The Parameter Estimation Method Of Rail Vehicle Lateral Suspension System Based On Q-GRBPF

As an important part of the rail vehicle system,on the one hand,the suspension system supports the carbody and the bogie;on the other hand,it provides driving in a straight line and curve,ensures the dynamic stability and provides comfort for the passengers in the vehicle.At present,the rail vehicle suspension system is mainly based on the way of the plan maintenance,which consumes a lot of time and money and could not find faults effectively.Therefore,the real-time monitoring for the rail vehicle suspension system has very important and practical significance.The evolving law of rail vehicle service performance is reflected by the change of the parameters,and in the real test,in order to estimate the parameters of the key components of the vehicle suspension system,it needs to decorate a small amount of sensors in the vehicle,and the failure could be found in time when the vehicles are running by comparing the actual and estimated values.For the key theory and technology problems of parameter estimation of the rail vehicle lateral suspension system based on Quasi-gaussian Rao-Blackwellised particle filter(Q-GRBPF),research work in this paper can be seen as follows:(1)The research status at home and abroad of the parameter estimation theory and its methods were combed systematically.The structure characteristics of the rail vehicle were analyzed,the parameter estimation of the rail vehicle lateral suspension system was a complicated nonlinear problem with many parameters and multimodal.On this basis,the parameter estimation methods based on the theory of the hybrid filter were pointed out which were the main development direction of the parameter estimation of the rail vehicle lateral suspension system.The theory on the one hand could make full use of the kalman filter(KF)which has the unique advantages to deal with linear state filtering;on the other hand could make full use of the particle filter(PF)or the quasi-gaussian particle filter(Q-GPF)which has unique advantages to process nonlinear filtering problem,and complementary advantages were formed in the way.(2)On the basis of the analysis of the basic theory of the KF,PF,and the Rao-Blackwellised particle filter(RBPF),the parameter estimation model based on RBPF was established.The effectiveness of the parameter estimation method based on RBPF was verified by an example.The theory of Q-GRBPF was optimized and the parameter estimation model based on Q-GRBPF was established on the basis of combing the basic theory of Q-GRBPF and aiming at the feature of parameter estimation.The effectiveness of the parameters estimation method based on Q-GRBPF and its parameter estimation model were verified by the example simulation.The superiority of parameter estimation method based on Q-GRBPF was verified in estimation accuracy and computational speed compared with the parameter estimation method based on RBPF.A relatively new parameter estimation method was formed.(3)The rail vehicle lateral dynamics model was established according to the principle of classical mechanics.The equation of state transition was established according to the defining method which was one of the continuous state discretization method.The observation equation was established according to the sensor configuration.The parameter estimate probability model of the augmented state space was built by combining the state transition equation and observation equation.(4)The parameter estimation method based on Q-GRBPF was combined with the rail vehicle lateral dynamics model and parameter estimation method based on Q-GRBPF for lateral suspension system of the rail vehicle was established.The validity of the model was verified through the real test and fault simulation based on SIMPACK,and its performance was compared with the parameter estimation method based on the RBPF for lateral suspension system of the railvehicle.The advantages of the parameters estimation method based on Q-GRBPF were verified in terms of the estimation accuracy and computational speed,and a relatively new parameter estimation method for the suspension system of the rail vehicle was formed.