A Caltech Library Service

DeepOPF: A Feasibility-Optimized Deep Neural Network Approach for AC Optimal Power Flow Problems

Pan, Xiang and Chen, Minghua and Zhao, Tianyu and Low, Steven H. (2022) DeepOPF: A Feasibility-Optimized Deep Neural Network Approach for AC Optimal Power Flow Problems. IEEE Systems Journal . ISSN 1932-8184. doi:10.1109/jsyst.2022.3201041. (In Press)

Full text is not posted in this repository. Consult Related URLs below.

Use this Persistent URL to link to this item:


To cope with increasing uncertainty from renewable generation and flexible load, grid operators need to solve alternative current optimal power flow (AC-OPF) problems more frequently for efficient and reliable operation. In this article, we develop a deep neural network (DNN) approach, called DeepOPF, for solving AC-OPF problems in a fraction of the time used by conventional iterative solvers. A key difficulty for applying machine learning techniques for solving AC-OPF problems lies in ensuring that the obtained solutions respect the equality and inequality physical and operational constraints. Generalized a prediction-and-reconstruction procedure in our previous studies, DeepOPF first trains a DNN model to predict a set of independent operating variables and then directly compute the remaining ones by solving the power flow equations. Such an approach not only preserves the power-flow balance equality constraints but also reduces the number of variables to be predicted by the DNN, cutting down the number of neurons and training data needed. DeepOPF then employs a penalty approach with a zero-order gradient estimation technique in the training process toward guaranteeing the inequality constraints. We also drive a condition for tuning the DNN size according to the desired approximation accuracy, which measures its generalization capability. It provides theoretical justification for using DNN to solve AC-OPF problems. Simulation results for IEEE 30/118/300-bus and a synthetic 2000-bus test cases demonstrate the effectiveness of the penalty approach. They also show that DeepOPF speeds up the computing time by up to two orders of magnitude as compared to a state-of-the-art iterative solver, at the expense of < 0.2% cost difference.

Item Type:Article
Related URLs:
URLURL TypeDescription ItemDiscussion Paper
Pan, Xiang0000-0002-6565-2339
Chen, Minghua0000-0003-4763-0037
Zhao, Tianyu0000-0002-9541-0197
Low, Steven H.0000-0001-6476-3048
Additional Information:This work was supported in part by the General Research Fund from Research Grants Council, Hong Kong under Grant 11203122, in part by the InnoHK initiative, The Government of the HKSAR, and Laboratory for AI-Powered Financial Technologies, and in part by the NSF through grants ECCS under Grant 1931662, Caltech Resnick, S2I funds, and Berkeley. The authors would like to thank T. Cui, W. Huang, Q. Lin, and A. Venzke for the discussions related to this study, and Y. Guo and the National Supercomputer Center in Jinan for providing GPU/CPU computing resources. They also would like to thank the anonymous reviewers for careful reading and the helpful comments.
Group:Resnick Sustainability Institute
Funding AgencyGrant Number
Research Grants Council of Hong Kong11203122
InnoHK initiativeUNSPECIFIED
Government of the HKSARUNSPECIFIED
Laboratory for AI-Powered Financial TechnologiesUNSPECIFIED
Resnick Sustainability InstituteUNSPECIFIED
University of California, BerkeleyUNSPECIFIED
Record Number:CaltechAUTHORS:20221010-454096500.22
Persistent URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:117306
Deposited By: Research Services Depository
Deposited On:14 Oct 2022 17:48
Last Modified:14 Oct 2022 17:48

Repository Staff Only: item control page