A computational analysis of the application of skewness and kurtosis to corrugated and abraded surfaces

In this work, we describe the results of our investigation into the relevance of skewness \nand kurtosis as measures of surface roughness. Two types of surfaces are computationally \ngenerated: abraded surfaces consisting of surface scratches, and corrugated \nsurfaces consisting of hemispherical features. It was found that abraded surfaces \ncould be well described by the skewness and kurtosis, which can both be specified by \nthe degree of coverage by the features on a surface. These two parameters showed \na large variation over the range of surfaces sampled. The root mean squared (RMS) \nslope and surface area ratio did not change significantly by comparison, and the RMS \nroughness changed significantly only for surfaces with a large variation of scratch \ndepths. A monotonic relationship was found to exist between skewness and kurtosis \nfor abraded surfaces composed mainly of smaller scratches. For corrugated surfaces, \nthe skewness and kurtosis were nearly constant for surfaces with RMS roughness values \nthat differed significantly. The RMS roughness, RMS slope, and surface area ratio changed significantly by comparison. No monotonic relationship was found between \nthe skewness and kurtosis for corrugated surfaces. This indicates that corrugated \nsurfaces are best described by the RMS roughness, RMS slope, and surface area ratio \nrather than the skewness and kurtosis.