Robust Short term Photovoltaic Power Prediction and Multimodal Data Fusion Based on Transformer ECANet GRU

Robust Short term Photovoltaic Power Prediction and Multimodal Data Fusion Based on Transformer ECANet GRU | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Robust Short term Photovoltaic Power Prediction and Multimodal Data Fusion Based on Transformer ECANet GRU Gao Yuan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9350775/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract In order to improve the accuracy and robustness of short-term photovoltaic power prediction, the paper proposes a hybrid neural network model (Transformer ECANet GRU) based on Transformer, ECANet, and GRU, combined with multimodal data fusion for short-term photovoltaic power prediction. The paper utilized the Australian DKASC photovoltaic power station and NREL dataset to integrate meteorological data, satellite cloud images, and regional load data. They captured long-range dependencies of time series using Transformer, enhanced spatiotemporal feature extraction using ECANet, and modeled short-term temporal dependencies using GRU. The experimental results show that the Transformer ECANet GRU model outperforms benchmark models such as SVR, GRU, TCN, CNN-GRU, and TCN-ECANet GRU in single step prediction performance on seasonal data. In terms of RMSE, MAE, and R² indicators, this model achieved the best values, with the highest improvement rates of 15.98%, 18.68%, and 0.84%, respectively. The results of multi-step prediction (3, 6, 9 steps) show that the model maintains an advantage in long sequence prediction, with RMSE and MAE reduced by 23.73% and 44.15% respectively in 9-step prediction. Cross dataset validation confirmed the model's generalization ability, and uncertainty quantification enhanced the reliability of prediction results. Conclusion: This method demonstrates excellent performance and robustness in short-term photovoltaic power prediction, providing a reliable prediction tool for smart grid scheduling. Photovoltaic power prediction Transformer ECANet GRU Multimodal data fusion Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 05 May, 2026 Editor invited by journal 21 Apr, 2026 Editor assigned by journal 13 Apr, 2026 Submission checks completed at journal 13 Apr, 2026 First submitted to journal 07 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9350775","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":635115460,"identity":"ae82cf1b-c465-4e9f-8758-62e3016a7a62","order_by":0,"name":"Gao Yuan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3klEQVRIiWNgGAWjYBAC/gYGNgkgLSfP33zgwIcKIrRIHIBoMTaccSzx4IwzRGgxYGBgBmKGxIYDOcaHeVuI0cJ++OBjnpo7iY0NZz4c4G1gkOcXO0BAC09awmGeY8+M25l7NxyQ3MFgOHN2An4thgdyDA7zsB2WbWw4u+GA4RmGBIPbBLQYnH8D1PLvMCPQLw8OJLYRo+UG0BbetsOKQC0MBw4So0XixrOEg3P7DoMC2eBgwxkJwn7h708+cODNt8OgqHz8+U+FjTy/NAEtGLaSpnwUjIJRMApGAXYAAPEjUiMdr+y9AAAAAElFTkSuQmCC","orcid":"","institution":"Wuhan Polytechnic","correspondingAuthor":true,"prefix":"","firstName":"Gao","middleName":"","lastName":"Yuan","suffix":""}],"badges":[],"createdAt":"2026-04-08 03:08:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9350775/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9350775/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109296098,"identity":"59f66e62-6b85-4744-8e8f-65573d904a17","added_by":"auto","created_at":"2026-05-15 08:45:31","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":682675,"visible":true,"origin":"","legend":"","description":"","filename":"paper.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9350775/v1_covered_228b9638-0eee-4f6c-8219-7077678f7896.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Robust Short term Photovoltaic Power Prediction and Multimodal Data Fusion Based on Transformer ECANet GRU","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Computing](https://link.springer.com/journal/10791)","snPcode":"10791","submissionUrl":"https://submission.springernature.com/new-submission/10791/3","title":"Discover Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Photovoltaic power prediction, Transformer, ECANet, GRU, Multimodal data fusion","lastPublishedDoi":"10.21203/rs.3.rs-9350775/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9350775/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn order to improve the accuracy and robustness of short-term photovoltaic power prediction, the paper proposes a hybrid neural network model (Transformer ECANet GRU) based on Transformer, ECANet, and GRU, combined with multimodal data fusion for short-term photovoltaic power prediction. The paper utilized the Australian DKASC photovoltaic power station and NREL dataset to integrate meteorological data, satellite cloud images, and regional load data. They captured long-range dependencies of time series using Transformer, enhanced spatiotemporal feature extraction using ECANet, and modeled short-term temporal dependencies using GRU. The experimental results show that the Transformer ECANet GRU model outperforms benchmark models such as SVR, GRU, TCN, CNN-GRU, and TCN-ECANet GRU in single step prediction performance on seasonal data. In terms of RMSE, MAE, and R² indicators, this model achieved the best values, with the highest improvement rates of 15.98%, 18.68%, and 0.84%, respectively. The results of multi-step prediction (3, 6, 9 steps) show that the model maintains an advantage in long sequence prediction, with RMSE and MAE reduced by 23.73% and 44.15% respectively in 9-step prediction. Cross dataset validation confirmed the model's generalization ability, and uncertainty quantification enhanced the reliability of prediction results.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConclusion: This method demonstrates excellent performance and robustness in short-term photovoltaic power prediction, providing a reliable prediction tool for smart grid scheduling.\u003c/p\u003e","manuscriptTitle":"Robust Short term Photovoltaic Power Prediction and Multimodal Data Fusion Based on Transformer ECANet GRU","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-14 14:06:32","doi":"10.21203/rs.3.rs-9350775/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-05-06T01:40:23+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-21T09:21:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-14T01:08:41+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-14T01:07:53+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Computing","date":"2026-04-08T02:55:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-computing","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Computing](https://link.springer.com/journal/10791)","snPcode":"10791","submissionUrl":"https://submission.springernature.com/new-submission/10791/3","title":"Discover Computing","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4b192476-84a3-41b7-aa94-533f10bdd8a6","owner":[],"postedDate":"May 14th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewersInvited","content":"36","date":"2026-05-06T01:40:23+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-14T14:06:32+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-14 14:06:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9350775","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9350775","identity":"rs-9350775","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}