Ultra-Short-Term Solar Irradiance Prediction Using an Integrated Framework with Novel Textural Convolution Kernel for Feature Extraction of Clouds

Lijie Wang, Xin Li (), Ying Hao and Qingshan Zhang
Additional contact information
Lijie Wang: Department of Electrical Engineering, Beijing Information Science and Technology University, Beijing 100192, China
Xin Li: Department of Electrical Engineering, Beijing Information Science and Technology University, Beijing 100192, China
Ying Hao: Department of Electrical Engineering, Beijing Information Science and Technology University, Beijing 100192, China
Qingshan Zhang: Department of Electrical Engineering, Beijing Information Science and Technology University, Beijing 100192, China

Sustainability, 2025, vol. 17, issue 6, 1-23

Abstract: Solar irradiance is one of the main factors affecting photovoltaic power generation. The shielding effect of clouds on solar radiation is affected by both type and cover. Therefore, this paper proposes the use of textural features to represent the shielding effect of clouds on solar radiation, and a novel textural convolution kernel of a convolutional neural network, based on grey-level co-occurrence matrix, is presented to extract the textural features of clouds. An integrated ultra-short-term solar irradiance prediction framework is then proposed based on feature extraction network, a clear sky model, and LSTM. The textural features are extracted from satellite cloud images, and the theoretical irradiance under clear sky conditions is calculated based on an improved ASHRAE model. The LSTM is trained with the textural features of clouds, theoretical irradiance, and NWP information. A case study using data from Wuwei PV station in northwest China indicate that the features extracted from the proposed textural convolution kernel are better than common convolution kernels in reflecting the shielding effect of clouds on solar irradiance, and integrating textural features of cloud with theoretical irradiance can lead to better performance in solar irradiance prediction. Thus, this study will help to forecast the output power of PV stations.

Keywords: solar irradiance prediction; textural convolution kernel; feature extraction; convolutional neural network; clear sky model; long short-term memory network (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2071-1050/17/6/2606/pdf (application/pdf)
https://www.mdpi.com/2071-1050/17/6/2606/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:17:y:2025:i:6:p:2606-:d:1613279

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().