MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations

This paper presents a deep-learning framework, Multi-load Generative Adversarial Network (MultiLoad-GAN), for generating a group of synthetic load profiles (SLPs) simultaneously. The main contribution of MultiLoad-GAN is the capture of spatial-temporal correlations among a group of loads that are se...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on smart grid 2024-03, Vol.15 (2), p.2309-2320
Hauptverfasser: Hu, Yi, Li, Yiyan, Song, Lidong, Lee, Han Pyo, Rehm, P. J., Makdad, Matthew, Miller, Edmond, Lu, Ning
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 2320
container_issue 2
container_start_page 2309
container_title IEEE transactions on smart grid
container_volume 15
creator Hu, Yi
Li, Yiyan
Song, Lidong
Lee, Han Pyo
Rehm, P. J.
Makdad, Matthew
Miller, Edmond
Lu, Ning
description This paper presents a deep-learning framework, Multi-load Generative Adversarial Network (MultiLoad-GAN), for generating a group of synthetic load profiles (SLPs) simultaneously. The main contribution of MultiLoad-GAN is the capture of spatial-temporal correlations among a group of loads that are served by the same distribution transformer. This enables the generation of a large amount of correlated SLPs required for microgrid and distribution system studies. The novelty and uniqueness of the MultiLoad-GAN framework are three-fold. First, to the best of our knowledge, this is the first method for generating a group of load profiles bearing realistic spatial-temporal correlations simultaneously. Second, two complementary realisticness metrics for evaluating generated load profiles are developed: computing statistics based on domain knowledge and comparing high-level features via a deep-learning classifier. Third, to tackle data scarcity, a novel iterative data augmentation mechanism is developed to generate training samples for enhancing the training of both the classifier and the MultiLoad-GAN model. Simulation results show that MultiLoad-GAN can generate more realistic load profiles than existing approaches, especially in group level characteristics. With little finetuning, MultiLoad-GAN can be readily extended to generate a group of load or PV profiles for a feeder or a service area.
doi_str_mv 10.1109/TSG.2023.3302192
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_proquest_journals_2929257137</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10208147</ieee_id><sourcerecordid>2929257137</sourcerecordid><originalsourceid>FETCH-LOGICAL-c245t-23f955dff5d311ab377b80178bacfff61a7c06c55adbdfc0b9cb17c5dc8f82f3</originalsourceid><addsrcrecordid>eNpNkL1PwzAQxS0EElXpzsBgiTnFH3Uds5UKAlILQ7Nbjj9oShoHOxn63-PSCnE33EnvvTvpB8AtRlOMkXgoN8WUIEKnlCKCBbkAIyxmIqNoji__dkavwSTGHUpFKZ0TMQJf66Hp65VXJisW749wAdPInlS0Bm4Obb-1fa3hUYdF8EMHC9vaoPrat3Bt-603cOnbWBsb6vYTbrokqSYr7b7zQTVJDME2v_54A66caqKdnOcYlC_P5fI1W30Ub8vFKtNkxvqMUCcYM84xQzFWFeW8yhHmeaW0c26OFddorhlTpjJOo0roCnPNjM5dThwdg_vT2S7478HGXu78ENr0URKRmnFMeXKhk0sHH2OwTnah3qtwkBjJI1SZoMojVHmGmiJ3p0htrf1nJyjHM05_AAOoc3g</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2929257137</pqid></control><display><type>article</type><title>MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations</title><source>IEEE Electronic Library (IEL)</source><creator>Hu, Yi ; Li, Yiyan ; Song, Lidong ; Lee, Han Pyo ; Rehm, P. J. ; Makdad, Matthew ; Miller, Edmond ; Lu, Ning</creator><creatorcontrib>Hu, Yi ; Li, Yiyan ; Song, Lidong ; Lee, Han Pyo ; Rehm, P. J. ; Makdad, Matthew ; Miller, Edmond ; Lu, Ning</creatorcontrib><description>This paper presents a deep-learning framework, Multi-load Generative Adversarial Network (MultiLoad-GAN), for generating a group of synthetic load profiles (SLPs) simultaneously. The main contribution of MultiLoad-GAN is the capture of spatial-temporal correlations among a group of loads that are served by the same distribution transformer. This enables the generation of a large amount of correlated SLPs required for microgrid and distribution system studies. The novelty and uniqueness of the MultiLoad-GAN framework are three-fold. First, to the best of our knowledge, this is the first method for generating a group of load profiles bearing realistic spatial-temporal correlations simultaneously. Second, two complementary realisticness metrics for evaluating generated load profiles are developed: computing statistics based on domain knowledge and comparing high-level features via a deep-learning classifier. Third, to tackle data scarcity, a novel iterative data augmentation mechanism is developed to generate training samples for enhancing the training of both the classifier and the MultiLoad-GAN model. Simulation results show that MultiLoad-GAN can generate more realistic load profiles than existing approaches, especially in group level characteristics. With little finetuning, MultiLoad-GAN can be readily extended to generate a group of load or PV profiles for a feeder or a service area.</description><identifier>ISSN: 1949-3053</identifier><identifier>EISSN: 1949-3061</identifier><identifier>DOI: 10.1109/TSG.2023.3302192</identifier><identifier>CODEN: ITSGBQ</identifier><language>eng</language><publisher>Piscataway: IEEE</publisher><subject>Classifiers ; Correlation ; Data augmentation ; Data models ; Deep learning ; Distributed generation ; Generative adversarial networks ; Iterative methods ; Load modeling ; load profile group generation ; Machine learning ; negative sample generation ; Predictive models ; Service areas ; synthetic data ; Training ; Transformers</subject><ispartof>IEEE transactions on smart grid, 2024-03, Vol.15 (2), p.2309-2320</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c245t-23f955dff5d311ab377b80178bacfff61a7c06c55adbdfc0b9cb17c5dc8f82f3</cites><orcidid>0000-0002-7301-4409 ; 0000-0003-3742-8740 ; 0000-0002-8796-7139 ; 0000-0002-6136-0371 ; 0000-0003-0125-0653</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10208147$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10208147$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Hu, Yi</creatorcontrib><creatorcontrib>Li, Yiyan</creatorcontrib><creatorcontrib>Song, Lidong</creatorcontrib><creatorcontrib>Lee, Han Pyo</creatorcontrib><creatorcontrib>Rehm, P. J.</creatorcontrib><creatorcontrib>Makdad, Matthew</creatorcontrib><creatorcontrib>Miller, Edmond</creatorcontrib><creatorcontrib>Lu, Ning</creatorcontrib><title>MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations</title><title>IEEE transactions on smart grid</title><addtitle>TSG</addtitle><description>This paper presents a deep-learning framework, Multi-load Generative Adversarial Network (MultiLoad-GAN), for generating a group of synthetic load profiles (SLPs) simultaneously. The main contribution of MultiLoad-GAN is the capture of spatial-temporal correlations among a group of loads that are served by the same distribution transformer. This enables the generation of a large amount of correlated SLPs required for microgrid and distribution system studies. The novelty and uniqueness of the MultiLoad-GAN framework are three-fold. First, to the best of our knowledge, this is the first method for generating a group of load profiles bearing realistic spatial-temporal correlations simultaneously. Second, two complementary realisticness metrics for evaluating generated load profiles are developed: computing statistics based on domain knowledge and comparing high-level features via a deep-learning classifier. Third, to tackle data scarcity, a novel iterative data augmentation mechanism is developed to generate training samples for enhancing the training of both the classifier and the MultiLoad-GAN model. Simulation results show that MultiLoad-GAN can generate more realistic load profiles than existing approaches, especially in group level characteristics. With little finetuning, MultiLoad-GAN can be readily extended to generate a group of load or PV profiles for a feeder or a service area.</description><subject>Classifiers</subject><subject>Correlation</subject><subject>Data augmentation</subject><subject>Data models</subject><subject>Deep learning</subject><subject>Distributed generation</subject><subject>Generative adversarial networks</subject><subject>Iterative methods</subject><subject>Load modeling</subject><subject>load profile group generation</subject><subject>Machine learning</subject><subject>negative sample generation</subject><subject>Predictive models</subject><subject>Service areas</subject><subject>synthetic data</subject><subject>Training</subject><subject>Transformers</subject><issn>1949-3053</issn><issn>1949-3061</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpNkL1PwzAQxS0EElXpzsBgiTnFH3Uds5UKAlILQ7Nbjj9oShoHOxn63-PSCnE33EnvvTvpB8AtRlOMkXgoN8WUIEKnlCKCBbkAIyxmIqNoji__dkavwSTGHUpFKZ0TMQJf66Hp65VXJisW749wAdPInlS0Bm4Obb-1fa3hUYdF8EMHC9vaoPrat3Bt-603cOnbWBsb6vYTbrokqSYr7b7zQTVJDME2v_54A66caqKdnOcYlC_P5fI1W30Ub8vFKtNkxvqMUCcYM84xQzFWFeW8yhHmeaW0c26OFddorhlTpjJOo0roCnPNjM5dThwdg_vT2S7478HGXu78ENr0URKRmnFMeXKhk0sHH2OwTnah3qtwkBjJI1SZoMojVHmGmiJ3p0htrf1nJyjHM05_AAOoc3g</recordid><startdate>20240301</startdate><enddate>20240301</enddate><creator>Hu, Yi</creator><creator>Li, Yiyan</creator><creator>Song, Lidong</creator><creator>Lee, Han Pyo</creator><creator>Rehm, P. J.</creator><creator>Makdad, Matthew</creator><creator>Miller, Edmond</creator><creator>Lu, Ning</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SP</scope><scope>7TB</scope><scope>8FD</scope><scope>FR3</scope><scope>KR7</scope><scope>L7M</scope><orcidid>https://orcid.org/0000-0002-7301-4409</orcidid><orcidid>https://orcid.org/0000-0003-3742-8740</orcidid><orcidid>https://orcid.org/0000-0002-8796-7139</orcidid><orcidid>https://orcid.org/0000-0002-6136-0371</orcidid><orcidid>https://orcid.org/0000-0003-0125-0653</orcidid></search><sort><creationdate>20240301</creationdate><title>MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations</title><author>Hu, Yi ; Li, Yiyan ; Song, Lidong ; Lee, Han Pyo ; Rehm, P. J. ; Makdad, Matthew ; Miller, Edmond ; Lu, Ning</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c245t-23f955dff5d311ab377b80178bacfff61a7c06c55adbdfc0b9cb17c5dc8f82f3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Classifiers</topic><topic>Correlation</topic><topic>Data augmentation</topic><topic>Data models</topic><topic>Deep learning</topic><topic>Distributed generation</topic><topic>Generative adversarial networks</topic><topic>Iterative methods</topic><topic>Load modeling</topic><topic>load profile group generation</topic><topic>Machine learning</topic><topic>negative sample generation</topic><topic>Predictive models</topic><topic>Service areas</topic><topic>synthetic data</topic><topic>Training</topic><topic>Transformers</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Hu, Yi</creatorcontrib><creatorcontrib>Li, Yiyan</creatorcontrib><creatorcontrib>Song, Lidong</creatorcontrib><creatorcontrib>Lee, Han Pyo</creatorcontrib><creatorcontrib>Rehm, P. J.</creatorcontrib><creatorcontrib>Makdad, Matthew</creatorcontrib><creatorcontrib>Miller, Edmond</creatorcontrib><creatorcontrib>Lu, Ning</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>CrossRef</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Mechanical &amp; Transportation Engineering Abstracts</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>Civil Engineering Abstracts</collection><collection>Advanced Technologies Database with Aerospace</collection><jtitle>IEEE transactions on smart grid</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hu, Yi</au><au>Li, Yiyan</au><au>Song, Lidong</au><au>Lee, Han Pyo</au><au>Rehm, P. J.</au><au>Makdad, Matthew</au><au>Miller, Edmond</au><au>Lu, Ning</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations</atitle><jtitle>IEEE transactions on smart grid</jtitle><stitle>TSG</stitle><date>2024-03-01</date><risdate>2024</risdate><volume>15</volume><issue>2</issue><spage>2309</spage><epage>2320</epage><pages>2309-2320</pages><issn>1949-3053</issn><eissn>1949-3061</eissn><coden>ITSGBQ</coden><abstract>This paper presents a deep-learning framework, Multi-load Generative Adversarial Network (MultiLoad-GAN), for generating a group of synthetic load profiles (SLPs) simultaneously. The main contribution of MultiLoad-GAN is the capture of spatial-temporal correlations among a group of loads that are served by the same distribution transformer. This enables the generation of a large amount of correlated SLPs required for microgrid and distribution system studies. The novelty and uniqueness of the MultiLoad-GAN framework are three-fold. First, to the best of our knowledge, this is the first method for generating a group of load profiles bearing realistic spatial-temporal correlations simultaneously. Second, two complementary realisticness metrics for evaluating generated load profiles are developed: computing statistics based on domain knowledge and comparing high-level features via a deep-learning classifier. Third, to tackle data scarcity, a novel iterative data augmentation mechanism is developed to generate training samples for enhancing the training of both the classifier and the MultiLoad-GAN model. Simulation results show that MultiLoad-GAN can generate more realistic load profiles than existing approaches, especially in group level characteristics. With little finetuning, MultiLoad-GAN can be readily extended to generate a group of load or PV profiles for a feeder or a service area.</abstract><cop>Piscataway</cop><pub>IEEE</pub><doi>10.1109/TSG.2023.3302192</doi><tpages>12</tpages><orcidid>https://orcid.org/0000-0002-7301-4409</orcidid><orcidid>https://orcid.org/0000-0003-3742-8740</orcidid><orcidid>https://orcid.org/0000-0002-8796-7139</orcidid><orcidid>https://orcid.org/0000-0002-6136-0371</orcidid><orcidid>https://orcid.org/0000-0003-0125-0653</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1949-3053
ispartof IEEE transactions on smart grid, 2024-03, Vol.15 (2), p.2309-2320
issn 1949-3053
1949-3061
language eng
recordid cdi_proquest_journals_2929257137
source IEEE Electronic Library (IEL)
subjects Classifiers
Correlation
Data augmentation
Data models
Deep learning
Distributed generation
Generative adversarial networks
Iterative methods
Load modeling
load profile group generation
Machine learning
negative sample generation
Predictive models
Service areas
synthetic data
Training
Transformers
title MultiLoad-GAN: A GAN-Based Synthetic Load Group Generation Method Considering Spatial-Temporal Correlations
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-06T02%3A51%3A23IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=MultiLoad-GAN:%20A%20GAN-Based%20Synthetic%20Load%20Group%20Generation%20Method%20Considering%20Spatial-Temporal%20Correlations&rft.jtitle=IEEE%20transactions%20on%20smart%20grid&rft.au=Hu,%20Yi&rft.date=2024-03-01&rft.volume=15&rft.issue=2&rft.spage=2309&rft.epage=2320&rft.pages=2309-2320&rft.issn=1949-3053&rft.eissn=1949-3061&rft.coden=ITSGBQ&rft_id=info:doi/10.1109/TSG.2023.3302192&rft_dat=%3Cproquest_RIE%3E2929257137%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2929257137&rft_id=info:pmid/&rft_ieee_id=10208147&rfr_iscdi=true