Skip to main navigation Skip to search Skip to main content

Design of electric dynamic load simulator based on recurrent neural networks

  • Mingyan Wang
  • , Ben Guo
  • , Yudong Guan
  • , Hao Zhang
  • Harbin Institute of Technology

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

Abstract

This paper describes the electric dynamic load simulator (DLS) driving by permanent magnet synchronous motor. It can reproduce desired load torque acting on loaded objective to test its performance. A simplified dynamic model is derived to clarify the causation of redundancy torque caused by the motion of loaded objective and illustrate the drawbacks of applying conventional control strategy. Real-time recurrent neural networks based iterative learning control strategy is adopted. It can restrain redundancy torque and improve the accuracy of load torque in spite of the nonlinearity and uncertainty in the system. The expected results have been obtained from simulation and experiment.

Original languageEnglish
Title of host publicationIEMDC 2003 - IEEE International Electric Machines and Drives Conference
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages207-210
Number of pages4
ISBN (Electronic)0780378172, 9780780378179
DOIs
StatePublished - 2003
EventIEEE International Electric Machines and Drives Conference, IEMDC 2003 - Madison, United States
Duration: 1 Jun 20034 Jun 2003

Publication series

NameIEMDC 2003 - IEEE International Electric Machines and Drives Conference
Volume1

Conference

ConferenceIEEE International Electric Machines and Drives Conference, IEMDC 2003
Country/TerritoryUnited States
CityMadison
Period1/06/034/06/03

Keywords

  • Actuators
  • Mathematical model
  • Mechanical sensors
  • Motion control
  • Nonlinear dynamical systems
  • Permanent magnet motors
  • Proportional control
  • Recurrent neural networks
  • Rotors
  • Torque control

Fingerprint

Dive into the research topics of 'Design of electric dynamic load simulator based on recurrent neural networks'. Together they form a unique fingerprint.

Cite this