Surface finish quality characterisation of machined workpieces using fractal analysis

Abstract

A new method on machined surface finish quality characterization using fractal analysis is proposed. This seems to be an improvement on Olaosebikan’s spectral analysis index method for surface finish assessment. Mathematical model based on disk count Monte Carlo ap-proach is developed and tested with simulated results from computer programme written in Fortran. Test cases involve five-finished machine surfaces (work pieces) that are ranked based on fractal dimensions obtained for the re-spective machined surface spectral trace. The work pieces, made using different machining operations (milling, grind-ing, etc.), have their quality of finishing described as a function of the machine operation that each workpiece passes through. The respective spectral fractal dimensions of six fractal images (A, B, C, D, E and F) were then ob-tained. The conjecture is that the ranked results will agree with ranking obtained by both CLA and spectral index methods. Contrarily, the ranked results disagreed with both CLA and spectral trace results. The new method seems superior to both CLA and spectral trace approaches since a higher accuracy and much less computation time is ob-served. The maximum percentage relative absolute differ-ence is 13.1%, and the computation time is as short as 3 minutes.