Driving Behavior Of Off-road Tire On Granular Terrain:GPU-based Simulation Research And Experimental Validation

Off-road vehicles usually drive on granular terrain.The study of the interaction between off-road tires and granular terrain is of great significance to the design of off-road vehicles,the optimization of traction performance and the assessment of the passability of complex road.The traditional finite element method(FEM)or discrete element method(DEM)can not only describe the interaction between off-road tires and granular terrain effectively,but also has the problem of time-consuming simulation calculation.Therefore,in this paper,the coupling method of discrete element and finite element(DEM-FEM)based on GPU parallel computing is used to accurately and efficiently simulate the driving behavior of off-road tire on granular terrain,in which the off-road tire uses the finite element method to describe its continuous medium characteristics,while the granular terrain uses the discrete element method to describe its granular characteristics.Firstly,an efficient and accurate contact detection algorithm,namely TSC algorithm,is proposed to judge the contact between sand particles and tire surface in simulation.The global search of TSC is suitable for the contact detection of tire-sand interaction.In the global search phase,according to the contact characteristics of tire movement on granular terrain,the potential contact pairs are quickly determined by using the method of dynamically changing the slave nodes and master segments within a moving contact region.The local search is based on the principle of inside-outside algorithm,and the contact region is divided into three types,i.e.surface contact region,edge contact region and node contact region.In the local search phase,the invalid contact pairs and repeated contact pairs can be quickly excluded from the multiple potential contact pairs of one particle according to the relevant judgment conditions.Therefore,TSC algorithm achieves efficient global search and accurate local search.Through three numerical examples,the accuracy and efficiency of TSC algorithm in dealing with the interaction between particles and complex structure are verified,and the efficiency of the algorithm in tire-sand simulation is also proved.Then,considering the actual interaction process of sand terrain and rubber tire,an improved contact model including the effect of sand grains shape and tread rubber deformation is proposed.In DE-DE contact model,the interaction between sand grains takes the form of surface contact instead of conventional point contact,where the contact calculation contains four interactions,i.e.normal force,tangential force,rolling resistance and twisting resistance.And in DE-FE contact model,the contact between tread rubber and sand grains is surface contact,which includes the rolling resistance and twisting resistance on thegrains caused by rubber deformation during contact.As a result,the complete and realistic evaluation of contact forces are accomplished.Next,a comparison of sandpile simulation of coarse particles and experiment is carried out to verify the effectiveness of the DE-DE contact model.At the same time,the improved contact model is applied to the simulation of tire-sand interaction.The simulation parameters of the discrete element model of pebble are obtained by triaxial compression test,and the effectiveness of the proposed method is verified by comparison with the indoor single-wheel test.Finally,aiming at the problem of time consumption in the simulation of tire-sand interaction,an efficient DEM-FEM based on GPU parallel computing is proposed to simulate the driving behavior of off-road tire on granular terrain.In this novel method,the main body of calculation is executed on GPU and programmed into an in-house developed code TSI-GPU.The new GPU-based computing framework consists of 14 Kernels,including efficient contact calculation between large-scale particles,novel contact calculation between particles and complex tread,contact calculation between particles and boundary wall,internal force calculation of finite elements and information update.The result of simulation analysis is consistent with the corresponding test,which verifies the effectiveness of GPU parallel computing method.At the same time,the discussion of computational efficiency shows that the proposed GPU-based DEM-FEM can be a powerful tool to simulate the interactions between an off-road tire and granular terrain.