Study On Parallel Computaition Of Discrete Element Method And Experiments On Dynamic Characteristics Of Ballasted Railway Track

The ballasted track is still the main surpport structure of railway transportation in China.With the continuous increase of train load and speed in recent years,the support capacity of ballasted track has been put forward with higher requirements.Under long-term train load,the ballast bed performs non-uniform settlement,reduction of elasticity,fouling and clogging,seriously influencing the regularity and stability of the track.In order to grasp the macro and meso dynamic evolution characteristics of ballast bed under the condition of long-term load,it is necessary to adopt methods of experiment and numerical simulation to analyze the dynamic characteristics and deterioration mechanism of ballast bed on the macro and meso scale,so as to provide theoretical guidance for the structural design and extend service life of ballast railway track.In this thesis,an improved spherical particle packing algorithm is proposed based on the advancing front approach,which can be used to construct the irregular shape of ballast rapidly.Discrete element method(DEM)is applicable to model the discontinuous distribution characteristics of ballast materials.The particle bonding model that can simulate the fragmentation of ballast and non-fragmentable clumped particle model are established to precisely model the microscopic characteristics,such as the size and shape of ballast particles,the grading and porosity of ballast bed.Based on GPU parallel algorithm,a program is developed to model the direct shear test of ballast aggregate and the ballast bed in engineering large scale.The main studies of this thesis includes:(1)An efficient geometrical packing algorithm for spherical particles based on the advancing front approach is proposed.In order to generate dense and isotropic particle packing samples in arbitrary closed packing domain,the method of spatial gridding,which can simplify the contact judgment in the process of particle packing,is adopted to improve the efficiency of advancing front approach.To obtain the relationship between geometric characteristics and particle size ratio,the average coordination number,solid fraction and structural symmetry of cubical standard packing samples are analyzed.Finally,the optimal gridding ratio and initial packing point are determined by analyzing the relationship between packing efficiency and these two impact factors.(2)The compression crushing test and discrete element numerical analyses of railway ballast particles are carried out.By measuring the load-displacement responses of ballast particles crushing process under the static compaction load,the tensile strength characteristics for three set of ballast are statistically analyzed and the KS fitting goodness test is conducted.The influence of average coordination number and elastic modulus of numerical samples on macro-mechanical properties of ballast materials are analyzed and determined.By uniaxial compression and Brazil splitting test and corresponding DEM simulations,the relationship between the macroscopic basic mechanical parameters of ballast and the microscopic mechanical parameters is studied,so as to determine the DEM calculation parameters.The static crushing behavior of ballast particles is analyzed by DEM numerical simulation.According to the numerical results,to reveal the crushing mechanism of ballast,the evolution of contact force chains and the growth of bond failure number during the process of ballast static crushing are analyzed.Besides,it shows that the ballast static crushing strength is accord with Weibull probability distribution which is consistent with the results obtained by the ballast crushing test.(3)The direct shear test and DEM analysis of influence of fine particles on the shear performance of ballast aggregates.A large direct shear apparatus was conducted to investigate the characteristics of the shear behaviors of ballast materials fouled with coal or sand particles.The test results show that increase in content of coal or sand particles leads to a decrease of shear performance of ballast aggregates,and coal particles make a more significant influence.The discrete element method is used to simulate the direct shear process of ballast aggregate and analyze the change law of contact numbers between ballasts.Moreover,the motion characteristics of ballast is carried out else to make a detail analysis on the microscopic mechanism of fouled ballast aggregate.Considering the difference between coal and sand particles,part of DEM calculation parameters is reselected to study and analyze the shear behavior of ballast aggregate fouled with coal particles,and finally obtain the similar results with corresponding tests.(4)The accumulated settlement and deterioration mechanism of ballast bed under the action of repeated load.The DEM numerical simulation model of ballast bed with engineering scale is developed based on the GPU parallel computing,and an efficient contact search algorithm for irregular ballast is established.Under the different load frequency and amplitude.the dynamic characteristics of ballast bed are studied.The numerical results show that the increase of load amplitude has a significant effect on the cumulative settlement of ballast bed.With the action of high-frequency load,the accumulated settlement of ballast bed will increase rapidly.In addition,the microscopic mechanism causing the accumulated settlement of ballast bed is studied by analyzing the contact sliding fraction and relative angular velocity of ballast particles.Finally,the research contents relating to this thesis are summarized and the main problems in further research are discussed and expected.