This paper discusses frequency domain identification for a MIMO active magnetic bearing (AMB) rigid rotor system based on measured open-loop frequency responses. The frequency responses measured by a digital controller include the effects of sensor, amplifier, and discretization dynamics. These dynamics affect the identification of poles of the AMB system as well as the achievable performance of the controller. A new reduced order frequency domain identification of a multi-input/multi-output (MIMO) AMB rigid rotor system is proposed, which implicitly includes these dynamics along with the rest of the identified plant. The proposed identification scheme consists of three distinct steps: i) sensor, amplifier and delay dynamics; ii) rigid modes; and iii) flexible modes. The proposed scheme directly tackles the open-loop MIMO frequency data and is based on an explicit MIMO parameterization. An accurate reduced order model, suitable for advanced control synthesis, can be obtained through the proposed identification scheme.Copyright © 2003 by ASME
[1]
R. G. Green,et al.
A high frequency stray-immune capacitance transducer based on the charge transfer principle
,
1988
.
[2]
Thomas F. Coleman,et al.
Optimization Toolbox User's Guide
,
1998
.
[3]
Raoul Herzog,et al.
Multivariable Identification of Active Magnetic Bearing Systems
,
1997
.
[4]
David S. Bayard,et al.
Multivariable Frequency Domain Identification via 2-Norm Minimization
,
1992,
1992 American Control Conference.
[5]
안형준.
A Study on system identification and vibration control of the AMB spindle for high speed precision machining using cylindrical capacitive sensors
,
2001
.
[6]
Thomas Kailath,et al.
Linear Systems
,
1980
.