Milling Parameters Optimization Based On Multi-layer Design And Multi-objective Grey Wolf Optimizer

Titanium alloy is a kind of hard to machine material and widely used in deep sea,medicine,aerospace and other fields.This paper presents the improved selfadaptive multi-objective grey wolf optimization algorithm to choose optimal cutting parameters for minimum surface roughness and maximum material removal rate considering the quality assurance and machining efficiency.The mathematical model is the basis of the optimization algorithm.Multilayer design(MLD)method based on computer experiment is adopted to design the physical test,which will not only reduce the number of test samples and test cost,but build a strong generalization objective function model with a small amount of data from MLD.To prove the robustness of MLD,Gaussian process regression model is used to train and predict the sample data based on the numerical simulation and physical test respectively.The prediction accuracy is taken as the measure to evaluate the design method,and the results of Taguchi design and response surface design are compared with MLD.Based on the multi-objective grey wolf optimizer(MOGWO),power-law attenuation model is employed to optimize the adjustment strategy of control parameter,and the fitness value is calculated to reflect the optimization effect in the iterative process.According to this,the population number of the algorithm is adjusted adaptively and dynamically to achieve the improvement of the selfadaptive multi-objective grey wolf optimizer(SA-MOGWO).The SA-MOGWO algorithm and standard MOGWO algorithm are tested by UF series test functions.The three evaluation indexes of generation distance,inverse generation distance and hypervolume are used to judge the convergence,uniformity and distribution of the two compared algorithms respectively.Finally,with the spindle speed,feed rate,axial depth and radial depth as design variables,16 sets of TC18 titanium alloy milling experiments are carried out based on MLD,and the relevant test data are obtained and a data-driven second-order polynomial regression model is established.The improved SAMOGWO is used to optimize the target model to make a balance between the quality and machining efficiency of the workpiece.According to the different needs of decision makers,the optimal cutting parameter combinations are selected by analytic hierarchy process,and verified by milling experiments.