Predicting the Dynamic Fracture of Steel via a Non Local Strain- Energy Density Failure Criterion
Kevin Schrum, Dean Sicking, Ronald Faller, John Reid
Predicting the onset of fracture in a material subjected to dynamic loading conditions has typically been heavily mesh-dependent, and often must be specifically calibrated for each geometric design. This can lead to costly models and even costlier physical testing. In response to this, a failure criterion was created based on the strain energy density (SED) of the material. Calculations to obtain the SED were developed to take advantage of a non-local length scale, wherein the sensitivity to mesh density was partially reduced. This method was applied to a steel coupon subjected to dynamic uniaxial tension. A one-time calibration was used to determine the material's critical SED in the non-local length scale. This length scale was dependent on the mesh density of the model and a prescribed magnifier, such that the failure criterion was a function of the length scale. Steel coupons were modeled and tested dynamically. Thicknesses of those coupons were varied and stress concentrations were included. Differing grades of steel were also employed. The non-local SED failure criterion provided consistent and accurate predictions, regardless of the changes in dimensions of the coupons.
Steel Fracture, Failure Criterion, Dynamic Loading Conditions, Peridynamics, Strain-Energy Density, Stress Concentration, LSDYNA, FORTRAN