FPV systems are expected to outperform ground-mounted PV (GPV) through water cooling, unconfirmed for closed-platform FPV in tropical reservoirs, where water temperature can exceed ambient air temperature. Platform type was isolated using a co-located, identical-hardware GPV reference, monitored for 12 months at hourly resolution in a Colombian hydroelectric reservoir. The GPV system generated 1407 kWh/kWp per year, compared to 1382 kWh/kWp for the FPV system, representing a shortfall of 25 kWh/kWp (1.75%), of which 12.4% is due to resistive losses in the FPV system’s longer wiring and the remaining 87.6% to thermal loss. The FPV modules operated at higher temperatures than the GPV modules during 70.5% of the monitored hours, at a mean differential of +1.17 °C. Water beneath the platform averaged 30.98 °C, above ambient throughout. Wind speed showed no cooling effect on the temperature difference between the FPV and GPV modules (Spearman: 0.090); airflow beneath the floating platform was not measured directly, and it is possible that the irregular bathymetry (depth of 6 to 38 m) limited renewal of the surface layer. Closed-platform FPV over warm water can act as a net heat source. The 25 kWh/kWp gap equals 567 USD/year (171 kWp) or 82,725 USD/25 years (1 MW), useful for prefeasibility assessments.
The reliability of offshore hybrid renewable energy systems (HRES) is critical due to seasonal environmental variability, and mechanical degradation. In this paper, a HRES that integrates wind, wave and solar power into an offshore platform is considered. The system is modeled using a Cyclic Non-Homogeneous Markov Chain (Cyclic NHMC), which unlike Homogeneous Continuous Time Markov Chains (HCTMC) with constant transition rates, the Cyclic NHMC incorporates season-dependent degradation, failure, maintenance, and environmental transitions. Building on this model, a Seasonal Energy Demand-Based Maintenance (SEDBM) strategy is introduced to allocate preventive maintenance according to seasonal demand criticality and environmental stress. The framework is validated through an experimental case study using in situ Aegean Sea environmental data, together with technical system data and expert-supported degradation and maintenance rates. System performance is assessed using Expected Energy Produced (EEP), Loss of Load Probability (LOLP), and Expected Energy Not Supplied (EENS). The results show that the Cyclic NHMC provides season-specific adequacy insights that are not captured by a homogeneous benchmark. Under SEDBM, EEP increases across all seasons, while LOLP and EENS are reduced during the critical seasonal periods. A modest increase in load loss in spring is observed, reflecting the deliberate reallocation of maintenance effort.
Fleets of light-duty vehicles with heterogeneous powertrain systems are increasingly common. While fleet electrification can reduce emissions, operational decisions such as vehicle-to-trip assignment remain a critical lever for CO2 mitigation. In mixed fleets, energy consumption and emissions are highly temperature-dependent, and operation under non-standard ambient conditions requires additional attention. This work presents an emissions-oriented vehicle-to-trip assignment planning framework demonstrated using a real-world mixed 15-vehicle fleet comprising two hybrid and one plug-in hybrid electric vehicles across multiple fleet applications. Application-specific representative drive cycles were developed using operational data collected from 10 fleet vehicles, while vehicle-specific tailpipe CO2 emission-rate models were trained for 15 fleet vehicles using a machine learning approach. For electrified powertrains, the effect of cold ambient temperature is considered in emissions. Using the resulting emission estimates, an optimization is implemented to identify minimum-emission assignments for daily schedules under eligibility, time-overlap, and battery recharging constraints. Across trip sequences examined, the minimum-CO2 assignment reduced average CO2 emissions by 6.9% at 20 °C, 11.0% at 0 °C, and 8.6% at −20 °C relative to the baseline greedy strategy, which prioritized the most fuel-efficient available vehicles. The results highlight the importance of integrating realistic drive cycles, data-driven emission modeling, and operational constraints to support low-carbon fleet assignment planning.
This paper investigates the scenario-based resilient tracking control problem for networked hybrid cascaded energy systems in constrained industrial networks. First, a tracking controller and a transmission mechanism are jointly designed to construct a networked hybrid cascade model that captures essential communication constraints and scheduling protocols. Then, by employing a hybrid Lyapunov framework, relaxed sufficient conditions are derived to guarantee exponential stability even under extreme network congestion scenarios. In addition, a benchmark example is used to quantitatively evaluate the conservatism of the proposed stability conditions, which shows that the maximum allowable transmission interval (MATI) is improved by up to 155.0% compared with existing results. Finally, the proposed approach is validated through scenario-based simulations on a pump-driven hybrid energy system, and the results confirm its effectiveness and operational resilience.
Accurate short-term forecasting of solar power and energy consumption is important for reliable grid operation, especially in systems with increasing renewable energy penetration. This paper proposes an entropy-guided CEEMDAN-VMD-CNN-BiLSTM framework for univariate short-term forecasting. The method first decomposes the original signal using CEEMDAN, then applies spectral-entropy-guided VMD refinement to selected components before using a CNN-BiLSTM network for prediction. To isolate the effect of decomposition, the proposed method is compared with SVD-VMD, EWT-VMD, EMD-VMD, and VMD-VMD under the same forecasting architecture. Experiments were conducted using chronological 70/15/15 train/validation/test splits on solar-power and energy-consumption datasets. The proposed CEEMDAN-VMD configuration achieved the lowest errors, with mMAPE = 0.6489, RMSE = 19.2061, and MAE = 9.2741 for solar power, and mMAPE = 0.2502, RMSE = 8.6063, and MAE = 6.7252 for energy consumption. Repeated-seed experiments and Friedman/Nemenyi statistical tests were used to assess robustness and rank differences among models. Computational analysis further showed that online inference is fast after model loading, with warmed inference latency remaining suitable for 5-minute solar-power and 15-minute energy-consumption forecasting. The results show that entropy-guided CEEMDAN-VMD improves forecasting accuracy while maintaining practical feasibility for short-term energy forecasting applications.
To achieve the carbon neutrality target, China is actively deploying large-scale photovoltaic projects with Northwest China serving as a key site for photovoltaic base construction. Focusing on Northwest China, this study evaluates ecological responses to photovoltaic deployment by combining multi-indicator ecological assessment, the Remote Sensing Ecological Index (RSEI), Propensity Score Matching (PSM) and Synthetic Difference-in-Differences (SDID). The results show that photovoltaic installation was generally associated with increases in vegetation cover and soil moisture. Grassland shows the most evident vegetation response, with Fractional Vegetation Cover (FVC) increasing by 0.048, whereas cropland exhibited the strongest moisture-retention response, with shallow soil moisture increasing by 0.0052 m3/m3. The estimated ecological responses associated with photovoltaic power installation shifted from negative to positive, both within and outside core ecological zones, suggesting that the relationship between photovoltaic deployment and ecological conditions became more positive over time. Under the existing distribution of photovoltaic power station sites, photovoltaic suitability and ecological conservation in Northwest China showed a generally coordinated spatial pattern. By addressing spatial conflicts between site selection and ecological conservation, this study examines the ecological responses of photovoltaic power deployment across macro spatial and temporal scales, which holds significant implications for achieving renewable energy transition towards carbon neutrality.
High-power solid-state laser systems suffer from excessive heating and thermal stress due to non-uniform heat generation induced by pump radiation. Efficient thermal management is therefore essential to sustain the optical performance and reliability of these devices. In this work, a geothermal heat pump (GHP) is proposed as a sustainable cooling solution to regulate the temperature of the laser rod and mitigate thermal stresses. A comprehensive energy, exergy, economic, and environmental (4E) analysis is conducted to evaluate the performance of the proposed system. A numerical model is developed and validated to investigate the conjugate heat transfer and thermal behavior of an end-pumped laser rod subjected to non-uniform heat generation. The results of the exergy analysis indicate that the GHP operates with low irreversibilities, achieving a high coefficient of performance (COP) of 3.984. The results further show that the proposed cooling system is capable to reduce the maximum temperature of the laser rod by 7.66% and decreases the maximum thermal gradient by 58.38% as the Reynolds number increases from 3031 to 121261. Additionally, the results showed that the GHP system offers significant operational and environmental advantages over the conventional system while simultaneously providing an additional heating capacity of 3.42 kW. Furthermore, the GHP system achieves a total reduction of 1495 kg of CO2 emissions, resulting in an environmental cost saving of USD 135 compared with the conventional system. Moreover, a parametric study is conducted to evaluate the sensitivity of the thermal response to key operating parameters, including flow rate, inlet temperature, heat generation rate, and rod length. Finally, the proposed system can be considered a promising and sustainable cooling solution due to its economic and environmental advantages.
The decarbonization of the industrial sector is essential to achieving global climate goals, given its significant contribution to greenhouse gas emissions. Industrial waste heat (IWH) recovery represents a promising strategy to enhance energy efficiency, reduce fuel dependency, and mitigate CO2 emissions. This study presents a comprehensive assessment of Morocco’s IWH recovery potential using three complementary approaches based on CO2 emissions, fuel energy consumption, and sectoral energy intensity. The analysis reveals that the recoverable IWH potential ranges between 7.26 and 26.96 PJ annually, with the non-metallic minerals, food and beverages, and chemical industries identified as the dominant contributors. Most recoverable heat lies within the 100–200 °C range, suitable for medium-temperature applications. To valorize this energy, a thermochemical energy storage (TcES) approach is proposed using phosphogypsum (PG) as a low-cost and abundant material. The CaSO4,0.5H2O/CaSO4,2H2O reaction system demonstrates favorable energy density (175 kJ/kg) and operational compatibility with industrial waste heat streams. Approximately 100 ktons of PG would be required to store the daily recoverable heat from Moroccan industries within this temperature range. This dual-benefit strategy not only enhances industrial energy efficiency but also promotes circular economy principles by converting an industrial by-product into a valuable thermochemical storage medium.
This paper presents a novel hybrid modeling framework that synergistically combines physics-based and machine learning models to achieve accurate surrogate prediction of hybrid solar–wind power generation. The methodology employs high-resolution meteorological data collected during 2024 in Salalah, Oman, incorporating enhanced physics-based models with temperature-dependent efficiency corrections for photovoltaic (PV) systems and realistic power curve characteristics for wind turbines. A rigorous time-series validation strategy ensures realistic performance evaluation under operational forecasting conditions, preventing data leakage and temporal dependencies. The Random Forest model serves as a computationally efficient surrogate of the physics-based model, learning to reproduce physics-based hybrid power outputs from meteorological inputs. On the held-out test set, the surrogate achieves R2 = 0.9951, RMSE = 0.4304 kW, and MAE = 0.0884 kW, substantially exceeding the standalone physics-based model (R2 = 0.95). These metrics quantify the fidelity of the machine learning surrogate in reproducing deterministic physics-based power outputs rather than forecasting accuracy against field-measured PV and wind generation data. Seasonal analysis across five distinct seasons reveals unique generation patterns, with the Khareef season exhibiting the lowest power generation due to reduced solar irradiance from extensive cloud cover. A quantitative framework for hybrid system sizing and load assessment provides practical tools for system design and optimization.
In this study, 3D-printed polymer microreactors were employed for real-time monitoring of biodiesel synthesis via transesterification of sunflower oil with methanol using potassium hydroxide (KOH) as a homogeneous catalyst. Four microreactor configurations, including straight channel and channels with 4, 8, and 12 bends, were fabricated from polyethylene terephthalate glycol (PETG) filament using fused filament fabrication (FFF). An integrated Fourier-transform infrared (FTIR) probe enabled continuous in situ monitoring of reaction conversion, eliminating the need for sampling. Calibration with batch-synthesized biodiesel solutions allowed quantitative determination of conversion, revealing key spectral bands at 1099 cm−1, 1196 cm−1, and 1436 cm−1. The 12-bend microreactor achieved the highest maximum conversion (64%), while maximum conversions of 57 % and 41 % were observed for the straight and eight-bend microreactors, respectively. The integrated FTIR probe demonstrated satisfactory linearity and enabled real-time tracking of biodiesel conversion. These results suggest that channel configuration alone does not fully determine biodiesel conversion under the investigated operating conditions and that factors such as flow stability and phase distribution may also influence reactor performance. Polymer microreactors with integrated FTIR probes offer a cost-effective, non-destructive platform for studying homogeneous catalytic reactions in microfluidics.
This paper presents a measurement-driven characterization of the outdoor ambient RF environment in Dubai to support deployment planning for battery-free IoT systems and RF energy harvesting. These findings provide a quantitative basis for band selection, placement prioritization, and feasibility assessment, and can establish more realistic simulation environments. A large-scale drive-test survey was conducted using a Keysight N9914C FieldFox spectrum analyzer and a broadband omnidirectional antenna, logging GPS-tagged received power (dBm) at 50–60 m intervals over almost 200 km of routes. Eleven bands spanning sub-1 GHz broadcast and cellular frequencies through 5 GHz Wi-Fi were measured. The results show that ambient RF availability is dominated by sub-1 GHz frequency windows, with the 791–862 MHz measurement window exhibiting the strongest received power results, reaching a maximum of −5.57 dBm and exhibiting power above −20 dBm for 4.30% of samples, and above −30 dBm for 38.36%. In contrast, higher-frequency bands show negligible exceedance at these levels, indicating limited availability of high-input-energy events. Spatial route maps reveal road-level variability with localized high-power segments along the sampled routes. The findings indicate that battery-free sensing is most feasible under a harvest-store-operate paradigm and establish a quantitative framework for band prioritization, measurement-informed placement, and realistic energy-availability assessment.
Transport decarbonisation is critical to achieving net-zero mobility, yet mitigation pathways vary globally. This study presents a regionalised life-cycle assessment using the Greenhouse gases, Regulated Emissions, and Energy use in Technologies (GREET) model to investigate how contrasting energy systems influence environmental performance across Southern Brazil, the United Kingdom (UK), and the United States of America (USA) Midwest over a 200,000 km vehicle lifetime. Results show that decarbonisation pathways cannot be universally prioritised and must align with regional grid characteristics. Compared with gasoline internal combustion engine vehicles, battery electric vehicles reduce greenhouse gas emissions by 47–70%, depending on the energy system. However, ethanol-fuelled vehicles in Brazil achieve emissions comparable to electric vehicles operating in fossil-intensive grids, highlighting low-carbon biofuels as a strong regional alternative. The proposed framework provides vital decision-support insights for designing context-specific transport policies based on regional fuel production.
The Hybrid Sulphur (HyS) cycle is attracting interest for decarbonising the sulphuric acid industry by integrating concentrating solar thermal (CST) technologies with the high-temperature acid splitting process. Several energy and cost studies have been performed and significant European Union (EU)-funded research projects have been implemented so far to bridge the gap towards the commercialisation of the HyS concept. However, even the concept development at the demonstration plant scale has not yet been achieved. Within the EU-HySelect project, this study presents the first demonstration-scale techno-economic assessment of the HyS plant’s solar thermal and thermochemical subsystems, utilising a detailed component-level costing approach to advance the Technology Readiness Level. The specific energy of solar SO2 production (at purity of 79.4 wt%, compressed at 1.3 MPa) and the total solar demonstration plant costs are calculated based on Aspen Plus at 15.62 MJ/kg (or 4.34 kWh/kg) and 3.4–6.1 million €2025, respectively. Unlike idealised models that ignore industrial constraints, this study aligns with practical standards to realistically assess implementation potential. It identifies critical bottlenecks in CST technology and the sulphuric acid splitting reactor, highlighting key pathways for scale-up. The reported cost estimates further establish a concrete baseline for future HyS research and industrial decarbonisation strategies.
Traditional point-forecasting model predictive control in wind-solar hydrogen systems suffers from substantial tracking misalignments, inducing frequent start-stop cycles and accelerated equipment degradation. However, a single deterministic forecast is inherently unable to represent the stochastic distribution of renewable outputs, limiting its effectiveness under high-variability conditions. To address this, this paper introduces a probabilistic forecasting framework that generates statistically reliable confidence intervals, explicitly quantifying renewable generation uncertainty. Next, a refined modeling approach for the alkaline electrolyzer array is established by explicitly incorporating non-linear production efficiency, dynamic operational boundaries, and differentiated degradation costs for cold and hot starts. Finally, an online rolling stochastic model predictive control strategy is formulated to dynamically balance green hydrogen yield and equipment lifespans by converting the empirical interval boundaries into adaptive chance constraints. Case studies demonstrate that the proposed stochastic strategy reduces average total start-stop operations by 5 and mitigates stack life degradation costs by up to 7,200 CNY, with critical cold-start cycles decreasing by an average of 3 per day. Furthermore, daily green hydrogen production increases by up to 141.5 kg, which lowers the average levelized cost of hydrogen significantly by 30.2 percent while increasing the average production revenue by 8.85 percent.
Agricultural parks face significant carbon emissions alongside insufficient local renewable energy accommodation capabilities. This paper develops a dispatch methodology for achieving low-carbon economic operation in an integrated energy system (IES) that combines biomass electrolysis for hydrogen production with heat-electricity-CO2 coupled greenhouse loads. Firstly, a biomass-assisted electrolyzer model grounded in electrochemical principles is incorporated to characterize the coupling among electrical input, biomass oxidation, thermal balance, and hydrogen production in the agricultural-park IES. Secondly, a greenhouse load model considering heat, electricity, and CO2 demands is incorporated to characterize the multi-energy coupling relationships and explore their collaborative potential for carbon emission reduction. Subsequently, a deterministic optimization scheduling model incorporating economic and environmental considerations is established. Finally, a distributionally robust optimization (DRO) method based on Kullback–Leibler (KL) divergence is adopted to address uncertainties in renewable energy generation and environmental parameters. Comparative analyses with deterministic optimization, chance-constrained stochastic optimization, and conventional box-set robust optimization demonstrate that the KL-divergence-based DRO method alleviates the excessive conservatism of worst-case robust optimization while retaining protection against distributional uncertainty. This method provides theoretical support for the low-carbon transformation of energy structures in agricultural parks.
