In this paper, solar irradiance short-term forecasts were performed considering time horizons ranging from 5 min to 30 min, under a 5 min timestep. Global Horizontal Irradiance (GHI) and Direct Normal Irradiance (DNI) were computed using Deep Neural Networks with One-Dimensional Convolution (CNN-1D), Long-Short Term Memory (LSTM) and CNN-LSTM layers on the benchmarking dataset FOLSOM, which is formed by predictors obtained by recursive functions on the clear sky index time series and statistical attributes extracted from images collected by a camera pointed to the zenith, characterizing endogenous and exogenous variables, respectively. To analyze the endogenous predictors influence on the accuracy of the networks, the performance was evaluated for the cases with and without them. This analysis is motivated, to our best knowledge, by the lack of works that cite the FOLSOM dataset using deep learning models, and it is necessary to verify the impact of the endogenous and exogenous predictors in the forecasts results for this specific approach. The accuracy of the networks was evaluated by the metrics Mean Absolute Error (MAE), Mean Bias Error (MBE), Root Mean Squared Error (RMSE), Relative root-mean-squared error (rRMSE), Determination Coefficient (R2) and Forecast Skill (s). The network architectures using isolated CNN-1D and LSTM layers generally performed better. The best accuracy was obtained by the CNN-1D network for a horizon of 10 min ahead reaching an RMSE of 36.24 W/m2, improving 11.15% on this error metric compared to the persistence model.
The world population is increasingly aware of environmental issues involving our planet. In search of solutions that will contribute to the conservation of the environment, you can use the renewable energies that are in great expansion, being the use of solar energy a solution for this purpose. This study aimed to evaluate the energy absorption capacity of a hybrid silver nanofluid and titanium dioxide used for direct solar absorption in a "Solar Wall." The experiments performed were exposed for 16 hours for several days. Various concentrations were analyzed, where a concentration of 23.2 ppm titanium dioxide was set, and silver varied to 0.40625 ppm, 0.8125 ppm, 1.625 ppm, and 3.25 ppm corresponding to a 3% molar fraction. 6%, 12% and 25% respectively. An analysis of the temperature profile was made, which showed a better utilization of the sample with a 12% molar fraction that obtained a temperature gain of 8.3 ° C corresponding to a 14.9% gain. An analysis of the stored energy ratio was analyzed, observing a good response of nanofluids at the early solar incidence and a maximum stored energy for the sample of 6%. The other metric analyzed was the specific absorption rate that reached a maximum value of 0.009293 KW / g. The work showed that for the preparation of hybrid nanofluid, not always increasing the silver concentration implied better results and through temperature profile analysis and energy analysis metrics, it was concluded that a silver concentration of 0.8125 ppm together With a concentration of 23.2 ppm (6% molar fraction) better gains are obtained for the use of nanofluid in the solar wall.
Topics related to the modeling of turbulent flow feature significant relevance in several areas, especially in engineering, since the vast majority of flows present in the design of devices and systems are characterized to be turbulent. A vastly applied tool for the analysis of such flows is the use of numerical simulations based on turbulence models. Thus, this work aims to evaluate the performance of several turbulence models when applied to classic problems of fluid mechanics and heat transfer, already extensively validated by empirical procedures. The OpenFOAM open source software was used, being highly suitable for obtaining numerical solutions to problems of fluid mechanics involving complex geometries. The problems for the evaluation of turbulence models selected were: two-dimensional cavity, Pitz-Daily, air flow over an airfoil, air flow over the Ahmed blunt body and the problem of natural convection between parallel plates. The solution to such problems was achieved by utilizing several Reynolds Averaged Equations (RANS) turbulence models, namely: k-ε, k-ω, Lam-Bremhorst k-ε, k-ω SST, Lien-Leschziner k-ε, Spalart-Allmaras, Launder-Sharma k-ε, renormalization group (RNG) k-ε. The results obtained were compared to those found in the literature which were empirically obtained, thus allowing the assessment of the strengths and weaknesses of the turbulence modeling applied in each problem.
Neste trabalho, previsões da média diária de irradiação solar global foram obtidas pela aplicação de algoritmos de aprendizagem de máquina em dois conjuntos de dados formados por variáveis exógenas (insolação, temperatura do ar, precipitação, etc), variáveis endógenas (série temporal da média diária de irradiação solar global) e variáveis temporais (ano, mês e dia da medição). A diferença entre os conjuntos de dados está relacionada ao fato de que em um se considera as intensidades dos fenômenos climáticos do El Niño e da La Niña como preditores para os modelos de aprendizagem utilizados, enquanto no outro não se considera. Desta forma, foi possível avaliar se a adição do preditor relacionado ao El Niño/La Niña contribui para uma melhor acurácia de previsão por parte dos modelos aplicados: Máquina de Aprendizagem Mínima, Regressão por Vetor Suporte, Florestas Aleatórias, K-Vizinhos mais Próximos e uma árvore de regressão com o uso de Bootstrap. As métricas de erro Erro Médio Absoluto, Erro de Viés Médio, Raiz do Erro Quadrático Médio, Raiz do Erro Quadrático Médio Relativo e Habilidade de Previsão foram utilizadas para a análise do desempenho dos algoritmos. A média aritmética da Raiz do Erro Quadrático Médio e da Habilidade de Previsão para o caso em que se considerou o El Niño/La Niña como atibutos foram de 40.78 W/m² e 7,87% , respectivamente. Já para o caso em que não se considera tais preditores os valores obtidos foram de 40.86 W/m² e 7.69%. Indicando que o uso destes preditores aumenta a acurácia de previsão dos algoritmos em questão.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.