Research On Prediction Modeling Of Surface Residual Stress And Optimization Of Process Parameters In Low-stress Turning Of Pure Iron Components

Due to the excellent performance in thermal conductivity and magnetic permeability,pure iron material has been used to manufacture the key components in the fields of national defense and military industry,energy and power,etc.However,the precision manufacturing requirements are difficult to meet in terms of machining surface quality owing to the difficultto-cut characteristics,i.e.,low hardness,high plasticity and high toughness,which are prone to cause large tensile residual stress on machined surfaces of related components.Therefore,it is of great significance to accurately build the prediction models of machining-induced surface residual stress and realize the low-stress machining of pure iron components by controlling and then minimizing surface residual stress through optimization of process parameters.This paper focuses on the turning process of pure iron components.The optimal geometric angles of cutting tool are selected.The effects of turning parameters on surface residual stress are revealed,and the prediction models of surface residual stress are established.Subsequently,the optimization of turning parameters are conducted,which further provides a technical reference for the low-stress machining of pure iron components.The main research contents of this thesis are summarized as follows.(1)The optimal geometric angles of cutting tool are selected.Simulation models concerning the turning process of pure iron material are established,and the orthogonal simulation groups with three factors and three levels are conducted considering the rake,relief and cutting edge angles of cutting tool.The effects regarding geometric angles of cutting tool on tensile residual stress in surface layer are investigated and the optimal combination is selected based on range analysis method.According to the measuring results regarding distribution of residual stress,the finite element model is validated to hold qualified simulation accuracy,and the optimal geometric angles of cutting tool are validated to be effective.(2)The prediction models of surface residual stress concerning the turning process of pure iron components are established.Regarding pure iron planar components as the research object,central composite experimental groups with three factors and five levels are performed considering cutting speed,feed rate and depth of cut under the optimal geometric angles of cutting tool,and machining-induced surface residual stress of each sample is measured.Subsequently,the effects of turning parameters on surface residual stress are investigated based on range analysis method,and the prediction models of surface residual stress in cutting and feeding directions are established respectively using support vector regression method.The comparison results indicate that the support vector regression models,which show prediction errors within 7.67%,have better prediction accuracy compared with the artificial neural network models based on radial basis function.(3)The optimization of process parameters for low-stress turning is conducted.Firstly,with the aim of minimizing surface residual stress,the improved particle swarm optimization algorithm is integrated with the support vector regression models,and the optimal combination of turning parameters is found,which is validated to indeed realize the minimization of surface residual stress.Besides,the comparison results indicate that the improved particle swarm optimization algorithm proposed shows better performance in convergence efficiency while keeping the same convergence accuracy compared with genetic algorithm and traditional particle swarm optimization algorithm.Subsequently,considering the minimization of surface residual stress and the maximization of material removal rate simultaneously,the fast nondominated sorting genetic algorithm with elite strategy is combined with the support vector regression models,and the Pareto front is obtained.Corresponding optimal combination of turning parameters is then selected according to the weighted comprehensive evaluation function,which is determined under the given weight ratio of optimization targets.(4)Graphic user interfaces are designed,developed and applied on MATLAB.The interfaces can output optimal combination of turning parameters for low stress based on experimental database input.Application results show that the interfaces can realize the prediction modeling of surface residual stress and obtain the optimal process parameters within given ranges accurately and conveniently.