Experimental and numerical study on the effect of strain rate to ductile damage

Bowden, Andrew Scott

Experimental and numerical study on the effect of strain rate to ductile damage

Master Thesis

2009

Abstract

Ductile fracture modelling is extensively used in the automotive, aerospace, aluminium and steel industries. However, these models are often only validated in a limited region of stress states, for example tensile failure by void growth but not shear. In addition, the predictions generally do not include strain rate or temperature effects. Quasistatic tests are often used in calibration, even though many applications such as automotive accidents and ballistic impact operate in the dynamic range. Thus the aims of this thesis were to develop a system to test the damage properties of materials at both quasistatic (≈ 1 s−1 ) and dynamic (> 1×103 s −1 ) strain rates, and then to determine the influence of strain rate to ductile fracture. From the literature the Bai-Wierzbicki damage model was identified as being applicable to the widest range of loading conditions. Thus tests to calibrate this failure locus were conducted on sheet specimens with notches cut into each to introduce non-axial stresses, resulting in a range of loading conditions. This testing procedure involved experimental testing combined with finite element analysis (FEA) to determine the stress and strain state at the position of fracture initiation. All specimens used material from the same sheet of mild steel. To break the dynamic specimens a tensile split Hopkinson pressure bar, or TSHB, was optimized and built. Hopkinson bars are the standard method of conducting high strain rate characterisation tests, however, there is no universal design to examine tensile deformation. The apparatus built used a tubular striker and produced a square input pulse with low noise as desired. Sheet specimens were glued into slotted sections of threaded bar, which in turn screwed into the split Hopkinson bars. This method was successful as in every case the specimens broke before the epoxy. FEA modelling techniques were optimized to minimize computation time. The most important was the use of infinite elements to simulate the bars which, when calibrated, were found to be the ideal method of modelling split Hopkinson bars. Ultimately it was found that strain rate does influence ductile damage. The dynamic specimens failed at a lower strain than the quasistatic equivalents. This indicates that, at high strain rates, fracture strain decreases with strain rate. In contrast, in the quasistatic range strain rate tends to decrease displacement to fracture and thus it is proposed that at quasistatic strain rates, fracture strain increases with strain rate. It is speculated that the degree that strain rate influences ductile fracture is related to the Lode angle, which is a measure of the third deviatoric stress invariant.

Keywords

mechanical engineering

Reference:

Collections

Masters

Full item page