Search Results (1,794)

High spatial resolution wind speed data is very important for wind farm planning, design, operation and maintenance. But due to cost, site and other factors, it is impossible to build a large number of anemometer towers to obtain high spatial resolution measured data. Therefore, this paper proposes a method for generating wind speed data in renewable energy bases based on physics-informed neural networks, which incorporates fluid mechanics control equations such as the Navier−Stokes equation as physical constraints into the model training process. The model’s input includes the wind speed data and the wind direction data of the anemometer towers as input, as well as the geographical difference data between the input anemometer towers and the output point, enabling to learn the mapping relationship between geographical differences and wind speed differences at different locations, achieving the goal of generating high spatial resolution wind speed data. Using normalized root mean absolute error (NMAE) to measure the model error, the average wind speed error and the average wind direction error of the proposed wind speed data generation method on different test sets are 8.28% and 10.50%, which is lower than that of BP neural network and graph convolutional neural network, and can provide more refined data support for wind turbine layout planning and wind farm power prediction of renewable energy bases. Full article

Consistent laser cutting quality is one of the problems associated with the nonlinearity of relationships between process parameters and output responses. This problem acquires particular importance when it comes to cutting advanced nanocomposites, which requires precise tuning. Despite the wide adoption of intelligent modelling, few studies have investigated the comparative efficiency of various approaches based on the use of the same dataset. In this research, the effectiveness of three models—Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Fuzzy Logic System (FLS)—was tested on experimental data related to the CO₂ laser cutting of ABS/CNT nanocomposites. Input parameters included laser power and cutting speed, whereas HAZ width, kerf width, and surface electrical resistivity were used as output data. Data was split into training, testing, and validation datasets; models were created using supervised machine learning. Model performance was evaluated using Root Mean Square Error (RMSE). Analysis of results showed that ANN demonstrated acceptable predictive capabilities, yielding correlation coefficients (R) close to 1 (≈0.99) and RMSE values of 0.2956 for HAZ, 0.2061 for kerf width, and 2.3655 for surface electrical resistivity. Prediction by means of FLS was able to identify general tendencies; however, it produced RMSE values of 0.4741 for HAZ, 0.6297 for kerf width, and 1.9258 for surface electrical resistivity. Finally, the ANFIS model proved to be the most reliable model, yielding the lowest RMSE values for HAZ (0.2784), kerf width (0.0450), and surface electrical resistivity (0.0905). In conclusion, this research shows that ANFIS can be used effectively for building models predicting laser cutting processes; therefore, it represents an approach worth using in future investigations in this field. Full article

Solar power plants are among the most widely used renewable energy sources today. Varying radiation levels from region to region, and similarly varying consumption depending on the user within a given region, make the optimal sizing of these plants challenging. In this study, a machine learning-based surrogate model for the real-time sizing optimization of solar power plants, trained with a completely original dataset, has been developed. In the first stage, 500 different solar power plant installation scenarios were synthetically generated and evaluated in HOMER, and the obtained optimal sizing outputs were used as training targets for the proposed surrogate model rather than real operational data. The results obtained by applying various machine learning methods to the generated dataset are presented comparatively. Among 7 different machine learning models, XGBoost, Gradient Boosting, and LightGBM demonstrated the best performance. The developed model achieved an average R² score of 0.9425 for a total of 3 targets, while target-specific performance showed R² scores of 0.9747 for inverters, 0.9365 for PV panels, and 0.9165 for batteries. This model serves as a computationally efficient surrogate of the HOMER optimization process, enabling high-accuracy real-time predictions while significantly reducing the computational burden associated with intensive mathematical calculations, iterative procedures, and complex search spaces. Full article

To address the insufficient accuracy of short-term load forecasting caused by the strong randomness of distributed energy output, variable electricity consumption patterns, and complex meteorological factors, this study proposes a load forecasting method that integrates K-means scene clustering and a Transformer–BiGRU–Attention (CTBA) hybrid deep learning architecture. Different from conventional Transformer–BiGRU hybrid forecasters that train a single global predictor across all operating conditions, the proposed CTBA framework first partitions daily load curves into representative scenes and then routes each sample to a scene-specific Transformer–BiGRU–Attention predictor, thereby reducing distributional heterogeneity before temporal modeling. First, the K-means algorithm is used to perform scene clustering on historical daily load curves, and the optimal number of clusters is selected according to the silhouette coefficient and downstream prediction performance. Subsequently, the CTBA model is trained separately for each clustering subset. The Transformer encoder captures the long-range global dependencies of load sequences through the self-attention mechanism, the BiGRU module extracts local bidirectional temporal fluctuation features, and the Attention mechanism further focuses on key time nodes such as morning and evening peaks while fusing multi-source data including historical load, day-ahead electricity price, and multi-dimensional meteorological factors. Experimental results based on the German ENTSO-E power dataset show that the coefficient of determination R² of the proposed model reaches 0.9893, with MAE, RMSE, and MAPE as low as 0.0141, 0.0187, and 3.92%, respectively, which are significantly improved compared to benchmark models such as SVR, LSTM, CNN, and TCN-BiGRU. Ablation experiments further demonstrate that removing the clustering, Transformer, BiGRU, or attention layer will degrade performance, thus verifying the effectiveness and superiority of the method in short-term load forecasting and providing an accurate solution for the short-term load forecasting of power systems. Full article

AI systems are entering criminal proceedings as evidence producers, risk assessors, and decision shapers, yet the procedural architecture of adversarial and mixed systems was built on the assumption that evidence originates from human actors whose reasoning can be reconstructed and challenged. This article introduces the concept of evidentiary authority—the power to determine what counts as reliable evidence and how much weight it carries—and argues that this authority is migrating from human decision-makers to algorithmic systems without adequate procedural safeguards. The article draws on forensic linguistics and comparative criminal procedure to examine two domains where this migration is most visible: generative AI, which can fabricate or manipulate the texts on which forensic authorship analysis depends, and predictive AI, which feeds opaque risk scores into judicial decisions at stages where adversarial scrutiny is weakest. A structural phenomenon, which the article terms the “inferential catalyst”, is identified: AI outputs that shape proceedings without entering the formal evidence record. These two domains are tested against seven principles of criminal procedure: free evaluation of evidence, immediacy, judicial independence, the right to a reasoned decision, adversarial proceedings, the right of confrontation, and the presumption of innocence. At each principle, the same structural problem recurs: the system presupposes human reasoning that AI outputs cannot provide and that existing procedural mechanisms cannot compel. Six safeguards are proposed as conditions for admissibility: algorithmic transparency, independent auditing, defence access to algorithmic expertise, admissibility standards for algorithmic evidence, enhanced justification obligations, and capacity building. Full article

Industrial waste heat recovery is an important approach for improving energy utilization efficiency and reducing environmental impacts. Thermoelectric devices can directly convert waste heat into electricity, but their practical application is limited by relatively low output power. Active water cooling can enhance the power generation performance of thermoelectric devices, but the pumping power may reduce the net output power. In this study, a water-cooling thermoelectric device is investigated under constant heat input conditions using three-dimensional numerical simulations and a semi-analytical prediction model. The effects of cooling water inlet temperature and flow rate on the thermal response, electrical output, heat transfer behavior, and net output power are systematically analyzed. The results show that increasing the cooling water flow rate increases the gross electrical power but also increases pumping power, resulting in an optimal flow rate of approximately 3 m/s to maximize the net output power. At inlet temperatures of 24 °C, 28 °C, and 32 °C, the maximum net output powers are 51.46 W, 49.89 W, and 48.68 W, respectively. A prediction model for cooling water input conditions is further developed based on energy balance and convective heat transfer correlations, and the predicted velocities agree with the numerical results with relative errors below 2%. Full article

Accurate, uncertainty-aware estimation of instantaneous wind turbine output is a prerequisite for integrating onshore assets into low-emission energy systems, where operational monitoring, energy-performance verification, and cooperative asset management depend on auditable digital representations of turbine behaviour. This study develops a Digital Shadow-based power-curve modelling framework on fourteen years of Supervisory Control and Data Acquisition records from an operational Vestas V52 onshore turbine (850 kW, Dundalk Institute of Technology, Ireland; 457,429 ten-minute records spanning 2006–2020) and benchmarks seven methods under identical preprocessing on a strict chronological hold-out (training 2006–2017; testing 2018–2020; n = 52,388). A parallel random 75/25 split is reported only as a within-distribution diagnostic; it quantifies an optimistic R² inflation of 0.003–0.027 depending on architecture. The Artificial Neural Network attains the best chronological performance (R² = 0.9924, BCa 95% confidence interval 0.9910–0.9931, RMSE = 19.79 kW); only the ANN and a one-dimensional Convolutional Neural Network with twenty-four-step wind-speed lags (R² = 0.9921) deliver clear positive skill against the IEC-style manufacturer power curve. Split-conformal calibration of a Quantile Regression Forest raises empirical 90% prediction-interval coverage from 0.534 to 0.904 at a width inflation from 30 to 51 kW. The framework qualifies as a Digital Shadow and is positioned, through a Horizon Europe Technology Readiness Level audit and an explicit mapping to ISO 50001:2018 Plan–Do–Check–Act energy management and Renewable Energy Community governance under Directive (EU) 2018/2001, as an auditable monitoring layer for cooperative onshore wind operations. The empirical evidence base is a single turbine; multi-turbine, multi-site replication is the natural follow-on validation. Full article

This study presents a computational and experimental aerodynamic characterisation of a full-scale 5.5 m diameter, six-sail horizontal-axis windmill of the traditional Cretan Lasithi type, equipped with flexible woven polyester sails that act as a passive load-control mechanism. Seventeen operating points spanning wind speeds of 2.3–18.3 m/s were simulated in OpenFOAM using a transient sliding-mesh Arbitrary Mesh Interface formulation with the k–ω SST turbulence closure on a 2.3 million cell grid, selected on the basis of a four-level grid convergence study. CFD simulations identify three distinct aerodynamic regimes: a drag-dominated high-TSR regime (λ > 2.1), a mixed lift–drag working range with peak loading near λ ≈ 1.4–1.5, and a deep-stall regime in which boundary-layer separation propagates from root to tip as λ falls below 1.0. Field measurements conducted at the Energy Systems Synthesis Lab of the Hellenic Mediterranean University in compliance with IEC 61400-12-1:2005(E) confirm that rotor speed stabilises passively at 55–58 RPM above 13 m/s without any active control mechanism; CFD predictions agree with measured power output within 8–12% across the 2–13 m/s attached-flow envelope. The combined evidence indicates that passive overspeed self-regulation is driven by aeroelastic sail deformation, reducing effective disc solidity at high wind speeds, a mechanism that rigid-geometry CFD correctly identifies in trend but cannot quantify in magnitude. The primary limitation of the present work is the rigid-sail assumption of the CFD model, which requires a two-way coupled fluid–structure interaction extension as a future step. Full article

The rapid transition toward renewable energy requires transparent and spatially explicit methods for identifying suitable photovoltaic (PV) development areas. This study develops a geospatial artificial intelligence (GeoAI) decision-support framework for PV site suitability assessment in Rajasthan, India. Eleven harmonized predictors were used: global horizontal irradiance (GHI), photovoltaic power output (PVOUT), temperature, wind speed, aerosol optical depth (AOD), elevation, slope, albedo, land use/land cover (LULC), distance to roads, and distance to power lines. Reference labels were generated from an explicit rule-derived suitability index, class thresholds, and exclusion logic; therefore, the machine-learning task was to reproduce a transparent suitability framework rather than to predict observed PV yield or project-level performance. Extreme Gradient Boosting (XGBoost) was compared with simpler baseline models, evaluated using random and spatial-block validation, and interpreted using SHapley Additive exPlanations (SHAP). Independent overlays with known solar-installation records, presence-background robustness testing, and uncertainty/sensitivity analysis were used to examine spatial plausibility, spatial autocorrelation, deterministic label effects, and parameter uncertainty. The resulting outputs include pixel-level suitability zones, contiguous candidate polygons, district-level capacity-oriented summaries, and planning-priority classes. The framework is intended as a risk-aware regional screening tool: high model agreement indicates consistency with the constructed suitability labels, while final project decisions require parcel-scale land, grid, environmental, social, and economic assessment. Full article

Accelerometer sensors and artificial intelligence (AI) are reshaping automated behavior monitoring in precision livestock management, yet their joint deployment on extensive rangelands is constrained by energy and bandwidth budgets. Low-Power Long-Range Wide-Area Network (LoRaWAN) collars address these constraints by compressing the raw tri-axial signal on the device into a single scalar per reporting interval, the Motion Index (MI). This onboard compression preserves enough signal to separate active behaviors but discards the per-axis and frequency content that fine-grained classification typically relies on. On a dataset of 9222 labeled observations from 24 cows across four breeds, MI distinguishes walking from grazing reliably but fails to separate ruminating from resting; both correspond to a stationary animal and yield near-zero, statistically indistinguishable distributions. Earlier MI-only models reached only about 65% four-class accuracy, and ruminating was commonly merged into resting. We show that much of this loss can be recovered by treating the MI stream as a time series. Session-aware lag features, rolling statistics, and an autoregressive previous-behavior feature lift four-class macro-F1 from 0.647 to 0.94, with per-class F1 of 0.95 for ruminating and 0.92 for resting (and at least 0.92 for every behavior). In autonomous deployment the previous behavior must be predicted rather than observed; for this setting we add a Viterbi sequence-decoding step that combines the classifier’s per-step outputs with a learned behavior-transition model, recovering a substantial part of the ruminating signal from the activity stream alone while keeping walking and grazing reliable. The gain is consistent across seven classifiers and four genetically distinct breeds, indicating that it is driven by the features rather than by a specific model. Full article

The high penetration of renewable energy has exposed integrated energy systems (IES) to stronger source-load uncertainties. Traditional scheduling methods that primarily pursue economic optimality often fail to account for system regulation margins, which may lead to excessive charging and discharging of energy storage systems, frequent fluctuations in unit output, and insufficient supply–demand matching capability under uncertain operating scenarios. To address these issues, this paper proposes a Flex-TD3 optimal scheduling method for IESs with embedded flexibility rules. First, a regional IES model incorporating photovoltaic generation, wind power, micro-gas turbines, gas boilers, electric chillers, waste heat recovery units, heat exchangers, and battery energy storage systems is established to describe the coupling relationships among electricity, heat, cooling, and gas flows, as well as the operational constraints of key devices. Second, active regulation flexibility indicators are constructed from the perspectives of system upward regulation capability, downward regulation capability, energy storage state health, and electro-thermal decoupling regulation margin. A comprehensive flexibility score is then formulated to characterize the system’s capability to cope with renewable energy fluctuations and load disturbances under the current operating state. Third, the flexibility indicators are embedded into the state space and reward function of the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm, and a rule-based physical feasibility mapping mechanism is introduced to modify the raw scheduling actions generated by the agent according to device operational constraints, thereby enhancing the physical consistency and operational safety of the scheduling strategy. Case study results show that, compared with traditional optimal scheduling methods, the proposed method achieves better overall performance in terms of training convergence speed, operational economy, and scheduling stability. It can effectively reduce system operating costs, improve renewable energy accommodation capability, and decrease renewable energy curtailment, supply shortages, and constraint violations. Under uncertain scenarios involving renewable energy prediction errors, load disturbances, and high renewable energy penetration, the proposed method still maintains favorable scheduling performance, demonstrating its effectiveness and robustness. Full article

This study applies a Long Short-Term Memory (LSTM) model to predict key air pollutants, i.e., sulphur dioxide (SO₂), nitrogen dioxide (NO₂), and particulate matter (PM_2.5), as well as the Air Quality Index (AQI) across South Africa using satellite-derived observations. The analysis focuses on comparing original pollutant fields with model-generated predictions for two consecutive days, highlighting both spatial patterns and predictive performance. Results reveal a persistent and intense pollution hotspot over the Mpumalanga Highveld, driven by coal-fired power generation and industrial activities. Elevated pollutant concentrations in this region translate into AQI levels ranging from Unhealthy to Very Unhealthy, while most other parts of the country remain within the Good category. Spatial comparison between original and predicted fields shows strong agreement, with only minor deviations in areas characterized by steep emission gradients and localized plumes. Quantitative evaluation using RMSE (0.020390) and MSE (0.000416) confirms the high accuracy of the predictive model, with error values remaining extremely low across all pollutants and AQI outputs. PM_2.5 exhibits the smallest errors (MSE = 4.230169 × 10⁻⁶), while slightly higher values for SO₂ (MSE = 2.628 × 10⁻⁴) and NO₂ (MSE = 1.39541 × 10⁻⁴) reflect the difficulty of capturing sharp spatial transitions associated with point-source emissions. Despite these localized discrepancies, the model demonstrates robust skill in replicating both pollutant magnitudes and AQI classifications. Overall, the findings indicate that machine-learning approaches offer a reliable, high-resolution tool for air-quality prediction in South Africa and have strong potential for supporting operational forecasting, exposure assessment, and environmental policy development. Full article

Wind turbine gearbox failures lead to substantial downtime and high maintenance costs. Although condition-monitoring systems are widely used, traditional life-extension methods that simply reduce power output often decrease revenue. Current research frequently treats life optimization and power generation independently, and as such lacks a quantitative link between control strategies and remaining useful life. To address this gap, this paper proposes a novel life-extension strategy that optimizes power generation by dynamically adjusting rotor speed and pitch angle. A transfer learning–long short-term memory model enhanced by multi-source information fusion is developed to predict remaining useful life accurately under conditions with limited fault data. Utilizing real operational data from 2 MW wind turbines in Northeast China, the study quantitatively analyzes the impact of variable-speed and pitch control. The results demonstrate that while both strategies extend life, variable-speed control offers superior effectiveness in improving remaining useful life. Furthermore, maximum power generation is achieved not at full capacity, but when the output is reduced to approximately 70% of the nominal power. At this optimal point, the proposed strategy increases power generation by up to 7.3%. This establishes a dynamic balance between operational safety and economic efficiency, overcoming the limitations of conventional methods. Full article

The escalating global demand for renewable energy has positioned solar photovoltaics (PV) as a critical technology for achieving net-zero emissions. However, PV efficiency is strictly limited by thermal degradation, where elevated operating temperatures significantly reduce power output and accelerate material aging. This review systematically evaluates the integration of advanced phase change materials (PCMs) as a passive thermal management solution. We analyze the transition from material-level innovations—including nano-enhanced PCMs, 3D conductive frameworks, and shape-stabilization—to system-level hybrid architectures such as liquid—PCM, heat pipe-fin, and thermoelectric generator (TEG) integrations. Synthesis of recent empirical data (2024–2026) demonstrates that optimized PCM composites can achieve PV temperature reductions of up to 32 °C and electrical efficiency enhancements exceeding 19%. Furthermore, techno-economic assessments reveal that these systems can reduce the levelized cost of energy (LCOE) by 5–15% and achieve energy payback times as short as 1.5 years. Finally, this paper identifies critical research gaps in long-term outdoor durability, AI-driven predictive modeling, and sustainable bio-based encapsulation, providing a strategic roadmap for the commercialization of next-generation solar thermal management systems. Full article

This study proposes an experimental data-driven hybrid prediction framework for hydraulic turbine performance using a Particle Swarm Optimization-enhanced Extreme Learning Machine (PSO-ELM). The performance of three hydraulic turbines, namely Pelton, Kaplan, and Francis turbines, was experimentally investigated under different jet-opening and guide-vane conditions. The experimental results showed that the Pelton turbine (PT) achieved its highest efficiency at low jet opening, whereas the Kaplan and Francis turbines performed better at higher guide-vane openings. The measured data includes 36 tests, which were then used to evolve and evaluate hybrid ML models for predicting hydraulic power and efficiency. Jet-opening or guide-vane position (25%, 50%, 75% and 100%) and rotational speed were used as input variables, while brake power and efficiency were used as output variables. The proposed PSO-ELM model was compared with other optimized ELM models, including Genetic Algorithms Extreme Learning Machine (GA-ELM), Differential Evolution Extreme Learning Machine (DE-ELM), and Whale Optimization Algorithm Extreme Learning Machine (WOA-ELM), as well as Particle Swarm Optimization–Adaptive Neuro-Fuzzy Inference System (PSO-ANFIS) and Particle Swarm Optimization–Multi-Layer Perceptron (PSO-MLP) models. The suggested method presents a hopeful structure for tackling the difficulties linked to performance evaluation, thus enabling a more dependable and effective use of energy resources. The main findings validate that a PSO-based structure reaching an R² value of 0.997 is more efficient in predictive tool performance optimization for hydropower systems. Full article