A numerical prediction of flexural strength probability for NBG-18 nuclear grade graphite using strength pair model

Abstract

In this work, a strength pair model has been proposed for the numerical prediction of flexural strength probability of NBG-18 nuclear grade graphite. The input to the proposed model is a random strength pair of tensile and compressive strengths whose value is based on its probability of occurrence in the experimental data. A finite element–based deterministic numerical approach has been implemented. To account for the large difference in tensile and compressive strengths, Drucker–Prager failure criteria has been implemented. The failure envelope of the Drucker–Prager failure criteria is assumed to have uniaxial fit with Mohr–Coulomb model in the principal stress space. A total of 292 simulations with random pairs of tensile and compressive strength are performed on a three-point bend specimen to obtain a set of flexural strength data. The flexural strength data obtained through numerical simulations are fitted using normal and Weibull distributions. The flexural strength probability obtained from the proposed model is found on conservative side. A goodness-of-fit test concludes that Weibull distribution fits the numerical data better than normal distribution.