Parameter Estimation of Solid Cylindrical Electromagnetic Actuator Using Radial Basis Function Neural Networks

The electro-magnetic actuator presents a solution for most of the technical problems of the traditional mechanical bearings since it ensures the total levitation of a body in space eliminating any mechanical contact between the stator and the levitated body. In practice there is lot of difference between theoretical force and actual force developed between the stator and rotor of an actuator and it varies with air gap and current. This difference is mainly due to different losses in the system. In the present work a correction factor is introduced to account for different losses. A radial basis function neural network (RBFNN) has been implemented to estimate the correction factor and validated with experimental values. The RBF network has been used to estimate the actuator parameters namely force, position stiffness and current stiffness. The RBF predicted values have been validated with the experimental values.