Adaptive variable bias control (AVBC) is an improved technique for controlling the currents supplied to the electromagnets in magnetic bearings. In AVBC, the bias currents are varied in such a way as to reduce overall power consumption.

In a magnetic bearing (see Figure 1), the magnetic forces depend nonlinearly upon the currents flowing in the electromagnet coils and on the distances between the rotor and the stator poles. Because the magnetic forces are attractive only and decrease with increasing distances, magnetic bearings are open-loop unstable and, consequently, closed-loop control is necessary for stable operation. A conventional control scheme for a magnetic bearing includes the decomposition of the total current into a constant bias component and a varying command component in order to linearize the relationship between the command component and the force delivered by the bearing. Usually, such classical feedback control techniques as proportional-derivative (PD) or proportional-integral-derivative (PID) are used.

Figure 1. A Magnetic Bearing utilizes magnetic forces to levitate a rotor without rotor/stator contact. For the sake of simplicity, electromagnets are shown on only one axis; to function properly, a bearing must include electromagnets on at least two axes.

Bias currents give rise to consumption of power, even when no loads are supported by the bearings. AVBC was developed as an alternative control technique in which bias currents, and thus power-consumption levels, are reduced. Another, closely related benefit of reductions in bias currents is reduction in overall generation and dissipation of heat (including eddy-current and hysteretic heating of rotors).

AVBC is based on the conventional bias method, but departs from that method according to the principle that by varying the bias current adaptively as needed, the power consumption in a bearing can be reduced while producing the desired force and maintaining the linearization afforded by using bias. The main static influence in changing the bias current is to change the static stiffness of the bearing. Dynamically, changing the bias current exerts two effects: the control laws depend on bias, and as the bias changes, the force-slew-rate limit changes. Inas- much as there are many possible ways to vary the bias, adaptive control is introduced. Adaptive control in this application entails monitoring the operation of the magnetic bearing and determining, from the information thus acquired, the best setting for the bias current.

Figure 2. This Eight-Pole Magnetic-Bearing Test Apparatus was used in experiments on AVBC.

In this approach, the bias setting is not based on typical parameter estimation; instead, it is based on the commanded current. More specifically, the adaptation law is that the bias current is required to equal the average of the amplitude of the command current over a suitable interval of time. The net result is that an AVBC controller is primarily a PD controller with (1) a bias current that is varied and is usually less than it would be in an equivalent conventional controller and (2) a control gain that varies slowly with time and with the bias current. Theoretical calculations and experimentation (see Figure 2) have shown that AVBC is nominally stable and that in comparison with conventional bias control, AVBC yields reduced power consumption and comparable control performance.

This work was done by Dexter Johnson and Gerald V. Brown of Glenn Research Center and Daniel J. Inman of Virginia Polytechnic Institute and State University. For further information, access the Technical Support Package (TSP) free on-line at under the Machinery/ Automation category.

Inquiries concerning rights for the commercial use of this invention should be addressed to

NASA Glenn Research Center
Commercial Technology Office
Attn: Steve Fedor
Mail Stop 4 - 8
21000 Brookpark Road
Ohio 44135.

Refer to LEW-16748.

Motion Control Tech Briefs Magazine

This article first appeared in the February, 2000 issue of Motion Control Tech Briefs Magazine.

Read more articles from the archives here.