Kyle N. Karlson

Sandia fracture challenge 2: Sandia California's modeling approach

The second Sandia Fracture Challenge illustrates that predicting the ductile fracture of Ti-6Al-4V subjected to moderate and elevated rates of loading requires thermomechanical coupling, elasto-thermo-poro-viscoplastic constitutive models with the physics of anisotropy and regularized numerical methods for crack initiation and propagation. We detail our initial approach with an emphasis on iterative calibration and systematically increasing complexity to accommodate anisotropy in the context of an isotropic material model. Blind predictions illustrate strengths and weaknesses of our initial approach. We then revisit our findings to illustrate the importance of including anisotropy in the failure process. Mesh-independent solutions of continuum damage models having both isotropic and anisotropic yields surfaces are obtained through nonlocality and localization elements.

Damage Evolution in 304L Stainless Steel Partial Penetration Laser Welds

Partial penetration laser welds join metal surfaces without additional filler material, providing hermetic seals for a variety of components. The crack-like geometry of a partial penetration weld is a local stress riser that may lead to failure of the component in the weld. Computational modeling of laser welds has shown that the model should include damage evolution to predict the large deformation and failure. We have performed interrupted tensile experiments both to characterize the damage evolution and failure in laser welds and to aid computational modeling of these welds. Several EDM-notched and laser-welded 304L stainless steel tensile coupons were pulled in tension, each one to a different load level, and then sectioned and imaged to show the evolution of damage in the laser weld and in the EDM-notched parent 304L material (having a similar geometry to the partial penetration laser-welded material). SEM imaging of these specimens revealed considerable cracking at the root of the laser welds and some visible micro-cracking in the root of the EDM notch even before peak load was achieved in these specimens. The images also showed deformation-induced damage in the root of the notch and laser weld prior to the appearance of the main crack, though the laser-welded specimens tended to have more extensive damage than the notched material. These experiments show that the local geometry alone is not the cause of the damage, but also microstructure of the laser weld, which requires additional investigation.

Combining Full-Field Measurements and Inverse Techniques for Smart Material Testing

Traditionally, material properties (such as Young’s modulus, yield stress, etc.) are determined from a series of simple tensile and shear tests performed under various temperatures, strain rates, etc. This process is time-consuming and expensive. Additionally, material properties determined from simplified stress states may not adequately describe material behaviour under more complex loading conditions. In the past decade, full-field deformation measurements such as Digital Image Correlation (DIC) and inverse techniques such as the Virtual Fields Method (VFM) have reached a maturity level that takes them from the realm of development to the realm of application. This paper will present a concerted effort beginning at Sandia National Laboratory to explore a new experimental protocol for high-throughput, high-quality material identification by combining DIC, full-field temperature measurements, and VFM. Our current thrust is focused on identifying visco-plastic material parameters of 304L stainless steel. In particular, we will discuss the question of uniqueness of the material model and how parameter covariance affects material identification.