Degree Name

Doctor of Philosophy


School of Mechanical, Materials, and Mechatronic, and Biomedical Engineering


In the pipeline industry, the Charpy impact test is the most common method for determining the fracture arrestability (or toughness) of a line pipe steel. Since the 1960s, line pipe steel has increased in both its stress capacity and toughness through advanced steel making and rolling processes. However, due to the lower finishing temperature, pronounced levels of microstructural anisotropy is generated. One consequence of anisotropy is the formation of weakly bonded planes along the rolling plane of the material. This can result in a phenomenon known as separations. Separations are fissures that form along the rolling plane of the material and lie perpendicular to the fracture plane. Separations are observed on all mechanical tests used to characterize a material’s properties (e.g. tensile, Charpy, DWTT, and full-scale burst tests). While separations have been observed on the fracture surfaces of line pipe steels since the 1960s, little research has been done to determine separations’ influence on the Charpy specimen’s ability to capture the effects of separations on full-scale fracture behavior in pipelines.

In this work, an API grade X80 line pipe steel, showing severe separations was used to quantify separation severity and investigate the reliability of Charpy specimens to accurately represent the separation phenomenon. This was accomplished by performing a number of Charpy tests along the longitudinal, transverse, and through-thickness orientations of the pipe. A set of novel Charpy specimens was also created to investigate the influence of separations on fracture characteristics. The novel Charpy specimens were created by introducing incisions along the rolling direction of the notched specimens. A set of specimens, containing one, two, and three incisions was manufactured and tested to simulate increasing levels of separation severity. Separation severity has been most commonly measured by the separation index metric. This metric considers the ratio of the total length of all separations to the fracture area. The separation index was compared with other metrics found in the literature as well as a newly defined metric introduced in this study, the separation area.

To further investigate the effect of separations on Charpy impact tests, a finite element model was created containing weak interfaces along the separation plane to simulate the formation of separations. The finite element model used the Gurson-Tvergaard-Needleman material behavior to simulate ductile fracture, while the cohesive zone model with a bilinear traction separation law was implemented to capture the brittle behavior of separations. The evolving stress state and geometry changes during full-sized Charpy impact testing was also evaluated for its influence on separation appearance and compared to Charpy specimens of lesser thickness as well as a similar thickness DWTT.

This study revealed that full-sized Charpy specimens, showing high levels of ductility, do not provide an adequate surrogate for assessing separation severity compared to full-scale fracture behavior. This is primarily due to the through-thickness constraint induced by the laterally expanding region as a result of the striker impact. For separation appearance in small-scale, laboratory tests to mirror that of full-scale fracture tests, a steady throughthickness stress state must be achieved for a majority of the fracture process.

Finite element analysis revealed that the specimen’s aspect ratio plays a prominent role on the striker’s ability to induce an artificial through-thickness constraint on the specimen. Lower aspect ratios provide a greater percentage of the fracture process to be influenced by a steady state through-thickness stress state.

By comparing several separation severity metrics to a newly defined separation area metric, which considers the total area of the specimen consisting of separations, the separation severity can be more easily estimated by the naked eye. The study showed that the separation area and commonly used separation index metric are causally related; however, the separation index requires a more detailed, time extensive process, which makes determining the separation area a more useful alternative.