Controller parameters optimization of large inertia and time-varying load system based on extremum seeking algorithm

Jiangbo Zhao, Feng Wei

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

Abstract

It is difficult to tune controller parameters for large inertia and time-varying load system. To solve this problem, a new controller parameters optimization strategy based on extremum seeking algorithm is proposed. This novel method is model-free. It utilizes data of experiments to get the gradient information of performance to control parameters, so as to tune the controller parameters. This paper analyzes the convergence of extremum seeking algorithm and proposes the high-pass filter parameters selection principles. Simulation and experimental results show that the system control performance is significantly improved using extremum seeking algorithm compared to traditional methods.

Original languageEnglish
Title of host publicationProceedings - 2013 International Conference on Mechatronic Sciences, Electric Engineering and Computer, MEC 2013
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages107-111
Number of pages5
ISBN (Electronic)9781479925650
DOIs
Publication statusPublished - 2013
Event2013 International Conference on Mechatronic Sciences, Electric Engineering and Computer, MEC 2013 - Shenyang, China
Duration: 20 Dec 201322 Dec 2013

Publication series

NameProceedings - 2013 International Conference on Mechatronic Sciences, Electric Engineering and Computer, MEC 2013

Conference

Conference2013 International Conference on Mechatronic Sciences, Electric Engineering and Computer, MEC 2013
Country/TerritoryChina
CityShenyang
Period20/12/1322/12/13

Keywords

  • Extremum seeking
  • Gradient information
  • Large inertia system
  • Parameters tuning

Fingerprint

Dive into the research topics of 'Controller parameters optimization of large inertia and time-varying load system based on extremum seeking algorithm'. Together they form a unique fingerprint.

Cite this