Hydraulic Design 101 - Air at Highpoint 5

[sskaff]

I was recently out at a project site where I saw a setup with a ton of problems. See attached picture (

http://files.engineering.com/getfil...-a265-47be-a015-956f6e21a547&file=AirTrap.jpg

). One major issue was a flaw in the piping design that had the discharge of the pump go up for a 10 foot horizontal section then back down below the eye of the impeller to an air eliminator with no high point vent. This line does not stay packed with fuel, and the high section of pipe will be drained of fuel after every loading event. The pump is generating about 42psig and flowing 360gpm through a 4" pipe.

So here is the question that me and my colleagues cannot agree on. What happens to the air in the highpoint? Does the fuel push the air down stream of does it just compress the air at the high point effectively restricting flow. Has anyone ever seen a simulation or video of a similar issue using transparent pipe to demonstrate why this is bad design principle?

 

[StevenHPerry]

I'm completely dodging your real question, but it looks like there's a vent at the top of the discharge elbow.

Problem solved? Or do the valves keep that portion of the line filled all the time?

- Steve Perry

http://www.linkedin.com/in/stevenhperry

This post is designed to provide accurate and authoritative information in regard to the subject matter covered. It is offered with the understanding that the author is not engaged in rendering engineering or other professional service. If you need help, get help, and PAY FOR IT.

 

[IRstuff]

6 gal/s in a 6" pipe --> 1 ft/s velocity

Assuming that the air eliminator is not a major pressure block, the two possibilities are:
> Most, if not all, the air will get pushed out
> The high point will have a mix of air and fuel at all times

The flow rate seems just high enough that it will most likely be the first alternative.

There is no way for the air to block the fuel, since the fuel is a liquid, and will, at the very least, flow along the bottom of the high point, strictly from gravity effects, with the air remaining above it, which is the second alternative. Seems to me that would depend on the flow rate; a slow flow will certainly tend toward the second alternative.

This is really no different than your bathroom faucet, if you think about it. You have a valve that allows water to flow up to a traverse, and then back down to the faucet opening. And you never have a problem with that, do you, even in the cases where the vertical piping to the second floor get drained, when you open the valve, all the air gets pushed out. It's actually a bit worse in my house, since the plumbing actually goes to the ceiling of the second floor before the dropping into the rooms where the faucets are.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[zdas04]

A 4-inch line needs 340 gpm to run full. At 360 gpm any air is going to be displaced pretty quickly.

When people say that a pipeline is "air locking" what they really mean is that there is a discontinuity in the liquid stream (usually at the top of a hill) and the flow on the backside of the hill cannot siphon the liquid that is on the front side over the hill. This results in the pump having to overcome the entire hydrostatic gradient of the height of the uphill. Air cannot block a pipe.

You didn't say what kind of problems you are having at this station. Is the air in the process a problem downstream of the station? If this is a fueling station, pushing the air from 30 ft of 4-inch into the vehicles doesn't seem major, but I could easily be wrong about that.

David

 

[sskaff]

Thanks for your feedback. We are not having any major issues with air, we just got into a discussion on the fundamentals of air in high points of pipelines and I started to question my fundamentals.

More info on the setup:
They are using a Gorman Rupp rotoprime pump with an integral air eliminator. We could not observe any air bleeding from the bulk air eliminator downstream of the pump. It is 20+ years old and could be malfunctioning. You can see air trapped in the high point of the valve stem indicator on the control valves (as expected for the location and no high point vent).

Q: So what are you really asking?
A: Does the majority of air remaining in the high section of pipe get pushed downstream into the air eliminator with the given pump conditions?
My colleagues are saying the pocket of air will continue to sit at the top of the pipe in the high section and compress as pressure builds. That does not seem likely to me. I did not have formulas or a formal document to back my position so I started to question my fundamentals.

 

[kacarrol]

A quick Reynolds number calculation puts your flow in the turbulent regime. I would say that your air is probably entrained in the medium over time and will eventually flow down the pipe.

But if you want a more definitive answer I think this book may be of some help. Although I can’t say for sure, as I haven’t read it in detail yet, and it’s on gravity flow systems:

http://www.aplv.org/files/docs/AirInPipesManual.pdf

As further viewing material Wolverine Tube has a video gallery up on their website that shows various two-phase flow patterns in horizontal tubes (again it’s not specific to your case but aimed at heat transfer):

http://www.wlv.com/products/databook/db3/DataBookIII.pdf

There is also this article on hydraulic jump:

http://article.pubs.nrc-cnrc.gc.ca/ppv/RPViewDoc?issn=1208-6029&volume=26&issue=3&startPage=368

Where you can clearly see the air being entrained in the medium as it makes the transition to pressurized conduit flow.

Regards,
K

 

[rcchap]

The criteria I have used to determine how a void travels in a piping system is to calculate the Froude number of the flow.

N_Fr = U / SQRT(gD)

where U is velocity, g is gravitational acceleration, and D is inside pipe diameter.

It has been shown experimentally that when the Froude number is greater than 1.0, the velocity is sufficient to sweep gas down a vertical pipe, so if you want to make sure that all gas gets swept out of a system, this is the criteria.

A Froude number of 0.55 is sufficient to sweep all gas out of a horizontal pipe into a plenum.

For Froude numbers less than 0.35, no gas will be swept down a vertical pipe.

For Froude numbers between 0.35 and 1.0, an increasing amount of the gas will be entrained vertically downward. The amount of gas entrained is a function of the Froude number.

 

[BigInch]

Regardless of whether the vapor or gas pockets go with the flow, or stay where they are, they will have the same pressure as the pipe at that point. Increasing pressure will reduce the volume of the gas pocket. If it is a vapor, a pressures inside the pipeline greater than the vapor pressure of the fluid will collapse the bubble to the liquid phase.

 

[cloa]

rrchap
For Froude numbers between 0.35 and 1.0, an increasing amount of the gas will be entrained vertically downward. The amount of gas entrained is a function of the Froude number.
I think you mean between 0.35 and 0.0

 

[rcchap]

I forgot to mention in my reply above that the Froude number criteria were developed for air or nitrogen in a water system. The concept would be the same for other fluids, but the values might be different.

These tests were also at lower pressures.

cloa, it is 0.35 to 1.0. At a Froude number of 0.0, there is no flow (velocity = 0). A Froude number of 0.35 is the threshold of where small amounts of gas will start to be pushed downward. As the flow rate (and velocity) are increased, the Froude number will increase, and more of the gas will be pushed downward.

There is a time element to this as well. For instance, if you had a flow rate that corresponded to a Froude number of 0.75, you could possibly remove most of the gas if you waited long enough. At a Froude number of 1.0, all of the gas will be pushed downward almost immediately.

 

[rconner]

If you achieve that magnitude of flow quantity/velocity through that 4” pipe, I believe (it appears along with several others) that the “air” you are talking about likely will not be there for long at subject high point once you start pumping. While I don’t know how many if any of your ”ton of problems” this addresses (and with the value of my opinion and about a buck and a half in our neck of the woods you could perhaps buy a Sunday paper!), I hope these comments are otherwise helpful. Have a good weekend.

 

[sskaff]

All problems relate to control valves, controls, and sequence of operation. No issues with air, more of a curiosity. I appreciate everyones feedback.

 

[katmar]

There is a large discrepancy between the answer given by zdas04 and the answers obtained using the methods proposed by kacarrol and rcchap.

I guess that David (zdaso4) is using the method proposed by Durand and M[á]rquez-Lucero in their article "Determining sealing flowrates in horizontal run pipes" published in Chemical Engineering, March 1998, pg 129 because using their formula I get exactly the same answer that David gave above. (This method is also given in Branan's "Rules of Thumb" Handbook.)

Perhaps the reason for the difference is that Durand and M[á]rquez-Lucero are answering a different question to the others. The Durand and M[á]rquez-Lucero article seems to be looking at the flow required to ensure that an open ended horizontal pipe will be completely full at the discharge point. The others seem to be looking for the flow required to flush the air out of a horizontal pipe when the end is sealed by either being turned down into a sump, or by a seal leg (goose neck) in the pipe.

If we convert the Durand and M[á]rquez-Lucero formula into the same format as the others and express it in Froude Number terms, it turns out that they require that Fr > 2.55 to ensure that the pipe is full. On the other hand the booklet written by Gilles Corcos (AirInPipesManual) referred to by kacarrol recommends Fr > 0.48 and rcchap gave Fr > 0.55. Actually Corcos (pg 57) says the theoretical value is 0.55 but that he has found by experiment that Fr > 0.48 is sufficient.

There have been many discussions here about the flow required to flush air bubbles down a vertical pipe, and it is well accepted that Fr > 1.0 is adequate for this. rcchap points out above that air begins to be drawn down a vertical pipe for Fr > 0.35 and increases as Fr increases towards 1.0. Corcos states that in practice Fr > 0.64 is sufficient to remove the air in a reasonable period of time for downflow in a sloped or vertical pipe.

In the light of the Fr No. requirements for a vertical pipe I would be surprised if Durand and M[á]rquez-Lucero's requirement of Fr > 2.55 is necessary if we just want to flush the air out of a sealed pipe. But I have never been able to follow the Durand and M[á]rquez-Lucero derivation myself and I would be a bit suspicious of whether we would ever require such a high flowrate, even for an open ended pipe.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[stanier]

This posting referred to fuel in the pipeline. I can recall reading a case of "what went wrong". Here a fuel line experienced a pump trip. When the operator hit the pump reset button the pump restarted. The resulting pressure spike caused detonation of the fuel and the resultant explosion caused havoc.

It was concluded that air was induced into the system by a leaking pump seal.

 

[BigInch]

Then again, high points in fuel lines, esp when tanks are at low levels, can often be dificult to avoid. Some also seem to think a high point above the maximum fuel tank level is desireable to prevent the tank emptying, if the fuel line should leak. I don't happen to think so because of the air trapping possibility that Stanier mentions and fostering vapor locks. The line should be designed not to leak, or with an alarm to close the tank valve in case of fuel line leaks, and/or with secondary containment.

 

[BigInch]

I also think that reliance on the Froude number alone to predict relatively high pressure flow will quickly run out of bounds. While it might be reliable in the suction line where lower pressures prevail, vapor volumes rapidly diminish with increasing pressure, becoming only 15% of their original volume at 100 psig, thereby greatly lessening the effect in the discharge. That would seem to indicate air entrainment in the pump station might be more of a pump efficiency issue than a flow issue. However it would still be important to remember that lower downstream pressures downstream would only tend to "move" the problem from the pump discharge area to the pipeline outlet region.

 

[ione]

The attached paper is a review of the bulk of available literature on the matter of “air in pipeline”. I have found chapter 6 to be particularly interesting.