Depressurisation of Cross-Country Line 1

[nichs2]

Hi,

I'm working on a problem, which I think should be relatively simple but I'm tying myself into knots over, so I thought posting here might help re-focus my tired brain. It can be embarrassing to keep going back to your boss...

The problem involves a large pipeline transporting unstabilised crude oil. In emergency situations, it may be necessary to quickly de-pressurise that line. To do this, there are 8 pressure control valves that can be fully opened to two surge relief tanks (4 to each tank). These pressure control valves have stops on them to limit the maximum flow. All entrants to the pipeline will be instructed to shutdown as part of the depressurisation procedure.

I need to calculate the maximum vapour flow-rate into the surge relief tanks (the vapour will go from there to flare), and the total liquid into the surge tanks over the course of the de-pressurisation. The vapour comes from flashing across the PCVs.

1) Maximum vapour flow. I think this will occur at the start of the depressurisation, when flowrate and pressure are at their highest. Is it appropriate just to calculate this by modelling flow through a valve,with pressure drop specified as (surge tank pressure - inlet pressure) and flow specified as the maximum the valves have been sized for? By modelling this, I will get a vapour flow out of the valve. What if I know that the flowrate coming down the pipeline could be higher than the maximum the valves have been sized for?

2) Total liquid flow into relief tanks. I think the liquid flowrate into the tanks will decrease over the course of the depressurisation. I thought I would model this as a stepwise decreasing inlet pressure (and therefore decreasing inlet flowrate), and calculate liquid flow out of the PCVs for each time step, then add them up. The problem is knowing how fast to decrease the pressure. I can pick a number, say a bar a minute, but this isn't based on much other than a bit of a chat with the ops guys. Is there a way I can calculate or model how fast the pressure upstream will decrease once I have opened the PCVs fully? I've tried in AFT fathom, modelling the inventory of the pipeline as a reservoir, but it seems to force me to describe it as a tank, and I run into problems. I'm also slightly confused as to how to represent different entrants shutting down at possibly different times (which is why i ended up with the simple 1bar/minute decrease, but I'm concerned that may not represent worst case).

How would the experienced posters on here approach this? I'm obviously not looking for exact answers, more a plan of attack!

Many thanks

 

[zdas04]

To start with, you are dealing with one of the most difficult fluids there is. That stuff is always so close to equilibrium that phases are changing all over the place all the time.

Since you can't be assured of how fast your suppliers are going to stop inflow, is it possible to model a control volume that includes a couple of hundred feet upstream and downstream of the valve station? In that subset you can look at staging inflows to the control volume, calculate the phase equilibrium, determine the flow through the valve, determine the makeup of fluids in the tail pipe, and model the flow in the header.

For inflows I would assume that no one shuts anything down for an hour (assume constant inflow upstream of your valves and calculate the material balance in the control volume to determine rate of pressure reduction), then assume one inflow is shut down each time step.

This might get you in a good direction. Good luck to you.

David

 

[nichs2]

Thanks for the kick in the right direction David - getting there now I think!