Hybrid approach to surface roughness evaluation in multistage machining processes

An assessment of surface quality in turned, ground and honed specimens is made by means of a computer-based processing of digitised surface profiles. Three different methods of surface finish characterisation, i.e. statistical, fractal and neural network-based approaches are examined and compared. Correlations between some representative roughness parameters and the fractal dimension (D) values estimated were found. Consequently, they can be converted to their corresponding roughness parameters, i.e. R-a, R-z and R-Deltaa. Finally, a set of parameters including the minimum surface finish data for machining of external cylindrical surfaces when using complex technological process, is proposed. (C) 2002 Elsevier Science B.V. All rights reserved.