Guillot, M., Fafard, M., Girard, S., Rahem, A. et al., "Experimental Exploration of the Aluminum Tube Drawing Process for Producing Variable Wall Thickness Components used in Light Structural Applications," SAE Technical Paper 2010-01-0222, 2010, https://doi.org/10.4271/2010-01-0222.
Tube drawing is a well known process involving at room temperature the reduction of diameter and wall thickness to obtain specified values. The initial tube is drawn into a die of a smaller opening and its thickness achieved by use of a mandrel. Usually, the mandrel has a land area which diameter defines by sizing the inside diameter of the final tube. Some structural components found in cars, aircrafts and other vehicles require bent or hydroformed tubes of lower weight. It is of interest to have tubes of varying axial or circumferential thickness so that to reduce overweight in low stressed areas and reinforce it otherwise. However, the production of tubes of varying thickness is more difficult in reason notably of higher metal flow stresses in the deformation zone and the need to control precisely the mandrel position during drawing. Axial thickness variation is obtained using a mandrel with stepped lands or with a slight taper while circumferential variation is achieved with a mandrel of desired internal or external shape (e.g. oval). In this paper, two techniques for axial tube wall thickness variation and one technique for circumferential variations are introduced and tested. First, the techniques to produce drawn tubes with thickness variations are presented. For testing, a small (335 kN) instrumented tube drawing machine is used. Details on this machine, process lubrication, monitored data and on the tooling implemented are also presented. Initial tubes are mainly AA6063 extrusions of 63.5mm O.D. and 2.6mm thick and the final outside diameter, i.e. the inside diameter of the die, is about 47.5 mm. AA6061 tubes are also drawn. Starting with drawing tests without mandrel, the natural flow of the tube and the drawing force involved are measured. Secondly, tubes of 4 different thicknesses are produced with a stepped mandrel and the strain hardening effect on mechanical properties established. Using a tapered mandrel, tubes of continuously varying wall thickness are tested. Higher local pressure in the die corner radius restricts proper lubrication in certain conditions but results are promising in most cases. We also study the effect of thickness rate of change along the tube. Finally, tests with a stepped oval mandrel provided good results for circumferential thickness variations. The dimensional quality is measured using a coordinate measuring machine and mechanical properties obtained from tensile tests in both initial and drawn tubes. Finally, despite some minor problems, the techniques proposed can efficiently produce tubes with thickness variations and have a very strong potential for industrial use.