| With the significant increase of train speed in high-speed railway,the dynamic interaction between the train and the track is obviously intensified,which leads to more and more prominent problems such as running safety,ride comfort and service safety of the track and subgrade.Moreover,due to the influence of external environmental factors and long-term dynamic load of the train,a series of structural performance degradation problems,e.g.,wheel–rail abnormal wear,damage and failure of components,subgrade cumulative settlement,etc.,will inevitably be derived,which will reduce the smoothness of rail surface,and then affect the dynamic performance of the vehicle–track–subgrade system.The vehicle–track–subgrade dynamic interaction is the root cause of this series of problems.In order to ensure the long-term safe and stable operation of high-speed railway,it is urgent to explore the dynamic coupled mechanism of the vehicle–track–subgrade system.However,the vehicle–track–subgrade system shows the characteristics of diverse components,complex dynamic interaction and large-scale solution,how to perform accurate and reasonable mechanical description and dynamic analysis is the focus and difficulty of current research.In view of this,aiming at the shortcomings of the existing vehicle–track–subgrade dynamic interaction model,this thesis proposes an accurate and reliable vehicle–track–subgrade coupled dynamics modeling method with high calculation efficiency and wide application range,and then establishes a more accurate and efficient high-speed vehicle–ballastless track-subgrade vertical coupled dynamics model.By applying the dynamics model,this thesis compares the dynamic behavior of the vehicle–track–subgrade system induced by different ballastless track types under high-speed running conditions,and analyzes the influence of three typical excitations such as wheel polygonal wear,rail weld irregularity and differential subgrade settlement on the dynamics responses of the vehicle–track–subgrade system,and carries out the research on intelligent detection of the track geometric irregularity based on the vertical accelerations of the vehicle system and meanwhile differential subgrade settlement identification.The relevant results have important theoretical significance and application prospect for the design,operation,and maintenance of the high-speed railway.Primarily,based on the vehicle–track coupled dynamics theory and the Green’s function method,a new method for the vehicle–track–subgrade coupled dynamics modeling is proposed,and a high-speed vehicle–ballastless track–subgrade vertical coupled dynamics model is established.The dynamics model divides the vehicle-track-subgrade system into the vehicle–rail subsystem and the sub-rail foundation–subgrade subsystem,and then adopts the fastener forces to transmit the dynamic interaction of the two subsystems in real time.On the one hand,the motion differential equations of the vehicle–rail subsystem are established based on the multi-rigid-body dynamics theory and the mode superposition method,and the dynamic interaction between the wheel and the rail is simulated by using the wheel–rail nonlinear contact force;on the other hand,the elaborated finite element models of the sub-rail foundation–subgrade subsystem of three types of ballastless tracks are established to obtain the system Green’s functions,and then the time-frequency hybrid Green’s function method is proposed to establish the motion integral equations of the sub-rail foundation–subgrade subsystem;meanwhile,the vehicle–rail subsystem and the sub-rail foundation–subgrade subsystem are coupled by the fastener forces.The accuracy of this dynamics model is verified by comparing with the field measured data;in addition,by comparing the calculation results and calculation efficiency with the dynamics model established by the finite element method with high modeling accuracy,it is proved that this dynamics model can greatly improve the calculation efficiency while ensuring the calculation accuracy.Then,taking CRTS II longitudinal slab ballastless track,CRTS III element slab ballastless track and double block ballastless track which are widely used in Chinese high-speed railway as the research objects,the time and spatial distribution characteristics of the Green’s functions of the sub-rail foundation–subgrade subsystem of different ballastless track types are analyzed,and the differences of dynamics responses of the vehicle–track–subgrade system caused by different track structures under high-speed running conditions are compared.Additionally,the attenuation law of the dynamics responses of the sub-rail foundation system along the track depth is studied.The results show that the displacement and velocity Green’s functions of the sub-rail foundation–subgrade subsystem of the three ballastless track types gradually attenuate along the time and space,but the attenuation rates are different.The difference among the three track structures has little effect on the wheel-rail vertical force,but has different effects on the vertical displacement of the rail,the vertical displacement,acceleration,and stress of the sub-rail foundation–subgrade subsystem.With the increase of the track depth,the vertical displacements of the sub-rail foundation–subgrade subsystem of the three types of ballastless tracks all show an approximately linear attenuation trend,while the vertical stresses show an approximately exponential attenuation trend.Further,from the perspective of the vehicle–track–subgrade system coupled dynamics,the dynamics responses characteristics of the vehicle–track–subgrade system caused by the excitations of three typical diseases,i.e.,wheel polygonal wear,rail weld irregularity and differential subgrade settlement,are studied in detail,and the influence law of the excitation parameters on the dynamics responses of the whole system is analyzed,and meanwhile,the limit values of the wave depths of the wheel polygon and rail weld under high-speed and safe running are given.Finally,a one-dimensional fully convolutional encoder-decoder deep neural network model is designed,and the vehicle system vertical accelerations excited by the track geometric irregularities are generated by using the dynamics model proposed in this thesis and are treated as the inputs of the network.Then,the network is trained by means of the supervised learning,and the intelligent detection of the track geometric irregularities based on the vehicle system accelerations is realized.Meanwhile,the influence of different network inputs on the prediction performance of the track geometric irregularity is studied,and the robustness of the network performance is analyzed.The results show that the vertical accelerations of the bogie frame and the car body present better prediction effect on the medium-long wavelength irregularity,while the vertical acceleration of the wheelset shows a better prediction effect on the short wavelength irregularity.When the three types of the accelerations are regarded as the network inputs at the same time,the overall prediction performance of the network reaches the best.Further,after obtaining the predicted track geometric irregularities,a time-frequency hybrid algorithm based on the wavelet transform and the Wigner–Ville distribution is adopted to identify the differential subgrade settlement hidden in the track geometric irregularities,and the effectiveness of the algorithm is proved by case study. |
