PSI - Issue 82

Abhijit Joshi et al. / Procedia Structural Integrity 82 (2026) 91–97 A. Joshi et al./ Structural Integrity Procedia 00 (2026) 000–000

95 5

2022). The creep parameters for the matrix shown in Table 1 were derived from the experimental results presented in (Joshi et al., 2025) and using Eqs. (5)-(6).

Table 1. Tensile creep parameters obtained from experiments and used in simulations

Creep parameters derived from tests 6.315 0.43

Creep parameters used in simulations 6.315 -0.57 8.04E-19

Temperature (°C)

500

1.88E-18

2.4. Validation of simulation approach The creep simulation approach in Abaqus was validated with the matrix-only model by comparing the simulation results against calculations. Using 150 MPa stress, and the material properties from Table 1, the expected creep strain at the end of 100 hours creep hold time was 7.425x10 -4 (mm/mm). The axial creep strain result CE22 from the matrix only model at the end of 100 hours creep hold time is shown in Fig. 4a, confirming the calculation result. The graph in Fig. 4b shows the evolution of creep strain with time from creep strain equation and from the FE simulations. A perfect match between the hand calculations and the FE model validates the simulation approach used in the analysis.

Fig. 4. Results from matrix-only model: (a) creep strain after 100 hours; (b) comparison of creep-strain evolution with time for calculations and from FE model. 3. Results from creep simulations Under the applied load of 15 N and nominal applied (macroscopic) stress of 150 MPa, the model with circular void using only elastic material properties showed a peak stress of 500.4 MPa around the hole as shown in Fig. 5a. This represents a stress concentration factor of 3.34 for this geometry. When this simulation was performed with the visco elastic matrix, i.e., accounting for the creep properties for the matrix, the peak stress in the first loading step reduced to 375.6 MPa (Fig. 5b), which is a substantial reduction of 25%. It is important to note that this reduction in stress occurred due to the viscous behaviour within just 0.01 hours (36 seconds). Another particularly important observation is that even though the macroscopic stress was relatively low – only 150 MPa, the local peak stresses around the stress concentration features were extremely high, exceeding the material yield strength. Such material behaviours can only be observed and appreciated with the micromechanical models, and this result highlights the importance of development of such models.

Fig. 5. Simulation results for axial stress after first loading strep: (a) in elastic model; (b) in visco-elastic model.