A Directed Acyclic Graph Neural Network for AC Optimal Power Flow

Zhenping Guo; Kai Sun; Byungkwon Park; Srdjan Simunovic; Wei Kang

doi:10.1109/PESGM52003.2023.10252547

A Directed Acyclic Graph Neural Network for AC Optimal Power Flow

Zhenping Guo, Kai Sun, Byungkwon Park, Srdjan Simunovic, Wei Kang

Multiscale Materials

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

Abstract

AC optimal power flow (OPF) is of great significance for power system security, reliability, and economy. As an NP-hard problem, its solution can be time consuming by traditional optimization techniques. For more efficient AC OPF algorithms, a Direct Acyclic Graph Neural Network (DAG-NN) is proposed in this paper, which enables an explicit design of a neural network utilizing the intrinsic structural information of the problem to be solved. The approach first reformulates an iterative Newton-Raphson based AC OPF algorithm as a compositional function, accordingly constructs a DAG, and then designs the neural network by realizing its each node by a shallow neural network. The paper also analyzes errors of the DAG-NN. The proposed approach is tested on a modified PJM 5-bus system.

Original language	English
Title of host publication	2023 IEEE Power and Energy Society General Meeting, PESGM 2023
Publisher	IEEE Computer Society
ISBN (Electronic)	9781665464413
DOIs	https://doi.org/10.1109/PESGM52003.2023.10252547
State	Published - 2023
Event	2023 IEEE Power and Energy Society General Meeting, PESGM 2023 - Orlando, United States Duration: Jul 16 2023 → Jul 20 2023

Publication series

Name	IEEE Power and Energy Society General Meeting
Volume	2023-July
ISSN (Print)	1944-9925
ISSN (Electronic)	1944-9933

Conference

Conference	2023 IEEE Power and Energy Society General Meeting, PESGM 2023
Country/Territory	United States
City	Orlando
Period	07/16/23 → 07/20/23

Keywords

AC optimal power flow
DAG
compositional function
directed acyclic graph
neural network

Access to Document

10.1109/PESGM52003.2023.10252547

Cite this

@inproceedings{c2b04e1dcffd457695aa7615c1471bd5,

title = "A Directed Acyclic Graph Neural Network for AC Optimal Power Flow",

abstract = "AC optimal power flow (OPF) is of great significance for power system security, reliability, and economy. As an NP-hard problem, its solution can be time consuming by traditional optimization techniques. For more efficient AC OPF algorithms, a Direct Acyclic Graph Neural Network (DAG-NN) is proposed in this paper, which enables an explicit design of a neural network utilizing the intrinsic structural information of the problem to be solved. The approach first reformulates an iterative Newton-Raphson based AC OPF algorithm as a compositional function, accordingly constructs a DAG, and then designs the neural network by realizing its each node by a shallow neural network. The paper also analyzes errors of the DAG-NN. The proposed approach is tested on a modified PJM 5-bus system.",

keywords = "AC optimal power flow, DAG, compositional function, directed acyclic graph, neural network",

author = "Zhenping Guo and Kai Sun and Byungkwon Park and Srdjan Simunovic and Wei Kang",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE Power and Energy Society General Meeting, PESGM 2023 ; Conference date: 16-07-2023 Through 20-07-2023",

year = "2023",

doi = "10.1109/PESGM52003.2023.10252547",

language = "English",

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

publisher = "IEEE Computer Society",

booktitle = "2023 IEEE Power and Energy Society General Meeting, PESGM 2023",

}

Guo, Z, Sun, K, Park, B, Simunovic, S & Kang, W 2023, A Directed Acyclic Graph Neural Network for AC Optimal Power Flow. in 2023 IEEE Power and Energy Society General Meeting, PESGM 2023. IEEE Power and Energy Society General Meeting, vol. 2023-July, IEEE Computer Society, 2023 IEEE Power and Energy Society General Meeting, PESGM 2023, Orlando, United States, 07/16/23. https://doi.org/10.1109/PESGM52003.2023.10252547

A Directed Acyclic Graph Neural Network for AC Optimal Power Flow. / Guo, Zhenping; Sun, Kai; Park, Byungkwon et al.
2023 IEEE Power and Energy Society General Meeting, PESGM 2023. IEEE Computer Society, 2023. (IEEE Power and Energy Society General Meeting; Vol. 2023-July).

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

TY - GEN

T1 - A Directed Acyclic Graph Neural Network for AC Optimal Power Flow

AU - Guo, Zhenping

AU - Sun, Kai

AU - Park, Byungkwon

AU - Simunovic, Srdjan

AU - Kang, Wei

PY - 2023

Y1 - 2023

N2 - AC optimal power flow (OPF) is of great significance for power system security, reliability, and economy. As an NP-hard problem, its solution can be time consuming by traditional optimization techniques. For more efficient AC OPF algorithms, a Direct Acyclic Graph Neural Network (DAG-NN) is proposed in this paper, which enables an explicit design of a neural network utilizing the intrinsic structural information of the problem to be solved. The approach first reformulates an iterative Newton-Raphson based AC OPF algorithm as a compositional function, accordingly constructs a DAG, and then designs the neural network by realizing its each node by a shallow neural network. The paper also analyzes errors of the DAG-NN. The proposed approach is tested on a modified PJM 5-bus system.

AB - AC optimal power flow (OPF) is of great significance for power system security, reliability, and economy. As an NP-hard problem, its solution can be time consuming by traditional optimization techniques. For more efficient AC OPF algorithms, a Direct Acyclic Graph Neural Network (DAG-NN) is proposed in this paper, which enables an explicit design of a neural network utilizing the intrinsic structural information of the problem to be solved. The approach first reformulates an iterative Newton-Raphson based AC OPF algorithm as a compositional function, accordingly constructs a DAG, and then designs the neural network by realizing its each node by a shallow neural network. The paper also analyzes errors of the DAG-NN. The proposed approach is tested on a modified PJM 5-bus system.

KW - AC optimal power flow

KW - DAG

KW - compositional function

KW - directed acyclic graph

KW - neural network

UR - http://www.scopus.com/inward/record.url?scp=85174738282&partnerID=8YFLogxK

U2 - 10.1109/PESGM52003.2023.10252547

DO - 10.1109/PESGM52003.2023.10252547

M3 - Conference contribution

AN - SCOPUS:85174738282

T3 - IEEE Power and Energy Society General Meeting

BT - 2023 IEEE Power and Energy Society General Meeting, PESGM 2023

PB - IEEE Computer Society

T2 - 2023 IEEE Power and Energy Society General Meeting, PESGM 2023

Y2 - 16 July 2023 through 20 July 2023

ER -

A Directed Acyclic Graph Neural Network for AC Optimal Power Flow

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this