Nonlinear modeling of bearingless permanent magnet synchronous motors with least squares support vector machiness

The flux linkage characteristic of the bearingless permanent magnet synchronous motor(BPMSM) is highly nonlinear,and the conventional mathematical model established by analysis method can not reflect the real characteristics of the BPMSM.Therefore,a novel modeling method is proposed for the BPMSM to take into account its nonlinearity more accurately by using the least squares support vector machiness(LSSVM).After the regression theory of the LSSVM is introduced,the LSSVM model of the BPMSM is built up by using the sampled data obtained from the experimental prototype with the finite elements method.Moreover,the LSSVM model is compared with the model based on neural network method.Simulation results show that the proposed model has desirable robustness and high accuracy.Finally,the optimal controller based on the modeling for the BPMSM is developed.