Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency

Cheolmin Kim; Kibaek Kim; Prasanna Balaprakash; Mihai Anitescu

doi:10.1109/PESGM40551.2019.8973468

Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency

Cheolmin Kim, Kibaek Kim, Prasanna Balaprakash, Mihai Anitescu

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

49 Scopus citations

Abstract

Power system operations under contingency need to solve large-scale complex nonlinear optimization problems in a short amount of time, if not real time. Such nonlinear programs are computationally challenging and time-consuming and thus do not scale with the size of the power system network. We apply a graph convolutional network (GCN) model, as a supervised learning model, for predicting an optimal load-shedding ratio that prevents transmission lines from being overloaded under line contingency (i.e., line tripping). In particular, we exploit the power system network topology in the GCN model, where the topology information is convoluted over the neural network. Using IEEE test cases, we benchmark our GCN model against a classical neural network model and a linear regression model and show that the GCN model outperforms the others by an order of magnitude.

Original language	English
Title of host publication	2019 IEEE Power and Energy Society General Meeting, PESGM 2019
Publisher	IEEE Computer Society
ISBN (Electronic)	9781728119816
DOIs	https://doi.org/10.1109/PESGM40551.2019.8973468
State	Published - Aug 2019
Externally published	Yes
Event	2019 IEEE Power and Energy Society General Meeting, PESGM 2019 - Atlanta, United States Duration: Aug 4 2019 → Aug 8 2019

Publication series

Name	IEEE Power and Energy Society General Meeting
Volume	2019-August
ISSN (Print)	1944-9925
ISSN (Electronic)	1944-9933

Conference

Conference	2019 IEEE Power and Energy Society General Meeting, PESGM 2019
Country/Territory	United States
City	Atlanta
Period	08/4/19 → 08/8/19

Funding

This material is based upon work supported by the U.S. Department of Energy, Office of Science, under contract DE-AC02-06CH11357.

Keywords

Graph covolutional network
alternating current power system
contingency analysis
machine learning
neural network

Access to Document

10.1109/PESGM40551.2019.8973468

Cite this

@inproceedings{43469788d1114b1b9bfd5ddffcac2c8c,

title = "Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency",

abstract = "Power system operations under contingency need to solve large-scale complex nonlinear optimization problems in a short amount of time, if not real time. Such nonlinear programs are computationally challenging and time-consuming and thus do not scale with the size of the power system network. We apply a graph convolutional network (GCN) model, as a supervised learning model, for predicting an optimal load-shedding ratio that prevents transmission lines from being overloaded under line contingency (i.e., line tripping). In particular, we exploit the power system network topology in the GCN model, where the topology information is convoluted over the neural network. Using IEEE test cases, we benchmark our GCN model against a classical neural network model and a linear regression model and show that the GCN model outperforms the others by an order of magnitude.",

keywords = "Graph covolutional network, alternating current power system, contingency analysis, machine learning, neural network",

author = "Cheolmin Kim and Kibaek Kim and Prasanna Balaprakash and Mihai Anitescu",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 IEEE Power and Energy Society General Meeting, PESGM 2019 ; Conference date: 04-08-2019 Through 08-08-2019",

year = "2019",

month = aug,

doi = "10.1109/PESGM40551.2019.8973468",

language = "English",

series = "IEEE Power and Energy Society General Meeting",

publisher = "IEEE Computer Society",

booktitle = "2019 IEEE Power and Energy Society General Meeting, PESGM 2019",

}

Kim, C, Kim, K, Balaprakash, P & Anitescu, M 2019, Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency. in 2019 IEEE Power and Energy Society General Meeting, PESGM 2019., 8973468, IEEE Power and Energy Society General Meeting, vol. 2019-August, IEEE Computer Society, 2019 IEEE Power and Energy Society General Meeting, PESGM 2019, Atlanta, United States, 08/4/19. https://doi.org/10.1109/PESGM40551.2019.8973468

Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency. / Kim, Cheolmin; Kim, Kibaek; Balaprakash, Prasanna et al.
2019 IEEE Power and Energy Society General Meeting, PESGM 2019. IEEE Computer Society, 2019. 8973468 (IEEE Power and Energy Society General Meeting; Vol. 2019-August).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency

AU - Kim, Cheolmin

AU - Kim, Kibaek

AU - Balaprakash, Prasanna

AU - Anitescu, Mihai

PY - 2019/8

Y1 - 2019/8

N2 - Power system operations under contingency need to solve large-scale complex nonlinear optimization problems in a short amount of time, if not real time. Such nonlinear programs are computationally challenging and time-consuming and thus do not scale with the size of the power system network. We apply a graph convolutional network (GCN) model, as a supervised learning model, for predicting an optimal load-shedding ratio that prevents transmission lines from being overloaded under line contingency (i.e., line tripping). In particular, we exploit the power system network topology in the GCN model, where the topology information is convoluted over the neural network. Using IEEE test cases, we benchmark our GCN model against a classical neural network model and a linear regression model and show that the GCN model outperforms the others by an order of magnitude.

AB - Power system operations under contingency need to solve large-scale complex nonlinear optimization problems in a short amount of time, if not real time. Such nonlinear programs are computationally challenging and time-consuming and thus do not scale with the size of the power system network. We apply a graph convolutional network (GCN) model, as a supervised learning model, for predicting an optimal load-shedding ratio that prevents transmission lines from being overloaded under line contingency (i.e., line tripping). In particular, we exploit the power system network topology in the GCN model, where the topology information is convoluted over the neural network. Using IEEE test cases, we benchmark our GCN model against a classical neural network model and a linear regression model and show that the GCN model outperforms the others by an order of magnitude.

KW - Graph covolutional network

KW - alternating current power system

KW - contingency analysis

KW - machine learning

KW - neural network

UR - https://www.scopus.com/pages/publications/85079034759

U2 - 10.1109/PESGM40551.2019.8973468

DO - 10.1109/PESGM40551.2019.8973468

M3 - Conference contribution

AN - SCOPUS:85079034759

T3 - IEEE Power and Energy Society General Meeting

BT - 2019 IEEE Power and Energy Society General Meeting, PESGM 2019

PB - IEEE Computer Society

T2 - 2019 IEEE Power and Energy Society General Meeting, PESGM 2019

Y2 - 4 August 2019 through 8 August 2019

ER -

Graph Convolutional Neural Networks for Optimal Load Shedding under Line Contingency

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this