PREDICTION OF SURFACE ROUGHNESS FOR END MILLING TITANIUM ALLOY USING MODIFIED PARTICLE SWARM OPTIMIZATION LS-SVM

It is difficult to construct the prediction model for titanium alloy through analyzing the formation mecha- nism of surface roughness due to the complicated relation between influential factors and surface roughness. A no- vel algorithm based on the modified particle swarm optimization （PSO） least square support vector machine （LS- SVM） is proposed to predict the roughness of the end milling titanium alloys. According to Taguchi method and features in milling titanium alloys, the influences of cutting speed, feed rate and axial depth of cut on surface roughness are investigated with the analysis of variance （ANOVA） of the experimental data. The research results show that the construction speed of the modified PSO LS-SVM model is two orders of magnitude faster than that of back propagation（BP） model. Moreover, the prediction accuracy is about one order of magnitude higher than that of BP model. The modified PSO LS-SVM prediction model can explain the influences of cutting speed, feed rate and axial depth of cut on the surface roughness of titanium alloys. Either a higher cutting speed, a lower feed rate or a smaller axial depth of cut can lead to the decrease of surface roughness.