f > fc. Parameter K defines slope of the sudden drop on the load - diameter reduction diagram and often is denoted as "accelerating factor". Two parts contribute to the increase of the void volume fraction in FE calculation with incorporated GTN yield criterion: one is the growth of the existing voids and the other is the nucleation of new voids during the external loading: growth nucleation f f f & & & + = , where p eq nucleation A f = e& & and p ii growth (1 f) f = - e& & (2) p eq e& is equivalent plastic strain rate and p ii e& is the plastic part of the strain rate tensor. Nucleation of the secondary voids led by strain increase is most frequently tried to be described using two approaches. The first one was defined by Gurland [10], and is determined by the model of continuous nucleation of new voids, so that the parameter A is constant. The second one was proposed by Chu and Needleman [11] and is based on hypothesis that void nucleation follows normal distribution. Although the second one has been much more used in investigation, it has been shown [6] that both approaches give similar results. RESULTS AND DISCUSSION Critical value of void volume fraction fc, corresponding to crack initiation in smooth specimen and crack growth initiation in CT specimen, was determined by combined experimental-numerical procedure. Void nucleation was defined by volume fraction of non-metallic inclusions. Initial void volume fraction f0 was determined by quantitative metallurgical analysis; nucleation of the secondary voids was not taken into account due to rather low presence of non-metallic inclusions in tested steel. Using optical microscope, three prepared samples of test material were examined; 100 fields of vision were made for each sample. Initial void volume fraction was determined as an average value of surface fraction of non-metallic inclusions for all fields of vision. For planimetric procedure of determination of volume fraction of nonmetallic inclusions, a semi-automatic measuring method was applied. Contouring of inclusion profiles and determination of surface fraction for each of the fields of vision were carried out using computer software. The inclusions were classified according to the procedure described in [12]. Numerical calculations of tension of smooth round and CT specimen were made according to the true stressstrain curve at 0°C and in accordance with ESIS TC8 round robin project [1]. For both calculations the large strain analysis with updated Lagrange procedure was applied. Plastic flow of the material was determined by GTN yield criterion (eqn. 1) with isotropic hardening. FE calculations did not incorporated void coalescence effect. The calculations for smooth specimen were made in two ways: by applying quadrilateral 4-noded and 8-noded FE with reduced integration. CT specimen was modelled only with quadrilateral 4-noded FE; 8noded FE were not used due to convergence problems. The calculation was made for plane strain conditions. Crack tip was modelled using refined mesh (0.4 x 0.4 mm), without singular FE. Dimensions of tested specimens and FE meshes are shown in Fig. 1. Figure 1: Tested specimens - dimensions and FE meshes
RkJQdWJsaXNoZXIy MjM0NDE=