Mathematical Model And Decoupling Control Of Bearingless Switched Reluctance Motor Under Non-ideal State

A bearingless switched reluctance motor (BSRM for short) is developed with the rotor which has both rotating and suspending function. Radial force windings are superimposed in stator together with rotary force winding. It not only has the merits of magnetic bearings, but also can work high axial utilization ratio, small capacity, and low power consumption. These features can demonstrate the motor’s potential applications in numerical control machines, industry robots, aerospace, and special electrical drive, etc. And it also has important practical significance for economic power improvement.There is complex electromagnetic coupling between rotating windings and bearing windings in bearingless switched reluctance motors. And magnetic saturation and radial displacement also tend to arise. So it is very difficult and complicated to gain an accurate mathematical model and realize decoupling of radial force and torque.So firstly, the characteristics of magnetic force in BSRM were analyzed with finite element method in this article. Secondly, a mathematical model under non-ideal state was computed based on Maxwell tensor method. Thirdly, the motor’s working state was divided into linear area and nonlinear area with a criterion for saturation, and two corresponding neural network inverse systems were established to decouple BSRM with an identification block online judging the motor’s working state and switching between these two neural network inverse systems. And lastly, three PID controllers were designed for pseudo-linear sub-systems. With the proposed method, the problem of insufficient consideration about varying working state can be conquered, which is a big limitation for those modeling and decoupling methods based on ideal no saturation assumption. And simulation results verified this method.