Research On Optimization For Curved Surface Milling Process Of Aluminum Alloy Mould

Aluminium mould is universally applied to industries of aircraft, optics, automobile, plastic products. Although aluminum alloy moulds possess the advantages of short release time, facilitated polishing, low material cost, fast process, the phenomenon of large chip surface friction, stick knives, tool wear result from increased plasticity will ultimately influence surface quality and processing efficiency. In terms of selecting appropriate machining parameter to improve the surface quality and processing efficiency, conventional approaches predominantly rely on experiment with the limitation of lacking optimization. Consequently, the research on the optimization of surface milling process will play an important role in shortening the processing time and improving the processing quality.In this paper, based on milling curved surface of 6061 Aluminum Alloy mould, building mathematical model of its machining process. By using the central composite design method designed milling experiment with four factors, five levels and selected the suitable of the cutter locus for milling. The factors includ feeding speed, cutting width, spindle speed,the cutting depth. After accomplishing milling experiment,measuring the surface topography of experimental sample.The 2D surface roughness(Ra),3D morphology roughness Sz and Sq(ISO 25178) are regarded as indicators of the quality of milling,Material removal rate(MRR) is regarded as a indicator of the efficiency of processing. Using nonlinear modeling method establish prediction model for the processing response included Ra, MRR, Sz and Sq. Then analysis influencing orderliness of cutting parameter to the processing response. Solving the prediction model can get the optimal cutting parameter selection range.It can be chosen suitable cutting parameter to satisfy the demands of different processing.The feasibility of milling process optimization is verified by experiments.The details of the paper are as follows:First, the response surface method (RSM), genetic algorithm optimization of the BP neural network (GA-BPNN), regression support vector machine (SVR) was applied to build the regression model and to analyze the influence of each factor on the processing response. The analysis shows that the main impact factors of Ra and Sq is spindle speed, MRR and Sz is the cutting width. In the established process response models, Ra model established by GA-BPNN has the highest accuracy with average prediction error is 3.18%. MRR model established by RSM has the highest accuracy with average prediction error is 3.18%. Sz and Sq model established by SVR have the highest accuracy with average prediction error is error is 10.64% and 2.69%.Secondly, choose the optimal response model, using the NSGA-II algorithm to optimize Ra、MRR and Sz、Sq、MRR by two-objective and three-objective optimization.So that can achieve the goal that keeping highly processing speed while reduce surface roughness.The Pareto-front solution sets offer the corresponding cutting parameters.Finally, in order to verify the reliability of the optimized results, the machining experiments are carried out using the optimized cutting parameters. Experimental results show that after optimization it also maintained a highly material removal rate, Ra decreased 11.99%, Sz decreased 27.09%, Sq decreased 2.25%. The relative error rate of actual value and predictive value of optimization for Ra, MRR, Sq are all less than 6%,Sz is 13%.