Surface roughness prediction and process parameter optimization of Ti-6Al-4 V by magnetic abrasive finishing

In order to effectively predict the surface roughness Ra of Ti-6Al-4 V material after magnetic abrasive finishing (MAF) process, and optimize the process parameters to improve the surface quality of the material. Firstly, diamond/Fe-based magnetic abrasive powders (MAPs) are prepared for the MAF process of Ti-6Al-4 V by using the gas–solid two-phase double-stage atomization and rapid solidification method. The effects of rotational speed of the magnetic pole, working gap, feed velocity of workpiece, and filling quantity of MAPs on the surface roughness efficiency are discussed. Secondly, the orthogonal experiment is designed. The prediction model of surface roughness based on gray wolf optimization (GWO) algorithm and support vector regression (SVR), which is constructed according to the experimental results. The simulation shows that the R2 of the optimized prediction model is 0.987456, and the MAPE is less than 1.99%. Finally, GWO algorithm is employed again to perform a global optimization search on the constructed prediction model. The optimal combination of process parameters is searched and verified, the surface roughness Ra is 0.098 μm, and the relative error is less than 2.82% compared with the model prediction. The comparison of surface morphology before and after MAF of Ti-6Al-4 V shows that the MAF technology combined with the prediction model based on GWO-SVR can effectively improve the surface quality of Ti-6Al-4 V.