Magnetic levitation of a one DOF system using simultaneous actuation and displacement sensing technique

This paper discusses the development of a displacement self-sensing estimator for a one degree of freedom (DOF) magnetically levitated system using the parameter estimation technique. Current demodulation technique is known to be capable of extracting the air gap length from the coil current with demodulation filters. However, one of its main disadvantages is that its output is a function of the air gap length and the duty cycle of the PWM amplifiers. It is therefore demonstrated here, through computer simulation and experimental investigation, that the self-sensing parameter estimation technique is capable of removing the variable duty cycle from the estimated output. It is composed of two identical demodulation filters, one coil inductance simulator and one PI convergence controller. Benefiting from the closed loop characteristics of the self-sensing parameter estimation, not only is the influence of the duty cycle removed, but the dynamic characteristics of the selfsensing system are also greatly enhanced. The design of the analogue circuitries implementing the algorithm of the self-sensing parameter estimator is described. Very good agreement in the static and dynamic calibrations of the estimator output is observed, when compared with a dial gauge and commercial eddy current probe. While the self-sensing active magnetic bearing (AMB) system is levitated, excellent signal tracking capability of the parameter estimator is noted.