Improved Direct-Model Predictive Control with a Simple Disturbance Observer for DFIGs

Mohamed Abdelrahem, Christoph Hackl, Jose Rodriguez, Ralph Kennel

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

In this paper, a computationally-efficient and robust direct-model predictive control (DMPC) technique for doubly-fed induction generators in variable-speed wind turbines systems is proposed. In order to avoid several predictions of the rotor currents, the reference voltage (RV) is directly computed from the reference currents, which notably reduces the calculation load. Furthermore, the disturbances due to parameters mismatches and un-modeled dynamics are considered in the RV calculation. Accordingly, the sensitivity to mismatches in the model parameters is avoided and a zero steady-state error is realized. Finally, based on the location of this RV, the quality function is evaluated for only two times to find the best switching state, which is applied to the power converter in the next sample. The performance of the suggested DMPC is experimentally validated and compared with the convention DMPC.

Original languageEnglish
Title of host publication2020 22nd European Conference on Power Electronics and Applications, EPE 2020 ECCE Europe
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9789075815368
DOIs
StatePublished - Sep 2020
Event22nd European Conference on Power Electronics and Applications, EPE 2020 ECCE Europe - Lyon, France
Duration: 7 Sep 202011 Sep 2020

Publication series

Name2020 22nd European Conference on Power Electronics and Applications, EPE 2020 ECCE Europe

Conference

Conference22nd European Conference on Power Electronics and Applications, EPE 2020 ECCE Europe
Country/TerritoryFrance
CityLyon
Period7/09/2011/09/20

Keywords

  • Doubly-fed induction generator
  • MPC (Model-based Predictive Control)
  • Robust control
  • Wind energy

Fingerprint

Dive into the research topics of 'Improved Direct-Model Predictive Control with a Simple Disturbance Observer for DFIGs'. Together they form a unique fingerprint.

Cite this