Affordable Access

An Experimental Test Procedure for Magnetic Model Identification of Multi-Three-Phase Induction Motors

Authors
  • Rubino, Sandro
  • Mandrile, Fabio
  • Tolosano, Luisa
  • Armando, Eric
  • Bojoi, Radu
Publication Date
Jan 01, 2023
Source
PORTO Publications Open Repository TOrino
Keywords
Language
English
License
White

Abstract

Multi-three-phase motor drives are experiencing increasing industrial development as these multiphase motors can be configured as multiple three-phase units operating in parallel. In this scenario, the literature reports several torque controllers for multi-three-phase motors to guarantee high performance in torque regulation accuracy and efficiency. However, these torque controllers almost often rely on the accurate knowledge of the magnetic model of the machine, i.e., the current-to-flux and current-to-torque relationships. When considering induction motors (IMs), most of the methodologies to evaluate flux and torque maps consist of their indirect computation from the parameters of the machine’s equivalent circuit. However, these parameters are almost always obtained by performing the no-load and locked-rotor tests, i.e., not evaluated in actual machine load conditions. In addition, if considering multi-three-phase IMs, no contributions dealing with the effective flux and torque maps in open-three-phase fault conditions are reported in the literature, making the post-fault analysis a topic not comprehensively investigated. Therefore, this paper proposes an experimental test procedure to directly identify flux and torque maps of a generic multi-three-phase IM. In this way, the magnetic model of the machine is evaluated by considering the actual operating conditions in terms of load/slip. Moreover, the proposed test procedure allows accurate machine analysis in open-three-phase fault conditions. Experimental results obtained on a 12-phase IM using a quadruple-three-phase stator winding configuration are shown, validating the proposed identification procedure.

Report this publication

Statistics

Seen <100 times