Design of Low Pressure Gas Gathering System 2

Why dont you start out by assessing the proposed pipeline system capacity under the new conditions e.g. using simulation software?

Your key issue on a wet gas system is likely to be liquid hold up and the inability to sweep out liquid slugs if they develop.

I did one of these low pressure gas systems a few years ago and with the turndown, ramp up and low flows and low pressures even a few psi made a difference and it wasn't easy to design for a full range of conditions.

If you've lowered inlet pressures you might need to lower compressor inlet pressure also to compensate and keep gas velocities quite high.

A steady state then a transient flow analysis is required, but you might need different sized lines / looping / parallel systems to cope with change of flows or composition over time.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

I've operated systems like this and designed them for people all over the world. There are two key learnings:
[ol 1]
[li]At these pressures, none of the empirical gas flow equations (i.e., AGA Fully Turbulent, Panhandle A, Weymouth, etc) come close to matching real flows. Their assumptions preclude their use in these conditions. The Isothermal Gas Flow equation still does pretty good, but you absolutely have to iterate each step until the Reynolds Number that went into the friction calc is within about 5% of the Reynolds Number of the computed flow. Fail to do this and your dP will be off by 50-100%. Transient models at these pressures are far worse than worthless--the flow will change flow regime on a milisecond time scale and the transient models are working on a minute time scale.[/li]
[li]EVERY SINGLE LINE MUST BE PIGGABLE. No exceptions. For flow lines from wells to trunks install Argus Pigging Valves (watch the video at the bottom of the link, the valves are pretty cool and only cost about 20% more than trunnion ball valves). Put in piggabale drips (I've got a design on my web page that you can have for free). With all this equipment you have to use it. My Pumpers watched the line pressure on their wells very closely and when the variability increased slightly they ran flow line pigs. Watch your system dP's. I recalibrated my pipeline model quarterly and flow lines that needed a lower efficiency to calibrate than they needed in the last calibration got their pigging schedule increased. Lines that showed an improved efficiency two quarters in a row got their frequency decreased. The pigging schedule ruled our lives since we treated the gathering system as our primary tool of reservoir management.[/li]
[/ol]
A line that looked like it was going to be properly sized (Using the Iso equation) at 50 psig wellhead pressure is probably still going to be fine at 40 psig wellhead pressure. If you are using another correlation then it will certainly tell you to go to bigger pipe. You probably should resist that.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. —Galileo Galilei, Italian Physicist

David, Could the "segment" length have something to do with the accuracy of the models you mentioned? Have you tried this with commercial software such as Schlumberger pipe-sim that they more or less sells as a "gas gathering system simulation tool"?

Yes, I've used a bunch of the programs (including PipeSim, which is far from horrible). The problem is the choice of correlations. Above 100 psig off-take I had really good results using the AGA Fully Turbulent equation with steel pipe in several of the programs (plastic pipe always puts you outside the fully turbulent line on the Moody Diagram), but as you drop pressure the impact of a small dP becomes too big a change in density for the "incompressible" assumption. This makes the horizontal friction line an invalid choice. Segmenting a line with AGA (or Panhandle etc) does not change the result (since friction is a function of ID instead of Reynolds Number in those empirical equations). For a simulator to match dP in a low pressure system it has to calculate friction factors explicitly. PipeSim doesn't (or at least it didn't 7 years ago when I was trying to do a project in it). Most simulators rely on one (or more, the best ones let you pick) of those empirical equations and don't allow friction factor to vary with Reynolds Number.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. —Galileo Galilei, Italian Physicist

Thank you all for the valuable feedback.

MortenA's point is noted on the need to redo the steady & transient simulation runs.

zdas04
"The pigging schedule ruled our lives since we treated the gathering system as our primary tool of reservoir management."
1. Do you mean here that Pigging Scehdule was such that pigging frequency too high to keep personnel busy?
2. Pigging here I guess you mean clean pigging. I understand Intelligent Pigging cannot be carried out in systems operating less than 150 psi, in which case would maintaining gathering pipeline network integrity be a issue here? Inspection techniques meant for non-piggable pipelines need to be used even though the pipelines are actually piggable. Considering the length & no. of pipelines, this clean pigging could be tedious.
3. Please indicate if Argus has a representative in the Middle East, if so contact details please.

I gave you a link to Argus, contact them yourself, I don't work for them.

"Pigging" means to run a pig. Period. If the pig is instrumented, then some call it smart pigging or intelligent pigging, which can be run at any pressure by the way, but of course the schedule was about making sure the accumulated water was removed. Call it "clean pigging" if you want, we called it "pigging".

When I said "pigging ruled our lives" I meant that it took priority over everything that was not currently on fire with significant risk of personal injury. If there was a scheduling conflict, pigging won.

I don't even know what your first point means. Everyone had plenty to do, but pigging took priority.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. —Galileo Galilei, Italian Physicist

@VERBabu,

When you do the simulation work you should pay attention to what correlation is used. You should propably go for one that uses isothermal flow with a calculated friction factor based on the Re and small segments. Maybe HYSYS with the "upstream/hydraulics" option that should be better suited for gas/gas condensate. I havnt tried that myself for a gas gathering systm. I would also recommend to stay in the SS regime i dont think there is much to gain from the trancient here. You have to be carefull when you specify your cases at least.

Step up to Stoner Pipeline Simulator. Will drop to 2 thousands of a second time steps when needed for fast transients, but works with only single phase gas, or liquid. 2 phase isn't very accurate anyway. Keep the pipeline clean of liquids and keep liquids out of the gas flow equations.

you must get smarter than the software you're using.