Natural Gas Pipeline pressure loss 2

[abolt]

I haven't done any fluids in a while so bear with me, and I apologize for the length of my post.

We have an NG pipeline that has an upstream pressure of 50 psig and a downstream pressure at the sales line of 45 psig (has been as low as 40 psig). The flow rate had been around 80 MCFD but as of yesterday was 73 MCFD. I needed to calc what the pressure drop should be. We are concerned that the line is pinched and/or has liquid in it. An aerial view of the pipeline shows that its basically straight, and I've included a horrible paint rendition of the elevation view (not to scale). The total length is 2520' and drops 45' over a span of about 300' and then runs mostly flat. Near the middle, the line was bored 3' deeper to go under a stream and then rises 3' on the opposite side of the stream to the original depth. I'm not sure over what distance the 3' rise/fall takes place. 2000' of the line is SDR 11 3" plastic pipe. Where they bored under the stream they used 520' of SDR 11 4" plastic pipe. There is 1 90* elbow in the line.

Heres what I did: Found the flow to be fully turbulent and used the Blasius eqtn for f. The mach number was <<< 1, therefore I assumed it was incompressible and isothermal. For the kf(pipe) I just used 4fl/D for 2000' of the 3" section and 520' of the 4" section. For the kf(fittings) I used the 3k method for the 90* elbow and modeled the 3" to 4" transition as a sudden expansion at an angle of 30* and the 4" to 3" as a sudden contraction of 30*. I then solved EBE for P2 and found there should be a slight increase in pressure to 50.3 psig. Basically undetectable on the gauge. I found some basic pressure drop calculators on the web and they all came up with similar pressure.

My questions are:
1) I assumed the pipeline to be straight. Was this correct? There are some long bends in the line from the elevation changes but I didn't know if they would contribute any losses and I'm not sure how I would even measure them. Is there some rule of thumb where where you can neglect the bend and assume it to be straight (some r/D ratio?)

2) I made no provisions for the 3" rise/fall to go under the stream, mostly because I didn't know how to handle it since the net elevation change is zero. Was it safe to assume that the rise/fall was symmetric and that the gain from the drop was canceled out by the loss from the rise?

3)Did I handle the 2 different diameters of line correctly? Namely using the 4fl/D and the transistions as contractions/expansions? I don't know how the pipes were connected, so I figured this would be a "worst case" scenario.

4) Does my thought process/assumptions and answer (50.3 psig) make sense? We suspect that there may be liquid laying in the pipe where it goes under the stream. We think there isn't enough pressure to purge the line.

5)I also forgot to include in my calcs that there is a riser in the line, so probably a straight through T is in there as well, but I doubt that would contribute any appreciable losses?

Any input is greatly appreciated. Thank you.

 

[BigInch]

1) I assumed the pipeline to be straight. Was this correct?

No but yes. Gradual bends are typically assumed lossless a lot, but actually some pressure drop could be taken typically with including an "efficiency or transmission factor" when you think there will be enough of them to make a difference.

2) elevation difference doesn't make much in natural gas until it is several thousand feet or more.

3)figure the pressure drop for each line segment of different diamters using each length and add them all together.

4) yes

5)more or less correct, but the difference is relative. It might be responsible for a couple of psi, relatively large, if you only have 50 to work with and your worried about a difference of 5 psi total. If it was a 500 psi line, I wouldn't worry much about 5 psi, but it's not, so you could be talking about 10% errors with that difference.

It might be the result of changing pressure drop with the temperature of the gas. Been any temperature changes lately and are you compensating for that in the flow equation you are using.

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[zdas04]

You didn't say what the source of the gas was. If it is well gas at around 50 psig, then you have to assume it is saturated and has several hundred lbm of water vapor per MMCF of gas. You can be confident that the 4-inch has standing water, probably a lot of it.

I recently had occasion to review the assumptions and boundary conditions of the major equations that people use. The AGA Fully Turbulent only works to the right of the fully-turbulent line on the Moody Diagram and HDPE never crosses that line. Panhandle A only works in a narrow range of Reynolds Numbers. Weymouth gets wanky with really smooth pipe.

That leaves the Isothermal Gas Flow equation which must be solved iteratively for friction factor. I find that it solves to within 5% in three iterations if I assume 10^6 for the first guess, and use the calculated Reynolds Number from the first guess for the second guess, and then repeat.

The Blasius Equation I'm familiar with (he was Schlicting's protege and hung his name on a bunch of stuff) is a shortcut to friction factors that was really useful with slide rules. If that is what you're talking about, just go to MathCad and do a find loop on the real equation.

David

 

[abolt]

It was about 40-45 *F when I went out and took the pressure readings. I don't know exactly when they took the previous readings and don't know what the weather was. It's very possible that it was much warmer. It's been near freezing one day and mid 60's the next around here lately. I just made the assumption that the gas would be the same as the ground, around 50 *F.

Yes this is well gas. I didn't know that you could assume well gas to be saturated. So it would be safe to say that the standing water in the line is contributing to the pressure loss?

The Blasius Equation I used was: f = 0.079*Re^(-1/4)

 

[zdas04]

Natural gas will be saturated at the last point there was a coherent gas liquid interface. At 60 psia and 60F, methane can hold 209 lbm/MMCF. At 0.08 MMCF/day that is about 2 gallons/day of liquid water if it all condensed (which it won't). Your 4-inch SDR 11 has an ID of 3.6 inches so the 520 ft would hold almost 300 gallons. It probably doesn't hold that much because as the line starts to fill up the velocity increases in the remaining space and drags some of the water with it. This increases the pressure drop dramatically.

The equation you used has about 4 pages of limitations and boundary conditions. I wouldn't use it for HDPE (pipe is too smooth). Go back to the Colebrook equation and solve it itteratively. I wrote a "Find" loop in MathCad in about 2 minutes that solves it slick.

David

 

[BigInch]

I think that is for very low pressure gas flow and it looks like that will give a pretty high friction factor for most turbulent flows. I'd say you're predicting a little more pressure drop than you should actually get in field readings.

There isn't anything that says you will have standing water there for sure, saturated gas means just carrying the maximum amount of vapor that it can hold at a given temperature. But obviously that means it is also sensitive to any temperature change; 1 degree F could precipitate liquids. You could be precipitating at low temperatures and not at higher temperatures. For that shor of a line, you most likely will not necessarily have gas temperatures equal to near-surface ground temperatures. I would think it would stay nearly the same temperature as it was when exiting the well head.

The other effect might be that your gas velocity sweeps out any liquids as soon as they are precipitated, thus keeping the line clean. I think 9 ft/sec should be enough to do it in a basically horizontal line. You could check a 2-phase flow correlation, Briggs-Brill formula, etc.

Is there a particular problem with water in the line. I tend to think the velocity is enough to sweep continuously, rather than move slugs of water. You seem to have enough pressure(?) to get where you're going and it's a plastic pipeline, so I assume pressure loss and corrosion arn't issues, but assuming things in these forums usually isn't a safe thing to do.



Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[abolt]

I don't have MathCAD, I'll just go back and do Colebrook by hand.


The well is producing both oil and gas (shallow well, ~2000'). As I understand it, the gas pressure and oil production are directly related. So if the pressure goes down the oil production goes down as well. That's why they are interested in knowing if the pressure drop they've been reading is "normal" or if there is a problem with line (pinched/full of water). They want to maximize the oil output from the well.

 

[zdas04]

I think you may have the P vs q relationship backwards. Lower wellhead pressure usually means more flow (both gas and oil).

I didn't calculate velocity, but BigInch's 9 ft/s is too slow to shift much water. There seems to be a physical change in gas' ability to move liquids around 11 ft/s.

If I have time in the morning I'll look at this line in MathCad.

David

 

[BigInch]

When the liquid restriction gets serious, the area available for gas flow decreases in the region. With that decreasing area comes a local velocity increase, which could then move the liquid along, even though the average cross sectional gas flow is at a considerably slower velocity. That's why a two phase flow correlation could give a better idea of what's going on there than either Blasius, CW, et al.

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[abolt]

I agree with you about the P and q, it makes much more sense the way you described it. It's very possible that there was some misunderstanding when they explained it to me.

Would two phase flow give a more accurate calculation? In theory, there should only be NG flowing through the line, we dont know for sure if there is standing water in the line. They did bleed the line at the end (where the pressure is low) and got some liquid out, only a couple gallons. So if there is any more water in there, it really isn't "flowing" with the gas, its just laying there, restricting it more than anything? If thats the case, and the velocity has increased due to the decrease in area, then the pressure has to decrease, correct?

I talked with a guy that used to be a well tender yesterday and he thought that we should box our line in from the sales line and check the pressure again to see if it pressures up to 50 psig. His theory was that the lower pressure was due to the draw from the sales line.

I really appreciate all the replies. I'm the only engineer on staff here and pretty green, so this place is a great help. Thanks again.

 

[BigInch]

Common misconception. There's no "draw down" from the sales line, like they are "sucking" your line or well pressure down. They may have reduced the backpressure they put on your line from 50 to 40, depending on how much they flow out of their system and thereby reduce the pressure in their system, which is seen at your outlet. Wells are just like an orifice, or any other fluid handling element. Reduce the downstream pressure (increasing the pressure drop driving flow across the element) and you get more flow going across them.

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[zdas04]

OK, I had a few minutes so I modeled it.

Assumptions:
1. The 3-inch flows reasonably monolithlically toward the and away from the 4-inch (i.e., any condensation will run toward the 4-inch from either upstream or downstream)
2. Pipe relative roughness 50E-6 ft
3. Clean new pipe efficiency 100%
4. Length of the pipe upstream of the 4-inch is about the same length as the length after the 4-inch
5. Flow rate 73 MCF/day
6. Pressures: Atmospheric, 14 psia, Upstream 50 psig, Downstream 45 psig
7. Temperatures: All assumed to be 60F
8. Specific gravity 0.65

Technique:
1. Solve for the downstream pressure and velocity in the section from the well (v=4.3 ft/sec, P(dwn)= 49.95 psig)
2. Solve for the upstream pressure and velocity in the final section (4.716 ft/sec, 45.05 psig)
3. Solve for the middle section, using the calculated upstream/downstream pressures from 1 and 2. Adjust pipe effeciency until flow rate equals 73 MCF/day. This is the key number and it worked out to 3.7%--a really really bad number, it should be over 80%.

Two things are happening: first, the accumulated water in the 4-inch is reducing the flow area (impossible to tell how much, but it is ok to assume an effective diameter that gives you 11 ft/sec or 1.39 inches effective diameter--I'll leave it to you to figure out the water height that gives you a flow area of 1.52 in^2); second, the gas flowing over the water is doing work on the water (think of a pneumatic wrench, 100 psig in, zero psig out, difference is work) by dragging it around and making white caps.

The MathCad sheet that I used for this has been pretty reliable, and the results were in the universe that I expected. I'd say that it is near certainty that your 4-inch trap is really close to full. Avoiding junk like that is why people hire engineers.

David

 

[BigInch]

Good stuff David.

As long as it's staying there and you can live with the slight pressure drop, is it a big problem, or will that pressure drop give you a lot less flow from the well?

Fit it with a ball launcher and receiver???

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[zdas04]

There is about 0.81 in of difference in ID. Spheres would stall out in the 4-inch. I would put in fittings to adapt up to 4-inch on the ends and put in a 4-inch Pigging Valve to run 4-inch medium density (Girard "Red" coating) poly pigs. They'll scrunch down to go through the 3-inch (eventually) and should expand quick enough to get some sweep effeciency in the 4-inch. It isn't perfect, but nothing is going to be with the pre-engineered train wreck they have in the middle of the system.

The dP is pretty small by historical standards, but historical standards mean less every year. A lot of the reservoirs that I work with are very sensitive to small changes in pressure drop. I looked at one well where the line pressure changed less than 5 psig and it went from 1.4 MMCF/day to zero. Over the next few months they were able to get it back to about 600 MCF/day, but never did get it back over a million. That kind of sensitivity is pretty rare (very rare in what is predominantly an oil field), but it happens. Most oil fields I've looked at would not have noticed a 5 psig dP across the system. Since the OP brought it up, I'm guessing that this field is at a cusp where 5 psig is the difference between economic and not.


David

 

[BigInch]

or at least somebody's paycheque

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[abolt]

Thank you David, you've been a tremendous help.

I'm not sure if the 5 psi difference is a break even point or not...I'm not privy to the economics of the pipeline. There was at one point in time a 10 psi difference, so I think they are more concerned with why there is that difference. The line goes under a road that has settled a little bit and they thought the line may have been pinched off, but looking at all of the great info here I'd say that's the least of their concerns. Why they put the 4" line in the middle is beyond me.

 

[BigInch]

They didn't have 3" and they thought bigger is better, which as you can easily see that isn't always so in gathering systems when high liquid condensation rates can choke off flow.

Like I said, I sure can't tell you what all the fuss is about. If it was a metal pipe and corrosion was a problem, or flow would increase if the inlet pressure was lower, OK, but otherwise a foot or two of water in a gathering system shouldn't mean the end of the world. Most production these days they're injecting water, or ... something worse.

Only put off until tomorrow what you are willing to die having left undone. - Pablo Picasso

 

[abolt]

We're relatively small, so they have time worry about everything. It's a running theme, but it's better than not caring.

So the very low pipe efficiency of ~3% is what led you to say that there is definitely water in the line? Is the 80% a general rule of thumb?

 

[zdas04]

Nope. When I'm evaluating a complex pipeline, if I'm above 80% on a pipe segment then I call that "good" and move on to finding soulutions for the stuff that is worse. Purely arbitrary. Purely one guy's approach.

David

 

[abolt]

I see. So there isn't really some formula or method that would make you say "yes there is water in the line". All you can really do is crunch the theoretical dP you should be experiencing in the given system and compare that to what is reality and then draw some reasonable conclusions as to why? In this case, we have a sag in the line that would obviously collect fluid, and the recent cooler weather has caused condensation from the presumably warmer gas?


It's amazing how quickly you forget stuff when you haven't used it in awhile..