Modeling of the influence of burnishing parameters on the surface roughness of rollers made of 42CRMO4 steel
PDF (Język Polski)


sliding burnishing
Hartley's plan
artificial neural networks
surface topography


The article presents the results of tests aimed at determining the influence of sliding burnishing parameters on the surface roughness of rollers made of 42CrMo4 steel. The burnishing process was performed using tools with a polycrystalline diamond tip. Before burnishing, the samples were turned on a tool lathe. The research was carried out according to Hartley's PS / DS-P: Ha3 plan, which enables the definition of a regression equation in the form of a second-order polynomial. Artificial neural network models were also used to predict the roughness of the surface of rollers after burnishing. The considered input parameters of the process included the values ​​of pressure, burnishing speed and feed speed. In all analyzed burnishing cases, the surface roughness value determined by Ra parameter decreased. The differences between the experimental data and the Hartley model did not exceed 24%. The best representation of the Hartley model was obtained for the burnishing parameters: feed f = 0.32 mm / rev, pressure P = 130 N and burnishing speed v = 180 rpm. Multilayer perceptrons were the best predictors of roller surface roughness. With Pearson's correlation coefficient R2 above 0.998, the mean absolute error did not exceed 0.005.

PDF (Język Polski)