Abstract
A new approach to modelling machined surface profiles is presented in this paper. The fractal characteristics of the ARMA model are first investigated. It is found that the ARMA model behaves in a fractal manner. Then, based on this result, the relationships between the parameters of the ARMA model and its fractal dimensions are studied by use of the multiple regression analysis. In this way, the parameters of the ARMA model of a machined surface profile can be decided by calculating its fractal dimension. The way the fractal dimensions of the machined surface profiles are affected by the machining conditions is also studied. Therefore using these results, the ARMA model of the machined surface profile can be derived directly from the machining conditions. The AIC (Akaike information criterion) is used as the selection criterion for the model obtained using this approach. Finally, modelling of cut surface profiles is carried out.