Multiphase flow modelling for hazard assessment of CO2 pipelines containing impurities

Written by Dr Solomon Brown, Postgraduate Researcher and Teaching Fellow, University College London

Working on UKCCSRC Call 1 Project: Multiphase flow modelling for hazard assessment of dense phase CO2 pipelines containing impurities. PI: Professor Haroun Mahgerefteh

This project funded by the UKCCSRC seeks to develop a rigorous model for the transient flow of CO2 following the failure of a dense phase pipeline. The thermodynamic properties of CO2 are such that decompression from dense phase will inevitably lead to phase transition in the fluid and hence two-phase flow. The failure of pipelines is most frequently in the form of puncturing, mostly occurring due to external interference with the pipeline (i.e. someone hitting it with a digger). During such a scenario the, especially in case of small punctures, the momentum of the flow will not be sufficient to fully entrain the vapour CO2 and phase stratification will occur. This has been observed experimentally in release tests performed by INERIS as part of the FP7 project CO2PipeHaz (, where a short transparent section was installed in the middle of the pipe. Fig. 1 shows photographs of flow patterns observed during puncture (b) and Full Bore Rupture (a) during the experiments.

Fig. 1 Flow patterns captured following the puncture (a) and FBR (b) of a CO2 pipeline courtesy of INERIS

Needless to say the modelling of this flow is extremely complicated; it’s known, for example, that in the initial stages of decompression there is a delay in the phase transition and that this will significantly impact the composition of the two-phase mixture. During the later stages of the depressurisation, when the phases begin to stratify, both phases will be at different temperatures and moving at different velocities. Furthermore, rapid moving waves will occur, e.g. a flashing front initiated upon failure will propagate into the pipe, and need to be accounted for using an appropriate numerical methodology capable of capturing them accurately and efficiently.

A model capable of accounting for the above phenomena is being tested. However key closure relations for describing the interaction between phases are not available, for example no experimentally determined flow regime map exists, even for pure CO2. One of the key objectives of the project is to develop and apply such correlations for various flow regimes using available experimental data.

Currently, the performance of simple closure relations is being assessed using experimental data recorded during a FBR test performed by Dalian University of Technology (DUT), also part of the CO2PipeHaz project. Fig. 2 shows photographs of the high velocity jet and the dispersion of the cloud formed respectively taken during the release. 

Fig. 2 CO2 outflow and dispersion during decompression courtesy of DUT