The effect of the riveting process and aging on the mechanical behaviour of an aluminium self-piercing riveted connection

Abstract

The joining of two aluminium plates with an aluminium self-piercing rivet, as discussed by Hoang et al. (2010), showed that heating to soften the plates to be joined might be used to avoid plastic compression or fracture in the rivet. In the present paper, the mechanical behaviour of self-piercing riveted connections using an aluminium rivet was investigated. Two U-shaped specimens in alloy AA6063-W, obtained by a solution heat treatment of the alloy in temper T4, were joined using an aluminium self-pierce rivet in alloy AA7278-T6. The mechanical behaviour of these connections was tested after 3 and 30 days of natural aging of the riveted specimens. Test results showed that the strength of the aluminium riveted joints tended to stabilize after three days of natural aging. In order to evaluate the process effect (i.e. the effect of the pre-straining of the plates in W temper as the result of the riveting process, and the subsequent natural age-hardening of the plates) on the mechanical properties of the riveted connection, a comprehensive material test programme was carried out. Test results revealed that there is an interaction between the pre-straining and natural aging which lowers the material properties in terms of the flow stress compared to the ‘virgin’ material, i.e. the curve obtained after heat treatment and aging only. Then, the process effect on the mechanical behaviour of the riveted connections was investigated more closely by using a 3D-model generated in LS-DYNA. The combined effect of the pre-straining and natural aging was accounted for in the numerical analyses by means of a user-defined constitutive model. Numerical analyses revealed that it is necessary to consider the combined effect of plastic deformation and aging in order to predict the mechanical behaviour with a reasonable accuracy.

Introduction

The self-piercing riveting (SPR) technology is today considered as one of the most important cold joining technologies in lightweight structures. Especially in the automotive industry, considerable knowledge has been obtained on steel riveted connections (Abe et al., 2006, Abe et al., 2009a, Mori et al., 2006, Porcaro et al., 2004, Porcaro et al., 2006a, Porcaro et al., 2006b, Porcaro et al., 2008 Sun et al., 2007); and the SPR technology has consolidated a firm position beside the welding technique. However, the combination of steel rivets and an aluminium car body makes recycling a challenge. The possibility of replacing a steel self-piercing rivet with an aluminium one in order to facilitate the recycling of an aluminium car body has thus been raised as an interesting topic, and has been discussed by Hoang et al. (2010) and Abe et al., 2009b.

The work of Hoang et al. (2010) revealed that when using an aluminium rivet, the strength of the rivet relative to the strength of the plate influenced the riveting process and thus the mechanical behaviour of the connection. Fig. 1a shows a case where rivet compression took place and no connection was formed between the upper and lower plates. Furthermore, Fig. 1b shows a case where the rivet penetrated the top plate and pierced into the bottom plate, but fracture of the rivet took place. In order to avoid these defects and to obtain a good riveted joint, the softening of the plates to be joined by, for instance, local heat treatment might be necessary just prior to the riveting process. However, such a heating will lower the material properties of the plate around the rivet and thus the mechanical strength of the connection. Furthermore, the latter may also be influenced by the age-hardening of the plate after being heat-treated and the plastic deformations generated in the plates during the riveting process. Thus, the understanding of the interaction between the results from the riveting process and the subsequent age-hardening of the plates to be joined is crucial in order to have a better insight into the mechanical behaviour of aluminium self-piercing riveting connections. Such an understanding will also be important for future development of heating equipment which is not available in the automotive industry today.

In the present project, local heating of the chosen alloy AA6063 in temper T4 could have been applied. However, as such equipment was not available at SIMLab the whole specimen to be joined was heat treated to W temper (solution heat treatment). In this way, no gradient in the material properties away from the rivet was present and thus the material properties in the whole plates could be controlled prior to the riveting process. Such an approach made it possible to better understand the effect of the riveting process and the subsequent age-hardening on the capacity and behaviour of a joint.

The present paper is then organized as follows. First, a test programme was established, focusing on the behaviour of two joint combinations, i.e. rivet in alloy AA7278-T6, plates in alloy AA6063-W, and two different dies (FM die and DZ die) after 3 days and 30 days of natural age-hardening. Extensive material tests were also carried out in order to understand the consequence of the pre-straining of the plates in alloy AA6063-W after subsequent natural aging. This included two series of tests, i.e. tension–tension tests, and rolling-tension tests. Finally, a numerical ‘through-process’ analysis of the mechanical behaviour of two riveted connections was performed, based on preceding material data. A user-defined material model was created in order to take into account the combined effect of the pre-straining of the plates in temper W due to the riveting process and the subsequent natural age-hardening on the final mechanical behaviour. All the numerical analyses were carried out by means of a 3D-model generated in the finite element LS-DYNA code.