State Estimation-based Active Control Strategy Of High-speed Railway Pantograph

The stable contact between the pantograph and the catenary is one of the main factors that impact the normal operation of the high-speed railway.The coupling quality has a decisive influence on the current collection quality(CCQ)of the locomotives.In general,the contact force between the pantograph and the catenary(CFPC)can stand for the CCQ to a great extent.With the speed increasing and the influence of external disturbances,the fluctuation of CFPC becomes more and more violent.Active control of pantograph is an effective way to decrease the CFPC fluctuation.However,previous studies were based on simple models,which resulted in inadequate validations.And the input limitation and the actuator delay were not considered.Based on above analyses,the study objects of this thesis are the high-speed pantograph,the single pantograph-catenary system(SPCS)and the double pantograph-catenary system(DPCS).The research target is to decarese the CFPC fluctuation.The main way is to actively control the pantograph.The research contents include modelling the pantograph-catenary system(PCS),state estimation of the pantograph,finite frequency control based on estimator and the controller design considering actuator delay.Firstly,the pantograph-catenary system model is constructed.The motion differential equation of the pantograph is built based on the Lagrange equation.Then its multibody model is constructed through analyzing the kinetic and potential energies of each bar.The three level mass model of the pantograph is built.The vertical displacement of the contact point on the catenary is expressed by a serious of sinusoidal function.Combing with the loads analysis of each element,the catenary model is built based on modal analysis method.The 3D nonlinear finite element model of the catenary is constructed,in which the messenger wire is modelled as nonlinear cable elements and the contact wire is modelled as nonlinear truss elements.The penalty function method is adopted to couple the pantograph model and the catenary model.Then the single pantograph-catenary system model and the double pantograph-catenary system model are built.The Europan Standard EN 50318 and the New Benchmark are adopted to validated the system model,and results show that the models correspond to these standards.Secondly,the state estimator of the pantograph is designed.Three cases for the nominal PCS model and non-nominal PCS model are considered,including that the measured data are randomly lost and the noise statistic is unknown and time-varying.The estimator that can consider the randomly missing measurements are introduced based on standard Kalman filter.For the case of the randomly missing measurements and the parameteric uncertainties,the robust recursive state estimation method(RRSEM)is introduced.An adaptive cubature Kalman filter(ACKF)is employed to estimate the pantograph states when the noise statistic is unknown.Furthermore,the adaptive cubature Kalman filter strong tracking filter(ACKFSTF)is adopted when the noise statistic is unknown and time-varying.Combining with the SPCS model and the DPCS model,these estimators are validated.And the results show that they can accurately estimate the pantograph states under different conditions.Thirdly,two robust controllers for the pantograph are designed.The control-oriented PCS model is built based on the analysis of the main reason of the CFPC fluctuation.The tracking error of the CFPC is augmented in the state space model.The control target is mathematically described.The multiobjective robust H? controller is presented based on state estimation.The relation between the wave length in the CFPC and the catenary structure is identified based on the power spectral density(PSD).To improve the control efficiency,the multiobjective robust H? finite frequency controller is presented based on the prior information of the CFPC and state estimation.The effective and robustness of these two controllers are validated on the condition of parametric uncertainties and external disturbances based on the nonlinear SPCS model and DPCS model.The results show that the multiobjective controller has an effective impact on the CFPC standard deviation,means,control force and vertical displacement of some critical points.The prior information of CFPC is employed in the finite frequency controller,which promotes the control efficiency.Lastly,the controllers with considering actuator time delay are designed based on state estimation.Depend on the length of the time delay,two robust controllers are presented on the condition of definite time delay and stochastic time delay.The efficiency of the controllers is validated based on the nonlinear DPCS model,and its robustness is validated based on the parameteric uncertainties.The simulation results show that the control strategies that are designed on the basis of explicit method with considering actuator time delay promote the control efficiency when the time delay exists based on the linear matrix inequalities(LMI)and Lyapunov theory.