ICF10B

analysed to gain more insight in the process of ductile fracture initiation in the smooth specimen. Table 1 shows the predicted area strains together with experimental results at low strain rates for a critical R/R0 value of 2.5. TABLE 1 COMPARISON OF NUMERICAL PREDICTIONS WITH EXPERIMENTAL VALUES FOR STRAIN FROM AREA REDUCTION AT FAILURE INITIATION. smooth R = 5 mm R = 1 mm R = 0.45 mm experiment 1.47 0.9 – 1.03 0.75 – 0.85 0.55 – 0.64 prediction 1.2 0.97 0.87 0.44 For the notched geometries a more consistent prediction of failure has been obtained. For all notch geometries a damage value of 2.5 has been obtained at fracture initiation. Smooth specimen modelling still results in a larger damage value at failure compared with notched specimen. An approximate failure criterion has been found however, underestimating the failure strain for a smooth specimen with a relative error of 20%. Considering energy absorption in the last stage of deformation prior to fracture of the specimen, this inaccuracy is acceptable. CONCLUSIONS & RECOMMENDATIONS From the experimental results has been observed that the failure strain from area reduction at fracture does not change significantly with the strain rate. In a first approximation this effect might be neglected. A decreasing notch radius also decreases the failure strain and increases the hydrostatic stress at the notch. A large influence has been found of the rolling direction of the plate material. The failure strain decreases strongly in the direction perpendicular to the rolling direction. Application of a ductile damage model according to Rice and Tracey results in inconsistent damage predictions at the experimentally determined fracture strain from area reduction. This is inherent to the applied formulation taking into account the triaxiality – plastic strain path for the points where fracture initiates. Another approach where a certain element volume has been taken into account results in consistent failure predictions for the notched specimen. Comparing this damage prediction at failure with a smooth specimen, an overestimation of damage has been found from the numerical results. However, this approach can be used for a first approximation of the failure behaviour of blast loaded panels in ship structures. A comparison of void sizes for a notched specimen with notch radius R = 5 mm and a smooth specimen before the onset of failure would be interesting, relating these values to finite element predictions. ACKNOWLEDGEMENTS This research was funded by the Technology Foundation STW, the TNO Prins Maurits Laboratory and the Delft University of Technology. REFERENCES 1. Rice, J.R., Tracey, D.M., (1969) On the Ductile Enlargement of Voids in Triaxial Stress Fields, Journal of the Mechanics and Physics of Solids, vol. 17, 201 2. Pape, G., Bakker, A., Janssen, M., (2001) Flow Stress of Steel at Large Strains and High Strain Rates, International Conference on Fracture, ICF 10, Conference Proceedings, Honolulu

RkJQdWJsaXNoZXIy MjM0NDE=