Model-free Adaptive Fault-tolerant Suspension Control For Maglev Train

The research of maglev train is helpful to improve the technical reserve of China’s rail transit field,enrich the diversification of transportation system,and help China to build a transportation power.Suspension system is the key technical equipment of maglev train,and its fault diagnosis and fault tolerant control methods have become the research hotspot at present.In this thesis,Model-free Adaptive Fault-tolerant Suspension control for maglev train is studied,considering the multiplicative and additive actuator fault,and the external disturbances.The main work is as follows:For the suspension system with actuator faults and external disturbances,the dynamic linearization method of model-free adaptive control theory is used to transform the suspension system into an equivalent dynamic linearization data model.The concepts of the system pseudo-partial derivative and total disturbance are introduced,and the relationship between the pseudo-partial derivative and multiplicative faults and the gain of the electromagnet is established,the external disturbances and additive faults are reduced to total disturbances.Two estimation algorithms for pseudo partial derivatives and total perturbations in the dynamic linearized data model of suspension systems are proposed.The first one is the pseudo partial derivative and total disturbance dispersion estimation algorithm,including the disturbance estimation algorithm based on disturbance observer and the pseudo partial derivative estimation algorithm based on projection algorithm,the second is the lumped estimation algorithm of pseudo partial derivative and total disturbance,the pseudo partial derivative and total disturbance are taken as the augmented gradient parameters,and the augmented gradient parameter estimation algorithm is designed based on the projection algorithm.In order to verify the effectiveness of the proposed algorithm,the estimation of pseudo-partial derivative and total disturbance of the system is equivalent to the electromagnet gain and total disturbance estimation,and compared with the electromagnet gain observer algorithm based on disturbance observer,the results show that the pseudo-partial derivative and total disturbance estimation algorithm is the best.For the nonlinear suspension system with unknown model,PID Suspension control scheme and Model Free Adaptive Suspension Control(MFASC)schemes were designed respectively.Multiplicative faults,additive faults and external disturbances of the suspension system are further considered,based on the dynamic linearized data model,the pseudo partial derivative and the centralized estimation algorithm of total disturbances,an improved Model Free Adaptive Fault Tolerant Suspension Control(MFAFTSC)was designed.In order to verify the effectiveness of the proposed levitation control scheme,the simulation is carried out in three cases: with fault and without disturbance,with fault and with disturbance,and with fault and with disturbance,the simulation results show that the MFAFTSC scheme has good control effect.