Artificial-intelligence method for the derivation of generic aggregated dynamic equivalent models

Eleftherios O. Kontis, Theofilos A. Papadopoulos, Mazheruddin H. Syed, Efren Guillo-Sansano, Graeme M. Burt, Grigoris K. Papagiannis

Research output: Contribution to journalArticle

9 Citations (Scopus)
25 Downloads (Pure)

Abstract

Aggregated equivalent models for the dynamic analysis of active distribution networks (ADNs) can be efficiently developed using dynamic responses recorded through field measurements. However, equivalent model parameters are highly affected from the time-varying composition of power system loads and the stochastic behavior of distributed generators. Thus, equivalent models, developed through in situ measurements, are valid only for the operating conditions from which they have been derived. To overcome this issue, in this paper, a new method is proposed for the derivation of generic aggregated dynamic equivalent models, i.e., for equivalent models that can be used for the dynamic analysis of a wide range of network conditions. The method incorporates clustering and artificial neural network techniques to derive robust sets of parameters for a variable-order dynamic equivalent model. The effectiveness of the proposed method is evaluated using measurements recorded on a laboratory-scale ADN, while its performance is compared with a conventional technique. The corresponding results reveal the applicability of the proposed approach for the analysis and simulation of a wide range of distinct network conditions.

Original languageEnglish
Pages (from-to)2947-2956
Number of pages10
JournalIEEE Transactions on Power Systems
Volume34
Issue number4
Early online date21 Jan 2019
DOIs
Publication statusPublished - 31 Jul 2019

Keywords

  • artificial neural networks
  • black-box modeling
  • clustering
  • dynamic modeling
  • measurement-based approach

Fingerprint Dive into the research topics of 'Artificial-intelligence method for the derivation of generic aggregated dynamic equivalent models'. Together they form a unique fingerprint.

Cite this