Study On Micro Scratch And Elliptical Vibration Cutting Of Titanium Alloy Based On Neural Network

With the development of science and technology,the studies on preparation of highperformance alloys have become the focus of scholars all over the world.Titanium alloy(Ti-6Al-4V)is widely used in variety of fields due to its excellent characteristics such as low density,high strength and thermal stability.However,some defects of Ti-6Al-4V such as low thermal conductivity and hardness may cause surface wear failure and fatigue fracture of Ti-6Al-4V parts,and these limit their further applications in various fields.Therefore,in order to study the surface friction and wear properties and processing characteristics of Ti-6Al-4V,in this thesis,the finite element analysis and artificial neural network are mainly used to study the micro-scratch process under quasi-static conditions and the elliptical vibration cutting process under high strain rate conditions.The main work and results are as follows.(1)For the study of the friction and wear properties of material surface,the friction coefficient is an important variable to characterize the surface characteristics of materials.Different from the macro scale,the ploughing effect cannot be ignored at the micro and nano scales.The friction coefficient of a material is not only related to the surface roughness and surface lubrication,but also related to the elastoplastic deformation.In the study of the relationship between scratch parameters,scratch morphology and friction coefficient,firstly,the quasi-static constitutive equation and shear friction coefficient of Ti-6Al-4V were obtained through micro-nano indentation and scratch experiments.Then,large-scale finite element simulations of the scratch process were performed.Finally,an artificial neural network model was constructed based on the results of the finite element simulations,and the relationships between scratch parameters,scratch morphology and friction coefficient were analyzed.The results show that the depth of the scratch,the elastic recovery height,the plowing height,and the plowing friction coefficient decrease with the increase of the tip radius,in particularly for smaller tip radius.The scratch depth and the elastic recovery height are independent of the shear friction coefficient,but the plowing height and plowing friction coefficient are increased as a result of an increase of shear friction coefficient.The established artificial neural network models accurately predicted the changes in scratch morphology and friction coefficient values caused by changes in scratch parameters,and the mathematical relationships between ploughing friction coefficient and scratch topography,indenter radius,and shear friction coefficient.(2)For the investigation of the processing performance of materials,cutting forces are significant variables in evaluation of the processing performance of materials.Different from conventional cutting,cutting forces in elliptical vibration cutting show certain reverse characteristics due to the elliptical vibration of tool.In the study of relationships between vibration and cutting parameters(vibration frequency,tangential amplitude,thrust amplitude and cutting speed)and cutting forces,in the beginning,the finite element model of Ti-6Al-4V elliptical vibration cutting was developed based on literature experiments.Further,large scale Ti-6Al-4V elliptical vibration cutting simulation were performed.In the last,based on the finite element simulation results,analysis of variance(ANOVA),regression analysis and artificial neural network were employed to statistically analyze and estimate the relationships between vibration and cutting parameters and cutting forces.When examining the simulation results,it is found that the tangential force decreased with increasing vibration frequency,tangential and thrust amplitudes,but with decreasing cutting speed.The positive and negative thrust forces decrease with increasing frequency,tangential amplitude,but with decreasing thrust amplitude and cutting speed.Meanwhile,ANOVA results indicate that vibration frequency,tangential amplitude,thrust amplitude and cutting speed all have statistical significance for the cutting forces with the reliability level of 95%,and the dominate parameter affecting the cutting forces is tangential amplitude with percentage contributions of 69.56%,66.03% and 62.83%,respectively.Compared with the first and second order regression models,the neural network is advanced in prediction of the relationships between cutting forces and parameters.The determination coefficients of the neural network model of main cutting force,positive and negative thrust forces are 0.999,0.995 and 0.995,respectively.