Spectral Approach on Multiscale Roughness Characterization of Nominally Rough Surfaces

This paper attempts to demonstrate a systematic spectral approach for the characterization of the topographic nature of rough surfaces. Multiscale roughness characterization reveals a panoramic structure of microgeometric features of engineering surfaces, and this is of practical importance in order to include length scale consideration in real contact problems. Surfaces with different levels of root mean square (rms) roughness values were prepared using mechanical finishing processes for this study. Both optical and stylus profilometry data were recorded and analyzed to plot autocorrelation and power spectral density functions (PSDFs) at five different cutoff bandwidths (BWs). Correlation distances were estimated by choosing normalized autocorrelation declination to 1/e as well as to 0.1. In most of the cases, these distances were found to be less than 10% of the corresponding cutoff lengths. Nature of power spectrum has been analyzed and discussion extended to the estimation of bandwidth limited fractal characteristics based on specific spectral information. Power spectral densities (PSD) and their higher moments were extensively used to compute roughness parameters of functional significance such as asperity curvature, asperity density, etc. Evolution of asperity sharpness and asperity density during finishing processes was demonstrated at par with their physical significances. The intrinsic bandwidth parameter as per Nayak's definition was estimated closely to be a value of two for all cases.