PSI - Issue 52

Jun Koyanagi et al. / Procedia Structural Integrity 52 (2024) 187–194 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

189

3

2. Studies for Nano-Scale Simulations 2.1. Molecular dynamics simulation for the quantitative prediction of experimental tensile strength of a polymer material (Koyanagi et al. 2020) So far, it has been very difficult to numerically simulate the quantitative failure of polymer materials with molecular dynamics simulations. In most cases, the estimated strength is much greater than that of reality(Koyanagi et al. 2019, Niuchi et al. 2017). This is partly because the failure of covalent bonding is usually not considered in molecular dynamics simulations. Koyanagi et al. (2020) suggest a method for estimating the experimental strength of polymer materials. We are aware that there are relationships between the strength obtained by MD and the volume and strain rate, which mimic real phenomena. The strength of polymer materials decreases with their volume, based on Weibull statistics, and increases with an increase in strain rate, based on viscoelastic characteristics. Hence, we simulate the failure behavior in MD with various volumes and strain rates, as shown in Figure 1. Based on these results, the experimental strength is estimated by extrapolating the strength with the assumption of real volume and real strain rate. The estimated result is close to that obtained from real experiments. However, this method might only be applicable to thermoplastic resins in which the failure of covalent bonding does not play an important role in discussing the failure. For this case we are now developing a new algorithm which simulates the failure of covalent bonding.

(a) Strength vs simulation volume (b) Strength vs strain rate Fig. 2 Relationship between failure strength and (a) simulation volume, the number of molecules and (b) strain rate. 2.2. Molecular dynamics simulation for evaluating fracture entropy of a polymer material under various combined stress state (Takase et al. 2021) Stress is second order tensor and there are six components so that it is not simple to define the fracture of material. But entropy is scalar. In this study, we assume the material fails when entropy reaches a critical value. It is desired to justify that the fracture entropy is the same independently of deformation mode, such as uniaxial, biaxial, triaxial, and shear stress state. Takase et al. 2021 examines the fracture entropy with varying stress state by MD simulations and the result is that the fracture entropy values are close values independently of the stress state as shown in Fig. 3. Full atom simulation is implemented by GROMACS with 570 thousand atoms. The assumed material is PA6. The fracture modes are uniaxial, biaxial, triaxial, and shear stress states. The entropy values are determined by dissipated energy divided by temperature, i.e. mechanically measured entropy. In these simulations, the failure of covalent bonding is not considered but this is reasonable for this material because the governing factor of the failure is entanglements of molecules. Furthermore, in this study, void size is proved as shown in Fig. 4. For this results, the void size increases with the deformation and entropy measured by thermodynamic point of view also increase simultaneously.