Method for solar resource assessment using numerical weather prediction and artificial neural network models based on typical meteorological data: Application to the south of Portugal

• Published in: Solar energy 2022-04, Vol.236, p.225-238
• Main Authors: Pereira, Sara, Abreu, Edgar F.M., Iakunin, Maksim, Cavaco, Afonso, Salgado, Rui, Canhoto, Paulo
• Format: Article
• Language: English
• Subjects: Adaptation; Artificial neural network; Artificial neural networks; Atmospheric models; Atmospheric monitoring; Bias; Climate models; Errors; Meteorological data; Neural networks; Numerical prediction; Numerical weather forecasting; Numerical weather prediction; Project planning; Radiation; Radiation measurement; Resource availability; Short wave radiation; Simulation; Site adaptation; Solar energy; Solar radiation; Systems design; Typical meteorological year; Weather forecasting
• ISSN: 0038-092X; 1471-1257
• DOI: 10.1016/j.solener.2022.03.003

•Method for improving solar radiation simulations from weather prediction models.•Artificial neural network models used as site adaptation method.•Site adaptation of solar resource that includes aerosol reanalysis data.•Method for solar resource assessment based on typical meteorological data.•Method validated for regional solar resource assessment in the south of Portugal. In this work a method for regional solar resource assessment based on numerical weather prediction (NWP) and artificial neural network (ANN) models is presented. The method was developed using typical meteorological and solar radiation data and applied to the location of Évora, Portugal with the goal of assessing solar global horizontal (GHI) and direct normal (DNI) irradiations with 1.25 km of horizontal resolution in the south of Portugal. The NWP model used was the research model Meso-NH and a site adaptation model was developed based in ANNs and using as inputs the simulated meteorological variables from Meso-NH and aerosol data from Copernicus Atmospheric Monitoring Services (CAMS) for the observation site. The resulting annual relative mean bias errors for GHI and DNI at Évora and typical meteorological year are of 0.55 % and 0.98 %, respectively, while the values for the original Meso-NH simulations are of 8.24 % for GHI and 31.71 % for DNI. The developed site adaptation model is applied to the region for the purpose of solar radiation assessment and validated using data from a network of solar radiation measuring stations scattered throughout the south of Portugal, showing relative mean bias errors of 2.34 % for GHI and 3.41 % for DNI, while the original Meso-NH simulations presents relative mean bias errors of 8.50 % and 29.54 %, respectively. These results allowed the generation of improved solar resource availability maps which are a very useful tool in solar resource assessment, the study of shortwave radiative climate, as well as project planning and solar system design and operation.