PSI - Issue 2_A

Takuhiro Hemmi et al. / Procedia Structural Integrity 2 (2016) 2230–2237 Author name / Structural Integrity Procedia 00 (2016) 000–000

2236

7

Stress tensor and strain rate used as the input data for microscopic and macroscopic model calculation were obtained From FEM analysis using nodal force releace technique in Abaqus 6.14 SIMULA(2014). Crack propagation velocity just before crack arrested in the FEM was assumed at 300m/s, based on experiences. We used 1/4 model, considering symmetry of the model. An example of stress distribution and deformation of a test piece calculated in FEM was shown in Fig.7.

Maximum Principal Stress [MPa]

700 500 300 100 0

Fig.7.Stress distribution of modelling

We fitted crack arrest length of test results with the length on the multiscale simulation of the crack arrest tests by changing coefficient of effective surface energy distribution. That is, we identified the value of coefficient C make the result of model calculations making consistent with the experimental results. We couldn’t confirm the dependency of temperature about energy distribution, and we used average coefficient under some temperature conditions. We showed representative value C of surface effective energy distribution compared with grain size in Fig8. As is shown below, the smaller grain size is, the higher effective surface energy is. These results are contrary to results of microscopic model on section 2. Therefore, it is revealed that it is necessary to revise the derivation method of effective surface energy. Table4. Temperature and coefficient of effective surface energy SA1 SA2 SA3 Temperature[K] 283 293 303 313 293 303 313 243 253 263 C 0.24 0.60 0.54 0.61 0.09 0.50 0.49 0.55 0.65 0.64 Average C 0.50 0.36 0.62

10 15 20 25 30 35 40 45 50 55 60

Grain size[μm]

0,3

0,4

0,5

0,6

0,7

Coefficient of effective surface energy C

Fig.8.Plot of coefficient of effective surface energy C and grain size