Integrations of Neural Networks and Transient Energy Functions for Designing Supplementary Damping Control of UPFC

Hung Chi Tsai, Jian Hong Liu, Chia Chi Chu*

*Corresponding author for this work

Research output: Contribution to journalJournal Article peer-review

19 Scopus citations

Abstract

This article presents an integrated scheme for supplementary damping control of unified power flow controllers (UPFCs) in order to mitigate low-frequency oscillations in power systems. The entire control system consists of the primary control and the supplementary control. In the primary control, the passivity-based control is considered in the current control loop of two converters for achieving more prompt and accurate dynamical responses. In the supplementary control, the transient energy function (TEF) approach will be explored first for designing the damping control of UPFCs. This control action can be considered as an extension of droop control used in the synchronous generator. In order to further cope with unmodeled dynamics of power systems and provide the online dynamic adaptation ability, neural networks approximated control actions are designated for online weight adjustments of the TEF-based supplementary control. Numerical simulations on three benchmark systems have been performed to validate the proposed control method for providing the extra damping and suppressing power swings even under severe operating conditions.

Original languageEnglish
Article number8807254
Pages (from-to)6438-6450
Number of pages13
JournalIEEE Transactions on Industry Applications
Volume55
Issue number6
DOIs
StatePublished - 01 11 2019
Externally publishedYes

Bibliographical note

Publisher Copyright:
© 1972-2012 IEEE.

Keywords

  • Damping control
  • energy functions
  • neural networks (NNs)
  • unified power flow controller (UPFC)

Fingerprint

Dive into the research topics of 'Integrations of Neural Networks and Transient Energy Functions for Designing Supplementary Damping Control of UPFC'. Together they form a unique fingerprint.

Cite this