Data-Driven Model Predictive Current Control for Synchronous Machines: a Koopman Operator Approach

Horacio M. Calderon, Issa Hammoud, Thimo Oehlschlagel, Herbert Werner, Ralph Kennel

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

3 Scopus citations

Abstract

In this paper, a data-driven continuous control set model predictive current control (CCS-MPCC) scheme for permanent magnet synchronous motors (PMSMs) is proposed. The model of the motor used in the model predictive control (MPC) strategy is obtained from collected measurements using the Koopman operator (KO) theory. Experimental results on a 500W PMSM show that the obtained model has yielded excellent prediction accuracy, and that it is capable of being incorporated within a real-time CCS-MPCC scheme in the sub-millisecond typically available sampling time for the current control loop of synchronous motors.

Original languageEnglish
Title of host publication2022 International Symposium on Power Electronics, Electrical Drives, Automation and Motion, SPEEDAM 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages942-947
Number of pages6
ISBN (Electronic)9781665484596
DOIs
StatePublished - 2022
Event2022 International Symposium on Power Electronics, Electrical Drives, Automation and Motion, SPEEDAM 2022 - Sorrento, Italy
Duration: 22 Jun 202224 Jun 2022

Publication series

Name2022 International Symposium on Power Electronics, Electrical Drives, Automation and Motion, SPEEDAM 2022

Conference

Conference2022 International Symposium on Power Electronics, Electrical Drives, Automation and Motion, SPEEDAM 2022
Country/TerritoryItaly
CitySorrento
Period22/06/2224/06/22

Keywords

  • Koopman operator theory
  • Synchronous machines
  • continuous control set model predictive control
  • data-driven modelling
  • online optimization

Fingerprint

Dive into the research topics of 'Data-Driven Model Predictive Current Control for Synchronous Machines: a Koopman Operator Approach'. Together they form a unique fingerprint.

Cite this