AC machine torque and stator flux estimation using a neural network based on the steady-state 2D field model

Author

Grzesiak, Lech M.

Author_Institution

Inst. of Control & Ind. Electrons., Warsaw Univ. of Technol.

Volume

1

fYear

1996

fDate

17-20 Jun 1996

Firstpage

358

Abstract

This paper presents a neural estimator of torque and stator flux, based on knowledge of stator current and speed. The suggested estimator is constructed from a multilayer feedforward neural network. Training sets are based on calculations using an FEM model of an induction machine. The steady-state performance of an AC motor has been assumed and the nonlinearity of the laminated core has been taken into consideration while modelling the problem. The end-region´s impedance has been considered by means of modification of the rotor cage conductivity

Keywords

electric machine analysis computing; feedforward neural nets; finite element analysis; learning (artificial intelligence); machine theory; magnetic flux; multilayer perceptrons; parameter estimation; rotors; squirrel cage motors; stators; torque; AC motor; FEM model; computer simulation; end-region impedance; laminated core; modelling; multilayer feedforward neural network; rotor cage conductivity; squirrel cage induction motor; stator flux estimation; steady-state 2D field model; steady-state performance; torque estimation; training sets; AC machines; AC motors; Feedforward neural networks; Impedance; Induction machines; Multi-layer neural network; Neural networks; Stator cores; Steady-state; Torque;

fLanguage

English

Publisher

ieee

Conference_Titel

Industrial Electronics, 1996. ISIE '96., Proceedings of the IEEE International Symposium on

Conference_Location

Warsaw

Print_ISBN

0-7803-3334-9

Type

conf

DOI

10.1109/ISIE.1996.548447

Filename

548447