Surface roughness prediction based on stepwise regression analysis

To reduce the magnetic resistance of the carbon steel shaft on the method of processing the carbon steel shaft of the magnetic particles, the abrasive is stored in a circular vessel in which a plurality of magnetic poles is contained, and the magnetic particles taken away by the workpiece can be in the magnetic pole for cyclic utilization. Although the self-update of magnetic abrasive is in a single magnetic pole, it is possible to prevent the loss of magnetic abrasives. Further, in the test of the axial part of the 20CrMnTi material, the particle size of the magnetic abrasive particles is independent variable with the mass ratio of the machining time, the workpiece rotation speed, the magnetic abrasive and the polishing liquid, and the surface roughness as the dependent variable. The prediction model of surface roughness is established by stepwise regression analysis. The accuracy of the prediction result is found through the test, and it is found that the absolute value of relative error of surface roughness can be controlled within 7%, which has good prediction ability.