研究目的
To propose two methodologies utilizing artificial neural networks (ANNs) to predict global horizontal irradiance in 1 to 5 minutes in advance from sky images without requiring cloud detection techniques.
研究成果
The proposed methodologies, utilizing ANNs and color information from sky images, achieve comparable prediction accuracy to existing methods with the advantages of requiring less expensive instruments and less computational efforts. The 1-step methodology with 4-axis is the most useful case among the proposed methods.
研究不足
The prediction accuracy degrades as the prediction minutes increase due to limited measurement duration and sky area covered by sampling points. The methodologies are suited for short-term predictions (1 to several minutes in advance).
1：Experimental Design and Method Selection:
The study proposes two methodologies for predicting global horizontal irradiance using ANNs, focusing on color information from sky images without cloud detection.
2：Sample Selection and Data Sources:
Sky images and global horizontal irradiance data were collected from July to September 2016, using a waterproof camera and a pyranometer.
3：List of Experimental Equipment and Materials:
A waterproof 12-megapixel camera with a fish-eye lens mounted on a 2-axis solar tracker and a pyranometer were used.
4：Experimental Procedures and Operational Workflow:
Sky images were taken at 20-second intervals, and global horizontal irradiance was measured minutely. Five successive sky images were used for ANN input data.
5：Data Analysis Methods:
The root mean square error (RMSE) was used to evaluate the accuracy of the solar irradiance prediction results.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
