相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Computer Science, Hardware & Architecture
Zahraa E. Mohamed et al.
Summary: Measurements of solar radiation significantly impact energy output ratios. Due to cost and measurement difficulties, many countries have developed alternative methods to estimate solar radiation. The advancement of machine learning algorithms provides an opportunity for improved accuracy.
Article
Meteorology & Atmospheric Sciences
Jared a. Lee et al.
Summary: Aerosol optical depth (AOD) is a major source of solar irradiance forecast error in clear-sky conditions. Improving the accuracy of AOD in NWP models like WRF can reduce errors in direct normal irradiance (DNI) and global horizontal irradiance (GHI), thereby improving solar power forecast accuracy in cloud-free conditions. This study analyzed clear-sky GHI and DNI using different aerosol representations in the WRF-Solar model, and compared them with high-quality irradiance observations. The results showed that WRF-Solar with GEOS-5 AOD had the lowest errors in clear-sky DNI, while WRF-Solar with CAMS AOD had the highest errors. Clear-sky GHI statistics did not differ much among the four models.
JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY
(2023)
Article
Environmental Sciences
Yunbo Lu et al.
Summary: This study developed and compared five high-speed and highly accurate hybrid models to simulate diffuse radiation in different regions of China. By combining multiple data sources and using various machine learning models, the proposed RTM-RF method exhibited superior performance. The uncertainty caused by the measurement errors of aerosol optical properties and land surface albedo was quantified, and the relative contributions of multiple variables to diffuse radiation were analyzed. AOD, SZA, and single-scattering albedo were found to contribute significantly more than other variables.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Computer Science, Information Systems
Hyojeoung Kim et al.
Summary: The renewable energy industry is expanding rapidly as a result of environmental pollution and rising prices of fossil fuels. Solar energy, which constitutes the largest proportion of renewable energy, is particularly affected by weather and climate change, making climate prediction essential. However, the Korea Meteorological Administration does not currently predict solar radiation, which is closely related to solar power, necessitating the need for solar radiation prediction technology. This study predicts solar radiation using extra-atmospheric solar radiation and three weather variables, comparing the performance of different single models in machine and deep learning.
Article
Geochemistry & Geophysics
Shailee Patel et al.
Summary: The study found that a high-resolution coupled physical-ecosystem model simulations can resolve most of the known surface and subsurface features of the Bay of Bengal, but with differences in salinity and chlorophyll-a, and negligible differences in temperature. The modeled chlorophyll-a concentration is correlated with nutrient supply to the base of the euphotic layer and mixed layer in the Bay of Bengal.
Article
Computer Science, Information Systems
Maozheng Pi et al.
Summary: In this study, a multichannel deep learning model named MC-WT-CBiLSTM is proposed to improve the prediction accuracy of global level irradiance. The model combines wavelet transform, convolutional neural network, and bidirectional long short-term memory and shows obvious advantages in predicting various time horizons.
WIRELESS COMMUNICATIONS & MOBILE COMPUTING
(2022)
Article
Energy & Fuels
Marcello Anderson F. B. Lima et al.
Summary: Solar energy is a significant renewable energy source that can contribute to global energy demand. However, its intermittent nature makes it challenging to integrate into the electrical system. In this study, we compared two machine learning techniques, deep learning (DL) and support vector regression (SVR), for solar forecasting. Our testing in Spain found that DL achieved a mean absolute percentage error of 7.9%, while SVR achieved 8.52%. Although DL had the best results, it is worth mentioning that SVR also performed well.
JOURNAL OF SOLAR ENERGY ENGINEERING-TRANSACTIONS OF THE ASME
(2022)
Article
Green & Sustainable Science & Technology
Dongyu Jia et al.
Summary: This study assessed three machine learning models for predicting global and diffuse solar radiation in eight Chinese cities. The results showed that coastal locations had higher prediction error values compared to inland locations. The SVM model outperformed the other models in all locations, followed by GLMNET and RF. The accuracy of solar radiation prediction was closely related to weather and pollution conditions.
Article
Green & Sustainable Science & Technology
Muhammed A. Hassan et al.
Summary: This study presents two novel ensemble forecasters that combine time delay and recurrent neural networks to predict the global horizontal irradiance, direct normal irradiance, and ambient temperature. The developed models outperformed persistence ones and showed comparable performances between daily and instantaneously calibrated models in different plants.
Article
Meteorology & Atmospheric Sciences
Chunlin Huang et al.
Summary: In this study, a system for estimating and nowcasting surface solar irradiance (SSI) using the FY-4A satellite was established. The system combines a hybrid estimation method and a cloud motion vector prediction model. Evaluation results based on measurements from a radiation station in China show that the system has high accuracy in estimating and forecasting global horizontal irradiance (GHI) and direct normal irradiance (DNI).
ADVANCES IN ATMOSPHERIC SCIENCES
(2022)
Article
Green & Sustainable Science & Technology
Xinghong Cheng et al.
Summary: This study improves the simulation of solar radiation by using satellite-based aerosol optical depth (AOD) data and investigates the impact of winter haze on solar radiation. The results show that the direct application of satellite-based AOD data effectively enhances the simulation of solar radiation.
Article
Green & Sustainable Science & Technology
Muhammad Sibtain et al.
Summary: This study develops novel hybrid prediction models for accurate prediction of global horizontal irradiance. The VMD-STA-525 hybrid model shows the highest prediction efficiency and the lowest error among all the models tested.
Article
Energy & Fuels
Chu Zhang et al.
Summary: A novel solar radiation prediction model based on wavelet transform, complete ensemble empirical mode decomposition with adaptive noise, improved atom search optimization and outlier-robust extreme learning machine is proposed in this study. The model utilizes denoising, decomposition, and optimization techniques to improve the accuracy and robustness of solar radiation prediction.
Article
Energy & Fuels
Hui-Min Zuo et al.
Summary: In this study, a short-term solar irradiance prediction model based on a deep learning network was proposed. By analyzing the relationship between cloud coverage and future solar irradiance, combined with meteorological parameters and historical data, the model achieved better performance compared to other models.
Article
Meteorology & Atmospheric Sciences
Zhigang Li et al.
Summary: Using a random forest model, this study analyzed solar radiation data from seven stations in the North China Plain. The results showed that total cloud cover and aerosol optical depth were the most important factors affecting global and diffuse radiation. The study also explored the interactions between boundary layer height, aerosols, and solar radiation in urban areas. Additionally, the joint analysis revealed non-linear relationships between solar radiation and input feature pairs.
ATMOSPHERIC RESEARCH
(2022)
Article
Green & Sustainable Science & Technology
Umit Agbulut et al.
Summary: The prediction of global solar radiation is essential for solar energy systems and investment policies. This study used four different machine learning algorithms and seven statistical metrics to predict daily solar radiation accurately. The artificial neural network algorithm was found to be the best performer among all models.
RENEWABLE & SUSTAINABLE ENERGY REVIEWS
(2021)
Article
Computer Science, Artificial Intelligence
Chun Sing Lai et al.
Summary: Solar radiation forecasting is critical in improving the performance of photovoltaic power plants, and a deep learning based hybrid method for 1-hour ahead Global Horizontal Irradiance (GHI) forecasting is proposed in this study. By utilizing deep time-series clustering and Feature Attention Deep Forecasting (FADF) deep neural network, the developed method achieves more accurate solar forecasting compared to existing models.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Sahbi Boubaker et al.
Summary: Various deep neural network models were developed for one day-ahead prediction of global horizontal irradiation (GHI) in Hail city, Saudi Arabia, using a dataset collected from NASA from 2000 to 2020. The models showed good performance, with a maximum correlation coefficient of 96% reached.
Article
Energy & Fuels
Muhammad Arif Budiyanto et al.
Review
Geography, Physical
Shunlin Liang et al.
INTERNATIONAL JOURNAL OF DIGITAL EARTH
(2019)
Article
Green & Sustainable Science & Technology
L. Benali et al.
Article
Green & Sustainable Science & Technology
M. Caldas et al.
Article
Thermodynamics
Yu Feng et al.
ENERGY CONVERSION AND MANAGEMENT
(2019)
Article
Energy & Fuels
Hadrien Verbois et al.
Article
Green & Sustainable Science & Technology
Enrica Scolari et al.
IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
(2018)
Article
Thermodynamics
Marius Paulescu et al.
ENERGY CONVERSION AND MANAGEMENT
(2014)
Article
Energy & Fuels
Maimouna Diagne et al.
Article
Energy & Fuels
Gordon Reikard
Article
Meteorology & Atmospheric Sciences
S.-Q. Peng et al.
