A Comparison of 3D Simulations Against Experiments for the Torsion Test

This paper describes simulation studies using the public domain version of the DYNA3D hydrocode of the torsion test, which by definition is a 3 dimensional test. The mesh resolution required to resolve the stress system and the general shear deformation within the gauge length of the specimen has been investigated. Furthermore the simulations have been compared in detail with experiments in terms of the shear stresslstrain curve and the general shah distribution over the gauge length under conditions of steady deformation for iron and RHA. Once the numerical technique has been proven then aspects of localisation and instability can be more thoroughly investigated under a well characterised loading scenario. The simulation of many ballistic events rely on accurate constitutive relations which are applicable under tensile, compressive and shear loading regimes at high strain rates and high strains. Many of these ballistic processes, such as shear plugging, are caused by shear localisation resulting from adiabatic shear phenomena. Thus the simulations must be capable of resolving this localisation behaviour in order to predict the shear plugging. Furthermore recent studies [l] have shown that the work hardening behaviour of iron and RHA is significantly less in torsion than in tension or compression. Indeed for a variety of RHAs tested the work hardening is almost zero. This behaviour should therefore be taken into account in the constitutive model. Since a ballistic event is a very complex process involving many mechanisms, it is useful to simulate a test which isolates the shear localisation to allow an in-depth evaluation of constitutive models and approaches. Such a test is the torsion test in which a specimen is deformed in pure shear, where shear localisation and instability in the form of adiabatic shear can be reproduced. This paper describes the development of a full 3D simulation capability of the torsion test using the public domain version of the lagrangian hydrocode DYNA3D [2] and preliminary comparison with torsion test experiments performed at Oxford University. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:19973127 C3-748 JOURNAL DE PHYSIQUE IV 2.DESCRIPTION OF TORSION TEST The torsion test performed at Oxford University consists of a thin walled specimen which is held fixed at one end and twisted at the other using a torsional Hopkinson bar, such that the gauge length of the specimen is under a state of pure shear. The specimen geometry is illustrated in Fig 1 and a full description of the apparatus is given in [3]. The measurements produced for comparison with the simulations are the dynamic shear stresslshear strain in the specimen gauge length. In addition high speed cameras are being used to monitor a deforming grid etched onto the specimen gauge length compare directly with experiments. Future work will attempt to produce and film shear instability in the specimen due to localisation using high speed cameras. Fillet 2.5 mm 0.25 mm b . 2.0 mm Fig 1 Section Through Specimen