Basics of How Magnetic Bearings Work
In the mid-twentieth century, a successful magnetic levitation
bearing was successfully demonstrated. This first successful
magnetic bearing utilized electromagnets to provide attractive
forces in the five degrees of freedom (with rotation being
the sixth). Active servo control stabilized the system by
using feedback signals from position sensors in each axis
of control to vary the currents flowing through the various
electromagnets.
Several individual electromagnets, usually from 8 to 12,
were arranged in a north-south-north-south configuration around
each end of a levitated shaft to provide radial support. This
design approach, which results in a multiplicity of magnetic
flux reversals around the circumference of the shaft, is known
as heteropolar. Most commercially available magnetic bearing
systems utilize this technology. A typical heteropolar magnetic
bearing system is shown in the figure below. The stator, composed
of an array of stationary electromagnets, generates powerful
attraction forces that suspend the ferrous rotor shaft in
the center of the magnetic field (with the help of an active
servo-control unit).
Typical Heteropolar Magnetic Bearing
A magnetic flux bias field is used to linearize the relationship
of force to control current. The bias is usually about 50%
of the full magnetic field capability of the bearing. The
position of the rotor is maintained dynamically through a
continuous feedback system, consisting of a position sensor,
a control loop, and a current amplifier for each axis of control.
This servo control system continuously varies the current
flow through the control coils and thus the intensity of control
flux, which modulates the air gap magnetic field so as to
keep the rotor in the correct geometric position. Typically
there are two radial bearings and one thrust bearing for a
complete 5-axis magnetic bearing system. Rotor laminations
are needed to allow for high bandwidth response and to reduce
eddy current loss and consequential rotor heating.
A typical magnetic bearing control system (see Figure 2 below)
is tuned to the required dynamic characteristics of the rotating
equipment, usually through the implementation of a proportional-integral
derivative (PID) control algorithm that can vary the bearing
stiffness and damping as a function of rotor disturbance frequencies.
The algorithm is based on the bearing transfer function (bearing
stiffness and damping factors). A typical transfer function
is illustrated in Figure 3, where the gain required to achieve
the desired bearing stiffness is shown as a function of frequency,
and the phase-lead required to provide the desired level of
damping is shown as a function of frequency.
Figure 1. Typical System Control Loop
Figure 2. Typical Bearing Transfer Function
The transfer function in Figure 3 provides a typical profile
of increasing stiffness at low frequency, with the stiffness
often approaching 108 lb/in at zero frequency. The notch shown
at about 800 Hz is a filter introduced to make the bearing
very compliant, either to avoid exciting a flexible mode at
that specific speed or to allow the rotor to spin at the notched
frequency without generating large imbalance loads. There
will be some rotor eccentricity while operating at the notched
frequency, with the orbit of motion determined by the amount
of imbalance in the rotor. Note that the bearing stiffness
drops off rapidly above the operating speed (about 20,000
rpm). This drop-off in gain is intentional to avoid exciting
resonant frequencies that exist above the operating range.
Alarms and trips can be set for a pre-determined bearing
load or a rotor deflection that exceeds a predetermined limit.
An alarm will provide the operator with audio and visual warnings,
while a trip can de-energize the drive unit.
To improve product reliability, an uninterruptible back-up
power supply can be provided to improve system reliability.
Typically, the capacity of a battery back-up system capable
of providing up to 30 minutes of uninterrupted operation is
less than one kilowatt-hour.
A catcher bearing (also known as an auxiliary, back-up or
touchdown bearing) is required to protect the rotor/stator
interface when the equipment is de-energized during assembly
or a shutdown, and in the event of a power loss or a severe
transient beyond the force capacity of the bearing. Typically
the catcher bearing is designed to be a readily replaceable
rolling element or sleeve type bearing, capable of surviving
5 to 100 lifetime drops from full speed.
The radial clearance between the magnetic bearing stator
and the rotor shaft are usually on the order of 20 to 50 mil
(G1 in Figure 4), and those of the catcher bearings are about
half of that value (G2). The clearances in any rotating seals
(G3) must be designed with the first two clearances in mind.
Alternatively the catcher bearing clearance can be reduced
somewhat if the machine process requires tighter seal clearances.
Typical Clearances of Magnetic Bearing Components
|
