Multiscale morphology of high-precision turning process surfaces

Abstract

The characterization of functional surfaces is done mainly by roughness which can be quantified by many parameters. In order to select relevant roughness parameters, a multiscale discriminant method is proposed and applied to characterize high-precision turned surfaces. First, surfaces are characterized by a single roughness parameter and secondly by a pair of roughness parameters. In all cases, the most relevant evaluation length is also determined for each parameter. The results obtained on four different samples show that the most relevant roughness parameter is Rk estimated on a 10μm evaluation length. The best pair of parameters is Δa and Ri, estimated respectively on 20μm and 100μm evaluation lengths. Rk well characterizes the microroughness, which seems to be mainly representative of the roughness of high-precision machined surfaces. However, only multiscale analyses with a pair of roughness parameters can characterize both the macroscopic and microscopic morphologies of machined surfaces.