Applicability of Statistical Flaw Distributions of Eglin Steel for Fracture Calculations

Abstract

Computational continuum codes can provide many details on the response of metals to high velocity impact and explosive loading events. However, most “production” level calculations use a homogeneous description of the metal. This is an incorrect representation since metals possess a microstructure whose details create variations in material strength and other properties such as strain to failure. Ultimately these variations influence the formation of fragments at the macroscopic level. The spatial scale of the microstructure is on the order of micrometers and is not readily accessible to current computational tools and resources for system level calculations. Rather than explicitly model the microstructure one can attempt to capture the effects of material non-homogeneity through the use of a statistical description. Specifically, a statistically compensated failure strain criterion can be used to simulate the non-homogeneity of a material. This technique has been used previously by the authors on a tungsten alloy with some success. In those experiments, tungsten alloy rings were subjected to explosive loading. This resulted in a stress state approaching uniaxial stress. Furthermore, the tungsten alloy had relatively low ductility. The combination of these two factors resulted in fragments that were formed by tensile failure. It is important to determine if this technique can be used on a more ductile material under a different stress state. Fragmentation data was available for explosively loaded cylinders of Eglin Steel-1 (ES-1). The combination of the cylinder geometry and a more ductile metal resulted in fragments formed by shear failure. The experiment was simulated using CTH for the explosive and Pronto3D for the ES-1. Comparisons of the cylinder calculation results are made to the experimental fragmentation data and the results analyzed show a viable path forward on the use of statistical descriptions of these continuum models.