Simulating running ductile fracture in CO2 pipelines

Blog with contributions from: Eskil Aursand, Stephane Dumoulin, Morten Hammer, Svend Tollak Munkejord, Håkon Ottar Nordhagen

According to the International Energy Agency (IEA), CO₂ Capture and Storage (CCS) is the third most important measure to limit the global warming by 2°C. An important part of the CCS chain is the transport of CO₂ from points of capture (e.g. power plants) to points of injection and storage. CO₂ transport will likely happen by pipeline and/or shipping. In the BIGCCS International CCS Research Centre, research has been performed on how to make CO₂ pipeline transport both safe and cost-efficient.

Under pressure

In order to transport sufficient amounts of CO₂ in reasonably large pipelines, the CO₂ must be compressed to a liquid state, at pressures around 100 atmospheres. One of the safety-aspects in high-pressure pipeline transport is the danger of an incident called running ductile fracture (RDF). In such an event, some small initial fracture in the pipeline can start running along the pipeline in both directions. In the worst case, this may potentially split hundreds of meters of pipeline, causing large and dangerous leaks.

(…) some small initial fracture in the pipeline can start running along the pipeline in both directions. In the worst case, this may potentially split hundreds of meters of pipeline, causing large and dangerous leaks.

Thicker steel is expensive

An important design-criterion for pipelines is then to make sure that such a running fracture will stop early (“arrest”), as opposed to run a long distance (“no arrest”), for the given transported fluid. This may be done by e.g. using stronger or thicker steel. However, stronger or thicker steel is expensive, so being overly safe will make the pipelines unnecessarily expensive, and thus make CCS implementation less likely to happen. Because of this, predicting the threshold between arrest and no-arrest is important to enable safe and cost-efficient CO₂ transport.

Old methods – incorrect results?

This has been an issue in natural gas transport for a long time, and semi-empirical two-curve methods have been developed to predict the mentioned threshold. However, there is an important difference compared to CO₂. Natural gas will expand from high pressures smoothly, while CO₂ will experience a liquid-to-gas phase transition. This has great consequences for how the pressure evolves in the event, and thus also for the RDF phenomenon. There are indications that the old methods give incorrect results for CO₂ transport, and that something new is needed.

In response to this, SINTEF Energy Research and SINTEF Materials & Chemistry has a collaborative effort in the BIGCCS Centre to develop a more advanced model for predicting RDF. This model aims to describe the phenomenon in much greated detail, using a tight coupling between structural/fracture dynamics and fluid/thermo dynamics.

Results will be presented on TCCS-8

Recent developments within this work will be presented on TCCS-8 on June 17. Below we see videos of two simulations, the first being no-arrest, and the second with arrest.

No arrest

Test #1: Visualization of a simulated running ductile fracture in a CO₂ pipeline. In this case, the combination of pipeline and fluid properties causes a running ductile fracture.

With arrest

Test #2: Visualization of a simulated running ductile fracture in a CO₂ pipeline. In this case, the combination of pipeline and fluid properties causes the fracture to arrest early

In these two tests, the pipelines are the same. So why does the fracture arrest in Test #2, but not in Test #1? The difference lies in the initial temperature and pressure of the CO₂. As shown in the figure below, this leads to different saturation-pressures, which is one of the most important driving forces behind RDF. Surprisingly and counter-intuitively, the test with the highest initial pressure (Test #2) is actually safer with respect to RDF, due to it leading to a lower saturation pressure. This illustrates the crucial importance of thermodynamic predictions when modelling RDF.

Surprisingly and counter-intuitively, the test with the highest initial pressure (Test #2) is actually safer with respect to RDF, due to it leading to a lower saturation pressure.

Find out more about our research on CO₂ transport in these blogs:

• Models, data and software for safe and efficient CO2 transport
• Why good models are important for safe and efficient CO2 transport
• Flow in CO2 pipelines

This work has been done with support from the BIGCCS Centre, performed under the Norwegian research program Centres for Environment-friendly Energy Research (FME). The authors acknowledge the following partners for their contributions: Gassco, Shell, Statoil, TOTAL, GDF SUEZ and The Research Council of Norway.