Numéro
J. Phys. IV France
Volume 01, Numéro C3, Octobre 1991
DYMAT 1991 - 3rd International Conference on Mechanical and Physical Behaviour of Materials under Dynamic Loading
Page(s) C3-931 - C3-936
DOI http://dx.doi.org/10.1051/jp4:19913131
DYMAT 1991 - 3rd International Conference on Mechanical and Physical Behaviour of Materials under Dynamic Loading

J. Phys. IV France 01 (1991) C3-931-C3-936

DOI: 10.1051/jp4:19913131

DEFORMATION AND FRACTURE ; MODELLING TECHNIQUES TO TAKE ACCOUNT OF STRAIN-RATE BEHAVIOUR AND STRESS STATE

D.C. BARTON1, J.L. STURGES2, M.S. MIRZA1 and P. CHURCH3

1  Dept. of Mechanical Engineering, University of Leeds, Leeds, GB-LS2 9JT, Great-Britain
2  Dept. of Construction, Leeds Polytechnic, Leeds, GB-LS2 8BU, Great -Britain
3  Ministry of Defence, RARDE, Fort Halstead, Sevenoaks, Kent, GB-TN14 7BP, Great-Britain


Abstract
Numerical simulations of the dynamic deformation and ductile fracture of notched tensile specimens have been carried out using the finite element hydrocode DYNA2D and advanced constitutive relations for both pure copper and Remko iron. The predictions of plastic flow were in good agreement with the observed shape at fracture of specimens tested dynamically in a unique impact test facility known as the 'flying wedge' which is capable of applying strain-rates in the range 102- 104 s-1. The DYNA2D results were used to give accurate predictions of the triaxial state-of-stress (defined as the ratio of mean stress P to effective stress Y) at the fracture initiation site as a function of initial geometry, mean strain and strain-rate. Experimentally-measured failure strains were then plotted against predicted values of P/Y for each material. For copper, failure strains decreased with increasing stress triaxiality P/Y but no significant strain-rate effect was apparent. In the case of iron, the same trend of reducing fracture strain with increase in P/Y was observed, but here increased strain-rate produced a marked effect. In fact, a ductile-brittle transition (confirmed by metallurgical examination) was seen to occur at the highest speed of testing and above a certain value of P/Y. Parameters for simple empirical fracture models were estimated from the plots of failure strain versus P/Y. When combined with the standard damage accumulation fracture criterion in DYNA2D, these models gave reasonable predictions of both the site of fracture initiation (ie the centre of the specimen) and the time-to-fracture when compared with fracture times measured on certain specimens using strain-gauges or high-speed photography. More fundamental ductile fracture models which allow for the statistical nature of void initiation and growth and the effect of voids on the yield surface are currently under investigation.



© EDP Sciences 1991