A total of nine full-scale fracture propagation tests with liquid or dense phase carbon dioxide (CO2) have been conducted by four research projects: CO2PIPETRANS, COOLTRANS, SARCO2B and CO2SafeArrest.

The semi-empirical Two Curve Model assumes that steady-state ductile fracture propagation can be described using an effective crack length. It was shown that a normalised effective crack length that separated the arrest and propagate data points from full-scale tests with subcooled water, water-air or gas could be defined on a plot of the normalised decompressed stress level (the hoop stress at the crack tip) versus the normalised toughness.

DNV-RP-F104 September 2021 Design and operation of carbon dioxide pipelines introduced a simple empirical model for establishing whether or not a running ductile fracture would arrest in a pipeline transporting liquid or dense phase CO2. The empirical model is also based on the aforementioned plot, but with data points from full-scale tests with dense phase CO2.

COOLTRANS and CO2SafeArrest have collaborated in a reanalysis of the raw data from each of their respective full-scale tests. A re-plotting of the data points confirms the appropriateness of the empirical model in DNV-RP-F104. It also provides an insight into a possible difference between the effective crack length for natural gas, and for liquid or dense phase CO2.

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