Multiparameter Identification for Active Magnetic Bearing With Uncertainties Based on a Coupled Nonlinear Model

Parameter identification enables active magnetic bearing (AMB) to obtain more state information, which can be used for control optimization, condition monitoring, and fault diagnosis for rotating machinery. Due to the limitation of working principle and modeling accuracy, the traditional parameter identification methods show inherent disadvantages in cost, applicability or precision. Thus, this article proposes an accurate multiparameter identification method without any instrumentation. This method takes into account the electromagnetic force coupling caused by the asymmetric magnetic circuit, the nonlinearity caused by the saturation effect, together with the uncertain parameters caused by machining and assembly errors, temperature, etc. First, the coupled nonlinear magnetic circuit model of the AMB is established by the reluctance network method, and the influence of nonideal factors on the electromagnetic force is treated. Second, a novel experimental method for parameter identification is presented. According to the equilibrium relationship between the electromagnetic force and the supporting load, the force balance equations of the rotor under different working conditions are obtained. Taking the supporting parameters and uncertain geometric parameters as the parameters to be identified, the multiparameter identification model is established. Finally, the proposed method is validated by the finite element method and experiments. The results show that the method can accurately identify the air gap length, rotor origin, static supporting load, and stiffness of the AMB.