Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Questions regarding pipeline pressure drop simulations

Status
Not open for further replies.
larryli2004

larryli2004

Mechanical
May 29, 2006
52
Recently I did some pipeline pressure drop simulations however the result sounds not positive. When I use the defaut value of the pipe internal roughtness which is 1e-5m, the result cannot match the field data. Can somebody help me out this issue? I can change the roughtness, but it shows 0.008 m which to me not reasonable.

Thanks.
 
Replies continue below

Recommended for you

Sort by date Sort by votes
If you are saying that using 0.008 m yields the correct delta p, it could be that the actual pipes are fouled, pitted, have poorly matched/welded/etc joints and fittings.

Or it could be that you missed some major "Minor" losses in your model. Any baffles, reducers, manifolds, etc?

Signature under construction, sorry about the mess - Steve
 
Upvote 0 Downvote
Field data is the reality of what is happening with your pipes.

I agree with LHA - I am also thinking there may be some losses missing from the model?

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?
 
Upvote 0 Downvote
Hate to say it, but its probably G.I.G.O. Have somebody else check your input data. You'd be surprized at what another set of eyes will come up with.

I've found that simulation programs, at least those on the major engineering and oil and gas companies approved lists, give results as good as the input data will let them.

OK, back to the flow. Gas or liquid? Or both?

Perhaps you have fouling, perhaps you have liquid accumulated in low points, maybe a valve is mostly closed, has a joint collapsed under a road crossing?, perhaps unaccounted "minor losses", but that would have to be a lot of them to add up to an average roughness of 0.008m, so that's probably not it. Good luck finding it! Start at the point where the head or ressure drop/meter length gets exceptionally large. Should be easy to see.

BigInch[worm]-born in the trenches.
 
Upvote 0 Downvote
Thanks for all of you. Basically I got the 0.008m roughness from history matching. The operating conditions are as below:

4" line pipe, 2" insulation, underground, no reducer, 4.8mm wall thickness, 12650 m length

the fluid is gas+oil, gas flow rate is 96000 STD m3/day, oil flow rate is 388 STD m3/day, for gas, 30% H2S, oil specific gravity can be assumed 0.87, pressure at both end 4400 kPa and 2100 kPa respectively. temp. 38 degree C

I use this data and got the 0.008 m roughness.
 
Upvote 0 Downvote
Good luck matching a model to the glop you're pumping. Even the much vaunted (and very expensive) Olga will give you unpredictable results with that much liquid in the gas (or that much gas in the oil). A multi-phase flow pressure-drop calculation is doing really well to end up +/-50% of the field-measured data.

The analysis is just not up to the task, and it may not be in our lifetime.

David
 
Upvote 0 Downvote
That is also what I thought. How is the pressure drop prediction for pure gas? I thought that would be better. I got another operating data, and it sounds like pipe roughness is smaller thann default value, i.e. 5e-5m
 
Upvote 0 Downvote
Pipe roughness, "efficiency", etc. are really just levers that you can pull to calibrate a model. The technique that has served me well is to get a set of quality data, then pull the levers until I match all inflow pressures to within a pretty tight tolerance (say within 5 psi on a system with more than 50 inflows). Then I can use the model for two things: (1) re-run it periodically with new data to see which nodes are getting worse to refine a pigging schedule; and (2) put in proposed projects to see how the system will react. The second use requires a recently calibrated system, but with the proper care you can get really good results on project performance that are within the same universe as the model said they should be. The first use is pretty rare because it is amazing how few people with operational responsibility for gathering lines also have the access to the model and/or the time/inclination to run it.

The ability for most models to calibrate on a fairly simple system is pretty good and the numbers kind of mean something. On a highly looped system with multiple off-system deliveries most fail miserably. I run what is probably the last functioning copy of MNET in the world because it can be calibrated to any system I've ever thrown at it. The company went out of business with a beta version of Window 3.1 compliant code on the streets. I've made the software security "dongle" do the last few system upgrades, but I don't think it will go from XP to Vista and I'm frantic to find a viable replacement. I've run 5 of the top name programs and all are "pretty", but none of them can meet my calibration criteria on my test case. I've got a salesman for one of them trying to match my field data and about a week ago he stopped calling with questions so I'm thinking that another program bit the dust.

You have to remember that a model is not reality. It is just a bunch of arithmetic trying to provide insight into how a piping system functions. If you can get a model calibrated, then for a brief period it can tell you very interesting things. BUT if your model disagrees with high-quality field data then the model is wrong. Simple as that.

David
 
Upvote 0 Downvote
Trying to correlate 2-phase flow with a single phase model?

In that case don't be too concerned about using any roughness that matches your pipeline flow & pressures under operational conditions. If it does that, its a good match.

It sounds like a simple Lockart-Martenelli model could do the job with parameter values that you might think are more reasonable. Let me know if you want to try my spreadsheet.

BigInch[worm]-born in the trenches.
 
Upvote 0 Downvote
It does fine (after I got new firmware for the dongle) in XP, the problem seems to be Vista not playing well with DOS. The dongle vendor doesn't have a work around for that yet.

David

David Simpson, PE
MuleShoe Engineering
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem
 
Upvote 0 Downvote
Zdas/BigInch,

I also have problems to get the quality data. The software I use also offer the bulk properties (single phase model) option, I may try that one.
 
Upvote 0 Downvote
What's a "Russian Work Around" sounds like a variant of "Russian Roulette" to me.

David

David Simpson, PE
MuleShoe Engineering
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

nesamani1983
  • Question
gas pipeline grid pressure reduction
Replies
3
Views
1K
georgeverghese
georgeverghese
nesamani1983
  • Locked
  • Question
Hydrotest pressure drop analysis 1
Replies
9
Views
361
nesamani1983
nesamani1983
Rodger Furey
  • Locked
  • Question
Fluid Flow Questions. Flow rate vs PSI drop. Min pressure diff across a solenoid valve 6
Replies
19
Views
518
Rodger Furey
Rodger Furey
Mark_B
  • Question
Calculating flow through (heating water) coil given only pressure drop and coil measurements 1
Replies
16
Views
7K
btrueblood
btrueblood
AMINATAISSA
  • Locked
  • Question
INQUIRY REGARDING RUPTURE DISC AND PRESSURE RELIEF VALVE IN MINE & WASH PLANT PIPELINE
Replies
3
Views
90
georgeverghese
georgeverghese

Part and Inventory Search

Sponsor

Back
Top Bottom