Abstract

Typically continuum damage mechanics (CDM) based constitutive models are applied deterministically where the uncertainty of experiments is not considered. This is also true for the Sine-hyperbolic (Sinh) CDM-based constitutive model where the model is calibrated to represent 50% reliability of creep data. There is a need to implement Sinh in a more stochastic manner. The objectives of this study is to incorporate the probabilistic feature in the Sinh creep damage model to reliably predict the minimum-creep-strain-rate, creep-rupture and creep deformation. This will be achieved using Monte-Carlo methods. Creep deformation data for 304 Stainless Steel is collected from literature consisting of tests conducted at 300 and 320 MPa at 600°C with five replicates. The replicate tests exhibited substantial scatter in the minimum-creep-strain-rate, stress-rupture, and overall creep deformation. Subsequently, upon calibration using the Sinh model, the material constants among the replicates varied. The trends of uncertainty carried by each material constant are studied. The interdependence of the material constants is evaluated to determine if the uncertainty carried by each material constant can be regressed using a co-dependence function. The Monte Carlo method was applied to determine the extent that the creep deformation curve varies taking into consideration the variability of the material constants. Monte Carlo simulations show that the predicted creep deformation persists within the bounds of the experimental data. A large number of Monte Carlo simulations using the Sinh model enabled the creation of credible reliability bands for the minimum-creep-strain-rate, stress-rupture, and creep deformation of 304 Stainless Steel. In future work, this statistical method will be applied to the variability of service conditions, pre-existing defects, and material constants to quantitatively establish the reliability of the Sinh model in simulating component-level creep deformation to rupture.

This content is only available via PDF.
You do not currently have access to this content.