Study on suspension rigidity control of electromagnetic suspension system based on NN-PID

Permanent magnets have inherent magnetism. Power consumption of suspension system can be reduced and large air-gap can be reached when permanent magnets are applied to maglev system. But permanent magnets cannot be well controlled and the load of the electromagnetic suspension system (hereafter, EMSS) is not constant, so electromagnets should be added. Therefore a hybrid structure made of permanent magnets and electromagnets is adopted. In addition, EMSS is a typical non-linear and hysteresis system, so itpsilas difficult to obtain its precise mathematic model and itpsilas also difficult to cater to its requirement of rapidity, stability and robustness by applying general PID or other control algorithms separately. Aimed at the characteristics of EMSS, the paper discusses the possibility of applying artificial neural network PID controller (hereafter, NN-PID) to maglev suspension system based on the induced mathematic model, and a composite control model, NN-PID, is introduced by applying P part to improve the systempsilas rapidity, applying NN part to improve the systempsilas dynamic characteristics and applying PI part which can erase the systempsilas static error to improve the systempsilas steady characteristics. The simulation proves that this composite controller can make the system have an excellent stability and control accuracy.