When a flowline is shut down, very quickly liquid settles into ‘dips’ and gas rises to the ‘humps’ and then the system gradually cools to the ambient temperature.  As gas has a much lower heat capacity than liquid, sections filled with gas cool much quicker than those filled with liquid.

Cooldown can be a cause for concern in subsea and especially deepwater developments due to the potential for hydrate formation.  Hydrates are crystalline compounds that form when water molecules encage light hydrocarbons such as methane, ethane, or propane at high pressures and low temperatures.  Figure 1 is a plot of a typical hydrate dissociation curve, to the left hand side of which hydrates may exist, this is often referred to as the hydrate envelope.

Figure 1 Typical Hydrate Curve

In most deepwater oil developments, hydrates are avoided during normal operation using insulation to maintain the fluid temperature above the hydrate dissociation temperature at the operating pressure.  During a shutdown however, if the fluids are held at pressure they may cool down into the hydrate envelope.

  Cooldown animation gif (826K)

This animation shows the results from a simulation of a multiphase flowline being shut in (at time zero) and allowed to cool to the ambient temperature.

Very soon after shut-in, the gas and liquid settle into peaks and troughs in the flowline geometry, respectively.  The geometry for this system was a horizontal flowline to which was added a saw-tooth undulation of 1m amplitude, 500m wavelength (i.e. distance between peaks).  Such an undulation is very gentle and may not be resolved in topography data.  However, even gentle undulations can influence where the liquids and gas settle and hence the cooldown result.  The different cool down rates arising because of the differing fluid heat capacities are seen as the animation progresses; the temperature profile of the flowline mimics the hold-up profile.

What can also be seen in this animation is how the liquid levels in each dip fluctuate and interact with each other.  In effect, the system is a series of manometers connected in series.

Consequently, in order to calculate exactly which section cools first one must predict exactly where the gas and liquid settle and how they move after shut-in.  To do this requires a detailed and accurate topographical profile and a transient flowline simulator that can accurately capture how these systems behave.  Unfortunately, accurate profiles of the pipeline in the as-built condition are seldom available even if we believe the results of the commercially available transient simulators!  Therefore, predicting the effect of gas-liquid distribution on cooldown behaviour is usually meaningless.

What is required is a pragmatic approach to the problem.  Such an approach has been developed by FEESA and is described in Life of Field Cooldown in a Deepwater Development Case Study.