Decarbonising heavy-duty transport, particularly rail, is critical to the UK’s 2050 net-zero target. This study assesses the feasibility of a green-hydrogen-powered rail system fuelled by agrivoltaics (APV), combining solar generation from APV, train-rooftop photovoltaics, and track-integrated photovoltaics (TIPV). QGIS determined optimal APV siting relative to rail infrastructure, with the Decision Support System for Agrotechnology Transfer (DSSAT), PVsyst and HOMER Pro coupled to evaluate agricultural, energy, and economic performance. The London Paddington to Penzance route was modelled using a 9-car Class-800 hydrogen train (8,080.9 kWh per leg, 440.8 kg of compressed hydrogen per journey, 7,053.1 kg daily fleet demand). Using existing overhead electrification between London and Newbury reduces annual electrical demand by 11% to 132.27 GWh, requiring 2,290 tonnes of green hydrogen annually. A 114.4 MWp APV array at Long Rock Depot, covering 198.3 ha, meets this demand, generating 151.00 GWh annually. DSSAT modelling showed faba bean yield rising by 66% at 30% ground coverage ratio (GCR), while sugar beet yield declined steadily with increasing GCR, indicating that agrivoltaic compatibility is crop-dependent rather than universal. HOMER Pro identified only a grid-supplemented architecture as commercially defensible: CAPEX £161.7 M, lifecycle net present cost (NPC) £263.1 M, IRR 18.39%, 5.41-year payback, and a levelised cost of hydrogen (LCOH) of £7.99/kg. Track- and vehicle-integrated photovoltaics both proved energetically marginal against real traction demand. Seasonal hydrogen storage, while technically feasible, was economically unviable, with LCOH rising to £28.11/kg. Seawater desalination was identified as the preferred water feedstock, avoiding conflict with domestic supply.
Rapid and reliable identification of multivariate geochemical anomalies is critical for delineating prospective mineralized zones and reducing uncertainty in mineral exploration targeting. Extended isolation forest (EIF) is a powerful unsupervised ensemble learning algorithm that efficiently isolates anomalies from high-dimensional geochemical datasets using randomly oriented hyperplane partitions. Previous studies have demonstrated the effectiveness of EIF in multivariate geochemical anomaly detection and mineral potential modeling. However, its performance can be significantly affected by stochastic variability arising from random partitioning and random subsampling during isolation tree construction, which may result in unstable anomaly patterns and inconsistent exploration targets in complex geological environments. To mitigate this limitation, we developed a robust unsupervised framework for the identification of multivariate geochemical anomalies associated with gold mineralization in the Southwestern Yilgarn Craton, Australia. The proposed framework integrates robust factor analysis (RFA), a Jaccard-based stability index and EIF to enhance the reliability and reproducibility of anomaly detection. RFA was first applied to compositional soil geochemical data to identify the most significant pathfinder elements associated with gold mineralization, which were subsequently used as input variables for the EIF model. The model was then optimized using a Jaccard-based stability criterion to ensure consistent anomaly detection across repeated independent runs. Model performance was assessed using area under the receiver operating characteristic curve (AUC). The obtained AUC value of 0.82 indicates strong predictive capacity, confirming that the generated anomaly map effectively delineates mineralization-related geochemical patterns and provides a reliable proxy for mineral prospectivity mapping. Overall, the proposed framework offers a robust and reproducible unsupervised approach for multivariate geochemical anomaly detection with strong applicability in both greenfield and brownfield mineral exploration settings.
Online fault detection and diagnosis (FDD) in photovoltaic (PV) plants must identify rare faults in imbalanced monitoring streams without using future information. We present a model-informed pipeline in which a multivariate autoregressive model with exogenous inputs (ARX) is calibrated on normal-only data and then frozen as a data-driven healthy-regime reference. Its one-step voltage predictions and signed residuals augment irradiance, temperature, and current measurements for five-class classification. The proposed ARX–NN processes this representation causally at each sample through a compact shared encoder with fault-classification and auxiliary voltage-estimation outputs. All classifiers are evaluated on identical chronological boundaries comprising an 80%/20% holdout and four expanding future-contiguous blocks.On the chronological holdout, ARX–NN attains macro F1=0.788±0.046, compared with 0.632±0.010 for NN using operating inputs and 0.650±0.039 when the ARX channels are masked; this ordering persists across all four later blocks. XGBoost using the same ARX-augmented representation is the benchmark accuracy leader at 0.912, compared with 0.542 using operating inputs. These results show that the complete ARX-augmented feature bundle is informative across classifier families, while positioning ARX–NN as a causal joint-output neural implementation rather than a universal replacement for XGBoost. At the selected event operating point, ARX–NN has shorter total false-alarm duration but more false-alarm episodes than XGBoost, indicating a distinct alarm profile. Because the benchmark lacks explicit timestamps and an external-plant test, conclusions are limited to later contiguous rows from the ordered stream.
Reliable object and keypoint detection on heliostats is essential for advancing automation and practical deployment of airborne condition-monitoring methods in concentrated solar thermal (CST) tower plants. Yet the wide variety of heliostat collector geometries in real systems limits the applicability of deep-learning approaches, and their transferability between plants remains unexplored. To address this gap, we compiled a database of representative real-world heliostat geometries and generated a large synthetic dataset using our previously published rendering framework. With this data, we evaluated three training strategies: geometry-specific baseline models trained from scratch, a universal model intended to generalize across all geometries, and fine-tuning approaches initialized from either the baseline or the universal model. Baseline models perform well on their respective geometries, while the universal model shows inconsistent performance and may not be sufficient as a stand-alone solution. Fine-tuning, however, consistently adapts the model to new geometries and achieves performance comparable to geometry-specific baselines, as shown by suitable metrics on real-world test datasets of three distinct collector types. In practice, an effective model for a new geometry can be obtained by rendering as few as 100 synthetic images and fine-tuning an available baseline or universal model for 2000 iterations. This procedure requires 20h on a single GPU and scales efficiently with additional computational resources. Overall, the results demonstrate that fine-tuning an existing heliostat detection model to the target domain provides a practical and reliable strategy for deploying deep models across the diverse heliostat collector geometries present in CST plants.
Photovoltaic (PV) radiometers based on crystalline silicon sensors are widely used as low cost instruments for solar irradiance monitoring in photovoltaic systems. However, their response depends on the angle of incidence (AOI) of incoming radiation, which introduces measurement uncertainties under real operating conditions. Although this effect is well recognized, its quantitative analysis remains limited. This work investigates the influence of AOI on the signal of a PV radiometer through a combined optical and electronic modeling approach. The optical response of the multilayer structure of the radiometer is first simulated using a matrix based method to determine the spectral transmittance as a function of AOI. These results are incorporated into PC1D simulations to obtain the angular-dependent spectral response of a crystalline silicon sensor. The radiometer signal is then calculated by integrating the simulated spectral response with solar spectra generated using the SMARTS (Simple Model of the Atmospheric Radiative Transfer of Sunshine) model for atmospheric masses from AM1 to AM10. Simulations show that the spectral response varies only slightly up to approximately 50°, followed by a marked change at larger angles. The solar spectrum also changes with atmospheric mass, introducing an additional dependence of the radiometer signal on AOI. Comparison with field measurements obtained using a PV radiometer shows good agreement with the simulated trends. The results demonstrate that the radiometer signal is determined by the combined effects of angle-dependent optical transmission and spectral variations in solar radiation.
Photovoltaic (PV) direct-current (DC) series-arc signals are highly stochastic and are readily affected by inverter switching noise and non-fault transients, including abrupt irradiance changes and load switching. To address this problem, this paper proposes a fault-detection method based on Log-Scattering and a residual Kolmogorov-Arnold network (ResKAN). First, a Cassie-Mayr hybrid arc model with dual energy-dissipation channels is developed, and a bounded 1/f-type stochastic component controlled by the fault-insertion signal is introduced to provide an equivalent representation of the conductance dynamics and stochastic texture during sustained arcing. Second, a windowing strategy based on an initial baseline and a fixed triggering threshold is used to locate valid analysis segments. Time-domain statistical features and multiscale Log-Scattering path features are then fused, while logarithmic compression reduces the influence of strong periodic harmonics and transient impulses on the feature responses. Finally, a ResKAN model comprising a base linear path, a B-spline nonlinear path, and residual connections is constructed to represent the complex nonlinear relationships between arc faults and non-fault disturbances. Evaluation on the simulated dataset developed in this study shows that the proposed method effectively distinguishes arc faults from normal operation, abrupt irradiance changes, and load switching, while achieving a favorable balance among detection performance, stability, and computational cost.
The performance of a novel hybrid photovoltaic–thermal evaporative (PVT–Ev) system, which consists of a semi-transparent photovoltaic–thermal (PVT) panel integrated into a solar chimney and an evaporative cooling cavity, has been investigated in this study. The proposed systemis designed to concurrently provide electrical power and improve indoor thermal comfort without any external energy dependency. The system performance has been evaluated for a residential building in the hot climate of Karbala, Iraq. Results reveal that the system produces about 2 kWh/day of electrical energy under peak-summer conditions. The relative humidity inside the building is within the ASHRAE comfort range (50–60 %). It achieves a net cooling energy savings of 11.57 kWh/day. Moreover, it can reduce the room temperature by 12.6 °C, and the average air velocity is below 0.9 m/s, which is comfortable for the occupants. The findings reveal that the PVT-Ev system is an efficient and sustainable solution for integrated power generation and passive cooling in hot climates.
The practical integration of phase change materials (PCMs) into thermal energy storage (TES) systems, such as domestic hot water tanks, is critically dependent on the packaging method, which affects both thermal stratification and heat storage capability. This study experimentally investigates the effect of three different PCM packaging techniques on the performance of a 450-liter vertically mantled hot water tank using 32.5 kg of paraffin (62-66 °C melting point). The methods compared are: (i) cylindrical capsules, (ii) a single central cylindrical rod, and (iii) a bulky cylindrical container. Thus, this study presents an experimental campaign is designed to systematically evaluate how PCMs should be integrated into a hot‑water thermal storage tank. By comparing multiple PCM configurations and operating conditions, the work provides a comprehensive assessment of their thermal performance, charging–discharging behavior, and overall suitability for enhancing sensible heat storage systems. The results for the charging phase show that the single container severely disrupts natural convection, promoting mixing and degrading the thermocline, which is quantified by a significant drop in the normalized Stratification factor. The encapsulated and rod methods, however, preserve stratification at levels comparable to the conventional (PCM-free) tank. During discharging, the encapsulated method suggests superior performance, increasing the usable hot water output by up to 20.6% (from 505 L to 609 L) compared to the conventional tank when charged at 80 °C and discharged at 10 L/min. Most significantly, all PCM configurations are 2-to-3 times more exergy-efficient than the conventional tank. This is because the PCM’s stable latent heat release preserves the thermal quality of the heat, avoiding the rapid temperature decay seen in the conventional tank. The study concludes that a distributed, high-surface-area encapsulated design is the optimal strategy, maximizing both the energy quantity and its thermodynamic quality.
Estimating solar irradiance over several years is essential for assessing large-scale photovoltaic and heliothermic projects. The usual way to achieve this is to adapt long-term satellite-based solar irradiance estimates for specific locations using short-term ground measurements. In this context, a novel site adaptation method is proposed for global horizontal irradiance (GHI) and direct normal irradiance (DNI) based on a two-level modelling strategy incorporating both clustered and unclustered features. These lead to global and local models, which are evaluated separately and in combination. Input variables are provided by the Copernicus Atmosphere Monitoring Service Radiation Service (CAMS) and ECMWF ERA5-Land. The method is tested at four locations – El Rosal and Salta (Argentina), Petrolina (Brazil), and Gobabeb (Namibia), addressing sites in the Southern Hemisphere with similar climates, but different latitudes and satellite viewing angles. A new non-supervised sky conditions classification method is proposed for the local models using the clear sky index, a variability index, and Kernel Density Estimation. This classification forms an integral part of the site-adaptation procedure by capturing local cloud-irradiance interactions. For GHI at 15-minute resolution, the method improves accuracy relative to CAMS at locations near the edge of the Meteosat Second Generation field of view satellites, with RMSE reductions of 26.2% (El Rosal) and 4.8% (Salta). For DNI in Petrolina and Gobabeb, the local and combination models outperform CAMS model, achieving RMSE reductions of 10.2% and 5.8%, respectively. Combination models present the more accurate results as they use the global, local, and CAMS outputs as input variables.
Rising silver (Ag) prices are driving efforts to reduce metallization cost in n-type tunnel oxide passivated contact (n-TOPCon) solar cells, which rely on Ag contacts on both sides. Copper (Cu) is an attractive alternative due to its low cost, abundance, and comparable conductivity, but its rapid diffusion in silicon (Si) during screen-printed metallization can increase recombination and degrade cell performance. Combining Cu metallization with the laser-enhanced contact optimization (LECO) process, which enables low-temperature firing (540 °C) of the screen-printed contacts, can significantly reduce Cu penetration into the Si substrate. This paper demonstrates 24.3% Cu-contacted n-TOPCon solar cells using a post-firing LECO process along with counterpart Ag-contacted cells of 24.5%. The metal-induced recombination current density (J0, metal) of the Cu contact on the rear side poly-Si was only 14 fA/cm2, about four times lower than Ag. Before LECO, contact resistivities of Ag and Cu contacts to n-TOPCon were 5.5 mΩ-cm2 and > 100 mΩ-cm2, respectively, which decreased to 1.7 and 19.7 mΩ·cm2 after optimized LECO. To compensate for higher contact resistivity of Cu, we increased the Cu coverage on n-TOPCon from 7% to 18% by adding more gridlines, yielding ∼ 24.3% efficiency. Quokka 3 device simulations validated the measured parameters, and power loss analysis showed that the ∼ 0.2% efficiency gap between Ag- and Cu-contacted cells is mainly due to higher contact resistivity and line resistance of Cu. Combining screen-printable Cu paste with LECO enabled high-efficiency Cu-contacted n-TOPCon cells, offering a viable pathway to reduce Ag in solar cell manufacturing.
Incomplete surface coverage, localized exposure of ITO protrusions, and interfacial nonuniformity in solution-processed SnO2 electron transport layers (ETLs) restrict the performance of hole-transport-layer-free (HTL-free) CsPbBr3 perovskite solar cells (PSCs). Herein, a planar homojunction SnO2/Mg-doped SnO2 bilayer ETL was fabricated by depositing an ultrathin Mg-doped SnO2 layer on the SnO2 surface to regulate the ETL/perovskite interface in an ITO/SnO2/CsPbBr3/carbon device architecture. This bilayer configuration improves surface coverage, interfacial contact, and energy-level alignment while preserving lattice compatibility within the SnO2-based system. The Mg-doped SnO2 top layer reduces surface cracks and pits, limits local ITO exposure, decreases roughness, and improves precursor wettability. These changes promote the formation of CsPbBr3 films with larger grains, fewer pinholes, better vertical morphology, higher crystallinity, and a near-stoichiometric composition. The PL, TRPL, EIS, dark J–V, XPS, and UPS results collectively indicate improved interfacial charge extraction, reduced defect-assisted non-radiative recombination, and more improved energy-level alignment at the ETL/perovskite interface. Consequently, the optimized HTL-free device achieved a champion PCE of 8.74% with a Voc of 1.520 V and retained 90.69% of its initial efficiency after 50 days of unencapsulated storage in air. This work indicates that introducing an ultrathin Mg-doped SnO2 overlayer effectively reconstructs the ETL surface and regulates the ETL/perovskite interface, providing an effective interfacial regulation strategy for developing efficient and stable HTL-free all-inorganic PSCs.
Building-integrated photovoltaic (BIPV) technology enhances renewable energy utilization and taps buildings’ energy savings potential, offering significant environmental benefits. However, conventional photovoltaic (PV) modules are limited by their monochromatic appearance, mainly installed on rooftops and leaving facades underutilized, impeding the advancement of BIPV facade systems. This study proposes a novel solution using highly transparent interference-type structural colour pigments (SCP) to achieve coloured PV modules, followed by computational and experimental methods to investigate SCP’s chromatic characteristics and coloured PV performance. The simulation of structural colours is first conducted by the transfer matrix method, which facilitates the synthesis and preparation of two distinct types of SCP, i.e. three-layer TiO2/mica/TiO2 (M/T) and five-layer SiO2/TiO2/mica/TiO2/SiO2 (M/T/S). The structural colours of SCP exhibit a clockwise shift on the horseshoe diagram with increasing coating thickness and show colour differences when surrounded by materials of different refractive indices, consistent with simulation results. The coloured glass sheet prepared by screen printing exhibits a high optical transmittance, with the M/T/S sample achieving approximately 67.0% ∼ 89.8% across 400 ∼ 1200 nm. Furthermore, the PV modules prepared with two types of SCP exhibited high-efficiency retention rates of at least 91.6% and 94.9%, respectively. The SCP feature simple preparation, adjustable colouration, and high transmittance, with minimal reflection outside the visible light band, meeting PV colourization requirements.
