Sensor runout compensation in active magnetic bearings via an integral adaptive observer

• An integral adaptive observer is proposed to compensate sensor runout in AMB systems.
• The Lyapunov method is used to prove asymptotic stability of the closed-loop system.
• The gains of the proposed observer are determined by solving an LMI problem.
• Simulations and experiments are performed on a three-pole AMB system.
• The proposed observer attenuates disturbances due to both runout and unbalance.

Sensor runout is one of the main sources of harmonic disturbances in active magnetic bearing systems. This type of the disturbance not only causes harmonic vibrations in the system but also changes the steady-state position of the axis of rotation from the geometric center of the AMB. In this paper, an integral adaptive observer is proposed to identify the dc and harmonic content of the sensor runout and to estimate the states of the system at the same time. The Lyapunov method is used to prove asymptotic stability of the proposed observer. Unlike the proportional observer which amplifies the measurement error, the sensor runout can be completely compensated when the states of the integral adaptive observer are used for feedback stabilization. It is shown that the proposed technique can also attenuate rotor displacements, when both sensor runout and mass unbalance disturbances are applied to the system. Simulation results have been presented for both cases to demonstrate the performance of the integral adaptive observer. Experimental results are also obtained by an AMB test rig, which confirm the effectiveness of the proposed method.