Generating Contextual Load Profiles Using a Conditional Variational Autoencoder

Chenguang Wang; Simon H. Tindemans; Peter  Palensky

doi:10.1109/ISGT-Europe54678.2022.9960309

Generating Contextual Load Profiles Using a Conditional Variational Autoencoder

Chenguang Wang, Simon H. Tindemans, Peter Palensky

Intelligent Electrical Power Grids

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

3 Citations (Scopus)

109 Downloads (Pure)

Abstract

Generating power system states that have similar distribution and dependency to the historical ones is essential for the tasks of system planning and security assessment, especially when the historical data is insufficient. In this paper, we described a generative model for load profiles of industrial and commercial customers, based on the conditional variational autoencoder (CVAE) neural network architecture, which is challenging due to the highly variable nature of such profiles. Generated contextual load profiles were conditioned on the month of the year and typical power exchange with the grid. Moreover, the quality of generations was both visually and statistically evaluated. The experimental results demonstrate our proposed CVAE model can capture temporal features of historical load profiles and generate ‘realistic’ data with satisfying univariate distributions and multivariate dependencies.

Original language	English
Title of host publication	Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022
Publisher	IEEE
Pages	1-6
Number of pages	6
ISBN (Electronic)	978-1-6654-8032-1
ISBN (Print)	978-1-6654-8033-8
DOIs	https://doi.org/10.1109/ISGT-Europe54678.2022.9960309
Publication status	Published - 2022
Event	2022 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe) - Novi Sad, Serbia Duration: 10 Oct 2022 → 12 Oct 2022

Publication series

Name	IEEE PES Innovative Smart Grid Technologies Conference Europe
Volume	2022-October

Conference

Conference	2022 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe)
Country/Territory	Serbia
City	Novi Sad
Period	10/10/22 → 12/10/22

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

CVAE
generative model
load profiles
synthetic data

Access to Document

10.1109/ISGT-Europe54678.2022.9960309

ISGT_2022__Generating_Contextual_Load_Profiles_Using_a_Conditional_Variational_Autoencoder-AAMAccepted author manuscript, 1.35 MB
Generating_Contextual_Load_Profiles_Using_a_Conditional_Variational_AutoencoderFinal published version, 2.9 MB

Cite this

Wang, C., Tindemans, S. H., & Palensky, P. (2022). Generating Contextual Load Profiles Using a Conditional Variational Autoencoder. In Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022 (pp. 1-6). (IEEE PES Innovative Smart Grid Technologies Conference Europe; Vol. 2022-October). IEEE. https://doi.org/10.1109/ISGT-Europe54678.2022.9960309

@inproceedings{bce7b85c1eb442bfa603f2a6b63370e2,

title = "Generating Contextual Load Profiles Using a Conditional Variational Autoencoder",

abstract = "Generating power system states that have similar distribution and dependency to the historical ones is essential for the tasks of system planning and security assessment, especially when the historical data is insufficient. In this paper, we described a generative model for load profiles of industrial and commercial customers, based on the conditional variational autoencoder (CVAE) neural network architecture, which is challenging due to the highly variable nature of such profiles. Generated contextual load profiles were conditioned on the month of the year and typical power exchange with the grid. Moreover, the quality of generations was both visually and statistically evaluated. The experimental results demonstrate our proposed CVAE model can capture temporal features of historical load profiles and generate {\textquoteleft}realistic{\textquoteright} data with satisfying univariate distributions and multivariate dependencies.",

keywords = "CVAE, generative model, load profiles, synthetic data",

author = "Chenguang Wang and Tindemans, {Simon H.} and Peter Palensky",

note = "Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.; 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe) ; Conference date: 10-10-2022 Through 12-10-2022",

year = "2022",

doi = "10.1109/ISGT-Europe54678.2022.9960309",

language = "English",

isbn = "978-1-6654-8033-8",

series = "IEEE PES Innovative Smart Grid Technologies Conference Europe",

publisher = "IEEE",

pages = "1--6",

booktitle = "Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022",

address = "United States",

}

Wang, C, Tindemans, SH & Palensky, P 2022, Generating Contextual Load Profiles Using a Conditional Variational Autoencoder. in Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022. IEEE PES Innovative Smart Grid Technologies Conference Europe, vol. 2022-October, IEEE, pp. 1-6, 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe), Novi Sad, Serbia, 10/10/22. https://doi.org/10.1109/ISGT-Europe54678.2022.9960309

Generating Contextual Load Profiles Using a Conditional Variational Autoencoder. / Wang, Chenguang; Tindemans, Simon H.; Palensky, Peter .
Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022. IEEE, 2022. p. 1-6 (IEEE PES Innovative Smart Grid Technologies Conference Europe; Vol. 2022-October).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Generating Contextual Load Profiles Using a Conditional Variational Autoencoder

AU - Wang, Chenguang

AU - Tindemans, Simon H.

AU - Palensky, Peter

N1 - Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2022

Y1 - 2022

N2 - Generating power system states that have similar distribution and dependency to the historical ones is essential for the tasks of system planning and security assessment, especially when the historical data is insufficient. In this paper, we described a generative model for load profiles of industrial and commercial customers, based on the conditional variational autoencoder (CVAE) neural network architecture, which is challenging due to the highly variable nature of such profiles. Generated contextual load profiles were conditioned on the month of the year and typical power exchange with the grid. Moreover, the quality of generations was both visually and statistically evaluated. The experimental results demonstrate our proposed CVAE model can capture temporal features of historical load profiles and generate ‘realistic’ data with satisfying univariate distributions and multivariate dependencies.

AB - Generating power system states that have similar distribution and dependency to the historical ones is essential for the tasks of system planning and security assessment, especially when the historical data is insufficient. In this paper, we described a generative model for load profiles of industrial and commercial customers, based on the conditional variational autoencoder (CVAE) neural network architecture, which is challenging due to the highly variable nature of such profiles. Generated contextual load profiles were conditioned on the month of the year and typical power exchange with the grid. Moreover, the quality of generations was both visually and statistically evaluated. The experimental results demonstrate our proposed CVAE model can capture temporal features of historical load profiles and generate ‘realistic’ data with satisfying univariate distributions and multivariate dependencies.

KW - CVAE

KW - generative model

KW - load profiles

KW - synthetic data

UR - http://www.scopus.com/inward/record.url?scp=85143751808&partnerID=8YFLogxK

U2 - 10.1109/ISGT-Europe54678.2022.9960309

DO - 10.1109/ISGT-Europe54678.2022.9960309

M3 - Conference contribution

SN - 978-1-6654-8033-8

T3 - IEEE PES Innovative Smart Grid Technologies Conference Europe

SP - 1

EP - 6

BT - Proceedings of 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2022

PB - IEEE

T2 - 2022 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe)

Y2 - 10 October 2022 through 12 October 2022

ER -

Generating Contextual Load Profiles Using a Conditional Variational Autoencoder

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this