Long short-term memory recurrent neural network for modeling temporal patterns in long-term power forecasting for solar PV facilities: Case study of South Korea
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Jung, Yoonhwa | - |
dc.contributor.author | Jung, Jaehoon | - |
dc.contributor.author | Kim, Byungil | - |
dc.contributor.author | Han, SangUk | - |
dc.date.accessioned | 2021-08-02T09:52:24Z | - |
dc.date.available | 2021-08-02T09:52:24Z | - |
dc.date.created | 2021-05-12 | - |
dc.date.issued | 2020-03 | - |
dc.identifier.issn | 0959-6526 | - |
dc.identifier.uri | https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/10661 | - |
dc.description.abstract | The sites selected for solar PV facilities significantly affect the amount of electric power that can be generated over the long term. Therefore, predicting the power output of a specific PV plant is important when evaluating potential PV sites. However, whether prediction models built with data from existing PV plants can be applied to other plants for long-term power forecasting remains poorly understood. In this case, topographical and meteorological conditions, which differ among sites and change over time, make it challenging to accurately estimate the potential for energy generation at a new site. This study proposes a monthly PV power forecasting model to predict the amount of PV solar power that could be generated at a new site. The forecasting model is trained with time series datasets collected over 63 months from 164 PV sites with data such as the power plant capacity and electricity trading data, weather conditions, and estimated solar irradiation. Specifically, a recurrent neural network (RNN) model with long short-term memory was built to recognize the temporal patterns in the time series data and tested to evaluate the forecasting performance for PV facilities not used in the training process. The results show that the proposed model achieves the normalized root-mean-square-error of 7.416% and the mean absolute-percentage-error (MAPE) of 10.805% for the testing data (i.e., new plants). Furthermore, when the previous 10 months' data were used, the temporal patterns were well captured for forecasting, with a MAPE of 11.535%. Thus, the proposed RNN approach successfully captures the temporal patterns in monthly data and can estimate the potential for power generation at any new site for which weather information and terrain data are available. Consequently, this work will allow planning officials to search for and evaluate suitable locations for PV plants in a wide area. | - |
dc.language | 영어 | - |
dc.language.iso | en | - |
dc.publisher | ELSEVIER SCI LTD | - |
dc.title | Long short-term memory recurrent neural network for modeling temporal patterns in long-term power forecasting for solar PV facilities: Case study of South Korea | - |
dc.type | Article | - |
dc.contributor.affiliatedAuthor | Han, SangUk | - |
dc.identifier.doi | 10.1016/j.jclepro.2019.119476 | - |
dc.identifier.scopusid | 2-s2.0-85076569765 | - |
dc.identifier.wosid | 000508829800062 | - |
dc.identifier.bibliographicCitation | JOURNAL OF CLEANER PRODUCTION, v.250, pp.1 - 13 | - |
dc.relation.isPartOf | JOURNAL OF CLEANER PRODUCTION | - |
dc.citation.title | JOURNAL OF CLEANER PRODUCTION | - |
dc.citation.volume | 250 | - |
dc.citation.startPage | 1 | - |
dc.citation.endPage | 13 | - |
dc.type.rims | ART | - |
dc.type.docType | Article | - |
dc.description.journalClass | 1 | - |
dc.description.isOpenAccess | N | - |
dc.description.journalRegisteredClass | scie | - |
dc.description.journalRegisteredClass | scopus | - |
dc.relation.journalResearchArea | Science & Technology - Other Topics | - |
dc.relation.journalResearchArea | Engineering | - |
dc.relation.journalResearchArea | Environmental Sciences & Ecology | - |
dc.relation.journalWebOfScienceCategory | Green & Sustainable Science & Technology | - |
dc.relation.journalWebOfScienceCategory | Engineering, Environmental | - |
dc.relation.journalWebOfScienceCategory | Environmental Sciences | - |
dc.subject.keywordPlus | SITE SELECTION | - |
dc.subject.keywordPlus | PREDICTION | - |
dc.subject.keywordPlus | GENERATION | - |
dc.subject.keywordPlus | OUTPUT | - |
dc.subject.keywordPlus | PLANTS | - |
dc.subject.keywordPlus | RADIATION | - |
dc.subject.keywordPlus | DATABASE | - |
dc.subject.keywordPlus | WIND | - |
dc.subject.keywordPlus | GIS | - |
dc.subject.keywordAuthor | Solar energy | - |
dc.subject.keywordAuthor | Photovoltaic power | - |
dc.subject.keywordAuthor | Long-term forecasting | - |
dc.subject.keywordAuthor | Long short-term memory | - |
dc.subject.keywordAuthor | Recurrent neural network | - |
dc.identifier.url | https://www.sciencedirect.com/science/article/pii/S095965261934346X?via%3Dihub | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
222, Wangsimni-ro, Seongdong-gu, Seoul, 04763, Korea+82-2-2220-1365
COPYRIGHT © 2021 HANYANG UNIVERSITY.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.