the learning task in the process physics of the electrolysis cell. CatBoost, LightGBM, and Random Forest were combined in a stacking ensemble with a linear regression meta-learner, attaining RMSE = 0.7451 °C and R2 = 0.9944 over two months of continuous industrial operation across seven cathode bars. Frequency-domain residual decomposition revealed why LightGBM—the aggregate-weakest base learner—received the dominant meta-learner weight (+2.15) while CatBoost—the aggregate-strongest—received a negative weight (−1.85): LightGBM uniquely minimized high-frequency error (1.06 vs. 1.29 °C for CatBoost). The ensemble advantage was spatially robust across all seven bars. The 0.75 °C RMSE establishes a noise floor for residual-based monitoring, demonstrating that process-informed feature engineering and frequency-resolved stacking ensemble learning deliver predictive accuracy suitable as a process monitoring baseline in high-temperature industrial environments under increasing renewable power penetration.

Abstract: This study investigates the incorporation of natural antimicrobial agents—tannin extracted from banana peels and curcumin extracted from turmeric—into natural rubber latex compounds for medical glove applications. The additives were introduced at concentrations of 0, 1.0, 1.5, and 2.0 wt%, resulting in a total of seven experimental formulations. The objective of this research is to evaluate the effects of bio-based additives on the structural, physical, and mechanical properties of the resulting glove films. Comprehensive characterization was performed and is reported in this study. The morphological and structural characteristics of the films were analyzed using scanning electron microscopy (SEM). Physical properties, including film thickness and solubility, were measured. Chemical functional groups were identified using Fourier transform infrared spectroscopy (FTIR), while phase composition and crystallinity were examined using X-ray diffraction (XRD). Surface wettability was evaluated through contact angle measurements using a non-destructive testing (NDT) approach. In addition, the mechanical properties of the rubber films were assessed, including tensile stress, tensile strength, and percentage elongation at break. The results provide insights into the influence of tannin and curcumin on the performance of natural rubber glove films, highlighting their potential as non-cytotoxic antimicrobial additives for enhancing the functionality and safety of medical gloves.

Abstract: There exists a huge quantity of papers written about deadlock prevention of Automated Manufacturing Systems (AMS) by means Petri net (PN)-based models in the world literature. In principle they consist in two approaches, namely based either on P-invariants or on siphons of Petri net (PN) models of AMS. The approach presented here makes use of the former approach simultaneously with the structure of deadlocks themselves acquired by means of the reachability tree (RT) of such models. This approach unifies structure of all deadlocks of the AMS PN model as to activity (i.e. operation) places in order to construct the basic matrix [L] used in P-invariant approach making the supervisor synthesis possible. Namely, the product of the matrix [L] with the incidence matrix [N] of the PN model directly yields the incidence matrix [N]_s of the supervisor which guarantees deadlock-freeness of the supervised system. In such a way the application of the only one supervisor is able to prevent all deadlocks in the actual model. The approach is verified by means of its application on quite a number of Petri net models of AMS of type S³PR, GS³PR and S⁴PR. In several cases, the results obtained by the proposed approach are compared with those obtained by the siphon-based approach. The comparison is based on the confrontation of the RTs of supervised models obtained by both approaches. Finally, the possibility of decomposition (or rather splitting) of the supervisor in the case of the structure of very complex S⁴PR PN models into two supervisors is shown, if in the case of using a single supervisor, the development of one of the two parts of the model (production lines) is suppressed at the expense of the other (as far as access to resources is concerned).