File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1016/j.epsr.2020.106732
- Scopus: eid_2-s2.0-85089080188
- WOS: WOS:000594662700004
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Modeling load forecast uncertainty using generative adversarial networks
Title | Modeling load forecast uncertainty using generative adversarial networks |
---|---|
Authors | |
Keywords | Generative adversarial networks (GAN) Probabilistic load forecasting Residual Uncertainty modeling Variation |
Issue Date | 2020 |
Citation | Electric Power Systems Research, 2020, v. 189, article no. 106732 How to Cite? |
Abstract | The integration of distributed energy resources (DER) increase the uncertainty of the load. Probabilistic load forecasting (PLF) is able to model these uncertainties in the form of quantile, interval, or density. However, the uncertainties are usually given individually for every single period which fails to capture the temporal variations across periods. Therefore, this paper proposes a generative adversarial network (GAN)-based scenario generation approach to model both the uncertainties and the variations of the load. Specifically, point forecasting is first conducted and the corresponding residuals are calculated. On this basis, a conditional GAN model is designed and trained. Then, the well-trained GAN model generates residual scenarios that are conditional on the day type, temperatures, and historical loads. Finally, the effectiveness of the uncertainty modeling by the generated scenarios is evaluated from different perspectives. Case studies on open datasets verify the effectiveness and superiority of the proposed method. |
Persistent Identifier | http://hdl.handle.net/10722/308819 |
ISSN | 2023 Impact Factor: 3.3 2023 SCImago Journal Rankings: 1.029 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Wang, Yi | - |
dc.contributor.author | Hug, Gabriela | - |
dc.contributor.author | Liu, Zijie | - |
dc.contributor.author | Zhang, Ning | - |
dc.date.accessioned | 2021-12-08T07:50:12Z | - |
dc.date.available | 2021-12-08T07:50:12Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Electric Power Systems Research, 2020, v. 189, article no. 106732 | - |
dc.identifier.issn | 0378-7796 | - |
dc.identifier.uri | http://hdl.handle.net/10722/308819 | - |
dc.description.abstract | The integration of distributed energy resources (DER) increase the uncertainty of the load. Probabilistic load forecasting (PLF) is able to model these uncertainties in the form of quantile, interval, or density. However, the uncertainties are usually given individually for every single period which fails to capture the temporal variations across periods. Therefore, this paper proposes a generative adversarial network (GAN)-based scenario generation approach to model both the uncertainties and the variations of the load. Specifically, point forecasting is first conducted and the corresponding residuals are calculated. On this basis, a conditional GAN model is designed and trained. Then, the well-trained GAN model generates residual scenarios that are conditional on the day type, temperatures, and historical loads. Finally, the effectiveness of the uncertainty modeling by the generated scenarios is evaluated from different perspectives. Case studies on open datasets verify the effectiveness and superiority of the proposed method. | - |
dc.language | eng | - |
dc.relation.ispartof | Electric Power Systems Research | - |
dc.subject | Generative adversarial networks (GAN) | - |
dc.subject | Probabilistic load forecasting | - |
dc.subject | Residual | - |
dc.subject | Uncertainty modeling | - |
dc.subject | Variation | - |
dc.title | Modeling load forecast uncertainty using generative adversarial networks | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1016/j.epsr.2020.106732 | - |
dc.identifier.scopus | eid_2-s2.0-85089080188 | - |
dc.identifier.volume | 189 | - |
dc.identifier.spage | article no. 106732 | - |
dc.identifier.epage | article no. 106732 | - |
dc.identifier.isi | WOS:000594662700004 | - |