Engineering

Abstract: Detailed chemical kinetics remains a major barrier to predictive computational fluid dynamics of chemically complex renewable fuels. This critical state-of-the-art review evaluates whether data-driven low-dimensional manifold methods have progressed from statistical compression to deployable reduced combustion models, with emphasis on biofuel applications. A reproducible evidence-identification and charting process classified 45 primary-evidence publications comprising 73 applications by fuel, reduction method, reconstruction and closure strategy, simulation framework, and validation tier. The evidence shows that global principal component analysis, its local, kernel, and higher-order variants, and autoencoder-based neural formulations can compactly reconstruct thermochemical states, while transported reduced coordinates have achieved a posteriori demonstrations in one-dimensional turbulence, Reynolds-averaged simulations, large-eddy simulations, direct numerical simulations, and engine-relevant ignition. The strongest solver-coupled evidence remains concentrated in methane, syngas, hydrogen-rich, hydrocarbon-surrogate, and ammonia–hydrogen systems. By contrast, biofuel-relevant ethanol and dimethyl ether studies remain predominantly a priori, despite promising compression of multistage and device-relevant thermochemical states. Across methods, the principal bottleneck has shifted from dimensionality reduction to closure accuracy, training-domain coverage, physical admissibility, differential diffusion, stiffness, and safe out-of-distribution response. Credible biofuel combustion simulation therefore requires closure-complete, physically constrained, uncertainty-aware models validated a posteriori in multidimensional and device-relevant configurations, with matched parent-chemistry accuracy, stability, memory, and end-to-end cost.

Abstract: Additive manufacturing (AM) has expanded the development of structures with high ge-ometric complexity and tunable mechanical behavior, particularly in mechanical met-amaterials. This study evaluates the mechanical performance of cellular structures pro-duced by fused deposition modeling (FDM) using polylactic acid (PLA) and polyethylene terephthalate glycol (PETG), considering standard specimens and three cellular architec-tures: conventional honeycomb, reentrant honeycomb, and EigenModes. The specimens were subjected to tensile, flexural, and compressive tests. PLA exhibited higher mechanical resistance in the standard specimens, reaching 1008 N in tensile loading, 41 N in flexural loading, and 99,150 N in compression. PETG showed lower maximum force values in most standard configurations but demonstrated better performance in selected cellular architectures. In the reentrant honeycomb geometry under compression, PETG reached 9,075 N, outperforming PLA by approximately 173%. This result indicates that ductility becomes decisive when the architecture activates deformation mechanisms typical of aux-etic mechanical metamaterials, such as inward folding, rib rotation, and cell-wall bending. Scanning electron microscopy (SEM) confirmed distinct fracture features, with PLA showing predominantly brittle fracture and interlayer separation, while PETG showed greater tolerance to voids and localized deformation. Overall, the results demonstrate that the mechanical response of FDM-printed metamaterials is governed by the interaction be-tween material behavior and cellular topology, rather than by polymer properties alone.

Abstract: Equipment platforms and large-span floors in buildings are often modeled as sandwich plates carrying discrete supported masses. If the in-plane support forces fluctuate periodically, bending stiffness is modulated in time and parametric instability can develop without direct transverse forcing, undermining vibration isolation. Prior work on viscoelastic sandwich plates has treated uniform layouts or isolated periodic design variables; how simultaneous spatial tailoring of face thickness, core moduli, and mass distribution interacts with biaxial longitudinal excitation is still open. Here we develop a coupled biaxial parametric stability formulation for periodically controllable viscoelastic sandwich plates with Kelvin–Voigt magnetorheological cores, extending the modeling framework in [1–3]. First-order shear deformation theory and Galerkin reduction yield a multi-degree-of-freedom system with periodic coefficients. Instability boundaries are obtained in one step through a direct eigenvalue procedure that couples Floquet theory, harmonic balance, and matrix eigenvalue analysis, avoiding branch-by-branch tracking. Finite element comparisons indicate that principal tongues appear near η ≈ 2ω_n, that matched periodicity in thickness and core modulus markedly increases stability margins, and that plate width, mass placement, and excitation waveform strongly reshape the safe operating region. The findings offer quantitative guidance for designing building-related sandwich foundations against parametric instability.

Abstract: Single metal droplets are widely used in electronic packaging and flexible circuit printing. However, the high density and interfacial tension of liquid metals make it difficult for the jet to stably generate single metal droplets, which greatly restricts the development of pneumatic drop on demand metal droplet ejection technology. In this paper, through experiments and numerical simulations, the jet motion states and internal fluid flow transition mechanisms during metal droplet generation are systematically investigated, and a dimensionless limiting jet length prediction model for stable single metal droplet generation is established. The results show that the dynamic competition and dominance transition between inertial force and interfacial capillary force fundamentally determine the jet motion states and internal fluid transition mechanisms. The pulse width is responsible for “activating” and “sustaining” the jet behavior, while the limiting jet length is primarily governed by the velocity provided by the supply pressure. The predictions of the dimensionless limiting jet length model are consistent with the experimental results, indicating that the model can effectively predict the limiting jet length for generating single metal droplets. The findings of this study are of significant importance for revealing the evolution mechanism of the jet during pneumatic drop on demand metal droplet generation and for achieving stable generation of single metal droplets, providing a theoretical basis for the further development of pneumatic drop on demand metal droplet additive manufacturing technology.

Abstract: To address the challenge of uneven coating deposition during air-assisted robotic spraying on horn-shaped surfaces, a gas–liquid coupled numerical model based on the Euler-Euler multiphase flow approach was developed and experimentally validated. Two dynamic spray trajectories (axial spraying and circumferential spraying) were designed to investigate the effects of spraying strategy, horn angle (45°–135°), spray distance, and spraying speed on film-forming characteristics. Results indicate that horn features alter the near-wall flow field through wall confinement effects, causing spray splitting and kinetic energy attenuation with significant anisotropy—stronger geometric constraints along the minor axis and weaker along the major axis. Axial spraying requires approximately 60% less time than circumferential spraying and yields more uniform coatings, making it the preferred strategy. During axial spraying, the central film thickness increases monotonically with horn angle (up to 147.7% over 45°–135°), while circumferential spraying achieves optimal uniformity at 67.5°–90° governed by competing geometric and temporal effects. The film thickness approximately follows a power-law decay with spraying speed (h ∝ 1/v). Pearson correlation analysis identifies horn angle as the primary variable controlling film thickness (r = 0.466), followed by spraying speed (r = −0.347) and spray distance (r = −0.325), with an optimal spray distance of 190–210 mm recommended. This study reveals the anisotropic regulation mechanism of horn features on the spray flow field and provides quantitative guidance for precise robotic spraying of complex-curved components.

Abstract: This study develops an integrated electromechanical model of a seed-removing device used in a saw-type cotton gin. The modeled machine unit comprises a squirrel-cage induction motor, an elastic-dissipative belt transmission, a perforated seed-removing tube rigidly connected to a ring gear, planet gears mounted on a fixed carrier, and an auger rigidly connected to the sun gear. The equations of motion were derived using Lagrange’s equations of the second kind, while the induction motor was represented by the dynamic characteristic proposed by A.E. Levin. The moments of inertia of the rotating components were identified experimentally by the acceleration method, and the resulting nonlinear ordinary differential equations were solved by a fourth-order Runge-Kutta scheme. The model reproduces the start-up, transient, and steady-state stages and makes it possible to evaluate angular velocities, torques, angular accelerations, power demand, and rotational irregularity. For the 3 kW, 735 rpm induction motor, the rated torque was 38.98 N·m, whereas the calculated peak starting torque reached 101.63 N·m, corresponding to a starting-torque ratio of 2.61. The transient process lasted approximately 3.5 s, and the maximum motor angular acceleration reached 2988.6 rad/s² at t = 2.25 s. Parametric calculations showed that the resistance moment of the perforated tube and the inertia of the auger exert the strongest influence on rotational irregularity, whereas the inertia and resistance of the planet gears have a comparatively weak effect. A reduction in the effective elastic-dissipative parameter of the belt drive from 17.7 to approximately 10.3 N·m/rad reduced the auger irregularity from 0.435 to 0.420 and decreased motor power consumption from about 2.55 to 2.50 kW. The proposed model provides a system-level framework for selecting drive parameters and limiting torsional oscillations in planetary-driven cotton-processing machinery.

Abstract: A unified experimental database and a reproducible Python-based harmonisation workflow are presented for local heat transfer analysis in rectangular-minichannel flow boiling. The database integrates 449 experimental source files and 64,385,791 point-level records covering six working fluids, multiple surface conditions, channel configurations, and orientations. The workflow combines template-based files, central-line in-frared wall-temperature data, geometric and operating metadata, local pressure and saturation-temperature reconstruction, bulk-fluid temperature interpolation, heat-loss correction, local heat transfer coefficient and Nusselt number calculation, operational regime labelling, and auditable quality-control flags. The database is characterised at point and source-file levels to identify regime imbalance, unequal experiment sizes, and heterogeneous coverage. A leakage-safe pilot benchmark is then conducted on streaming-sampled subcooled and saturated subsets using Random-Forest regressors with source-file-level train/test separation. The strongest result is obtained for subcooled Nusselt number (R² = 0.88), followed by subcooled heat transfer coefficient (R² = 0.80); saturated targets are more difficult (R² = 0.46 for Nu and 0.41 for heat transfer coefficient). These results are treated as feasibility screening rather than final model ranking. The proposed database architecture and validation strategy establish a repro-ducible foundation for controlled physical interpretation and extended machine-learning assessment.

Abstract: Free-wake vortex-ring models are the simplest way of describing a uniformly loaded actuator disc consistently with the Euler equations, i.e. including the radial velocity that accompanies slipstream contraction or expansion. This paper examines the numerical behaviour of this model class using an independent open-source FORTRAN 90/95 implementation and a 1:1 Python replica, for the propeller case (c_T = 1) and the Betz case (c_T = -8/9). Three results are reported. First, the two convergence measures in common use—the residual of the wake (sheet) equations and the deviation of the power coefficient c_P from momentum theory—are shown not to be equivalent: the latter has a discretisation floor and is not monotone, so it is unsuitable as a stopping criterion, and accuracy figures obtained with it are sometimes misleading. Second, the discrete Kelvin–Helmholtz saw-tooth mode is stabilised by a damping factor proportional to √z, which reaches the residual floor within a few hundred instead of 10⁴ iterations; the remaining c_P fluctuation band reflects the unresolved disc-edge region and is removed by node collocation with a spacing-proportional vortex kernel, which converges to a unique fixed point with machine-level residuals and \( c_P-(-16/27)=+4.7\times10^{-5} \). Third, the converged solutions show a bounded edge strength with fitted exponent \( a=0.00\pm0.04 \) in \( \gamma\propto s^{a} \), differing from both the \( s^{-1/2} \) spiral and the constant-\( \gamma \) proposals in the literature; we explain why single-valued, regularised sheet discretisations cannot decide this question.

Abstract: Accurate thermal modeling of friction stir welding (FSW) requires correct representation of heat generation and heat partitioning at the workpiece/tool interface. However, most published models are validated against temperature measurements from only one side of this interface, so agreement with experiment does not guarantee that interfacial heat transfer is correctly represented. This study evaluates whether one-sided temperature validation is sufficient, using steady-state FSW of AA 6061-T6 aluminum with an H13 steel tool. Temperatures were measured with thermocouples embedded in both the workpiece and the tool, and an Eulerian thermomechanical model was developed in ForgeNxt. The viscoplastic friction coefficient and the workpiece/tool heat transfer coefficient were calibrated against tool temperatures only, workpiece temperatures only, and both simultaneously. Tool only calibration reproduced tool temperatures within 2.5% but overpredicted workpiece temperatures by 18% on average; workpiece only calibration achieved 5% average workpiece error but underpredicted tool temperatures by 26%. Sensitivity analysis showed that workpiece temperatures were governed primarily by the friction coefficient, while tool temperatures were sensitive to both friction coefficient and heat transfer coefficient. No single parameter pair reproduced both temperature sets, indicating that one-sided validation can produce misleading agreement and that two-sided validation is necessary to achieve accurate interfacial heat partitioning.

Abstract: Accurate three-dimensional digitization of cultural stone surfaces is essential for the documentation, conservation, and quantitative assessment of architectural and archaeological heritage. Although commercial laser scanning systems provide high measurement accuracy, their cost, limited portability, and dependence on external power supplies often limit their use in many practical applications. This work presents the design, implementation, and experimental validation of a low-cost, portable laser triangulation scanner for the autonomous digitization of small- and medium-sized stone surfaces. The system integrates a lightweight motorized positioning platform, interchangeable laser line sensors, embedded motion control, wireless communication, and dedicated acquisition and visualization software into a compact device. Static calibration demonstrated high measurement accuracy along both the vertical and horizontal directions, while experimental validation under controlled laboratory conditions confirmed the capability of the system to reconstruct complex surface topographies and to determine areal roughness parameters in accordance with ISO 25178. The scanner successfully digitized granite specimens with different surface finishes, preserving fine morphological features relevant to quantitative surface characterization. The proposed system provides a reliable and cost-effective alternative to conventional scanning solutions and constitutes a versatile platform for the digital documentation, monitoring, and quantitative characterization of cultural stone surfaces, with potential applications in a broad range of non-contact surface metrology tasks.

Abstract: Mackinac Island in the state of Michigan is one of the most visited destinations during summertime in the United States. The island attracts over a million tourists in the short period of 5 months. The only commute to the island is using ferries that travel to the island from Mackinaw City and St. Ignace. In 2022, a study was conducted on the potential bene-fits of electrification of the ferry from a technical, environmental and economic standpoint. The results from the study concluded that while the retrofitting of the Mackinac Island ferry to an electric ferry was feasible based on components available in the market, electri-fication would lead to an increase in the CO2 emissions. The economic analysis showed that the electric ferry would result in significant financial losses to any company operating the ferry. The overarching conclusion was that an electric ferry powered by Michigan’s electric grid was not the optimum solution for that time based on both carbon emissions as well as Total Cost of Ownership (TCO). This study builds on the work completed in 2022 to determine whether an electric ferry powered by local solar energy can help reduce CO2 emissions. The analysis includes siz-ing the solar array based on the solar radiation received by Mackinaw City, the CO2 emis-sions in the manufacture of solar panels and the addition of battery energy storage as a means of energy storage between ferry trips. Ferry data used for this analysis is the same as that collected for the original study, although the appropriate parameters relating to Michigan’s electricity grid and emissions from battery manufacturing have been updated based on technological changes and the addition of different energy sources on the Mich-igan grid. The study also includes economic considerations for all potential electri-fication scenarios discussed in this work.

Abstract: Thermal error remains a primary bottleneck limiting the machining accuracy of preci-sion CNC machine tools. As a proactive, source-level countermeasure, thermal design has become increasingly critical for enabling next-generation high-performance ma-chine tools. This paper presents a critical and systematic review of machine-level thermal design methodologies, categorizing existing approaches into three principal technical routes: temperature control, material improvement, and structural optimiza-tion. For each route, we critically examine the underlying theoretical foundations, representative implementations, and reported effectiveness, with particular emphasis on the persistent gap between academic research and industrial practice. Critically, we find that existing thermal design efforts overwhelmingly target the magnitude of thermal deformation, while neglecting its spatial pattern, a deficiency that fundamen-tally undermines the effectiveness of subsequent thermal error compensation. In re-sponse to this critical gap, we argue that future thermal design should shift its para-digm from “amplitude minimization” to “deformation mode regularization”, actively shaping the spatial distribution and temporal evolution of thermal deformation to make it more predictable, repeatable, and readily compensable. This review concludes by outlining a forward-looking framework that integrates thermal mode analysis, dig-ital twin-based thermal state perception, and design-for-compensation principles, of-fering both theoretical foundations and practical guidelines for the thermal design of next-generation high-precision machine tools.

Abstract: We used High Power Impulse Magnetron sputtering (HiPIMS) to deposit Cr-Al-C-N thin films. Experiments were done by means of a commercial PVD unit equipped with a segmented sputter target. The target consisted of an upper half of Cr and a lower half of Al, respectively. While Cr-Al-N ternary coatings made by PVD are well understood, adding carbon to refine coating properties and applicability is much less explored. We added carbon in reactive sputter mode to possibly reduce internal stress and to add a friction reducing component. By varying the acetylene reactive gas flow and using the segmented target approach, different Al : Cr : C ratios could be sampled in an efficient way. Depending on positioning of the samples, Cr : Al ratios could be varied between about 4 : 1 and 1 : 2 while three different levels of carbon concentration were studied (0, 11, and 25 atomic % of coating composition including nitrogen). It was found that adding carbon on the first level is ef-fective in increasing hardness for coatings on the Cr rich side achieving maximum plastic hardness values over 40 GPa with a drop in hardness for higher Al contents. The cubic CrN phase with mostly 200 oriented planes was detected for all variants. With increasing shares of Al and C the crystallite size shrinks rapidly and the x-ray reflection of the 200 plane has less intensity which in-dicates a nano-crystalline coating structure. A turning test in stainless steel showed decreasing flank wear with higher Al content but a benefit of adding carbon could not be confirmed for the concen-trations sampled.

Abstract: Three-dimensional concrete printing (3DPC) enables the fabrication of complex artificial coral reef structures, yet the role of binder composition in controlling mechanical anisotropy, water absorption, and alkali leaching remains insufficiently understood. This study compares two regional formulations: Mix_Vn (Vietnam, 40% fly ash) and Mix_Cn (China, 30% fly ash + 20% ground granulated blast-furnace slag, GGBS). Compressive strength was evaluated in three orthogonal orientations, water absorption was measured in tap and seawater, and pH evolution was monitored under four immersion regimes. Mix_Cn exhibited higher and more isotropic strength, reduced water absorption, and slower alkali release compared with Mix_Vn. In particular, GGBS substitution improved pore connectivity and stabilized pH near 9.0 under renewed seawater, whereas Mix_Vn exceeded pH 10.5 within two days in static tap water. These findings provide the first comparative evidence that GGBS incorporation simultaneously mitigates anisotropy and durability concerns in 3DPC, making Mix_Cn the recommended formulation for marine ecological applications and contributing to low-carbon concrete construction through substantial Portland cement replacement by industrial by-products.

Abstract: The research focuses on the dynamic calibration of off-the-shelf inertial measurement unit (IMU) triaxial accelerometers, sensors generally used for kinematic analyses in sport and biomedical engineering. The assessment of Xsens Movella DOT sensors' performance, reliability, and limitations is presented, providing a metrological basis for their application in wearable monitoring systems. The metrological characterisation was performed to quantify the sensors’ dynamic response, their bandwidth, and measurement repeatability and reproducibility within the range of interest. Three units from the same batch were tested along three orthogonal axes under controlled excitation conditions, using a laser Doppler vibrometer as reference. The experimental protocol included harmonic excitations in the 10–45 Hz range, harmonic excitation up to 130 Hz to quantify limitations and potential errors of the sensors when measuring signals out of the nominal bandwidth, and random excitation, limited in the frequency range up to about 40 Hz, to validate their applicability in a generic dynamic environment. Thus, the acquired signals were analysed in both the time and frequency domains: in particular, the Frequency Response Function (FRF) between the IMU accelerometers and the reference system was measured, along all three measurement directions, and the corresponding Power Spectral Densities (PSDs) were computed. A numerical optimisation procedure was then applied to model the acquired FRF, providing an estimation of the FRF complex function, allowing for correction in general dynamic applications. One major result was that dynamic compensation is mandatory within the nominal bandwidth, given the attenuation of the measured amplitude of about 25% at the maximum frequency of the bandwidth; moreover, aliasing error occurs if the excitation frequency is above the Nyquist frequency, introducing frequency-dependent errors and misleading results. Thus, the proposed methodology and correction model, together with the highlighted instrumental effects, allow for accurate acceleration measurements by using the tested Xsens Movella DOT sensors in an induced and transmitted vibration scenario, although the defined methodology can be more generally extended to similar devices and instruments, aiming for proper dynamic characterisation.

Abstract: Engineering laboratory pedagogy is often framed through binaries such as collaboration versus autonomy, structure versus flexibility, and in-person participation versus alternative modes of engagement. This exploratory mixed-methods study examines how these preferences coexist within individual students rather than treating them as mutually exclusive orientations. Data were drawn from an end-of-term self-reflection survey completed by 50 students in an undergraduate engineering laboratory course at a public university in the US. Closed-ended responses were analyzed through theoretically defined cross-classifications, descriptive comparisons, effect-size estimates, and exact tests, while open-ended responses were used to interpret the resulting preference configurations. The findings revealed substantial overlap among seemingly contradictory preferences. Among students who preferred structured, in-person laboratory sessions, 41.9% also expressed strong interest in greater spatial-temporal flexibility, while 46.5% rated temporal flexibility as highly important. Similarly, 51.1% of students who felt comfortable contributing in groups nevertheless perceived students who prefer working alone as disadvantaged. Students who frequently completed coursework outside standard hours were also especially likely to value temporal flexibility. These findings suggest that engineering students cannot be meaningfully divided into fixed categories of collaborative versus independent or traditional versus flexible learners. Instead, inclusive laboratory pedagogy may be better understood as a configurable learning ecology that preserves structure and social connection while providing meaningful autonomy, flexibility, and multiple pathways for participation.

Abstract: Fresh paddy transported by barge in the Mekong Delta can accumulate respiration heat during journeys longer than 24 h, accelerating quality deterioration. This study evaluated forced aeration using a 1 m² × 2.5 m laboratory model (approximately 1.4 t) and field trials on a 60 t barge with aerated and non-aerated compartments. Fresh paddy (24.1 ± 1.35% wet basis) was aerated at an average superficial air velocity of 0.053 m s⁻¹, equivalent to 129 m³ h⁻¹ t⁻¹. Grain temperature, moisture content, airflow, static pressure, air enthalpy, and milling quality were measured. In field trials, aeration reduced grain temperature to approximately 29.0 °C after 6 h, close to ambient temperature (29.3 °C), whereas non-aerated paddy reached 37.8 °C. The temperature difference of approximately 10–11 °C was maintained during transportation. Mean specific heat removal was 616 kJ h⁻¹ t⁻¹, and cumulative thermal exposure decreased from 243.8 to 13.1 °C·h, corresponding to 94.6% suppression. Moisture content and most quality indicators did not differ significantly, while chalkiness decreased from 7% to 4%. Forced aeration can therefore stabilize high-moisture paddy during barge transportation and reduce heat-related quality loss.

Abstract: Project CIPACTLI (Compact Integrated Polymer Amphibian for Coastal and Littoral Investigation) evaluates the operational limits of rare-earth magnetic connections in a cannon-inserted amphibious robotic chassis designed for search-and-rescue (SAR) and environmental monitoring. Two prototype iterations were systematically compared across movement speed, structural loading, and ballistic survivability under combustion-cannon deployment. The first prototype employed a tensegrity-inspired, serpentine soft robot with segmented inflatable buoyancy modules connected via magnetic couplings. A 2³ factorial design evaluated three factors: core tendon material (nylon versus steel), terrain (terrestrial versus aquatic), and magnet type (flexible magnetic tape versus rare-earth neodymium N52 magnets). Results showed no significant effect of magnet type on movement speed (p = 0.507), with tendon material dominating performance (partial η² = 0.637). Both magnet types achieved 100% survival under impulsive cannon launch, indicating that chassis compliance attenuates peak shock below the brittle-fracture threshold of neodymium magnets. The second prototype eliminated magnetic couplings and inflatable actuators in favor of a monolithic PLA body with integrated buoyancy. This design demonstrated a 6.96-fold increase in compressive load capacity (p = 0.001) and equivalent ballistic survivability, though with a 24.9% reduction in aquatic movement speed. Tethered deployment trials validated mechanical viability for maritime SAR applications. These findings demonstrate that rare-earth neodymium magnets confer no performance advantage over flexible magnetic tape within the tested operational envelope, and that structural integration—not magnet grade—governs resilience under impulsive loading. The results counsel against blanket substitution of rare-earth magnets into impact robotics without task-specific justification, supporting expanded evaluation of polymer-bonded composites and ferrite-bonded tapes as lower-criticality alternatives.

Abstract: Miniaturized tests can offer significant material and time savings while still providing representative results comparable to macro-scale tests for relevant applications. Moreover, in challenging fields like hazardous materials research in nuclear, chemical, and similar industries, miniaturized testing presents an even more compelling and practical solution to counter safety, specimen preparation, and material disposal constraints. With an emphasis on optimization to guarantee maximum stress concentration, and that a uniform stress distribution, and random location of failure, always occur within the gage, this study investigates the bending fatigue behavior of meso-scale Krouse specimens. Initially theoretical and finite element analyses were performed for geometry optimization. Subsequently, the optimized geometry was experimentally validated through constant amplitude, load controlled bending fatigue tests with stress ratio, . Both theoretical framework and experimental validation confirmed the effective optimization of the specimen design for stress behavior. Furthermore, the S-N curve generated from the experiments demonstrated consistent patterns when compared with analytical and literature data, proving the validity of the optimized specimen design. The scientifically validated specimen design serves as a novel and innovative approach that could be applied in applications where small size, and bending loads are critical.

Abstract: Subway tunnels exacerbate fire hazards, revealing a critical lacuna regarding the spatio-temporal coupling of fire progression and human egress. To address this, this study proposes and validates an Integrated Spatio-Temporal Risk Assessment Model to explicitly quantify survivability thresholds under complex fire dynamics. The framework synergizes Large-Eddy Simulation (LES) based Computational Fluid Dynamics with agent-based pedestrian trajectory modeling within a 3D tunnel featuring a 2% longitudinal gradient. Evaluating 9.5 MW and 12 MW fire energies, the model assessed Single-Sided Evacuation (SSE), Double-Sided Evacuation (DSE), and Sprinkler-Assisted Single-Sided Evacuation (SSE-S) across 2,160 agents. Hazards were quantified by continuously resolving Fractional Effective Dose (FED) indices, 60°C boundaries, and 500 ppm CO fronts. Simulations reveal the gradient induces a severe stack effect, accelerating toxic dispersion and yielding temperatures exceeding 1200°C. Consequently, SSE engendered fatal bottlenecks (FED: 15.33), whereas DSE optimized pedestrian flux, capping peak FED at 0.52. Crucially, while active suppression (SSE-S) extinguished flames within 105 seconds, thermodynamic cooling induced a paradoxical loss of smoke buoyancy, causing toxic layers to stratify at the breathing zone. Ultimately, while DSE and SSE-S are paramount for survivability, water-based suppression generates localized toxicological risks, necessitating the integration of low-level smoke detection and extraction architectures in future subterranean designs.