Photovoltaic Power Forecasting Using Sky Images and Sun Motion

Solar energy adoption is moving at a rapid pace. The variability in solar energy production causes grid stability issues and hinders mass adoption. To solve these issues, more accurate photovoltaic power forecasting systems are needed. In intra-hour forecasting, the most challenging issue is high ou...

Descripción completa

Detalles Bibliográficos
Autores: Berresheim, Arne, Agudo Martínez, Antonio
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2024
País:España
Institución:Consejo Superior de Investigaciones Científicas (CSIC)
Repositorio:DIGITAL.CSIC. Repositorio Institucional del CSIC
OAI Identifier:oai:digital.csic.es:10261/388085
Acceso en línea:http://hdl.handle.net/10261/388085
https://api.elsevier.com/content/abstract/scopus_id/105001505997
Access Level:acceso abierto
Palabra clave:Deep Learning
Photovoltaic Power Estimation
Sky Images
Sun Tracking
id ES_1dc320c8a8a59e48e27e7d696cdc38c8
oai_identifier_str oai:digital.csic.es:10261/388085
network_acronym_str ES
network_name_str España
spelling Photovoltaic Power Forecasting Using Sky Images and Sun Motion Berresheim, Arne Agudo Martínez, Antonio Deep Learning Photovoltaic Power Estimation Sky Images Sun Tracking Solar energy adoption is moving at a rapid pace. The variability in solar energy production causes grid stability issues and hinders mass adoption. To solve these issues, more accurate photovoltaic power forecasting systems are needed. In intra-hour forecasting, the most challenging issue is high output fluctuations due to cloud motion, which can occlude the sun. Using ground-based sky images, this paper proposes two convolutional neural network models for intra-hour nowcasting and forecasting that incorporate physical information on sun motion and cloud coverage by means of the sun area mean pixel intensity. Particularly, our models exploit that information instead of relying exclusively on photovoltaic output history data as it is standard in state of the art. Taking advantage of sun position and cloud coverage information, we were able to reduce the overall root mean squared error for the nowcasting task, making the model more accurate especially during cloudy days, and obtaining competitive results on forecasting. Moreover, our models are more robust against artifacts such as occlusion and noisy observations. This work has been supported by the project MoHuCo PID2020- 120049RB-I00 funded by MCIN/ AEI /10.13039/501100011033. Peer reviewed Institute of Electrical and Electronics Engineers http://hdl.handle.net/10261/388085 https://api.elsevier.com/content/abstract/scopus_id/105001505997
title Photovoltaic Power Forecasting Using Sky Images and Sun Motion
spellingShingle Photovoltaic Power Forecasting Using Sky Images and Sun Motion
Berresheim, Arne
Deep Learning
Photovoltaic Power Estimation
Sky Images
Sun Tracking
title_short Photovoltaic Power Forecasting Using Sky Images and Sun Motion
title_full Photovoltaic Power Forecasting Using Sky Images and Sun Motion
title_fullStr Photovoltaic Power Forecasting Using Sky Images and Sun Motion
title_full_unstemmed Photovoltaic Power Forecasting Using Sky Images and Sun Motion
title_sort Photovoltaic Power Forecasting Using Sky Images and Sun Motion
author Berresheim, Arne
author_facet Berresheim, Arne
Agudo Martínez, Antonio
author_role author
author2 Agudo Martínez, Antonio
author2_role author
topic Deep Learning
Photovoltaic Power Estimation
Sky Images
Sun Tracking
topic_facet Deep Learning
Photovoltaic Power Estimation
Sky Images
Sun Tracking
description Solar energy adoption is moving at a rapid pace. The variability in solar energy production causes grid stability issues and hinders mass adoption. To solve these issues, more accurate photovoltaic power forecasting systems are needed. In intra-hour forecasting, the most challenging issue is high output fluctuations due to cloud motion, which can occlude the sun. Using ground-based sky images, this paper proposes two convolutional neural network models for intra-hour nowcasting and forecasting that incorporate physical information on sun motion and cloud coverage by means of the sun area mean pixel intensity. Particularly, our models exploit that information instead of relying exclusively on photovoltaic output history data as it is standard in state of the art. Taking advantage of sun position and cloud coverage information, we were able to reduce the overall root mean squared error for the nowcasting task, making the model more accurate especially during cloudy days, and obtaining competitive results on forecasting. Moreover, our models are more robust against artifacts such as occlusion and noisy observations.
publishDate 2024
format article
status_str acceptedVersion
url http://hdl.handle.net/10261/388085
https://api.elsevier.com/content/abstract/scopus_id/105001505997
eu_rights_str_mv openAccess
publisher Institute of Electrical and Electronics Engineers
institution Consejo Superior de Investigaciones Científicas (CSIC)
collection DIGITAL.CSIC. Repositorio Institucional del CSIC
reponame_str DIGITAL.CSIC. Repositorio Institucional del CSIC
instname_str Consejo Superior de Investigaciones Científicas (CSIC)
_version_ 1878432196816011265
publishDateSort 2024
author_browse Agudo Martínez, Antonio
Berresheim, Arne
publisherStr Institute of Electrical and Electronics Engineers
score 6,9303427

Similares en LA Referencia