Publication:
A Moment in the Sun: Solar Nowcasting from Multispectral Satellite Data using Self-Supervised Learning

Simple item page
dc.contributor.authorBansal, Akansha Singh
dc.contributor.authorBansal, Trapit
dc.contributor.authorIrwin, David
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.contributor.departmentUniversity of Massachusetts Amherst
dc.date2023-09-24T09:21:07.000
dc.date.accessioned2024-04-26T16:35:46Z
dc.date.available2024-04-26T16:35:46Z
dc.date.issued2022-01-01
dc.description.abstractABSTRACT Solar energy is now the cheapest form of electricity in history. Unfortunately, signi.cantly increasing the electric grid’s fraction of solar energy remains challenging due to its variability, which makes balancing electricity’s supply and demand more di.cult. While thermal generators’ ramp rate—the maximum rate at which they can change their energy generation—is .nite, solar energy’s ramp rate is essentially in.nite. Thus, accurate near-term solar forecasting, or nowcasting, is important to provide advance warnings to adjust thermal generator output in response to variations in solar generation to ensure a balanced supply and demand. To address the problem, this paper develops a general model for solar nowcasting from abundant and readily available multispectral satellite data using self-supervised learning. Speci.cally, we develop deep auto-regressive models using convolutional neural networks (CNN) and long short-term memory networks (LSTM) that are globally trained across multiple locations to predict raw future observations of the spatio-temporal spectral data collected by the recently launched GOES-R series of satellites. Our model estimates a location’s near-term future solar irradiance based on satellite observations, which we feed to a regression model trained on smaller site-speci.c solar data to provide near-term solar photovoltaic (PV) forecasts that account for site-speci.c characteristics. We evaluate our approach for di.erent coverage areas and forecast horizons across 25 solar sites and show that it yields errors close to that of a model using ground-truth observations.
dc.description.pages251–262
dc.identifier.doihttps://doi.org/10.1145/3538637.3538854
dc.identifier.urihttps://hdl.handle.net/20.500.14394/22970
dc.relation.ispartofe-Energy '22: Proceedings of the Thirteenth ACM International Conference on Future Energy Systems
dc.relation.urlhttps://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1009&context=elevate_pubs&unstamped=1
dc.source.statuspublished
dc.subjectnowcasting
dc.subjectmultispectral
dc.subjectSolar energy
dc.subjectphotovoltaic forecasts
dc.subjectconvolutional neural networks
dc.subjectlong short-term memory networks
dc.subjectCivil and Environmental Engineering
dc.subjectElectrical and Computer Engineering
dc.subjectElectrical and Electronics
dc.subjectEnvironmental Engineering
dc.subjectPower and Energy
dc.titleA Moment in the Sun: Solar Nowcasting from Multispectral Satellite Data using Self-Supervised Learning
dc.typearticle
dc.typearticle
digcom.contributor.authorisAuthorOfPublication|email:akansha.singh803@gmail.com|institution:University of Massachusetts Amherst|Bansal, Akansha Singh
digcom.contributor.authorisAuthorOfPublication|email:trapitbansal@gmail.com|institution:University of Massachusetts Amherst|Bansal, Trapit
digcom.contributor.authorIrwin, David
digcom.identifierelevate_pubs/5
digcom.identifier.contextkey31609304
digcom.identifier.submissionpathelevate_pubs/5
dspace.entity.typePublication
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
A_moment_in_the_sun.pdf
Size:
1.01 MB
Format:
Adobe Portable Document Format
Thumbnail Image
Name: A_moment_in_the_sun.pdf
Size: 1.01 MB
Kind: Adobe PDF
Collections
ELEVATE Publications