This paper introduces an adaptive bias current control method for an active magnetic bearing (AMB). The bearing force is analyzed theoretically, and the dynamic performance of the magnetic bearing for various bias currents is discussed. Then power consumption is analyzed and the optimum bias current that minimizes power consumption is derived. A novel optimization method using a steepest descent method is proposed. This requires less computing power than the former optimization method using a recursive Fourier transform algorithm. Experimental results show that the optimized bias current can be achieved by the proposed method. However, the dynamics of the rotor is affected by the bias current variation. In order to overcome this problem, the effects of parameter errors are investigated and correction methods are introduced. Experimental results show that the rotor dynamics are not affected by the variable bias current if the parameters are corrected. Results are also presented for machine run-up and run-down.
