Air in pipelines and associated losses 2

[dsg1985]

Simple question with a probably not-so-simple answer: How do you generally account for the losses in pipelines caused by air pockets, that (as far as I've been told) can intruduce a restriction in the pipe (by reducing the effective cross sectional area) at the point where they're located?

 

[stanier]

Here is your starting point.

http://www.arivalves.com/ARTICLES/AIR VALVES/Air in Pipelines Lit review.pdf

 

[BigInch]

You wouldn't normally do that, since it is an undesireable condition, so you would design the pipe diameter to give you a velocity high enough to sweep those pockets along. The length of an air pocket would also not normally be sufficiently long to give you a really big non-recoverable pressure loss either. If you think you might have a long enough length of air pocket, you could determine the head loss based on flow in a partially empty pipe using a velocity calculated on the partially full cross-section and the hydraulic radius, as is typically done for sanitary sewer lines not flowing full.

If you are continuously introducing large quantities of air and not using air elimination valves, you need to start considering the possibilities of two-phase flow, which may have many forms of flow across a partially to full cross-section depending on the gas to liquid ratios and pipe slopes encountered in your pipeline at various points. If you get there, you are definitely out of the simple-solution range. Search "2-phase flow regimes" and you'll get some pictures of what you're up against.

 

[stanier]

Ventomat has some very good articles on removal of air from pipelines @

http://www.ventomat.com

 

[ione]

From page 11 on of the attached paper you can find useful information.
This paper has been recently posted by eng-tips member kacarrol in another thread.

 

[dsg1985]

Thanks for the responses.

One of the pipelines I'm considering goes down a mine.

One problem related to air entrainment is that this pipeline, which may be considered essentially as a pipe running down a very long hill (drift/mains), is half DN125 pipe and half DN140. This causes below atmospheric pressures, which I'm worried will lead to air coming out of solution too frequently.

Why half/half? We found using DN140 would reduce the transient pressures enough (by lowering velocity) in the higher pressure areas to allow us to continue using HDPE, which isn't available in > PN25 rating (otherwise for the lower portions we'd use DN125 PN30 or whatever). Using PRVs isn't an option as we actually require high pressures in the lower sections (the pipe travels back upwards for another 150m vertical / 4km horiz after the bottom of the 'hill').

The issue is, as doing some math showed me, if you have a downhill pipeline that goes from small to big at a high enough flow you can get negative pressures generated, especially at the junction where small pipe meets big pipe. And I've read this may cause air to come out of solution (or get sucked into the pipe if fittings are loose).

We could simply use DN140 for the whole run, which I'm considering doing if lower pressures lead to air in the pipeline and cause problems. The only reason I didn't is because of the cost saving over kms of pipe.

 

[katmar]

A detailed analysis of a very similar situation was done in thread378-207481 . The article referred to by ione and kacarrol should answer your concerns about air pocket flushing, and the tread I have referred to shows how to calculate the point pressures along the pipeline. In that thread the situation also was that the pressure went below atmospheric at the end of the smaller diameter pipe.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

Assuming a steady state pressured pipe flow,

Where small pipe meets big pipe, according to the Bernoulli equation, velocity slows and pressure increases. Being that the larger diameter is at a lower elevation, the static pressure component would be higher too.

What it sounds like to me, you did say "pipeline flowing downhill", and assuming that the flow is downhill in this pipe, is that you probably are not getting too much steady state flow, probably because of a fully open end or an outlet valve open to much, thus giving you no, or still a too low an outlet backpressure to allow steady state flows. When you are drawing out a higher flowrate than what is flowing in, vacuum conditions can be developed in the lower sections. If outlet back pressures were to be controlled properly to allow steady state flow across the full pipe cross-section, you would be able to maintain pressures above vacuum. I think you probably have steady state two phase flow in the lower sections, where air is backflowing into the pipeline from the outlet until the water flow from the source gets blocked enough to push the air out again. Just a guess and I may be wrong, but I think it allows the physics to agree with what you are apparently describing.

 

[ione]

Maybe I’m wrong but what if the two sections were changed? What I mean is to have the higher diameter line (DN140) in the first part followed by a reducer to the lower diameter (DN 125) line.
The entrance and exit effects should in part affect the flow rate, but the main contribution to head loss is always given by the friction loss which should be the same in this case. This way you should have a gentle slope in the first segment and the a steeper line in the second segment, instead a “more consuming” head in the first segment. I think this way the issue of having negative pressures generated could be moved downstream.

 

[BigInch]

Frictional pressure loss is more likely to be equal to static pressure gain with the smaller diameter on the steepest slopes, thus maintaining flowrate without negative pressures. It usually makes more sense to have the resulting sum of both static pressure buildup and frictional pressure drop when different diameters are used to move the same steady state flowrate follow as closely to the terrain slope as possible in order to try to keep the required pipe operation pressures at minimums and flowing across full pipe cross sectional areas. Deviation by too much in either direction either increases the required design pressure of the pipe, or tends to casuse low pressures, vacuums and cascade flows in partially full segments.

 

[ione]

In my previous post when I used the term slope, I was not making reference to the physical pipeline profile (which should follow the soil’s profile). I rather was making reference to the slope of the hydraulic grade line, which shows how pressure goes over the pipeline. The hydraulic grade line will have a slope in the first segment of DN140 pipe that is gentler than those characterizing the segment of DN125 pipe. I consider that inverting the diameter of the segments of this composed line could be beneficial, though slightly affecting the flow rate.

 

[BigInch]

Yes I know. The hydraulic grade line should follow the actual grade as nearly as possible to minimize flow and pressure ranges. Putting the gentlest hydraulic grade on the steepest profile may change things significantly. I wouldn't want to suggest it without knowing exact profiles and flowrate ranges, but I wouldn't discount the fact that you could be right either. Just too many unknowns to say.

 

[dsg1985]

Thanks again for the responses.

I should have mentioned the hydraulic grade for this pipeline under normal operating conditions does follow the grade of the earth fairly closely and negative pressures will not occur. However, under normal operating conditions the flow is throttled.

If someone were to come along and fully open the throttling valve at the bottom of the pipeline the negative pressures will occur (hydraulic grade becomes steeper than earth, etc). This is probably more likely than not, i.e. someone wants to throttle the flow further and asks 'Joe' to adjust the valve, but Joe doesn't know which way to turn it and takes a guess... and fully opens the valve....

If this means people have to go along and bleed the air out of the line at the high points - fine - but is it possible that this pipeline, already operating close to the limit of the pipes rating in terms of pressure, could overpressurise? Perhaps as a result of collapsing vapor cavities...

 

[katmar]

It is easy to limit the maximum flow by installing a restriction at the pipeline outlet. This could be a fixed orifice, or a globe valve with the handle locked (or even welded) in position. Then 'Joe' is able to select any flow rate he likes, up to the predefined maximum.

But as BigInch has said, until we can calculate the likelihood of air being drawn in and the minimum pressures attainable, we are proposing solutions to unknown problems.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

The maximum steady state (ignoring all transient possibilities) pressure of a pipeline flowing downhill is usually the static pressure (no flow) at the lowest point in the pipeline. When you open the end valve, increasing friction losses as flow increases will reduce the original static pressure levels, so it is likely that you will not have any overpressures during steady state flow. On the other hand, it is very possible that transient pressures could reach very high levels, especially if the end valve can be closed quickly. Since with this pipeline you could have voids formed from either air, or from water vapor pockets that develop under negative pressure, eliminating air will not guarantee that you do not ever have the voids formed containing water vapor. Only controlling the pressure at all points to be greater than the water's vapor pressure will guarantee that. Controlling the pressure to be sure that all points in the pipeline will always be greater than vapor pressure must be done using that end valve. Basically that end valve should never be allowed to flow out any more than what the pipeline can draw in, considering any dynamic delays involved in changing the inflow rate by only adjusting the outflow rate. That requires that the outlet valve position be changed slowly, possibly very slowly. In that manner you do not develop any transient conditions and instead can treat the changes in flowrates as simply a slow step change between steady states.

 

[stanier]

At the discharge of this line you could fit a back pressure control valve with a maunal valve downstream. The PCV could be set , locked in a box etc and then then manual valve used by the operator. That way even if opened fully the flow would not oncrease and cause the problems you mention.

As for fast closing the manual valve you could fit a gearbox. Even then I have seen ingenious operators make up a yoke that fits into a pneumatic powered drill. They use this to close the valve quickly because otherwise it takes toooo long!

 

[ione]

dsg1985,
You’ve got a valid input by fellow katmar (make impossible for “Joe” to create flow conditions which can be detrimental or catastrophic).


You’ll probably be aware of the stuff in the attached paper anyway......

http://www.itacanet.org/eng/water/Section 2 Gravity/gravity_flow_pipelines.pdf