Surface roughness modelling considering uncertainty in measurements

Surface finish tends to be decisive in a large number of applications and, in general, it must be corrected by means of finishing operations. It is also highly important that the characteristics required from the products obtained should be determined beforehand, and on this basis the operating conditions that most closely suit the materials to be employed and their characteristics should be chosen. Factorial design will be employed to study the effects of these operating conditions. Although factorial design is a known technique for analysing the effect of variables over an objective function, the uncertainty of these variables is often ignored. This means that the true behaviour of the objective function cannot be explained. Most of the previous investigators have studied the effect of cutting variables on surface roughness and many different models have been proposed, but they do not take uncertainty in the measurements into accountand this uncertainty is often ignored - in order to perform the regression analysis. In this paper, a development of models that would allow us to determine the surface quality of parts, obtained by turning processes, is carried out, using the response surface methodology and considering the uncertainty due to the process and the measuring instruments. In addition, a methodology for determining the capability of manufacturing processes is presented.