Surface topography in general is not easy to characterize due to a great number of different features that appear on it. It is still more challenging for machined textured surfaces that are of high functional significance for tribological purposes. For practical reasons, there is a need to describe such surfaces using only a small number of parameters. Which of them represent surface details the best is still an open issue. To find out which parameters can be the most suitable in that case, three groups of machined textured surfaces were prepared. They were plateau-honed cylinder surfaces made of gray cast iron, steel, and bronze surfaces with isolated dimples and steel surfaces after abrasive blasting followed by lapping. All of them were measured by means of a white light interferometer. Different parameters and relationships were evaluated and based on them correlation and regression analyses were used. The basic description contained statistically independent parameters that can be used in production control, while the wider description in scientific research. In general, parameters of random surfaces were more intercorrelated than those of surfaces with isolated dimples. As was found for the basic description of random two-process surfaces, five parameters were enough while description of textured surfaces with isolated oil pockets needed six. In wider, scientific description, regardless the surface type seven parameters contained the necessary information about the surface. It was also proved that a pair of parameters, the emptiness coefficient Sp/Sz and Sq/Sa, can describe the shape of the ordinate distribution of machined textured surfaces better than, for example, skewness Ssk and kurtosis Sku, commonly used for that purpose.
Keywords: correlation; parameters; regression; surface texture; textured surfaces; two-process surfaces.