Intumescent flame-retardant coatings (IFRCs) can modify the thermal degradation and mass-loss behavior of wood. This study experimentally evaluated the mass-loss characteristics of Fagus sylvatica and Pinus rigida under coated and uncoated conditions, using data from four independent cone calorimeter experiments conducted at 50 kW m−2 in accordance with ISO 5660-1:2015. The experiments generated 845 observations, and we assessed key mass-loss characteristics, including total mass loss, t50, peak instantaneous mass-loss rate, and time to burnout. For these individual experiments, the data indicate that the coating reduced total mass loss in both wood species. Total mass loss declined from 86.87% to 80.93% for F. sylvatica and from 89.07% to 81.92% for P. rigida. For F. sylvatica, the coating also delayed the onset of rapid mass loss from 25 to 90 s and reduced the fraction of rapid mass loss from 36.6% to 14.9%. Four machine-learning models, CatBoost, NGBoost, Random Forest, and HistGB, were subsequently evaluated for predicting specimen mass using default and Bayesian-optimized hyperparameters. NGBoost showed favorable relative performance under 5-fold pointwise cross-validation, reflecting within-trajectory predictive performance, yielding an MAE of 0.167 g, MSE of 0.074 g2, and R2 of 0.999769. Under a leave-one-run-out validation framework, the HistGB showed notable run-level generalization within the evaluated set, achieving an overall R2 of 0.8941 and MAE of 4.5554 g. These results suggest that the experimental–ML framework can yield consistent predictions within the specific experimental domain
Hydrothermal liquefaction (HTL) offers a promising route for simultaneously recovering energy from sewage sludge (SS) while mitigating the burden of conventional sludge management.. However, SS-derived biocrude is typically rich in lighter fractions, limiting its suitability as a substitute for petroleum middle-to-heavy distillate fuels. This study investigated rice husk (RH) co-feeding for adjusting total solids (TS) loading and modifying the fuel-range distribution. Co-HTL experiments were conducted with a 12-L batch reactor at 320 °C with a 60 min retention time. The 1:0.5 SS:RH ratio (TS = 14.5 wt%) provided the most favorable process condition, yielding 42.5 wt% biocrude and a 12.1% synergistic enhancement. Elemental analysis and Van Krevelen analysis indicated enhanced carbon densification associated with deoxygenation. GC–MS showed that co-HTL shifted the biocrude composition toward aromatic and ester compounds while reducing NSO-heterocycles and phenolics, whereas TGA demonstrated that over 50% of the biocrude was distributed within middle-to-heavy distillate fractions, indicating a pronounced shift toward diesel- and fuel–oil-range products. The co-produced biochar (HHV = 10.4 MJ/kg) exhibited potential as both a supplementary solid fuel and a potential nutrient-rich soil amendment. A preliminary TEA for the Dasherkandi Sewage Treatment Plant estimated a minimum fuel selling price of 1.2 USD/kg and a simple payback period of 9.4 years, indicating promising economic potential for the proposed pathway. The results show that RH co-feeding can provide a practical way of controlling TS while favorably modifying biocrude fuel-range characteristics, linking SS valorization with renewable fuel production and advancing circular resource recovery in wastewater treatment systems.
Reliable renewable-resource characterization is critical for designing cost-effective and climate-resilient microgrids, particularly in rural regions where conventional planning often relies on outdated meteorological averages. This study proposes a forecast-integrated framework that couples deep-learning-based meteorological forecasting with techno-economic optimization of a hybrid renewable microgrid for a rural educational facility in Bangladesh. A 40-year dataset (1984–2024) of solar irradiance, wind speed, and ambient temperature was used to develop and validate six forecasting models: SVM, LSTM, BiLSTM, GRU, CNN–LSTM, and CNN–BiLSTM. The CNN–BiLSTM model achieved the best performance, with R2 values of 0.980, 0.961, and 0.959 for solar irradiance, wind speed, and temperature, respectively. The resulting forecasts were integrated into HOMER Pro to evaluate seven microgrid configurations incorporating photovoltaic, wind, battery storage, hydrokinetic generation, and grid resources over a 25-year planning horizon. The optimal PV–wind–battery–grid configuration achieved a renewable fraction of approximately 90.9%, a cost of energy of $0.0471/kWh, and a net present cost of $17,225, while reducing annual CO2 emissions by approximately 80.5% relative to the base case. A site-specific assessment further showed that hydrokinetic generation is technically feasible but economically non-optimal under the available river-flow conditions. Sensitivity analysis identified wind-turbine cost as the most influential component-level economic factor. The proposed framework provides a data-driven approach for robust renewable-energy planning, technology screening, and low-carbon rural electrification.
A three-dimensional numerical investigation of a Swiss-roll micro-combustor is performed, focusing on hetero-/homogeneous combustion of premixed hydrogen–air using detailed gas-phase and surface reaction kinetics. At higher wash-coat factor, oxygen adsorption dominates surface reactions, blocking active catalytic sites and reducing key radicals such as H and OH. Gas-phase kinetics exhibit non-monotonic behaviour: low wash-coat factors (effective catalytic-area factor) enhance chain-branching reactions, moderate value suppresses them, and higher value for wash-coat factors favor chain-termination processes. Thermal analysis identifies F = 0.5 as the condition with maximum catalytic heat release and minimal wall–gas temperature difference. However, under the baseline wall material (stainless steel) thermal conductivity (=12 W·m−1·K−1) and fixed operating conditions, the highest emitter efficiency occurs at F = 0.01. For uniform temperature combustion regime at F = 1, emitter efficiency remains largely insensitive to inlet velocity under stoichiometric conditions, while lean mixtures show a slight decline with increasing flow rate. In addition, increasing equivalence ratio and wall thermal conductivity significantly improves the emitter efficiency, enabling a growth up to 62% By replacing the baseline material (λs = 12 W·m−1·K−1) with superior thermal conductivity material (λs ≅ 30 W·m−1·K−1), while maintaining lower combustion temperatures and broader material compatibility.
Plastic waste pyrolysis can complement conventional recycling within integrated circular-economy systems, although reported performance varies with experimental conditions and life-cycle assessment methods. This structured analytical review examines a core corpus of 70 peer-reviewed publications from 2014 to 2025. Five reviews published in 2026 are considered separately to provide context for recent developments without altering the predefined analytical period. The comparison of technical, energy, environmental, fuel-quality, and deployment indicators accounts for differences in feedstock, functional unit, and system boundary. Continuous, heat-integrated configurations report process-energy recovery approaching 90% under favourable pilot-scale or refinery-integrated conditions. The reported greenhouse-gas reduction relative to incineration is approximately 59% when product-substitution credits are included and approximately 31% when these credits are excluded. Sensitivity analysis indicates that the general pathway-level Product Quality Score ranking remains stable across the tested criterion weights, while the leading fluidised-bed configuration changes when product quality receives greater weight than yield and readiness. The applicability of these findings depends on electricity and heat supply, collection and sorting systems, contaminant control, product requirements, and environmental permitting. Pyrolysis is therefore most relevant for suitably prepared mixed and contaminated plastic streams within an integrated waste-management system.
Batteries are essential energy storage devices that enhance the reliability and efficiency of renewable energy systems. The Battery Management System (BMS) is necessary for ensuring the battery’s consistency, safety, performance improvement, and efficiency. It is able to discriminate the discharging and charging current, to provide the warning information very early and manage the batteries connected economically. In case of large scale BMS; it needs complex wiring setup, expensive hardware, air-conditioners and regular maintenance with man-power. To address these problems, the Cloud Integrated Battery Management System (CIBMS) has been proposed to monitor battery characteristics continually and that the proposed controller use sophisticated computational techniques to predict the battery’s State of Charge (SOC) and State of Health (SOH). It stores the bulk amount of measured data to the Amazon Web Services (AWS) 1 GB RAM and 40 GB Space cloud. The proposed system regulates the battery charging from solar PV and the maximum discharge rate. The hardware setup has been implemented in the institutional laboratory and tested for solar powered lead acid batteries. The sensors connected to Internet of Things (IoT) devices continuously collect and store real-time battery statistics in cloud databases. The robust machine learning method of the Support Vector Regression (SVR) technique has been used to examine these data to estimate the SOH with the values of 99.14% to 79.87% corresponding to 100% − 80% of actual SOH and SOC with the values of 99.61% to 11.29% corresponding to 100% − 10% of actual SOC. Additionally, it anticipates the battery’s age, fault, and maintenance needs and best manages the battery’s charging and discharging limits. The proposed CIBMS has been executed using various algorithms and the results are validated with the actual measurement.
Reliable instantaneous high-power supply without compromising long-duration autonomy remains a key challenge for lunar base microgrids. This paper proposes a hybrid microgrid architecture integrating a helium-xenon closed Brayton cycle (CBC) nuclear power system with a battery energy storage system (BESS). A multi-physics co-simulation platform is developed in OpenModelica, incorporating reactor point kinetics, turbomachinery, a permanent magnet synchronous machine (PMSM), and battery electrochemistry. Compared with RELAP5, the steady-state errors are within 3% for core parameters and within 7% for all main parameters, and the transient response agrees under a 30% heat-source-power perturbation introduced. A state-of-charge (SOC) threshold-triggered hierarchical coordinated control strategy decouples the fast-response battery layer from the slow-regulation reactor layer under a representative lunar load profile. The strategy maintains the DC bus voltage deviation within ±2.0% and the transient core outlet temperature deviation within ±20 K. An NSGA-II optimization framework is then constructed using the battery parallel-string count and SOC dispatch thresholds as decision variables, with total system mass and reactor power regulation effort as objectives. The optimized scheme reduces total system mass by 46.2 kg and drives reactor power regulation effort toward zero. When the peak-to-valley power ratio increases to 133%, mass saving rises to 123.3 kg; however, extending the peak-load duration from 3 h to 6 h weakens or reverses this benefit. This paper provides a reusable methodological framework for lightweight design and energy management of space nuclear-storage hybrid power systems.
Parabolic Trough Collectors (PTCs) are of pivotal importance in harnessing solar power for thermal energy. Nevertheless, the performance of PTCs has been largely impaired by thermal loss and inefficient solar irradiance absorption. The main objective of the present study was to propose a numerical simulation of a PTC system utilizing a novel hybrid nanofluid and new flow inserts. In the present study, the fundamental equations for mass, momentum, and energy conservation were numerically solved using the k-ε turbulence model. A second-order scheme was used in the numerical simulation. The mesh independence study was conducted, and the numerical simulation was validated using experimental measurements. The isolated and interaction effects of Al2O3; SiO2; TiO2 + MWCNT/water as nanofluids and compared them to twisted tape, corrugated tube, and fin-type inserts. Results showed that the TiO2 + MWCNT/water nanofluid coupled with a fin insert posted better thermal performance, with improvements by 29.34% and 20.40% in heat transfer over traditional water-based systems, supported by CFD simulations and experiments. In addition, the Nusselt number and friction factor peaks increased by 15.26% and 10.86%, and the size of the collector was minimized by 39.33%. The combined CFD-experimental approach developed during this study offers a dependable approach to nanofluid-based PTC optimization as well as useful design recommendations for high-efficiency solar thermal systems of the future.
The transition to low-carbon urban transport has stimulated interest in hydrogen fuel cell electric buses (FCEBs) as alternatives to conventional compressed natural gas (CNG) vehicles. A hydrogen fuel cell bus and a CNG bus were evaluated using coupled well-to-wheel (WTW) fuel-cycle and techno-economic analyses excluding vehicle and infrastructure manufacturing, with vehicle behavior simulated in Simcenter Amesim under the New York City and SORT cycles. The analysis considers steam methane reforming (SMR) and grid electrolysis, fuel cell stack degradation, grid carbon intensity, and a midlife power-plant replacement scenario. At beginning of life, the FCEB reduces WTW greenhouse gas (GHG) emissions by 58 % with SMR hydrogen and 47 % with grid-electrolysis hydrogen relative to the CNG bus. When degradation and a replacement at 25,000 operating hours are propagated across the 12-year service period, the distance-weighted reductions are 51.2 % and 39.1 %, respectively. The corresponding total costs of ownership (TCOs) are 1.88 million USD for SMR and 2.10 million USD for electrolysis, compared with 1.39 million USD for CNG, yielding premiums of approximately 35 % and 52 %. In 100,000 Monte Carlo realizations, the FCEB is more expensive in 99.99 % of cases under SMR and all cases under electrolysis. The results provide a strong operational-phase environmental case for FCEBs, while showing that purchase price, dispensed hydrogen cost, stack aging, and potential midlife replacement remain substantial economic barriers.
As microgrids are increasingly deployed for off-grid electrification, energy management systems (EMS) are required to jointly optimize operating cost, equipment degradation, and reliability. Although multi-objective EMS optimization has attracted increasing interest, few studies explicitly enforce reliability and storage-cyclicity requirements as hard constraints while treating equipment wear as an optimization objective. This paper formulates the 24-hour dispatch of an islanded PV–WT–BESS–H2–DG microgrid as a constrained bi-objective optimization problem minimizing operating and degradation costs, with D=72 decision variables and three hard constraints: LPSP≤2%, battery state-of-charge cyclicity, and hydrogen-tank cyclicity. Four constrained multi-objective evolutionary algorithms (CMOEAs), namely BiCo, NSGA II, C-TSEA, and CMOSMA, are benchmarked over 20 independent runs under three representative meteorological scenarios and further evaluated under two additional renewable-deficit and renewable-surplus conditions. BiCo achieved the highest median hypervolume in all three baseline scenarios (0.857, 0.126, and 0.842 for SC1-SC3), with statistical analysis confirming significant advantages over C-TSEA and CMOSMA, while NSGA II remained the closest competitor. At the BiCo best-compromise operating point, the cost of energy (COE) reached as low as 0.0598 €/kWh under best conditions, with a CO2 intensity of 0.0630kgCO2/kWh. Sensitivity analysis identified fuel price and battery cost as the main drivers of operating and degradation cost, while removing hydrogen storage increased diesel consumption by up to 48% under renewable-surplus conditions. These findings show the effectiveness of the proposed constrained bi-objective framework for day-ahead energy management of islanded hybrid microgrids.
Radiative cooling has emerged as a zero-energy and sustainable strategy for mitigating excessive energy consumption in the cooling sector. Among various radiative cooling architectures, polymer–particle composite paints have recently attracted attention for their environmental durability, material accessibility, scalability, and compatibility with conventional coating methods and diverse substrates. However, most existing studies and reviews have evaluated these materials primarily as optical systems, emphasizing solar reflectance, thermal emissivity, and cooling performance, while giving comparatively less attention to the formulation-structural architectures-durability principles that govern their practical performance. This review approaches polymer–particle radiative cooling composites fundamentally as paints, governed first by coatings-formulation science and only secondarily by optical design. Accordingly, it highlights the importance of systematically examining these paints through the lens of paint formulation, microstructure, rheology, mechanical integrity, processing, durability, and manufacturability. The review first summarizes the fundamental physical principles governing radiative heat transfer, cooling power, and the light-scattering mechanisms relevant to these paints. It then examines key formulation parameters that link formulation to performance, including pigment–binder relationships, particle characteristics, and standard figures of merit. From a materials perspective, this work examines the roles of polymeric binders (thermoplastic, thermosetting, and dual-binder systems), together with inorganic pigments/fillers, in the optical performance, mechanical durability, environmental stability, adhesion, and long-term outdoor performance. Furthermore, it introduces a structural taxonomy of packed-particle, porous-particulate, and hierarchical multi-scale architectures, systematically compares their optical performance, mechanical integrity, and manufacturing complexity, and analyzes the role of coating thickness in tuning these performance metrics. The influence of deposition and manufacturing routes on coating morphology, scalability, and performance reproducibility is also critically examined. Beyond material and structural design, the review evaluates durability assessment, outdoor testing methodologies, climate-specific formulation requirements, embodied carbon, cost, and environmental considerations. After surveying established and emerging applications, this review concludes with a technology readiness level (TRL) assessment of these paints. It identifies specific barriers to commercialization, including the need for long-term outdoor durability data, manufacturing and quality-control challenges at large scale, and insufficient life-cycle environmental assessment and cost analyses. By integrating optical physics with paint-formulation science, durability, manufacturing, and sustainability considerations, this review charts a path toward deploying polymer-based particle-composite radiative cooling paints as sustainable and commercially viable cooling solutions.
The progressive displacement of conventional synchronous generation by variable renewable energy sources reduces grid inertia and intensifies frequency stability challenges in interconnected power systems. This study presents a fractional-order model predictive control (FO-MPC) strategy for load frequency control (LFC) in a two-area interconnected power system incorporating wind generation in Area 1 and photovoltaic (PV) generation in Area 2, alongside reheat thermal generation, with practical nonlinearities including generation rate constraints (GRC) and governor dead bands (GDB). The controller introduces fractional-order integral (order λ) and fractional-order derivative (order μ) operators into the model predictive cost function, extending the degrees of freedom for frequency deviation minimisation beyond integer-order formulations. Seven design parameters — the prediction horizon, control horizon, fractional orders λ and μ, and three cost-weighting coefficients — are simultaneously optimised using the grey wolf optimiser (GWO) with the integral of time-weighted absolute error (ITAE) as the objective function. Simulation results for a 0.1 per-unit step load perturbation in Area 1 demonstrate that FO-MPC reduces the frequency deviation undershoot in Area 1 to 0.0312 Hz, a reduction of 35.1% relative to integer-order MPC (IO-MPC), 49.9% relative to GWO-optimised fractional-order proportional-integral-derivative (FOPID) control, and 65.0% relative to GWO-optimised proportional-integral-derivative (PID) control. The tie-line power deviation is simultaneously reduced to 0.0098 per unit. The settling time of 4.2 s under FO-MPC is less than half the 11.3 s obtained under FOPID control. Under stochastic renewable generation profiles and ± 25% parametric uncertainty, FO-MPC sustains consistent frequency recovery, with an ITAE increase of only 12.3% across the full range of perturbed conditions, confirming the reliability of the proposed scheme.
This paper proposes a new efficient optimization framework for siting, sizing and power factor optimization of photovoltaic (PV) and wind turbine (WT)-based distributed generations (DGs) in the presence of capacitor banks. In the proposed framework, the weighted sum of energy loss and voltage deviation index is defined as the objective function and minimized subject to several technical constraints. To solve the optimization problem, two variants of crow search algorithm (CSA), three variants of exploration–exploitation (E2) and two variants of particle swarm optimization (PSO) are employed and the results are compared. For the first case study (IEEE 69-bus distribution network), simulation results demonstrate that differential CSA (CSAd) achieves more accurate and robust results than other investigated methods. In addition to the importance of power factor optimization, it is observed that in the absence of capacitor banks, two WTs (with power factors of 0.81 and 0.85) are allocated to the network and no PV installation is suggested. In the presence of capacitor banks, two WTs with unity power factors and a relatively small PV with a power factor of 0.81 are suggested for installation. The proposed framework is also evaluated on a real distribution network and the results demonstrate its effectiveness.
Effective heat removal from photovoltaic cells to the working fluid remains still a techno-economic challenge in photovoltaic/thermal (PV/T) systems, which directly affects both electrical and thermal performance. In this work, internal wire-coil inserts within the fluid tubes are investigated as a passive heat-transfer enhancement technique in a PV/T system. The presence of wire coils promotes flow mixing, decreases the thickness of the thermal boundary layer, and enhances heat transport. Beyond technical performance, this study also provides a comprehensive evaluation of the economic and environmental implications of such enhancement techniques. A numerical model of a multi-layer PV/T configuration equipped with a flat-plate collector was developed using ANSYS Fluent. Three configurations, i.e., a plain PV, a conventional PV/T, and a wire-coil-enhanced PV/T (PV/T + WC), were simulated under the climatic conditions of Karaj, Iran, based on ten-year average meteorological data. The results indicate that the integration of a thermal collector improves electrical efficiency by a relative 2.9 %, while the addition of wire coils further enhances it to 4.9 % compared to the plain PV module. It is also demonstrated that using wire coils increases the annual average thermal efficiency from 32 % to 35.7 %. Although reconstruction of a PV system to a PV/T counterpart may increase the electrical levelized cost of energy (LCOE) from 0.028 to 0.042 USD/kWh, the LCOE of overall energy (i.e., heat and electricity) of the PV/T and PV/T + WC systems is 0.0165 and 0.0155 USD/kWh, respectively, which is much lower than that of the standalone PV system. In addition, environmental assessment based on net CO2 mitigation reveals that the incorporation of a thermal collector increases emission reduction by 118.4 %, while the addition of wire coils provides a further 7 % improvement. Overall, the findings demonstrate that integrating wire-coil inserts within PV/T collectors offers a cost-effective and practical approach for enhancing heat transfer, improving energy efficiency, and reducing environmental impact.
Recently, sustainable energy sources have attracted significant attention from researchers, developers, companies, and investors. One of these sources is solar energy, particularly concentrated solar power (CSP) which is integrated with other conventional sources of energy. Small-scale solar Brayton turbines suffer from severe inlet temperature fluctuations due to varying solar radiation, demanding fast-adaptive control systems for higher station performance. This paper presents a Reinforcement Learning (RL) agent capable of controlling Fuel-to-Air Ratio (FAR) and mass flow rate (ṁ) to maximize electrical power output. An accurate surrogate model was built using Random Forest based on 80 operating points from ANSYS VISTA, achieving exceptional accuracy (R2 = 0.999). The agent was trained using the Proximal Policy Optimization (PPO) algorithm for 30,000 time steps. Results show that the agent automatically learned to fix FAR at its minimum value (0.02) and raise ṁ to its maximum (0.07 kg/s). Sensitivity analysis revealed that changing FAR from 0.02 to 0.08 changes power by less than 1.11 %, and feature importance ranking confirmed that FAR has a negligible impact (importance < 0.1 %) on power output within the VISTA model and the analysed operating range. Compared to a well-tuned PID controller, the RL agent achieves a 36.4 % reduction in fuel consumption and 13.5 % improvement in the turbine’s output power. Moreover, it outperformed random and fixed policies by 18 % and 12 % in power improvement and 46 % and 49 % in fuel savings, respectively. These findings open the door to radically simplifying control systems for small-scale solar Brayton turbines and achieving higher performance for the power station system.
This research presents a new approach to enhance the efficiency of solar thermal systems using a double-pipe heat exchanger with a parabolic solar reflector with multiple thermal enhancement techniques. The phase change material (PCM) is paraffin wax (RT35-HC) that fills the inner tube of the heat exchanger. To address the inherently low thermal conductivity of paraffin, single-walled carbon nanotubes (SWCNTs) are dispersed within the PCM, while cross-shaped fins are strategically embedded to further enhance the melting and solidification processes. The outer tube is filled with water and improved by adding copper porous foam to develop the heat transfer significantly through combined convection and conduction. A thermoelectric generator (TEG) module is also mounted on the outer surface of the water tube to convert the waste thermal energy to electricity. SolTrace is utilized to accurately simulate the solar flux distribution from the parabolic reflector, and ANSYS FLUENT is employed to perform a comprehensive unsteady three-dimensional simulation to evaluate the thermal behavior and energy conversion efficiency of the system for various designs. Results indicate that the addition of SWCNT nanoparticles and cross-shaped fins in the PCM zone boosts the thermal efficiency, the stored energy, and the liquid fraction significantly by about 53.32%, 63.12%, and 74.21%, respectively. Using porous foam in the water zone only, without fins, causes considerable enhancements of 2.64% in the overall efficiency, 24.74% in the stored energy, and 28.66% in the liquid fraction. When all techniques are combined, the system shows significant enhancements with the overall efficiency, energy storage capacity and phase change progress being improved by factors of 1.67, 2.09 and 2.74 respectively, compared to the baseline case. The system behavior in the solidification step, without solar input, was also evaluated. It was observed that the use of porous foam reduces the freezing time by 16.17% with respect to fin-only configurations while the fastest solidification (3420 s) is obtained when all enhancement features are applied. The results highlight the effectiveness and novelty of combining multiple enhancement techniques in a solar PCM-TEG hybrid system that has the potential to enable more efficient, compact and multifunctional solar energy storage technologies.
Atmospheric CO2 concentrations now exceeding 420 ppm are accelerating global warming, prompting urgent interest in solar-driven CO2 conversion as a carbon–neutral fuel production strategy. Among emerging photocatalyst platforms, Metal-Organic Frameworks (MOFs) offer exceptional tunability through their designable organic linkers, metal ions and porous structure. Yet the complex, nonlinear relationships between their structural properties and catalytic performance remain poorly understood. Herein, we present a machine learning framework trained on 417 experimental data points compiled from the MOF-based photocatalytic CO2 reduction literature. The dataset was randomly split into training (80%) and testing (20%) to predict and interpret the formation yields of four target products: CH3OH, HCOOH, CO, and CH4. Six tree-based ensemble models namely CatBoost, XGBoost, LightGBM, GBDT, Random Forest, and AdaBoost, were systematically compared across six performance metrics. SMOGN Augmentation was applied only to the training set. CatBoost model achieved a training R2 of 0.988 with an RMSE of 50.21, while maintaining strong generalization on the independent test set with a test R2 of 0.966 and an RMSE of 83.59. Furthermore, validation using an independent dataset also confirmed the excellent performance of the CatBoost model, yielding an R2 of 0.996 with an RMSE of 30.15 for the training set and an R2 of 0.976 with an RMSE of 70.07 for the test set. CatBoost was selected for downstream interpretation using SHAP (SHapley Additive exPlanations) analysis. SHAP enabled the quantitative identification of the catalyst properties and reaction conditions that are strongly associated with product formation. These findings establish an interpretable, data-driven roadmap for the rational design of MOF photocatalysts, enabling targeted optimization of both activity and product selectivity for sustainable solar fuel production.
The continuous growth in global energy demand, together with environmental pollution caused by fossil fuel consumption and concerns over resource depletion, has significantly increased interest in renewable energy sources, particularly biomass-based bioenergy. This study develops a novel multi-period resilient biomass supply chain network design model. A closed-loop network structure is incorporated, in which recoverable water and generated nutrients during the feedstock-to-bioenergy conversion processes are assumed to be reused to support Jatropha curcas L. cultivation. The proposed supply chain network is assumed to be exposed to disruptions affecting facility capacities and transportation links. To enhance system resilience, several resilience strategies are embedded in the network design, including multiple transportation modes, facility dispersion, multiple sourcing, facility fortification, and capacity expansion. A two-stage optimization framework is proposed. In the first stage, candidate biorefinery locations are screened using geographic information system-based spatial analysis. In the second stage, a multi-period mixed-integer programming model is developed to optimize the supply chain configuration. To capture uncertainties associated with demand, feedstock yield, and the impacts of disruptions on facility capacity and transportation availability, a multi-stage scenario-based stochastic programming approach is employed. This framework enables adaptive decision-making over time under uncertainty. A real-world case study is conducted to evaluate the applicability and performance of the proposed model. Computational results indicate that the multi-stage stochastic programming approach provides improved performance and greater adaptability compared to the conventional two-stage stochastic programming method. Furthermore, numerical analyses confirm the significant contribution of resilience strategies and the closed-loop network structure to enhancing supply chain profitability.
Static reflectors below a fixed photovoltaic module raise its output at almost no cost, yet their benefit is usually reported as one percentage from a short campaign and scaled to the year, with the reflector treated as a uniform boost. This work builds a first-principles optical model of a booster reflector, validates it against a controlled field sweep, and derives design guidance transferable beyond the test site. The campaign, at a Mediterranean site, swept reflector tilt, length and material, pairing every reading with a concurrent bare-panel reference on the same meter. A height-resolved ray trace with a diffuse view-factor term predicts the irradiance added along the module. A five-parameter electrical model of the half-cut cell topology then converts that profile into power, so series-string mismatch is computed per configuration rather than assumed. The model reproduces the measured mirror response across four tilts and two lengths to 0.84 percentage points, at a geometric transfer factor of 0.81. Mismatch proves configuration dependent, holding the power gain to between 0.6 and 0.9 of the average-irradiance gain rather than to one fixed fraction. A rigid glass mirror follows flat-reflector optics; a thin aluminum sheet does not, its tilt-rising gain being consistent with mild concentration from surface non-flatness. Measured gains reach 4.7 to 10.4 percent in autumn , and the model concentrates them in the months when the Lebanese grid is most stressed, giving an annual 2.9 to 5.1 percent. Optimum reflector tilt follows 62 degrees minus site latitude.
