Analysis of the surface roughness obtained in a friction spinning process based on empirical models

Abstract

In this paper, an new way to analyze and empirically model the surface roughness of a flange geometry after a friction spinning process is presented. The friction spinning process is an innovative incremental forming technology. This new process combines thermomechanical friction elements of the friction welding process within a conventional metal spinning process. The friction allows a self-inducted heat generation and as a consequence a defined increase of formability. It is possible to produce multifunctional, complex parts from standard tubes and sheets. The process thus readily meets the demands placed on efficiency and the manufacturability of complex lightweight components. By choosing the appropriate process parameters, e.g., axial feed rate or relative motion, the contact conditions between the tool and the workpiece can be influenced in a defined way. A further advantage is the feasibility of influencing the grain structure and the hardness in a locally defined manner. It offers the possibility to manufacture finished components with the required surface roughness. Since not all parameter settings of the parameters rotation speed, head radius, feed, and tool infeed width lead to a measurable surface of the component, a gradient boosting machine will be used as classifier for predicting the valid areas of the design space. Furthermore, the influence of the tool temperature is taken into account. For the empirical modeling of the spatial distribution of the surface roughness, methods from the design and analysis of computer experiments are employed.