Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics

This paper presents, for the first time, a framework for Kolmogorov-Arnold Networks (KANs) in power system applications. Inspired by the recently proposed KAN architecture, this paper proposes physics-informed Kolmogorov-Arnold Networks (PIKANs), a novel KAN-based physics-informed neural network (PI...

Full description

Saved in:
Bibliographic Details
Main Authors: Hang Shuai, Fangxing Li
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Open Access Journal of Power and Energy
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10843279/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832540459187044352
author Hang Shuai
Fangxing Li
author_facet Hang Shuai
Fangxing Li
author_sort Hang Shuai
collection DOAJ
description This paper presents, for the first time, a framework for Kolmogorov-Arnold Networks (KANs) in power system applications. Inspired by the recently proposed KAN architecture, this paper proposes physics-informed Kolmogorov-Arnold Networks (PIKANs), a novel KAN-based physics-informed neural network (PINN) tailored to efficiently and accurately learn dynamics within power systems. PIKANs offer a promising alternative to conventional Multi-Layer Perceptrons (MLPs) based PINNs, achieving superior accuracy in predicting power system dynamics while employing a smaller network size. Simulation results on test power systems underscore the accuracy of the PIKANs in predicting rotor angle and frequency with fewer learnable parameters than conventional PINNs. Specifically, PIKANs can achieve higher accuracy while utilizing only 50% of the network size required by conventional PINNs. Furthermore, simulation results demonstrate PIKANs’ capability to accurately identify uncertain inertia and damping coefficients. This work opens up a range of opportunities for the application of KANs in power systems, enabling efficient dynamic analysis and precise parameter identification.
format Article
id doaj-art-5c149b6426944a578822cf8553c9eef7
institution Kabale University
issn 2687-7910
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Open Access Journal of Power and Energy
spelling doaj-art-5c149b6426944a578822cf8553c9eef72025-02-05T00:01:21ZengIEEEIEEE Open Access Journal of Power and Energy2687-79102025-01-0112465810.1109/OAJPE.2025.352992810843279Physics-Informed Kolmogorov-Arnold Networks for Power System DynamicsHang Shuai0https://orcid.org/0000-0001-5087-699XFangxing Li1https://orcid.org/0000-0003-1060-7618Department of Electrical Engineering and Computer Science, The University of Tennessee, Knoxville, TN, USADepartment of Electrical Engineering and Computer Science, The University of Tennessee, Knoxville, TN, USAThis paper presents, for the first time, a framework for Kolmogorov-Arnold Networks (KANs) in power system applications. Inspired by the recently proposed KAN architecture, this paper proposes physics-informed Kolmogorov-Arnold Networks (PIKANs), a novel KAN-based physics-informed neural network (PINN) tailored to efficiently and accurately learn dynamics within power systems. PIKANs offer a promising alternative to conventional Multi-Layer Perceptrons (MLPs) based PINNs, achieving superior accuracy in predicting power system dynamics while employing a smaller network size. Simulation results on test power systems underscore the accuracy of the PIKANs in predicting rotor angle and frequency with fewer learnable parameters than conventional PINNs. Specifically, PIKANs can achieve higher accuracy while utilizing only 50% of the network size required by conventional PINNs. Furthermore, simulation results demonstrate PIKANs’ capability to accurately identify uncertain inertia and damping coefficients. This work opens up a range of opportunities for the application of KANs in power systems, enabling efficient dynamic analysis and precise parameter identification.https://ieeexplore.ieee.org/document/10843279/Kolmogorov-Arnold Networks (KANs)power system dynamicsdeep learningswing equationphysics-informed neural network (PINN)
spellingShingle Hang Shuai
Fangxing Li
Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
IEEE Open Access Journal of Power and Energy
Kolmogorov-Arnold Networks (KANs)
power system dynamics
deep learning
swing equation
physics-informed neural network (PINN)
title Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
title_full Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
title_fullStr Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
title_full_unstemmed Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
title_short Physics-Informed Kolmogorov-Arnold Networks for Power System Dynamics
title_sort physics informed kolmogorov arnold networks for power system dynamics
topic Kolmogorov-Arnold Networks (KANs)
power system dynamics
deep learning
swing equation
physics-informed neural network (PINN)
url https://ieeexplore.ieee.org/document/10843279/
work_keys_str_mv AT hangshuai physicsinformedkolmogorovarnoldnetworksforpowersystemdynamics
AT fangxingli physicsinformedkolmogorovarnoldnetworksforpowersystemdynamics