FME SOLAR Norsk forskningssenter for solkraft

• Moschetti, Roberta & Nocente, Alessandro (2026). Modelling and validation of ventilated roofs with BIPV: A case study of a zero-emission building in Norway. Journal of Physics: Conference Series (JPCS). ISSN 1742-6588. 3302(1), s. 1–10. doi: 10.1088/1742-6596/3302/1/012026. Fulltekst i vitenarkiv Vis sammendrag

  As climate change leads to increasing temperatures and precipitation, building envelopes must adapt to ensure durability. Meanwhile, improving energy efficiency requires integrating new technologies directly into the building envelope. In this context, building-integrated photovoltaics (BIPV) are increasingly being adopted. However, the hygrothermal performance of BIPV components in cold, humid Nordic climates remains insufficiently characterised. This study develops and validates a computational model of the ventilated roof with BIPV at the Zero Emission Building Laboratory, a full-scale office laboratory in Trondheim, Norway. The model is implemented in COMSOL Multiphysics and simulates coupled heat and moisture transport in the ventilated air cavity, accounting for turbulent airflow, surface-to-surface radiation, moisture transport in air, and heat transfer in both solids and fluids. Boundary conditions are defined through available in situ measurements. Model validation uses temperature and relative humidity data from two sensors located near the inlet and outlet of the air cavity during two representative periods with different climatic conditions. Temperature predictions show CV(RMSE) values of approximately 6–16% and NMBE between -10% and +8%. Relative humidity predictions show CV(RMSE) values of about 8–9% at the inlet and 20–24% at the outlet, with NMBE ranging from -13% to +8%. Pearson correlation coefficients range from 0.85 to 0.99, indicating generally good agreement in the temporal evolution of the variables. The results provide insights into the coupled thermal and moisture behaviour of ventilated BIPV roofs in cold climates and support the design of climate-resilient and energy-efficient building envelopes for Nordic conditions.

• Garcia, Alfredo Sanchez & Dimd, Berhane Darsene (2026). Enhanced Bifacial Photovoltaic Power Prediction through Procedural Training and Comprehensive Irradiance Data. Energy and AI. 25, s. 1–14. doi: 10.1016/j.egyai.2026.100830. Fulltekst i vitenarkiv Vis sammendrag

  This work investigates the benefits of an incremental training strategy designed to emulate realistic operational updates, and the impact of using complete irradiance inputs — including direct normal, diffuse horizontal, and global horizontal irradiance — for bifacial photovoltaic power prediction. Using data from a Nordic climate setup with four bifacial photovoltaic modules equipped with dual-axis sun trackers, three classical machine learning models and one long short-term memory model were trained to predict instantaneous and short-horizon power output. The proposed incremental strategy was found to be sensitive to training-month order; to address this, multiple runs with randomized month-order sequences were performed. Zero-shot predictions, corresponding to months not used in training, produced median coefficient-of-determination values ranging from 0.380 to 0.964 depending on model and module orientation. Models trained with global horizontal irradiance alone exhibited lower accuracy and greater sensitivity to training data selection. Feature-importance analysis further showed that direct normal and diffuse horizontal irradiance contributed substantially to prediction accuracy, particularly under low-sun and snow-affected conditions typical of Nordic environments.

• Gorjão, Leonardo Rydin & Nygård, Heidi Samuelsen (2026). Modelling the Nordic power system expansion with PyPSA-Eur: A comparison with the Nordic Grid Development Perspective 2025. International Conference on the European Energy Market. ISSN 2165-4077. s. 1–6. doi: 10.1109/EEM68581.2026.11589650. Fulltekst i vitenarkiv Vis sammendrag

  We compare the Nordic Grid Development Perspective 2025 (NGDP) with a least-cost transition pathway for the Nordic power system obtained with PyPSA-Eur for the period 2025-2050 under a net-zero CO2 constraint. The model co-optimises generation, storage, and transmission expansion with hourly system operation while constraining the Nordic generation mix to remain consistent with NGDP milestone ranges. Overall, the optimised capacity pathways reproduce the broad NGDP expansion of wind and solar power, with Denmark emerging as the main locus of offshore wind and solar deployment, Finland as a major site for onshore wind expansion, and Norway and Sweden relying more strongly on existing hydropower and nuclear capacity. The main discrepancy between the two pathways does not lie in the broad buildout of variable renewables, but in the timing and role of hydrogen. Whereas the NGDP treats hydrogen as an early and steadily increasing component of industrial demand and system development already from 2030, PyPSA-Eur delays substantial hydrogen deployment until 2045-2050, when it becomes costeffective primarily as a flexibility and seasonal storage option in a strongly variable-renewable-energy-dominated system. This contrast suggests that early hydrogen build-out in the Nordics is not a robust outcome of cost optimisation alone, but depends strongly on assumptions regarding industrial demand growth, sector coupling, and anticipatory infrastructure planning. We therefore argue that hydrogen remains the principal strategic uncertainty in the Nordic transition, with the NGDP representing a more anticipatory planning perspective and PyPSA-Eur a more conservative least-cost timing of deployment.

• McCulloch, Manuela; Myhre, Stine Fleischer; Remund, Jan; Jouttjärvi, Sami; Murkowska, Marta Irena & Bouramdame, Ayat-Allah [Vis alle 12 forfattere av denne artikkelen] (2026). Integrating Solar PV into Energy, Capacity and Balancing Markets: An International Comparative Analysis of Market Design and Regulation. International Conference on the European Energy Market. ISSN 2165-4077. doi: 10.1109/eem68581.2026.11589789. Fulltekst i vitenarkiv
• Nygård, Magnus Moe; Riise, Heine Nygard; Eriksen, Erling Ween & Martsein, Erik Stensrud (2026). Nowcasting based on spatio-temporal string-level power measurements in utility-scale photovoltaic power plants. Solar Energy. ISSN 0038-092X. 310. doi: 10.1016/j.solener.2026.114480. Fulltekst i vitenarkiv Vis sammendrag

  A novel method is proposed to estimate cloud motion vectors (CMVs) from endogenous photovoltaic (PV) power plant energy yield time-series data. The method is tested using three months of 1-min mean aggregated power output data collected from 8000+ string-sets in a 150+ MW PV power plant. The CMVs are used as input to a cloud speed persistence PV power forecast which is benchmarked against a smart persistence PV power forecast. Forecasts issued for the entire power plant achieve an overall forecast skill of 0.16 at a forecast horizon of 1-min. If the forecast is issued for smaller sections of the power plant with foresight provided by the remaining sections, an increase in forecast skill up to 0.47 is demonstrated. The results emphasize the potential that can be unlocked if highly resolved spatio-temporal energy-yield time-series data are coupled with exogenous data, for instance, captured through satellite imagery and all-sky imagers.

• Øgaard, Mari; Eriksen, Erling Ween; Dahlioui, Dounia; Rønneberg, Sigrid; Müller, Thore & Riise, Heine Nygard (2026). Analyzing and modeling snow loss and snow loss accumulation in ground-mounted photovoltaic systems. EPJ Photovoltaics. 17. doi: 10.1051/epjpv/2026004. Fulltekst i vitenarkiv Vis sammendrag

  Snow can cause significant energy loss in photovoltaic systems, making it essential to account for in performance modeling for all snow-prone regions. Previous snow loss research has been conducted primarily for roof mounted systems, and there is a lack of reported snow losses and validation of snow loss models for ground mounted systems. The aim of this work is to quantify snow loss and validate the Marion snow loss model for ground mounted PV systems using data from two scientific test sites in Norway. The quantification of snow losses indicates that ground mounted systems have lower snow losses compared to what have previously been documented for roof mounted systems in Norway and similar climates. Additionally, monofacial and bifacial arrays experienced similar losses. At the same test site, the bifacial array’s monthly losses were, on average, 0.2 percentage points higher than those of the monofacial array. In the snow loss model validation, the impact of snow data quality was evaluated. The global weather snow data sources assessed (ERA5-Land and ERA5 global) had a large share of snowfall events not giving snow accumulation. These events were linked to both small snowfalls and higher temperatures. The default pvlib implementation of the Marion model typically underestimates losses, but the error in relative yearly losses is less than 0.5 percentage points. The error in the modeling is mainly related to periods where the temperature is close to zero, when both the snow input data and the prediction of snow shedding are more uncertain. Ground interference due to large snow depths also leads to slower shedding and underestimation of losses in the model.

• Manni, Mattia; Nocente, Alessandro & Lobaccaro, Gabriele (2025). A comparison of machine learning and calibrated physical models for solar energy analysis in snowy regions. Journal of Physics: Conference Series (JPCS). ISSN 1742-6588. 3140(3), s. 1–7. doi: 10.1088/1742-6596/3140/3/032014. Fulltekst i vitenarkiv Vis sammendrag

  This study investigates the potential of machine learning (ML) tools to enhance sub-hourly photovoltaic (PV) power simulations in high-latitude locations with snowy conditions. A comparative analysis was conducted between calibrated physical PV model chains and ML tools to assess their ability to estimate snow losses (i.e., PV power output reduction due to snow accumulation). An ensemble of 100 physical PV model chains was implemented to probabilistically estimate the generated power. Additionally, five ML methods – gradient boosting regressor, linear regression, decision tree regressor, random forest regressor, and multi-layer perceptron regressor – were used both to i) calibrate the physical model ensembles (PME) and to ii) directly simulate PV power output. Model training, calibration, and validation were performed using observational data from the PV modules installed at the ZEB Test Cell Laboratory in Trondheim, Norway. PV power output was monitored with a 15-minute resolution from November 2023 to April 2024. The findings demonstrated that ML models, particularly random forest regressor, could outperform PME in capturing complex meteorological interactions such as snow accumulation, melting, and irradiance variability. This highlights the potential of ML tools to reduce uncertainties in solar analyses and support the design of more resilient and adaptive PV systems for the Subarctic and Arctic regions.

• Serrato, Jose Tirado; Garcia, Alfredo Sanchez & Maximov, Serguei (2025). Parametrization of I-V characteristics of solar cells via asymptotics solutions of the super-ellipse model. Solar Energy. ISSN 0038-092X. 300, s. 1–12. doi: 10.1016/j.solener.2025.113797. Fulltekst i vitenarkiv Vis sammendrag

  The computation of I-V characteristic curves is vitally important for tasks related to the design, sizing, and simulation of photovoltaic modules, as it allows them to evaluate the performance of different PV technologies under changing weather conditions. A simple and computationally efficient way to calculate I-V curves involves using the explicit super-ellipse model. However, this model is complicated to parameterize, as it requires numerically solving complicated systems of nonlinear equations. In response to this problem, this paper presents analytical methodologies for parameterizing the super-ellipse model. Two approaches are developed: one utilizing the maximum power point (MPP) and another employing two measured points from the I-V curve. The methods use asymptotic approximations to solve the equations, avoiding computationally intensive numerical techniques. The model is validated with experimental corresponding to six PV technologies and 18 weather and irradiance conditions using RMSE as the evaluation metric. The results show that the asymptotic solutions achieve performance comparable to numerical methods while offering simplicity and computational efficiency, presenting a RMSE of less than 0.0705 W for the methodology that uses the MPP and 0.027 A for the one that uses the measurement of two points on the I-V curve. In addition, it was verified that this model complies with the IEC EN 50530 standard. This study also shows the dependencies of Super-ellipse model parameters with irradiance and temperature. The proposed approaches address limitations of prior models, and enhance the representation of I-V curves under diverse operational scenarios.

Se alle arbeider i NVA

• Rønneberg, Sigrid; Seiffert, Christoph; You, Chang Chuan; Thomassen, Bent; Roosloot, Nathan Abel & Otnes, Gaute (2026). Color Degradation of Wall-Mounted BIPV Modules in the Nordics Compared with Temperate Climates. Fulltekst i vitenarkiv Vis sammendrag

  Color is a key feature of building-integrated photovoltaic (BIPV) systems, and both developing a range of colors, optimizing color for performance, and ensuring the stability of the color appearance are of great interest in order to increase the uptake of BIPV systems in new and renovated buildings. This work investigates the color degradation of 15 modules set up in a south-facing wall-mounted BIPV mock-up system situated in Kjeller, Norway, compared with identical modules that have been installed in a temperate climate in France. The Norwegian modules consist of 5 groups of 3 modules each, and include 4 different colors made with two coloring techniques, namely printed glass and colored foil. The colors made with printed glass are light and dark grey, while those made with colored foil are white and grey. The color of each BIPV module in Norway was characterized prior to installation in 2021 with a portable spectroradiometer in 5 different standardized points. This work repeats the same color characterization for the wall-mounted modules, giving changes in visual color perception of the modules in terms of color uniformity, color variability, and with respect to parameters such as coloring technology, module temperature, energy yield, and climate. The outdoor measurements will be performed in spring 2026, and will also include similar measurements on identical BIPV modules that were installed in southern France. By comparing the colors of the modules in Norway and those installed and operated in France, this work will determine the potential effect of climate on change in color perception of BIPV modules after more than four years of operation.

• McCulloch, Manuela; Myhre, Stine Fleischer; Remund, Jan; Jouttjärvi, Sami; Murkowska, Marta Irena & Bouramdame, Ayat-Allah [Vis alle 12 forfattere av denne artikkelen] (2026). Integrating Solar PV into Energy, Capacity and Balancing Markets: An International Comparative Analysis of Market Design and Regulation. Fulltekst i vitenarkiv
• Zheng, Huilan; Waszak, Maryna; Lam, An Ngoc; Marthinsen, Astrid; Ulyashin, Alexander & Ma, Xiang (2026). A Simple Statistical Modelling Approach for Evaluating Diverse PV Module Performances under Nordic Environments. Fulltekst i vitenarkiv Vis sammendrag

  A simple nonlinear statistical model using irradiance and module temperature is established to predict the performance of diverse PV technologies with > 95% accuracy, providing an efficient tool for benchmarking PV systems under Nordic environmental conditions.

• Wiig, Marie Syre; Nygård, Magnus Moe; Sødahl, Elin Dypvik & Kjeldstad, Torunn (2026). Accurate temperature modelling in energy yield assessments for large scale tracker systems. Fulltekst i vitenarkiv
• Nygård, Magnus Moe; Melandsør, Jonas Gran & Wiig, Marie Syre (2026). Reducing Interrow Shading Losses in Utility-Scale Photovoltaic Power Plants: From Power Output Data to Terrain-Optimized Tracking. Fulltekst i vitenarkiv Vis sammendrag

  The present work demonstrates an end-to-end framework that enables: (1) estimation of the underlying terrain of a photovoltaic (PV) power plant from interrow shading signatures present in power output time-series data; (2) optimization of tracker angles to limit interrow shading losses; and (3) separation of these losses into mitigable and non-mitigable components. The framework is evaluated using operational data collected from a row of 60 horizontal single-axis trackers (HSATs) in a utility-scale PV power plant. The estimated terrain profile is validated against independent LiDAR measurements and shows good agreement. The results demonstrate that interrow shading losses can almost be eliminated for the investigated row, providing a 2.5% improvement in overall power output relative to conventional flat-terrain backtracking. With an increasing amount of utility-scale PV being installed in nonuniform terrain, the proposed framework could enable plant operators to reduce both the costs associated with dedicated LiDAR campaigns and the power losses caused by terrain-induced interrow shading.

• Melandsør, Jonas Gran; Wiig, Marie Syre & Nygård, Magnus Moe (2026). Improved energy production by optimization of backtracking for PV power plants in rolling terrain. Fulltekst i vitenarkiv
• Eriksen, Erling Ween; Riise, Heine Nygard & Nygård, Magnus Moe (2026). Nordic Benchmark of ASI-based Nowcasting Solutions. Fulltekst i vitenarkiv Vis sammendrag

  Presentation of preliminary results of a benchmark of solar nowcasting solutions

• Secic, Una; Gorjão, Leonardo Rydin & Engan, Lill Mari (2026). Flicker-Related Voltage Changes from Short-Term PV Variability in MV Grids. Fulltekst i vitenarkiv Vis sammendrag

  Power systems are moving from synchronous generation towards inverter-connected renewable generation, raising concern regarding the effects on voltage quality. Solar PV contributes little mechanical inertia, and its active-power output follows irradiance conditions. Rapid cloud movements can therefore become rapid active-power changes. In radial medium-voltage grids, these changes may appear as voltage changes at the point of interconnection.

• Tang, Kai; Ma, Xiang; Syvertsen, Martin & Safarian, Jafar (2026). EoL PV Silicon: Metallurgical Viability for Closed Loop Feedstock. Fulltekst i vitenarkiv
• Vik, Eskil; Conti, Michela; Fjeld, Joakim Christoffer; Karlsen, Magnus Sandviknes; Kildemo, Morten & Reenaas, Turid Dory (2026). Optimizing Pulsed Laser Deposition of Doped TiO2 for Intermediate Band Solar Cells. Fulltekst i vitenarkiv
• Tang, Kai; Ma, Xiang; Låstad, Jonas & Safarian, Jafar (2026). End-of-Life PV Silicon: Metallurgical Viability for Closed‑Loop Feedstock. Fulltekst i vitenarkiv
• Vik, Eskil; Conti, Michela; Fjeld, Joakim Christoffer; Karlsen, Magnus Sandviknes; Kildemo, Morten & Reenaas, Turid Dory (2026). Optimizing Crystalline Quality of Cr + N Co-Doped TiO2 Thin Films for Intermediate Band Solar Cells. Fulltekst i vitenarkiv
• Myhre, Stine Fleischer & Olkkonen, Ville (2026). The role of PV in Norway towards 2050. Fulltekst i vitenarkiv
• Hebnes, Tuva & Øgaard, Mari (2026). Experience from IR campaigns in PV power plants in Norway. Fulltekst i vitenarkiv
• Xu, Yijiang; Ganesan, Nagarajan Somi; Dimd, Berhane Darsene; Ma, Xiang; Marthinsen, Astrid & Stokkan, Gaute [Vis alle 7 forfattere av denne artikkelen] (2026). Monitoring, Performance Analysis, and Mitigation of PV Reliability Issues in Nordic Installations. Fulltekst i vitenarkiv
• Nygård, Magnus Moe; Riise, Heine Nygard & Eriksen, Erling Ween (2025). Cloud motion detection and short-term PV power forecasting using spatially resolved production data. Fulltekst i vitenarkiv
• Nygård, Magnus Moe; Aarseth, Bjørn Lupton; Otnes, Gaute; Wiig, Marie Syre & Marstein, Erik Stensrud (2025). Towards a cost-priority number for module replacements informed by energy yield time series and infrared thermography data. Fulltekst i vitenarkiv Vis sammendrag

  In this presentation, we combine infrared thermography (IRT) and energy yield time series data to detect thermal signatures in photovoltaic (PV) modules in a 75 MWp PV power plant and assess the associated power loss over nearly a decade of operation in the field. An IRT scan carried out after 5 years of operation reveals that 0.2% of the PV modules in the power plant are affected by thermal signatures which corresponds to a 0.06% reduction in yield from the power plant. Moreover, the power loss associated with thermal signatures seems to develop nonlinearly over time which underlines the importance of continuous monitoring and informed operations and maintenance (O&M) of the power plant. Thus, the results presented is an important contribution to the available literature on degradation mode statistics and the temporal development of their associated power loss. This information is essential to determine a cost-priority number (CPN) and enable a cost-based approach to prioritization of replacements of PV modules suffering from such issues over other O&M tasks.

• Marstein, Erik Stensrud (2025). Solkraft som eksportnæring - hvor står vi? Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Solenergi for fremtiden - Insentiver for høsting på næringsbygg. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Fra vekst til bråstopp - Hvor skal den norske solbransjen? Glass og fasade. ISSN 0802-295X. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Leker seg med BIPV. Glass og fasade. ISSN 0802-295X. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Mer solstøtte i Oslo. [Journal]. Elmagasinet. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Norge på jumboplass på sol globalt. [Journal]. Montel News. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Solkraft kan bli Norges neste store eksporteventyr. [Journal]. EnergiAktuelt. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Forskningssenteret FME SUSOLTECH. Glass og fasade. ISSN 0802-295X. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Mystiske funn i solkraftanlegg skaper uro. [Journal]. NrK . Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Slik rammer finansstormen solbransjen. [Internet]. Energi og klima. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Solkraft i Norge og FME SOLAR. Glass og fasade. ISSN 0802-295X. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Verden blir mer og mer soldrevet. [Journal]. forskning.no. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Mer areal brukes til golfbaner enn til sol- og vindkraft i flere land. [Journal]. forskning.no. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Verden blir soldrevet - himla fort. [Journal]. Titan. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Forsker advarer: Norge henger etter i den globale solkraftrevolusjonen. [Journal]. ITB Aktuelt. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Håper batteriløsninger kan redde solkraftmarkedet. [Journal]. NrK Innlandet. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud & Selj, Josefine Helene Krogh (2025). Roadmap for the Norwegian PV industry. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). The Norwegian PV industry. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Norsk forskningssenter for solkraft. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Veikart for norsk solbransje mot 2030. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Status for solkraft og BIPV. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). En kort oppdatering fra solindustrien. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Status og potensial for solkraft i Norge. Fulltekst i vitenarkiv
• Dahlioui, Dounia; Schading, Steve; Høye, Ingar Alvaro & Vehus, Tore Sandnes (2025). Performance Evaluation of Installed Bifacial PV Modules: Towards the Ground Albedo Enhancement. doi: 10.4229/EUPVSEC2025/4DV.4.16. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud; Klyve, Øyvind Sommer; Nygård, Magnus Moe; Olkkonen, Ville & Riise, Heine Nygard (2025). Quantifying the Benefits of Hybrid PV-wind Power Plants. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud; Fagerström, Jonathan; Ollas, Patrik & Archer, Dan-Eric (2025). Assessing Cold-Weather Reliability of PV + BESS. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Solbransjen som eksportindustri. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Reliability of PV + BESS. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Batterisystemenes rolle. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). The state of the PV industry in 2025. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). The solar (PV) industry. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Design and optimization of hybrid PV + wind power plants. Fulltekst i vitenarkiv
• Garcia, Alfredo Sanchez & Dimd, Berhane Darsene (2025). Deep Learning-Based Solar Irradiance Decomposition Models for Nordic Regions. doi: 10.4229/EUPVSEC2025/4BV.3.8. Fulltekst i vitenarkiv Vis sammendrag

  This work presents a comparative evaluation of machine learning (ML) and deep learning (DL) models for solar irradiance decomposition in Nordic regions, where traditional empirical models often struggle. Using data from the Alpha Centauri outdoor test facility in Trondheim, Norway, the present work benchmarks the performance of Histogram-based Gradient Boosting (HGB), Artificial Neural Networks (ANN), and Long Short-Term Memory (LSTM) networks against the Erbs model. Results indicate that HGB performs best on the initial evaluation set, achieving strong R² scores for both DNI and DHI, while requiring minimal computational resources. However, in a zero-shot prediction scenario using independent data, HGB's performance drops significantly, suggesting overfitting to seasonal patterns. ANN maintains the highest accuracy for DNI decomposition, capturing nonlinear dependencies more effectively, whereas LSTM shows mixed results, particularly underperforming in DNI estimation.

• Marstein, Erik Stensrud (2025). FME SOLAR in a nutshell. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Reliability of PV + BESS. Fulltekst i vitenarkiv
• Marstein, Erik Stensrud (2025). Reliability of PV + BESS. Fulltekst i vitenarkiv
• Øverås, Aslak & Kjeldstad, Torunn (2025). Fornybaren Podcast - Sol vokser rask globalt, men Norge henger etter. doi: https:/www.fornybarnorge.no/fornybaren/?gad_source=1&gad_campaignid=10262923139&gclid=CjwKCAiAoNbIBhB5EiwAZFbYGLhxXVRGuZnNpVRYW_kFO9B5f-mgJtSsb2p-rt51Fn4UQbIXpRlOBRoCGN0QAvD_BwE. Fulltekst i vitenarkiv
• Øgaard, Mari; Eriksen, Erling Ween; Dahlioui, Dounia; Rønneberg, Sigrid; Müller, Thore & Riise, Heine Nygard (2025). Analyzing and Modeling Snow Loss in Ground-Mounted PV Systems. Fulltekst i vitenarkiv
• Riise, Heine Nygard; Mehlhoop, Anne Catriona; Roosloot, Nathan Abel; Nygård, Magnus Moe; Eriksen, Erling Ween & Walpita, Harsha Lakmal [Vis alle 9 forfattere av denne artikkelen] (2025). Nature-Inclusive Operation and Maintenance of PV Parks – An Expert Elicitation of PV Performance and Ecological Impacts. Fulltekst i vitenarkiv
• Honningdalsnes, Erlend Hustad (2025). Matproduksjon og solenergi = sant. Fulltekst i vitenarkiv
• Øgaard, Mari; Eriksen, Erling Ween; Dahlioui, Dounia; Rønneberg, Sigrid & Riise, Heine Nygard (2025). Analyzing and modeling snow loss in ground-mounted PV systems. Fulltekst i vitenarkiv
• Garcia, Alfredo Sanchez; Hendawi, Rania Tayseer Atieh; Schøn, Hendrik & Sabatino, Marisa Di (2025). Correlating Structure Loss and Operational Conditions in Czochralski Silicon Ingot Growth using Machine Learning. Fulltekst i vitenarkiv Vis sammendrag

  This work investigates the relationships between process parameters and the occurrence of structure loss in Czochralski silicon ingots using machine learning. Subsets of features are identified from a dataset of over 14,000 ingots and are used to train random forests to predict structure loss with high accuracy. Multiple rounds of feature importance analysis and refinement are conducted to isolate the process parameters that may have the most significant impact in the occurrence of structure loss. Partial dependence analysis is employed to examine how variations in particular parameters might affect the likelihood of structure loss happening. The results show that the most predictive features of structure loss are primarily recorded late in the process. These features are often influenced by manual interventions or reflect the outcome of structure loss itself. In contrast, early-stage parameters exhibit limited predictive power, suggesting that either early indicators of structure loss are not captured in the available data or that structure loss originates from events occurring later in the growth process. While not predictive in a preventive sense, the model effectively detects deviations from normal operation, thereby demonstrating the value of machine learning for uncovering complex patterns in manufacturing processes data.

• Stein, Joshua S.; Maugeri, Giosuè; Whitney, Erin; Jahn, Ulrike; Tonita, Erin & Granlund, Alexander [Vis alle 18 forfattere av denne artikkelen] (2026). Photovoltaics and Energy Security in the Greater Arctic Region. IEA PVPS. ISSN 9781764290296. doi: 10.69766/jxhm9635. Fulltekst i vitenarkiv

Se alle arbeider i NVA