AI misleiden: zonne-energie voorspellen van een installatie die nog niet bestaat
Stel, je installeert een nieuwe fotovoltaïsche installatie en je wilt jouw elektrische auto opladen wanneer de zon schijnt en zo geld besparen. Eigen opgewekte elektriciteit verbruiken is niet enkel goed voor je portemonnee, maar ook voor het elektriciteitsnet en het klimaat. Om dit eigen verbruik te maximaliseren, moet je wel op voorhand kunnen voorspellen hoeveel zonne-energie er doorheen de dag opgewekt wordt. Hoe weet je de dag voordien hoeveel je zonnepanelen gaan opbrengen? Zeker wanneer de zonnepanelen juist geïnstalleerd zijn en er nog geen data beschikbaar is. Met mijn thesis toon ik aan dat AI modellen getraind met virtuele data de oplossing kan zijn.
Voorspellen van zonne-energie: fluitje van een cent?
De meeste mensen denken dat zonne-energie uitsluitend afhangt van de hoeveelheid licht. Voorspel het licht van de zon, voorspel de zonne-energie. Zeker niet fout, echter niet helemaal compleet. De hoeveelheid nuttige elektrische energie die opgewekt kan worden door een zonnepaneel hangt af van een hele hoop factoren, onder andere de positie van de zon, de temperatuur van het paneel en de windsnelheid. Er bestaan vergelijkingen die al deze factoren in rekening nemen. Recent onderzoek toont echter aan dat het beter zou kunnen zijn om artificiele inteligentie (AI) te gebruiken, de hype van het moment.
Moet er nog data zijn?
Wie AI-modellen zegt, zegt data. AI-modellen gebruiken een gigantische hoeveelheid data om daarop te trainen. Bijvoorbeeld, een model met de taak om katten van honden te onderscheiden krijgt duizenden foto’s van honden en katten en bijhorend een label van hond of kat. Aan de hand van deze enorme berg trainingsdata, kan een AI model nieuwe input data, namelijk ongeziene foto’s van honden of katten, juist classificeren als kat of hond. Het is dus duidelijk dat deze AI-modellen, om succesvol te kunnen zijn, eerst genoeg trainingsdata moeten hebben.
Het probleem stelt zich wanneer een eigenaar van een nieuwe installatie de energie van de volgende dag wilt voorspellen. Er is nog geen data, dus hoe kan de energie dan voorspeld worden? De eerste optie is om helemaal af te zien van AI-modellen en fysische modellen te gebruiken die werken met eerdergenoemde fysische vergelijkingen, verder fysische modellen genoemd. Alleen geloofde ik dat het nog beter kon. Wat als we ergens data vinden waar het model op kan trainen en zo die fysische modellen kan verslaan?
PVGIS: de ridder op het witte paard
PVGIS staat voor de Photovoltaic Geographical Information System. Het is een initiatief van de Europese Unie die op hun website meerdere tools aanbiedt met betrekking tot zonnepanelen. Een van deze tools is het voorzien van virtuele data voor een fictieve zonnepaneelinstallatie. Dit wil zeggen dat wanneer jij de locatie, oriëntatie en het aantal van de zonnepanelen meegeeft, de tool de energieproductie van die fictieve installatie geeft vanaf het jaar 2005-2020. Hiervoor gebruiken ze een publieke database vol met satellietbeelden. Aan de hand van deze satellietbeelden kan de hoeveelheid zonlicht bepaald worden en fysische modellen worden dan gebruikt om de hoeveelheid zonne-energie te schatten. In mijn thesis gebruikte ik deze virtuele data om het AI model te trainen. Deze techniek wordt transfer learning genoemd.
Transfer Learning
Transfer learning is een populaire techniek in onderzoek naar AI. Het is een techniek ontwikkeld om performante AI-modellen te verkrijgen voor een taak waar weinig of geen data voor beschikbaar is. Als er een taak bestaat die vergelijkbaar is met de taak die je wilt dat het model uitvoert en waar wel veel data voor beschikbaar is, kan je die data gebruiken om je model al de juiste richting uit te sturen. Je kan dit het beste vergelijken met een kind dat het veel makkelijker zal hebben om te leren fietsen met een normale fiets nadat hij eerst heeft leren fietsen met steunwieltjes of een driewieler. Dit is dus hetzelfde principe als voor transfer learning: een model dat al heeft getraind om bloemen in afbeeldingen te detecteren, zal ook beter zijn in het detecteren van rozen in vergelijking met een model dat op nog niets heeft getraind. Zo kan ook een AI-model, getraind op de virtuele data van PVGIS, de zonne-energie van een nieuwe installatie voorspellen. Alleen blijft de vraag of dit model het beter kan voorspellen dan de fysische modellen.
AI wint het gevecht tegen de fysica
Het antwoord op die vraag luidt volmondig ja. Na het AI-model en fysische model te vergelijken voor werkelijke zonnepaneelinstallaties blijkt dat het AI-model significant performanter is. De reden hiervoor is tweeledig. Ten eerste gebruiken beide modellen weersvoorspellingen van belangrijke parameters zoals temperatuur, hoeveelheid licht, enz. Elke persoon heeft al in levende lijve ondervonden dat weersvoorspellingen fouten kunnen bevatten. Hiervan zijn de meeste onmogelijk te voorspellen. Andere fouten zijn het gevolg van de manier waarop het weer voorspeld wordt. Het voordeel van een AI model is dat het deze fouten kan ontdekken wanneer het traint op virtuele data. Met enige kennis kan het model deze fouten corrigeren in de toekomst en helpen om energie van de nieuwe installatie nauwkeuriger te voorspellen. Ten tweede, recente literatuur bevestigt dat het complexe proces van de omzetting van licht naar elektriciteit beter gevat kan worden door AI-modellen dan door fysische modellen. Verder toont het onderzoek dat voorgaande data over een installatie niet noodzakelijk is. Er zijn echter nog enkele kanttekeningen die moeten gemaakt worden.
Wat nu?
Een meer accurate voorspelling leidt er toe dat het gebruik van andere elektrische apparaten beter kan afgestemd worden op het ritme van de zon. Zo wordt meer eigen zonne-energie gebruikt wat kan leiden tot een lagere energiefactuur. Verder onderzoek moet aanwijzen welk effect dit exact heeft op de factuur. Ook de open beschikbaarheid van kwaliteitsvolle data maakt onderzoek in dit domein moeilijk en vraagt zeker verdere aandacht.
Bibliografie
[1] J. S. Kikstra, Z. R. J. Nicholls, C. J. Smith, J. Lewis, R. D. Lamboll, E. Byers, M. Sandstad, M. Meinshausen, M. J. Gidden, J. Rogelj, E. Kriegler, G. P. Peters, J. S. Fuglestvedt, R. B. Skeie, B. H. Samset, L. Wienpahl, D. P. van Vuuren, K.-I. van der Wijst, A. Al Khourdajie, P. M. Forster, A. Reisinger, R. Schaeffer, and K. Riahi, “The IPCC Sixth Assessment Report WGIII climate assessment of mitigation pathways: From emissions to global temperatures,” Geoscientific Model Development, vol. 15, no. 24, pp. 9075–9109, Dec. 2022.
[2] “CO2 Emissions in 2022 – Analysis,” https://www.iea.org/reports/co2- emissions-in-2022, Mar. 2023.
[3] United Nations Environment Programme, Emissions Gap Report 2023: Broken Record – Temperatures Hit New Highs, yet World Fails to Cut Emissions (Again). United Nations Environment Programme, Nov. 2023.
[4] “Electricity Market Report 2023 – Analysis,” https://www.iea.org/reports/electricity-market-report-2023, Feb. 2023.
[5] “2023 Levelized Cost Of Energy+,” https://www.lazard.com/researchinsights/2023-levelized-cost-of-energyplus/.
[6] IEA, “Solar,” https://www.iea.org/energy-system/renewables/solar-pv.
[7] Winckelmans, “Fluvius grijpt in om uitval zonnepanelen te voorkomen | De Standaard,” De Standaard, Apr. 2024.
[8] ELIA, “CCMD,” https://www.eliagroup.eu/en/ccmd.
[9] “Capacity tariffs explained – gridX,” https://www.gridx.ai/knowledge/capacitytariffs, Apr. 24.
[10] V. Z. Gjorgievski, S. Cundeva, and G. E. Georghiou, “Social arrangements, technical designs and impacts of energy communities: A review,” Renewable Energy, vol. 169, pp. 1138–1156, May 2021.
[11] T. Hong, P. Pinson, Y. Wang, R. Weron, D. Yang, and H. Zareipour, “Energy Forecasting: A Review and Outlook,” IEEE Open Access Journal of Power and Energy, vol. 7, pp. 376–388, 2020. 75 Bibliography
[12] M. P. Almeida, O. Perpiñán, and L. Narvarte, “PV power forecast using a nonparametric PV model,” Solar Energy, vol. 115, pp. 354–368, May 2015.
[13] W. Khan, S. Walker, and W. Zeiler, “Improved solar photovoltaic energy generation forecast using deep learning-based ensemble stacking approach,” Energy, vol. 240, p. 122812, Feb. 2022.
[14] J. G. da Silva Fonseca Jr., T. Oozeki, T. Takashima, G. Koshimizu, Y. Uchida, and K. Ogimoto, “Use of support vector regression and numerically predicted cloudiness to forecast power output of a photovoltaic power plant in Kitakyushu, Japan,” Progress in Photovoltaics: Research and Applications, vol. 20, no. 7, pp. 874–882, 2012.
[15] S. E. Razavi, A. Arefi, G. Ledwich, G. Nourbakhsh, D. B. Smith, and M. Minakshi, “From Load to Net Energy Forecasting: Short-Term Residential Forecasting for the Blend of Load and PV Behind the Meter,” IEEE Access, vol. 8, pp. 224 343–224 353, 2020.
[16] K. Bot, A. Ruano, and M. d. G. Ruano, “Short-Term Forecasting Photovoltaic Solar Power for Home Energy Management Systems,” Inventions, vol. 6, no. 1, p. 12, Mar. 2021.
[17] R. P. Masini, M. C. Medeiros, and E. F. Mendes, “Machine learning advances for time series forecasting,” Journal of Economic Surveys, vol. 37, no. 1, pp. 76–111, 2023.
[18] K. Bi, L. Xie, H. Zhang, X. Chen, X. Gu, and Q. Tian, “Accurate mediumrange global weather forecasting with 3D neural networks,” Nature, vol. 619, no. 7970, pp. 533–538, Jul. 2023.
[19] W. Guo, L. Che, M. Shahidehpour, and X. Wan, “Machine-Learning based methods in short-term load forecasting,” The Electricity Journal, vol. 34, no. 1, p. 106884, Jan. 2021.
[20] A. Alcañiz, D. Grzebyk, H. Ziar, and O. Isabella, “Trends and gaps in photovoltaic power forecasting with machine learning,” Energy Reports, vol. 9, pp. 447–471, Dec. 2023.
[21] K. Zaouali, R. Rekik, and R. Bouallegue, “Deep Learning Forecasting Based on Auto-LSTM Model for Home Solar Power Systems,” in 2018 IEEE 20th International Conference on High Performance Computing and Communications; IEEE 16th International Conference on Smart City; IEEE 4th International Conference on Data Science and Systems (HPCC/SmartCity/DSS), Jun. 2018, pp. 235–242.
[22] A. Sharma and A. Kakkar, “Forecasting daily global solar irradiance generation using machine learning,” Renewable and Sustainable Energy Reviews, vol. 82, pp. 2254–2269, Feb. 2018. 76 Bibliography
[23] Z. Huang, W. Xu, and K. Yu, “Bidirectional LSTM-CRF Models for Sequence Tagging,” Aug. 2015.
[24] Y. Zhang, Q. Liu, and L. Song, “Sentence-State LSTM for Text Representation,” May 2018.
[25] L. Wang, M. Mao, J. Xie, Z. Liao, H. Zhang, and H. Li, “Accurate solar PV power prediction interval method based on frequency-domain decomposition and LSTM model,” Energy, vol. 262, p. 125592, Jan. 2023.
[26] J. Zhang, Y. Chi, and L. Xiao, “Solar Power Generation Forecast Based on LSTM,” in 2018 IEEE 9th International Conference on Software Engineering and Service Science (ICSESS), Nov. 2018, pp. 869–872.
[27] L. Visser, T. AlSkaif, and W. van Sark, “Operational day-ahead solar power forecasting for aggregated PV systems with a varying spatial distribution,” Renewable Energy, vol. 183, pp. 267–282, Jan. 2022.
[28] L. Visser, T. AlSkaif, J. Hu, A. Louwen, and W. Van Sark, “On the value of expert knowledge in estimation and forecasting of solar photovoltaic power generation,” Solar Energy, vol. 251, pp. 86–105, Feb. 2023.
[29] D. V. Pombo, M. J. Rincón, P. Bacher, H. W. Bindner, S. V. Spataru, and P. E. Sørensen, “Assessing stacked physics-informed machine learning models for co-located wind–solar power forecasting,” Sustainable Energy, Grids and Networks, vol. 32, p. 100943, Dec. 2022.
[30] E. Commission, “Photovoltaic Geographical Information System (PVGIS),” https://joint-research-centre.ec.europa.eu/photovoltaic-geographicalinformation-system-pvgis_en, Feb. 2024.
[31] R. M. Goody and Y. L. Yung, Atmospheric Radiation: Theoretical Basis. Oxford University Press, Dec. 1995.
[32] R. H. Inman, H. T. C. Pedro, and C. F. M. Coimbra, “Solar forecasting methods for renewable energy integration,” Progress in Energy and Combustion Science, vol. 39, no. 6, pp. 535–576, Dec. 2013.
[33] P. Lauret, R. Alonso-Suárez, J. Le Gal La Salle, and M. David, “Solar Forecasts Based on the Clear Sky Index or the Clearness Index: Which Is Better?” Solar, vol. 2, no. 4, pp. 432–444, Oct. 2022.
[34] D. Yang, W. Wang, C. A. Gueymard, T. Hong, J. Kleissl, J. Huang, M. J. Perez, R. Perez, J. M. Bright, X. Xia, D. Van Der Meer, and I. M. Peters, “A review of solar forecasting, its dependence on atmospheric sciences and implications for grid integration: Towards carbon neutrality,” Renewable and Sustainable Energy Reviews, vol. 161, p. 112348, Jun. 2022. 77 Bibliography
[35] Y. Chu, Y. Wang, D. Yang, S. Chen, and M. Li, “A review of distributed solar forecasting with remote sensing and deep learning,” Renewable and Sustainable Energy Reviews, vol. 198, p. 114391, Jul. 2024.
[36] A. G. Amillo, T. Huld, and R. Müller, “A New Database of Global and Direct Solar Radiation Using the Eastern Meteosat Satellite, Models and Validation,” Remote Sensing, vol. 6, no. 9, pp. 8165–8189, Sep. 2014.
[37] C. Voyant, G. Notton, S. Kalogirou, M.-L. Nivet, C. Paoli, F. Motte, and A. Fouilloy, “Machine learning methods for solar radiation forecasting: A review,” Renewable Energy, vol. 105, pp. 569–582, May 2017.
[38] G. Huang, Z. Li, X. Li, S. Liang, K. Yang, D. Wang, and Y. Zhang, “Estimating surface solar irradiance from satellites: Past, present, and future perspectives,” Remote Sensing of Environment, vol. 233, p. 111371, Nov. 2019.
[39] B. V, “Das Problem der Wettervorhersage, betrachtet vom Standpunkte der Mechanik und der Physik,” Meteor. Z., vol. 21, pp. 1–7, 1904.
[40] E. Lorenz, J. Hurka, D. Heinemann, and H. G. Beyer, “Irradiance Forecasting for the Power Prediction of Grid-Connected Photovoltaic Systems,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 2, no. 1, pp. 2–10, Mar. 2009.
[41] “Irradiance Data Methodology,” http://solcast.com/irradiance-datamethodology.
[42] P. A. Jimenez, J. P. Hacker, J. Dudhia, S. E. Haupt, J. A. Ruiz-Arias, C. A. Gueymard, G. Thompson, T. Eidhammer, and A. Deng, “WRF-Solar: Description and Clear-Sky Assessment of an Augmented NWP Model for Solar Power Prediction,” Bulletin of the American Meteorological Society, vol. 97, no. 7, pp. 1249–1264, Jul. 2016.
[43] M. Bai, P. Yao, H. Dong, Z. Fang, W. Jin, Xusheng Yang, J. Liu, and D. Yu, “Spatial-temporal characteristics analysis of solar irradiance forecast errors in Europe and North America,” Energy, vol. 297, p. 131187, Jun. 2024.
[44] F. J. L. Lima, F. R. Martins, E. B. Pereira, E. Lorenz, and D. Heinemann, “Forecast for surface solar irradiance at the Brazilian Northeastern region using NWP model and artificial neural networks,” Renewable Energy, vol. 87, pp. 807–818, Mar. 2016.
[45] K. Bakker, K. Whan, W. Knap, and M. Schmeits, “Comparison of statistical post-processing methods for probabilistic NWP forecasts of solar radiation,” Solar Energy, vol. 191, pp. 138–150, Oct. 2019.
[46] N. O. a. A. A. US Department of Commerce, “Breakthrough article on the First Climate Model,” https://celebrating200years.noaa.gov/breakthroughs/climate_model/. 78 Bibliography
[47] L. Fara, A. Diaconu, D. Craciunescu, and S. Fara, “Forecasting of Energy Production for Photovoltaic Systems Based on ARIMA and ANN Advanced Models,” International Journal of Photoenergy, vol. 2021, p. e6777488, Aug. 2021.
[48] S. I. Vagropoulos, G. I. Chouliaras, E. G. Kardakos, C. K. Simoglou, and A. G. Bakirtzis, “Comparison of SARIMAX, SARIMA, modified SARIMA and ANN-based models for short-term PV generation forecasting,” in 2016 IEEE International Energy Conference (ENERGYCON), Apr. 2016, pp. 1–6.
[49] D. Lee and K. Kim, “Deep Learning Based Prediction Method of Long-term Photovoltaic Power Generation Using Meteorological and Seasonal Information,” Journal of Society for e-Business Studies, vol. 24, no. 1, Mar. 2019.
[50] E. Kim, M. S. Akhtar, and O.-B. Yang, “Designing solar power generation output forecasting methods using time series algorithms,” Electric Power Systems Research, vol. 216, p. 109073, Mar. 2023.
[51] S. Atique, S. Noureen, V. Roy, S. Bayne, and J. Macfie, “Time series forecasting of total daily solar energy generation: A comparative analysis between ARIMA and machine learning techniques,” in 2020 IEEE Green Technologies Conference(GreenTech), Apr. 2020, pp. 175–180.
[52] B. Li, X. Chen, and A. Jain, “Enhancing Power Prediction of Photovoltaic Systems: Leveraging Dynamic Physical Model for Irradiance-to-Power Conversion,” Feb. 2024.
[53] M. J. Mayer and D. Yang, “Pairing ensemble numerical weather prediction with ensemble physical model chain for probabilistic photovoltaic power forecasting,” Renewable and Sustainable Energy Reviews, vol. 175, p. 113171, Apr. 2023.
[54] D. Markovics and M. J. Mayer, “Comparison of machine learning methods for photovoltaic power forecasting based on numerical weather prediction,” Renewable and Sustainable Energy Reviews, vol. 161, p. 112364, Jun. 2022.
[55] A. Gensler, J. Henze, B. Sick, and N. Raabe, “Deep Learning for solar power forecasting — An approach using AutoEncoder and LSTM Neural Networks,” in 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC). Budapest, Hungary: IEEE, Oct. 2016, pp. 002 858–002 865.
[56] M. Massaoudi, I. Chihi, H. Abu-Rub, S. S. Refaat, and F. S. Oueslati, “Convergence of Photovoltaic Power Forecasting and Deep Learning: State-of-Art Review,” IEEE Access, vol. 9, pp. 136 593–136 615, 2021.
[57] W. Liu and Z. Mao, “Short-term photovoltaic power forecasting with feature extraction and attention mechanisms,” Renewable Energy, vol. 226, p. 120437, May 2024. 79 Bibliography
[58] K. Wang, X. Qi, and H. Liu, “A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network,” Applied Energy, vol. 251, p. 113315, Oct. 2019.
[59] R. A. Rajagukguk, R. A. A. Ramadhan, and H.-J. Lee, “A Review on Deep Learning Models for Forecasting Time Series Data of Solar Irradiance and Photovoltaic Power,” Energies, vol. 13, no. 24, p. 6623, Jan. 2020.
[60] A. Agga, A. Abbou, M. Labbadi, Y. E. Houm, and I. H. Ou Ali, “CNNLSTM: An efficient hybrid deep learning architecture for predicting short-term photovoltaic power production,” Electric Power Systems Research, vol. 208, p. 107908, Jul. 2022.
[61] M. Zounemat-Kermani, O. Batelaan, M. Fadaee, and R. Hinkelmann, “Ensemble machine learning paradigms in hydrology: A review,” Journal of Hydrology, vol. 598, p. 126266, Jul. 2021.
[62] M. AlKandari and I. Ahmad, “Solar power generation forecasting using ensemble approach based on deep learning and statistical methods,” Applied Computing and Informatics, vol. ahead-of-print, no. ahead-of-print, Jan. 2020.
[63] D. Chakraborty, J. Mondal, H. B. Barua, and A. Bhattacharjee, “Computational solar energy – Ensemble learning methods for prediction of solar power generation based on meteorological parameters in Eastern India,” Renewable Energy Focus, vol. 44, pp. 277–294, Mar. 2023.
[64] O. Boussif, G. Boukachab, D. Assouline, S. Massaroli, T. Yuan, L. Benabbou, and Y. Bengio, “Improving day-ahead Solar Irradiance Time Series Forecasting by Leveraging Spatio-Temporal Context,” Oct. 2023.
[65] L. Yang and A. Shami, “On hyperparameter optimization of machine learning algorithms: Theory and practice,” Neurocomputing, vol. 415, pp. 295–316, Nov. 2020.
[66] M. F. Tahir, M. Z. Yousaf, A. Tzes, M. S. El Moursi, and T. H. M. El-Fouly, “Enhanced solar photovoltaic power prediction using diverse machine learning algorithms with hyperparameter optimization,” Renewable and Sustainable Energy Reviews, vol. 200, p. 114581, Aug. 2024.
[67] F. Zhuang, Z. Qi, K. Duan, D. Xi, Y. Zhu, H. Zhu, H. Xiong, and Q. He, “A Comprehensive Survey on Transfer Learning,” Proceedings of the IEEE, vol. 109, no. 1, pp. 43–76, Jan. 2021.
[68] K. Weiss, T. M. Khoshgoftaar, and D. Wang, “A survey of transfer learning,” Journal of Big Data, vol. 3, no. 1, p. 9, May 2016.
[69] S. J. Pan and Q. Yang, “A Survey on Transfer Learning,” IEEE Transactions on Knowledge and Data Engineering, vol. 22, no. 10, pp. 1345–1359, Oct. 2010. 80 Bibliography
[70] M. Weber, M. Auch, C. Doblander, P. Mandl, and H.-A. Jacobsen, “Transfer Learning With Time Series Data: A Systematic Mapping Study,” IEEE Access, vol. 9, pp. 165 409–165 432, 2021.
[71] M. Abubakr, B. Akoush, A. Khalil, and M. A. Hassan, “Unleashing deep neural network full potential for solar radiation forecasting in a new geographic location with historical data scarcity: A transfer learning approach,” The European Physical Journal Plus, vol. 137, no. 4, p. 474, Apr. 2022.
[72] H. Sheng, B. Ray, J. Shao, D. Lasantha, and N. Das, “Generalization of solar power yield modeling using knowledge transfer,” Expert Systems with Applications, vol. 201, p. 116992, Sep. 2022.
[73] Y. Tang, K. Yang, S. Zhang, and Z. Zhang, “Photovoltaic power forecasting: A hybrid deep learning model incorporating transfer learning strategy,” Renewable and Sustainable Energy Reviews, vol. 162, p. 112473, Jul. 2022.
[74] A. Genovese, V. Bernardoni, V. Piuri, F. Scotti, and F. Tessore, “Photovoltaic Energy Prediction for New-Generation Cells with Limited Data: A Transfer Learning Approach,” in 2022 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), May 2022, pp. 1–6.
[75] Z. Tang, Y. Tang, A. Qiao, J. Liu, and J. Gao, “Transfer Learning Based Photovoltaic Power Forecasting with XGBoost,” in 2023 Panda Forum on Power and Energy (PandaFPE), Apr. 2023, pp. 1781–1785.
[76] E. Sarmas, N. Dimitropoulos, V. Marinakis, Z. Mylona, and H. Doukas, “Transfer learning strategies for solar power forecasting under data scarcity,” Scientific Reports, vol. 12, no. 1, p. 14643, Aug. 2022.
[77] T. Zhang, K. Zheng, and Q. Chen, “Short-Term Photovoltaic Prediction Based on Transfer Learning and LSTM,” in 2023 IEEE 7th Conference on Energy Internet and Energy System Integration (EI2), Dec. 2023, pp. 3438–3443.
[78] G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, and L. Yang, “Physics-informed machine learning,” Nature Reviews Physics, vol. 3, no. 6, pp. 422–440, May 2021.
[79] K. Kashinath, M. Mustafa, A. Albert, J.-L. Wu, C. Jiang, S. Esmaeilzadeh, K. Azizzadenesheli, R. Wang, A. Chattopadhyay, A. Singh, A. Manepalli, D. Chirila, R. Yu, R. Walters, B. White, H. Xiao, H. A. Tchelepi, P. Marcus, A. Anandkumar, P. Hassanzadeh, and n. Prabhat, “Physics-informed machine learning: Case studies for weather and climate modelling,” Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 379, no. 2194, p. 20200093, Feb. 2021.
[80] G. S. Misyris, A. Venzke, and S. Chatzivasileiadis, “Physics-Informed Neural Networks for Power Systems,” in 2020 IEEE Power & Energy Society General Meeting (PESGM). Montreal, QC, Canada: IEEE, Aug. 2020, pp. 1–5. 81 Bibliography
[81] Q. Paletta, Y. Nie, Y.-M. Saint-Drenan, and B. Le Saux, “Improving cross-site generalisability of vision-based solar forecasting models with physics-informed transfer learning,” Energy Conversion and Management, vol. 309, p. 118398, Jun. 2024.
[82] P. G. Loutzenhiser, H. Manz, C. Felsmann, P. A. Strachan, T. Frank, and G. M. Maxwell, “Empirical validation of models to compute solar irradiance on inclined surfaces for building energy simulation,” Solar Energy, vol. 81, no. 2, pp. 254–267, Feb. 2007.
[83] R. Perez, R. Seals, P. Ineichen, R. Stewart, and D. Menicucci, “A New Simplified version of the Perez diffuse irradiance model for tilted surfaces,” Solar Energy, vol. 39, pp. 221–231, Dec. 1987.
[84] P. Ineichen, R. R. Perez, R. D. Seal, E. L. Maxwell, and A. Zalenka, “Dynamic global-to-direct irradiance conversion models,” ASHRAE transactions, vol. 98, no. 1, p. 354, 1992.
[85] E. L. Maxwell, “A quasi-physical model for converting hourly global horizontal to direct normal insolation,” Solar Energy Research Inst., Golden, CO (USA), Tech. Rep. SERI/TR-215-3087, Aug. 1987.
[86] M. K. Fuentes, “’ A Simplified Thermal Model for Flat-Plate Photovoltaic Arrays.”
[87] T. A. Olukan and M. Emziane, “A Comparative Analysis of PV Module Temperature Models,” Energy Procedia, vol. 62, pp. 694–703, Jan. 2014.
[88] K. S. Anderson, C. W. Hansen, W. F. Holmgren, A. R. Jensen, M. A. Mikofski, and A. Driesse, “Pvlib python: 2023 project update,” Journal of Open Source Software, vol. 8, no. 92, p. 5994, Dec. 2023.
[89] P. Ineichen and R. Perez, “A new airmass independent formulation for the Linke turbidity coefficient,” Solar Energy, vol. 73, no. 3, pp. 151–157, Sep. 2002.
[90] F. Antonanzas-Torres, R. Urraca, J. Polo, O. Perpiñán-Lamigueiro, and R. Escobar, “Clear sky solar irradiance models: A review of seventy models,” Renewable and Sustainable Energy Reviews, vol. 107, pp. 374–387, Jun. 2019.
[91] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, May 2015.
[92] Y. Yu, X. Si, C. Hu, and J. Zhang, “A Review of Recurrent Neural Networks: LSTM Cells and Network Architectures,” Neural Computation, vol. 31, no. 7, pp. 1235–1270, Jul. 2019.
[93] S. Hochreiter and J. Schmidhuber, “Long Short-term Memory,” Neural computation, vol. 9, pp. 1735–80, Dec. 1997. 82 Bibliography
[94] Y. Tian and Y. Zhang, “A comprehensive survey on regularization strategies in machine learning,” Information Fusion, vol. 80, pp. 146–166, Apr. 2022.
[95] T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama, “Optuna: A Nextgeneration Hyperparameter Optimization Framework,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, ser. KDD ’19. New York, NY, USA: Association for Computing Machinery, Jul. 2019, pp. 2623–2631.
[96] D. Berrar, “Cross-Validation,” in Encyclopedia of Bioinformatics and Computational Biology. Elsevier, 2019, pp. 542–545.
[97] M. Schnaubelt, “A comparison of machine learning model validation schemes for non-stationary time series data,” FAU Discussion Papers in Economics, Working Paper 11/2019, 2019.
[98] A. Dobos, “PVWatts Version 5 Manual,” National Labratory of U.S. Department of Energy, Golden, Technical Report NREL/TP-6A20-62641, 1158421, Sep. 2014.
[99] J. Kratochvil, W. Boyson, and D. King, “Photovoltaic array performance model.” Sandia National Laboratories, Albuquerque, Tech. Rep. SAND2004- 3535, 919131, Aug. 2004.
[100] T. Huld, G. Friesen, A. Skoczek, R. P. Kenny, T. Sample, M. Field, and E. D. Dunlop, “A power-rating model for crystalline silicon PV modules,” Solar Energy Materials and Solar Cells, vol. 95, no. 12, pp. 3359–3369, Dec. 2011.
[101] I. de la Parra, M. Muñoz, E. Lorenzo, M. García, J. Marcos, and F. MartínezMoreno, “PV performance modelling: A review in the light of quality assurance for large PV plants,” Renewable and Sustainable Energy Reviews, vol. 78, pp. 780–797, Oct. 2017.
[102] D. Yang, S. Alessandrini, J. Antonanzas, F. Antonanzas-Torres, V. Badescu, H. G. Beyer, R. Blaga, J. Boland, J. M. Bright, C. F. M. Coimbra, M. David, Â. Frimane, C. A. Gueymard, T. Hong, M. J. Kay, S. Killinger, J. Kleissl, P. Lauret, E. Lorenz, D. van der Meer, M. Paulescu, R. Perez, O. PerpiñánLamigueiro, I. M. Peters, G. Reikard, D. Renné, Y.-M. Saint-Drenan, Y. Shuai, R. Urraca, H. Verbois, F. Vignola, C. Voyant, and J. Zhang, “Verification of deterministic solar forecasts,” Solar Energy, vol. 210, pp. 20–37, Nov. 2020.
[103] “What Is Random Forest?” https://www.ibm.com/topics/random-forest, Oct. 2021.
[104] F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, and É. Duchesnay, “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, no. 85, pp. 2825–2830, 2011. 83 Bibliography
[105] A. Jain, H. Patel, L. Nagalapatti, N. Gupta, S. Mehta, S. Guttula, S. Mujumdar, S. Afzal, R. Sharma Mittal, and V. Munigala, “Overview and Importance of Data Quality for Machine Learning Tasks,” in Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. Virtual Event CA USA: ACM, Aug. 2020, pp. 3561–3562.
[106] G. P. Compo, J. S. Whitaker, P. D. Sardeshmukh, N. Matsui, R. J. Allan, X. Yin, B. E. Gleason, R. S. Vose, G. Rutledge, P. Bessemoulin, S. Brönnimann, M. Brunet, R. I. Crouthamel, A. N. Grant, P. Y. Groisman, P. D. Jones, M. C. Kruk, A. C. Kruger, G. J. Marshall, M. Maugeri, H. Y. Mok, Ø. Nordli, T. F. Ross, R. M. Trigo, X. L. Wang, S. D. Woodruff, and S. J. Worley, “The Twentieth Century Reanalysis Project,” Quarterly Journal of the Royal Meteorological Society, vol. 137, no. 654, pp. 1–28, 2011.
[107] H. Hersbach, B. Bell, P. Berrisford, S. Hirahara, A. Horányi, J. Muñoz-Sabater, J. Nicolas, C. Peubey, R. Radu, D. Schepers, A. Simmons, C. Soci, S. Abdalla, X. Abellan, G. Balsamo, P. Bechtold, G. Biavati, J. Bidlot, M. Bonavita, G. De Chiara, P. Dahlgren, D. Dee, M. Diamantakis, R. Dragani, J. Flemming, R. Forbes, M. Fuentes, A. Geer, L. Haimberger, S. Healy, R. J. Hogan, E. Hólm, M. Janisková, S. Keeley, P. Laloyaux, P. Lopez, C. Lupu, G. Radnoti, P. de Rosnay, I. Rozum, F. Vamborg, S. Villaume, and J.-N. Thépaut, “The ERA5 global reanalysis,” Quarterly Journal of the Royal Meteorological Society, vol. 146, no. 730, pp. 1999–2049, 2020.
[108] S. Rasp, S. Hoyer, A. Merose, I. Langmore, P. Battaglia, T. Russel, A. SanchezGonzalez, V. Yang, R. Carver, S. Agrawal, M. Chantry, Z. B. Bouallegue, P. Dueben, C. Bromberg, J. Sisk, L. Barrington, A. Bell, and F. Sha, “WeatherBench 2: A benchmark for the next generation of data-driven global weather models,” Aug. 2023.
[109] J. Muñoz-Sabater, E. Dutra, A. Agustí-Panareda, C. Albergel, G. Arduini, G. Balsamo, S. Boussetta, M. Choulga, S. Harrigan, H. Hersbach, B. Martens, D. G. Miralles, M. Piles, N. J. Rodríguez-Fernández, E. Zsoter, C. Buontempo, and J.-N. Thépaut, “ERA5-Land: A state-of-the-art global reanalysis dataset for land applications,” Earth System Science Data, vol. 13, no. 9, pp. 4349–4383, Sep. 2021.
[110] Copernicus Climate Change Service, “Complete ERA5 global atmospheric reanalysis,” 2023.
[111] ECMWF, “TIGGE archive,” https://confluence.ecmwf.int/display/TIGGE.
[112] R. Swinbank, M. Kyouda, P. Buchanan, L. Froude, T. M. Hamill, T. D. Hewson, J. H. Keller, M. Matsueda, J. Methven, F. Pappenberger, M. Scheuerer, H. A. Titley, L. Wilson, and M. Yamaguchi, “The TIGGE Project and Its Achievements,” Bulletin of the American Meteorological Society, vol. 97, no. 1, pp. 49–67, Jan. 2016. 84 Bibliography
[113] R. Hagedorn, R. Buizza, T. M. Hamill, M. Leutbecher, and T. N. Palmer, “Comparing TIGGE multimodel forecasts with reforecast-calibrated ECMWF ensemble forecasts,” Quarterly Journal of the Royal Meteorological Society, vol. 138, no. 668, pp. 1814–1827, 2012.
[114] O. Antonić, J. Križan, A. Marki, and D. Bukovec, “Spatio-temporal interpolation of climatic variables over large region of complex terrain using neural networks,” Ecological Modelling, vol. 138, no. 1, pp. 255–263, Mar. 2001.
[115] Centre for Environmental Data Analysis, “NWP-Euro: Operational Numerical Weather Prediction (NWP) output from the European Atmospheric High Resolution Model; part of the Met Office Unified Model (UM).”
[116] ——, “NWP-Global: Operational Numerical Weather Prediction (NWP) output from the UK Met Office Global Atmospheric Unified Model (UM). Centre for Environmental Data Analysis.”
[117] D. Reinert, Prill, H. Frank, M. Denhard, M. Baldauf, C. Schraff, C. Gebhardt, C. Marsigli, J. Förstner, G. Zängl, and L. Schlemmer, “DWD Database Reference for the Global and Regional ICON and ICON-EPS Forecasting Systemb,” Deutscher Wetterdienst, Offenbach Am Main, Tech. Rep., Apr. 2024.
[118] DWD, “NWP forecast data,” https://www.dwd.de/.
[119] “Open Climate Fix,” https://openclimatefix.org/.
[120] Open Climate Fix, “Dwd-icon-global.”
[121] ——, “Dwd-icon-eu.”
[122] “Dissemination schedule - Data and Charts - ECMWF Confluence Wiki,” https://confluence.ecmwf.int/display/DAC/Dissemination+schedule.
[123] G. Lentze, “Model upgrade increases skill and unifies medium-range resolutions,” https://www.ecmwf.int/en/about/media-centre/news/2023/modelupgrade-increases-skill-and-unifies-medium-range-resolutions, Jun. 2023.
[124] “ECMWF,” https://www.ecmwf.int/, Mar. 2024.
[125] R. Lam, A. Sanchez-Gonzalez, M. Wilson, P. Wirnsberger, M. Fortunato, F. Alet, S. Ravuri, T. Ewalds, Z. Eaton-Rosen, W. Hu, A. Merose, S. Hoyer, G. Holland, O. Vinyals, J. Stott, A. Pritzel, S. Mohammed, and P. Battaglia, “Learning skillful medium-range global weather forecasting,” https://www.science.org/doi/10.1126/science.adi2336.
[126] C. Skamarock, B. Klemp, J. Dudhia, O. Gill, Z. Liu, J. Berner, W. Wang, G. Powers, G. Duda, D. Barker, and X.-y. Huang, “A Description of the Advanced Research WRF Model Version 4.3,” 2021. 85 Bibliography
[127] M. Optis, A. Kumler, G. Scott, M. Debnath, and P. Moriarty, “Validation of RU-WRF, the Custom Atmospheric Mesoscale Model of the Rutgers Center for Ocean Observing Leadership,” Tech. Rep. NREL/TP–5000-75209, 1599576, Feb. 2020.
[128] B. K. Blaylock, “Herbie: Retrieve Numerical Weather Prediction Model Data,” Zenodo, May 2024.
[129] Open Climate Fix, “Uk_pv.”
[130] A. Alcañiz, A. V. Lindfors, M. Zeman, H. Ziar, and O. Isabella, “Effect of Climate on Photovoltaic Yield Prediction Using Machine Learning Models,” Global Challenges, vol. 7, no. 1, p. 2200166, 2023.
[131] “PV power databases,” https://www.tudelft.nl/ewi/over-defaculteit/afdelingen/electrical-sustainable-energy/photovoltaic-materialsand-devices/dutch-pv-portal/pv-power-databases.
[132] P. Zippenfenig, “Open-Meteo.com Weather API,” Zenodo, Jan. 2024.
[133] KMNI, “Code rood voor sneeuwstorm Darcy,” https://www.knmi.nl/over-hetknmi/nieuws/code-rood-voor-seeuwstorm-darcy, Feb. 2021. [134] M. Office, “Severe winter weather and storm Darcy, February 2021,” Met Office, London, Tech. Rep., Feb. 2021.
