Electrical generator or motor structure – Dynamoelectric – Rotary
Reexamination Certificate
2000-06-07
2002-08-06
Waks, Joseph (Department: 2834)
Electrical generator or motor structure
Dynamoelectric
Rotary
C310S06800R
Reexamination Certificate
active
06429561
ABSTRACT:
BACKGROUND AND SUMMARY OF THE INVENTION
The present invention relates to magnetic bearing systems and methods for controlling magnetic bearing systems. More specifically, this invention describes a system and a method for moving the rotor position setpoint away from the center point between two opposing bearings and closer to one of the bearings.
Magnetic bearings provide a host of advantages over traditional bearings. Chief among these are decreased frictional loss leading to increased efficiency and the possibility of increased rotational speeds and increased component life. However, the control of the magnetic bearings often proves problematic.
A typical magnetic bearing system, shown in
FIG. 1
, detects the position of the rotor shaft
19
using either Hall sensors or inductive sensors
16
. These sensors create an output voltage that is proportional to the magnetic flux, which is inversely proportional to the air gap width between the sensors and the rotor. The position signal is sent to a main controller, which compares the position to a pre-determined setpoint
13
and emits an output current proportional to the change in the bearing current that is necessary to bring the shalt back to the setpoint. The controller output current often passes through current amplifiers
18
that then emit the correctly scaled and conditioned bearing current. In state of the art bearings, the bearings operate in groups of at least one dual-magnet pair
20
and
21
. That means that as the current and therefore the force in one is increased, the current and force in the other is decreased by a similar amount. This methodology allows for twice the response of a non-dual-magnet pair. In order to enact the dual-magnet scheme, it is necessary to supply a bias current to the bearing pair. In this way, the current in each bearing can either be decreased or increased, whereas in a bearing without a bias current, no decrease is possible as the current is zero in the normal equilibrium state. It should be noted that some systems operate the bearings independently of each other in a non-dual magnet pair. In such a system, only the bearing that needs to exert the corrective force on the rotor position is activated.
In current magnetic bearing systems, the main dynamic controller typically operates using proportional-integral-derivative (PID), proportional-integral (PI), or proportional-derivative (PD) algorithms. There are many forms of the control algorithm equation, but all of them utilize the error signal, defined as the difference between the actual position and the position setpoint, along with combinations of the integral and derivative of the error signal. Each of these three components of the controller algorithm are multiplied by a separate gain that determines to what degree each term has control over the outputted correction signal. One form of the PID equation is given in Equation 1 with the transfer function given in Equation
G
(
t
)
=
K
TG
[
K
P
e
(
t
)
+
K
I
∫
e
(
t
)
ⅆ
t
+
K
D
ⅆ
ⅆ
t
e
(
t
)
]
[
1
]
G
(
s
)
=
K
TG
[
K
P
+
K
I
s
+
K
D
s
]
[
2
]
Equation 2 is the Laplace transform of Equation 1. The Laplace transform associates one function with a simpler function of another variable. Often, this is used to switch between the time and frequency domains.
Additional filters such as low pass filters, notch filters, or lag-lead filters are sometimes employed to filter out high frequency noise and improve system stability.
Magnetic bearings are inherently unstable without dynamic control in place. One method of increasing the system stability is to decrease the bias current. However, by decreasing the bias current, the current in one of the dual-set bearings can approach zero under a lesser bearing load. Once the bearing current is at zero, the overall control of the magnetic bearings becomes highly non-linear because the system has effectively switched from utilizing a dual-magnet pair of bearings to utilizing only single independent bearings. Therefore, it would be preferable to use a method that decreases the bias current to increase system stability while maintaining both bearing currents above zero.
Additionally, different PID, PI, PD and filter constants are desirable for the dynamic controller for start-up and run situations. Start-up involves a large sudden transient as the bearing is levitated. Therefore, a set of parameters that can handle a large transient is needed. However, during operation, a different set of tuning parameters may be desirable in order to better track the steady-state disturbances in the system and thereby achieve better control over the bearings. These two sets of parameters may be very different, so a method of switching between two controller settings would be very beneficial.
In addition to the standard PD, PI, and PID control algorithms, over the past ten years, researchers have investigated several nonlinear control methods for magnetic suspension systems, including magnetic bearings. One technique is based on nonlinear feedback linearization, which gives excellent performance, but requires an accurate nonlinear design model as well as a central processing unit (CPU) capable of computing the nonlinear control algorithm. Variable structure control methods have also been investigated, but these methods are not suitable for use with current amplifiers because the control output is a high frequency switching signal. Many nonlinear controllers cannot be retrofit to commercial magnetic bearing systems, which are set up for PID-type control. Hence, a control system with improved stability based on linear models would be highly advantageous.
U.S. Pat. No. 6,023,115 uses a series of short duration voltage pulses to increase the force in radial electromagnets at several points during start-up as a means of preventing dragging in motors at their resonant frequencies. However, this does not provide steady-state position variation improvement and the added bias current may decrease the stability of the system.
U.S. Pat. No. 5,760,510 uses a CPU to determine the frequency spectrum of the position oscillations and then outputs a current to create a magnetic flux that counters each separate frequency component, but the position displacement is something that he seeks to eliminate.
U.S. Pat. No. 5,703,424 details a method of correcting for air gap fluctuations that occur in the natural operation of the system. The inventors do this partially by reducing the bias current. However, though he realizes the importance of reducing the bias current for increased stability, he does not provide for a means of reducing the bias current below the natural limit of the bearing system.
U.S. Pat. No. 5,471,106 uses the variable flux that results from the natural movement of the rotor during normal operation as an input in the determination of a current that will reject disturbances of varying frequencies, but the position displacement is again something the inventors seek to eliminate.
The above inventions all assume a central position for the rotor equidistant between two dual-magnet bearings, and they all vary the current as a means of producing the variable flux and hence the variable force on the rotor necessary to maintain that central position. However, a controller that varied the rotor position setpoint is a means of varying the flux and force would have several advantages, assuming the rotor was under a unidirectional load. First, since force is related directly to current and inversely to the air gap width, for a constant current a smaller air gap would produce a greater force. Thus, if a non-central position is used, a bearing magnet could produce a greater force than the force it was designed for. Secondly, given a constant force, a smaller air gap would require a smaller current. This would mean that smaller windings could be used in the bearing magnets. Third, as t
Back Dwight D.
Cole Gregory S.
Davis Russell W.
Hung John
Crowell & Moring LLP
Mainstream Engineering Corporation
Waks Joseph
LandOfFree
Magnetic bearing system and method of controlling magnetic... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Magnetic bearing system and method of controlling magnetic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Magnetic bearing system and method of controlling magnetic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2949357