PSI - Issue 13

Gyo Geun Youn et al. / Procedia Structural Integrity 13 (2018) 1305–1311 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

1307

3

with the hardening exponent n =0.586 for unaged CF8A. By applying Eq. 5, the damage model parameters ( W f , ω c ) can be determined from monotonic tensile and C(T) test. The detailed procedure to determine damage model parameters are given in Youn et al. The determined multi-axial fracture strain energy is

  

  



(6)

p

W

3400exp 2.379 −

280

=

+

m

f

e 

with critical damage value ω c =0.65 for 0.15mm mesh size at the crack propagation area. However, 0.15mm mesh size is too small for pipe simulation. It is known from Nam et al that the critical damage value is dependent on mesh size but the multi-axial fracture strain energy is not affected by the mesh size. Therefore, the critical damage value, ω c for a larger element size should be found. Fig. 1 shows the FE modelling of CF8A 0.5T C(T) which has mesh size of 0.6mm ( L e =0.6mm) at the crack propagation area. By using the model shown in Fig. 1, FE simulation was performed and the results are shown in Fig. 2. The results are compared with the test data and the FE simulation results from the model which has the meh size of 0.15mm ( L e =0.15mm) at the crack propagation area. Fig. 2a and Fig. 2b show the load-LLD, J-R result, respectively. And they show good agreement with test data. From the result, it can be determined that critical damage value, ω c , is 0.35 ( ω c =0.35) when L e =0.6mm.

Fig. 1. FE modelling for 0.5T C(T) specimen of CF8A with mesh size 0.6mm at crack propagation area

3. Numerical prediction of pipe fracture behavior 3.1. Description of virtual pipe test

In this study, virtual four-point bending pipe tests with different load ratios were applied. FE analysis with multi-axial fracture strain energy damage model explained in Section. 2 is used. The dimensions of the pipes used for virtual tests are shown in Fig. 3. The dimensions are same as that of the pipe from Battelle pipe test. Also, the displacement applied for the virtual pipe tests are same as that of pipe experiment performed by Battelle. The applied displacement for monotonic loading is 33mm and the applied displacements for load ratio R = 0 and R = -1 are shown in Fig. 4. The only difference between the virtual pipe test and the pipe test performed from Battelle is that the test material is changed from A106 Gr. B (288 o C) to unaged and aged CF8A (room temperature). The material property for unaged and aged CF8A (room temperature) are suggested in Youn et al.

Fig. 2. FE simulation results of unaged/aged CF8A 0.5T C(T) under monotonic loading condition: comparison of L e =0.6mm with test data and L e =0.15mm (a)load-LLD comparison (b) J-R comparison