Publication detail
APPLICATION OF LWR COMPENSATOR FOR ACTIVE MAGNETIC BEARING CONTROL
BŘEZINA, T. PULCHART, J. KREJSA, J.
Czech title
APPLICATION OF LWR COMPENSATOR FOR ACTIVE MAGNETIC BEARING CONTROL
English title
APPLICATION OF LWR COMPENSATOR FOR ACTIVE MAGNETIC BEARING CONTROL
Type
conference paper
Language
en
Original abstract
The active magnetic bearing (AMB) control is a complex problem, as it is fast, unstable and nonlinear system. The contribution deals with the simulation verification of AMB control using conventional PSD controller with nonlinear forward compensator. The automated universal approximator based on locally weighted learning (namely Receptive Field Weighted Regression – RFWR method) is used as the compensator. The paper shows the improvement of control quality criterion over conventional controller (approx. 2.5x) and furthermore the results of test of robustness against overtraining, when control is not disturbed by noise; test of robustness against the noise; test of robustness against the action delay and test of robustness against noise and action delay combination.
Czech abstract
The active magnetic bearing (AMB) control is a complex problem, as it is fast, unstable and nonlinear system. The contribution deals with the simulation verification of AMB control using conventional PSD controller with nonlinear forward compensator. The automated universal approximator based on locally weighted learning (namely Receptive Field Weighted Regression – RFWR method) is used as the compensator. The paper shows the improvement of control quality criterion over conventional controller (approx. 2.5x) and furthermore the results of test of robustness against overtraining, when control is not disturbed by noise; test of robustness against the noise; test of robustness against the action delay and test of robustness against noise and action delay combination.
English abstract
The active magnetic bearing (AMB) control is a complex problem, as it is fast, unstable and nonlinear system. The contribution deals with the simulation verification of AMB control using conventional PSD controller with nonlinear forward compensator. The automated universal approximator based on locally weighted learning (namely Receptive Field Weighted Regression – RFWR method) is used as the compensator. The paper shows the improvement of control quality criterion over conventional controller (approx. 2.5x) and furthermore the results of test of robustness against overtraining, when control is not disturbed by noise; test of robustness against the noise; test of robustness against the action delay and test of robustness against noise and action delay combination.
Keywords in English
Active magnetic bearing, control, forward compensator, RFWR
RIV year
2004
Released
07.07.2004
Publisher
Czech Technical University in Prague , Faculty of Mechanical Engineering Prague, Czech Republic
Location
Praha
ISBN
80-01-03088-1
Book
3rd Internatiomal Congress on Mechatronics MECH2K4
Pages count
15
BIBTEX
@inproceedings{BUT12459,
author="Tomáš {Březina} and Jaroslav {Pulchart} and Jiří {Krejsa},
title="APPLICATION OF LWR COMPENSATOR FOR ACTIVE MAGNETIC BEARING CONTROL",
booktitle="3rd Internatiomal Congress on Mechatronics MECH2K4",
year="2004",
month="July",
publisher="Czech Technical University in Prague , Faculty of Mechanical Engineering Prague, Czech Republic",
address="Praha",
isbn="80-01-03088-1"
}