10 replies to this topic

[lsoaresfernandes]

Hello,

 

I need to perform a dynamic simulation of the following system:

 

 

A air vent added at the upper part of a siphon, where the relative vacuum formed sucks air in from the atmosphere. I attached below a spreadsheet with the screen shot of the steady state simulation, but when I try to turn it to dynamic, I get unrealistic pressure values.

 

I am specifying the pressure at all inlet streams (pictures also attached).

 

 

If someone can help me how to properly perform this simulation I would really appreciate. 

Attached Files

•  HYSYS_Prints.xlsx   119.97KB   30 downloads

[Pilesar]

I've modeled dynamic systems in other software. The information supplied with the question is sparse and I won't be able to help much. I don't find much dynamic element to the model with only one control valve. It looks to me that this could as well be kept in steady state as a case study. Are you able to get results in steady state under different conditions? If not, then the dynamic model will have little chance. If you 'need' the dynamic model for some reason, then perhaps putting a control valve on the air flow might help in troubleshooting.

[lsoaresfernandes]

Hello,

 

I am having problems to understand the sizing criterion for a vent line (vacuum breaker line) placed at the upper part of a siphon just upstream overboard discharge. This line is part of the sea water system, of a FPSO, so there is a pump upstream the siphon and several heat exchangers.

 

I believe the siphon exists so it can provide back pressure to the pump and remove trapped air from the upstream pipes. But I am having difficulties trying to understand how to proper size the vent line (vacuum breaker line) at the upper part of the siphon.If my understanding is correct, there must be an equilibrium pressure slightly below atmospheric pressure, which should suck air into the system. This air will then be discharged along with water downstream of it. 

 

 

I checked some norms and it looks like Norsok P002 give some information, but to a froude number up to 0.7 (mine is 0.81 and line diameter cannot be increased).I was asked to perform some calculations using OLGA to understand if we can use this line, but I still don´t get what the main outputs would be.

 

 

If someone can help me answering the following questions, I would really appreciate:

 

1) What do we seek for when sizing vertical downwards lines before overboard to a froude number less than 0.7?

 

2) I believe the vacuum breaker line at the top of the syphon is required to avoid pulsatile flow due to vacuum formed, but how should I size it?

 

3) FInally, it looks to me that this criterion of Froude less than 0.7 is used for gravity driven flow. Should this be the case?

 

 

I attached a draft of the process.

Thanks a lot for any helps.

 

 

Attached Files

•  schematic.pptx   34.11KB   30 downloads

[breizh]

Hi,

Consider the links underneath to support your work :

https://www.valmatic...Valves_4-18.pdf

 

https://www.bermad.c...-Air-Valves.pdf

 

Hope this is helping you

Breizh

[katmar]

Although the piping setup that you describe is very frequently used, I have never come across a useful design procedure. Nevertheless, we need to be able to design these and I will explain below how I have approached the problem. Hopefully others will chime in with comments to improve this method.

One thing that I have seen in successfully operating plants, and also in the limited academic studies I have found, is that the siphon breaker piping can be of significantly smaller diameter than the water pipe.

Dealing with your questions in the order that you asked them:

1. I have not come across any recommendation to keep the Froude number below 0.7 for arrangements like yours. What I have seen (and used successfully myself) is the recommendation to keep the Froude number below 0.31 to prevent a siphon from ever being established. At this low Froude number air bubbles are not entrained by the water flowing downward. But this criterion would not be applicable to your case where you do want to establish a siphon. It is most often used in tail pipes from barometric condensers and similar arrangements.

In your case I would recommend keeping the Froude number above 0.7, and close to 1.0 if possible. This ensures that any air in the pipe is entrained by the water and removed from the piping. At Froude numbers between 0.3 and 0.7 pulsation and vibration is possible as the air and water try to flow in opposite directions.

2. The approach I have taken in sizing the vacuum breaker line is to say that I would allow a small vacuum of around 5 kPa (0.7 psi) to form where the vacuum breaker joins the water pipe and then size the vacuum breaker line for a volumetric flow rate of air the same the volumetric flow of water in the main pipe. Under these conditions air will be pulled into the line in preference to lifting water over the siphon, but when the water is being pumped the vacuum will never form and no air will be drawn into the pipe.

You have not given the size of the water pipe, but for the sake of having a worked example I will assume that it has an ID of 300 mm. You can change this to the correct value in your own calculations.

A Froude number of 0.81 in a 300 ID pipe gives a velocity of 1.4 m/s and a flow rate of 100 liter/s. A pipe to allow 100 liter/s of air, assuming a length of 5 m and a pressure drop of 5 kPa, would have to be 52 mm ID. So you could use a 2" pipe unless you are being ultra-cautious and then you could use a 3" vent. These numbers "feel" to be in the correct proportion to the water pipe.

The 5 kPa limit I used for the allowable vacuum would prevent water being siphoned up more than 0.5 m. You can adjust this value if your arrangement is different.

To be completely thorough you should check what the pressure would be at the joint of the vent and the main pipe during normal pumping. If this pressure is below atmospheric then some air will be drawn in and you should re-calculate the pressure drop for 2-phase flow downstream of the vent. A couple of iterations would probably be necessary to come to an answer.

3. As I stated earlier, I have not seen a recommendation to keep to a Froude number of less than 0.7. Values that I have seen are 0.31 to prevent a siphon, and in long water lines that follow the lie of the land and can have local high points it is sometimes recommended to keep the Froude number above 0.6 to flush out any air pockets that may form.

[lsoaresfernandes]

Hello Katmar,

Thank you a lot for your detailing answer. Since I´m obviously not a specialist in the field, I still have some doubts on your commnents:

 

1) Recent Norsok norm (P-002) indicates a Froude number of 0.7. Nevertheless, it seems that it was always used as 0.3. 

I don´t get what you mean by preventing a siphon to ever being established. Do you mean a pipe with full cross-section occupied by water? In my case I don´t think this is possible. Vertical downard line to atmosphere is longer than 10m (around 30m), so I believe this would create a vacuum at the higher part of the pipe, where water vaporization would happen and flow would be intermittent. In my mind, that´s why the vacuum break line is added.

 

2) You mentioned that "To be completely thorough you should check what the pressure would be at the joint of the vent and the main pipe during normal pumping. If this pressure is below atmospheric then some air will be drawn in and you should re-calculate the pressure drop for 2-phase flow downstream of the vent. A couple of iterations would probably be necessary to come to an answer"

 

Since the pipe is discharding to atmosphere at a lower level, I believe pressure there will always be below atmospheric and Indeed would expect some air inlet. 

 

What I don´t get is that you mentioned "when the water is being pumped the vacuum will never form and no air will be drawn into the pipe.". Isn´t this exactly the situation? I would expect air to be sucked into the pipe at this exactly condition.

 

 

3) Going back to the Froude number, I still have some difficulties to understand what would happen above and below let´s say the froude number limit of 0.3.. Am I right to say that below 0.3 you would stablish a connection between atmosphere (air) at the lower part to the upper part? Forming something like a swirl at the top vessel if that´s the case? 

 

 

 

Thank you a lot for your answers, I am feeling a little lost with what seems to be a simple engineering design.

[katmar]

I do not have NORSOK P-002 but in the old P-001 the Froude number was given as 0.3. I can't say that 0.7 is a mistake without having seen the context.

My previous comments were mainly aimed at sizing the vent pipe to function as a siphon break to prevent the flow from continuing on after the pump was turned off. The drop of 30 m puts a completely different slant on the problem and there will be a vacuum drawn at the top of the loop all the time if the vent is not adequate. This could cause vaporization of the water and result in vibration.

If the water is being discharged at 30C or less then a pressure of less than 5 kPa absolute would have to be generated to get the water to boil. I would not design for such a low pressure because the air flowing in through the vent would reach very high velocities and would be very noisy.

Looking at some rough numbers it appears that with the very low pressures that could potentially develop and the resulting air velocities it would probably be best to make the vent the full size of the water line. You should however do a full pressure balance to determine the pressure at the top of the loop so that you can estimate the flow (and velocity) of air into the system and to make sure that the vent does not result in a water shower for everyone on board.

Regarding your question about the 0.3 limit for the Froude number - If the bottom (discharge) of the pipe is not sealed (i.e. not under water) then with a Froude number of 0.3 it is unlikely that the pipe cross section will be full of water. This could quite possibly result in the scenario you have described where there is a continuous air path from the bottom of the pipe to the top. In this situation the pressure at the top of the loop (and all the way up the downleg) would be atmospheric. There certainly must be some flow rate below which this continuous air path exists, but I have not seen a Froude number specified to ensure this. However, without hands-on experience I would not rely on this occurring and I would include the full-sized vent.

The 0.3 limit is usually given for pipes that are water sealed at the bottom and the water velocity down the pipe just has to be low enough to allow air bubbles to de-entrain from the liquid. I do not know that 0.3 is sufficient to ensure that the air path is continuous over the whole 30 m. At the bottom of the 30 m pipe I am confident that the cross section would not be full of water - you need a Froude number of around 2.0 to guarantee that. But there could be some point higher up in the pipe where the water velocity is lower and the cross section is liquid-filled.

[lsoaresfernandes]

Thanks for the answer. I think it worth me elaborate some more about the problem..

 

We are designing a platform to one of our clients. In order to assure the system will work properly an air vent must be positoned at the upper part of the siphon. I understand the concern of avoiding a siphon effect, but in my mind this air vent is there to assure a pressure close to atmospheric at the upper part of the siphon, so it can provide backpressure to the control valves downstream the pump.

 

 

This situation started because the sea water flowrates are extremely high, so even for a 52inch pipe, the froude number is high (around 0.8). They use Norsok to approve or not such system (vertical lines froude should be less than 0.3 and horizontal lines following manning flow for 75% filled pipe). After some discussions, we came out with the idea to add a funnel (see updated attached figure) between the siphon and the discharge. Nevertheless, they still want evidences this system would work without too much vibration or risks of water spillage.

 

I am struggling to understand why in such system we should assure that the froude is smaller than 0.3.. After a lot of research what came to my mind was what I said above, that by doing so, the system would be self vented, so it would have a continuously air path from the lower discharge point up to the air vent.. but from what you said, froude smaller than 0.3 would not assure that.

 

Nevertheless, I think it would not be that difficult to calculate the air inflow at the air vent by an iterative method as discussed above. The only problem with this methodology is that it assumes there is a pressure drop from the air vent up to the discharge. That would not be the case if the continuous air path is formed, as you mentioned. 

 

The second issue is how can we assure water would flow from the funnel to discharge offboard. Since the horizontal line has a slope, my understanding was that the water level at the funnel + pipe downstream of it had to be high enough to equalize friction drop along the line.. it would be a simple bernoulli calculation.. I don´t get why they are worried with froude number at this part of the pipe. Maybe there is some concern with air entrainment/vibration? I just don´t get it..

 

 

 

Since pipes above 52 inch are not commercial, it would be way more expensive to increase the size of it.. that´s why we are being pushed to get this system to work with a 52inch line.

 

Attached Files

•  schematic.pptx   34.82KB   18 downloads

[katmar]

I just want to make it absolutely clear that although I have successfully designed siphon breakers in the past (using the method described above) I do not have practical experience with ship overboard discharge systems. So all my comments are based on first principle deductions. If you can get advice from someone who has actually done this before it would be better. However, just because something has been done in a certain way before it does not mean that that is the best or only way. And it is an interesting challenge anyway so I will carry on but please keep in mind that I have not designed this type of system before and you should check your final design with someone who has relevant experience.

Having a Froude number of 0.8 in a 52" pipe gives a velocity of 2.9 m/s and a flow rate of 3.85 m3/s. You don't want to get something on this scale wrong!

I don't see the benefit of adding the funnel system. It means you have 2 design problems where you previously had only one. Design procedures are well established for partially filled sloped pipes, but are you able to account for how the rolling and pitching of the ship will affect the slope and instantaneous flow?

But I think that using the Manning's sloped pipe method can give us some very useful insights into the problem. To get a flow of 3.85 m3/s in a 52" pipe running full requires a slope of only 0.5%. And to get the pipe to run with a depth of half the diameter requires a slope of only 2.5%. So whereas I was previously unsure of whether the 52" pipe would be running with the cross section full of water, these calculations give a very strong indication that at these flow rates a 52" pipe will have a continuous air path from the bottom to the top.

My only reservation is that a Froude number of 0.8 is getting close to the point where a siphon will form quickly and it is likely that a lot of air will be entrained and the vent pipe will be sucking in a lot of air. The flow rate of 3.85 m3/s gives a superficial velocity (i.e. assuming full cross section) of 2.9 m/s so if the cross section is actually only 25% full the water velocity will be close to 12 m/s and will drag a lot of air with it. The vent size required would be much bigger than for a siphon breaker and I would think that a vent half the diameter of the main pipe would be reasonable.

It is interesting to also consider the other extreme - i.e. where we are sure there is no air path from bottom to top. If the static head is exactly balanced with the friction head in a vertical pipe while running under gravity it would require the pipe to run completely full. Under this condition there would be no pressure gradient in the pipe and (although filled with water) the pressure would be atmospheric at all points of the downleg. No air would be drawn down the vent by pressure difference but some air would surely be entrained by the venturi effect as the water passed the vent. A pipe sized on this basis would only be about 500 mm diameter and would be significantly cheaper, but if it has not been used and proven before the risk would be high and I would not try it without confirmation that it has worked previously.

Overall I think your 52" pipe will work well if provided with a conservatively sized vent and I cannot see the value of increasing the pipe size to 76" just to have a Froude number of 0.3.

[lsoaresfernandes]

I think the benefit of adding the funnel is that it will decrease the distance the flow has to travel before discharge. This would imply in a smaller inflow of air at the vent.. for high Froude I would not expect any free air path or pressure equalized from the discharge at the funnel up to the air vent. 

 

From the funnel up to the discharge, where the flow is purely gravitational, I don´t see any problem of a smaller froude number, do you?

 

I think we will perform some simulation using OLGA just to have a better view of the pressure distribution, hold up, etc. 


Still, I can´t understand why the norm sugests a froude smaller than 0.3 for vertical drain pipes or manning flow for semi-horizontal pipe.. In my mind, for the simple purpose of discharging water, that would not be a problem.. I would only have to assure a pressure head to account for friction drop at the semi-horizontal part of the pipe.

[Art Montemayor]

To all members interested in this thread and advocates of better and safer fluid transport designs and installations:

 

I have followed this thread because of my interest in the subject and my respect for the expertise and knowledge of our some of our fellow members such as Katmar, Breizh, Pilesar and others.  I read the documents that Breizh presented and found that they contained information that was, in my opinion, very useful and applicable to the subject of this thread.

 

Although I found typos, grammatical and spelling mistakes, as well as bad formatting and data organization, I feel some of you may be interested in reading the material that I have copied, formatted, corrected, and organized.  The Italian authors have, in my opinion, put forth some valuable and insightful information that may help our members in their future piping design and field installations.  A lot of what they point out has been experienced by me in piping installations and modifications.

 

The document may still be subject to further improvements to make it easier to read and understand the subject matter and I would appreciate any comments or suggestions on the matter.  I found the original Italian brochure difficult to follow, read, and study because of the language errors and layout of illustrations.  I hope you find my revision easier to follow.

 Working with air valves to prevent syphon & vacuum.docx   9.68MB   56 downloads