sizing properrly spilout from open channel & sizing the vertical pipe

[Mike4chemic]

Hi all

I have a practical question that causes me to realize that 2 subject are not fully clear to me.
One is the subject of designing lines for vertical gravity flow, and the other is open channel flow.

Please bear with me as this might be a bit long…

First I will give the raw data:
We have a power plant, which has an air cooler as a condenser.
This air cooler has a collector that collects all the sub-cooled liquid after the condensation to a big header. The header size – usually 24” diameter, and 95 m long .
The header is completely horizontal (no slope at all) and from the header it is being drained from one central pipe ( or 2 smaller, symmetric pipes ) to the pump.


We would like to maintain level in the collector: from one hand not to flood it fully ( it will create backpressure in the condenser ) ,and from the other hand – to have at least enough level to assure: No gas bubbles coming into the pump .

I read your recent threads on the subject ( thread: 378-81608,387-87442 ), and also the articles mentioned there ( Simpson ,”sizing piping for process plants” from Chemical Engineering magazine, June 17, 1968 ,and P.D Hills , “designing piping for gravity flow “ , 1983 )

I have to say I am still a bit confused and here are my questions :

First, the practical questions :
1) How do I decide the proper diameter of the vertical pipe from the collector? (to assure both level in the collector and gas bubbles not reaching the pump ?
Is it only a matter of Froude calculation or also other considerations (such as open channel hydraulics ) are needed?
2) How will I know weather to preffer one big pipe or 2 smaller symmetric pipe instead?

And to the more general questions:

3) Should I design the vertical pipe to have first a self-venting section, and then reduce it to full-flooded flow? Or full flooded flow from the collector and on?
4) If I have a case of vertical pipe from the collector, then a long horizontal pipe and then another vertical pipe – should I really only care about the last vertical pipe before the pump ,or I need to make sure the first vertical pipe is also fully flooded?
5) Is it indeed correct to relate to this situation as gravity flow, or is it not a gravity flow because the pump in the end is use as some kind of a seal ?

6) What happens in such a long horizontal pipe (with no slope ) ?
Do we have one level at the beginning, and when opening a valve, and the flow starts- what causes the water to flow? Does a slope start to form from the sides to the center? Does it stay the same level, but inertia is dictating the flow anyway?
What will dictate the final level in the collector eventually? (assuming the flow in and out is constant )
What am I missing here and not understanding fully regarding a flow in an open channel?

That is it.

Hope it wasn’t too tedious ….

Thanks in advance

 

[katmar]

The header may be horizontal, but the water level inside will not be (assuming the pipe is large enough to not run full). There has to be a slope to generate a difference in head, or there will be no flow.

It would be very difficult to design the vertical pipe so that it will always have a level in it to prevent air getting to your pump. The right way to do it would be to drain from the horizontal header down into a surge vessel and then pump from the surge vessel. And then control the level in the surge vessel with an automatic valve on the discharge of the pump.

I have seen it done without the surge vessel, but it was always pulling air into the pump.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Mike4chemic]

Katmar.

Thank you for you quick response.

I still have a few questions though :

1) how do I decide if 2 outlets are better than one ? should I employ some open-channel calculations here? Does it even matter?
2) according to what u say, designing vertical flow to a pump is always problematic without a tank ? ( requiring a pump that can handle low NPSH )?
Can't I instead design a horizontal section of the pipe at the bottom that will have a similar effect like a tank?

So actually, all the suggestions in P.D Hills article ( regarding the needed head for self-venting,or full-flooded ) are won't enough to make sure no bubbles reach the pump?

I only ask this because in our company they did it years without a surge vessel, and bought a very low NPSH pump ( 2 ft , although the air condenser height is sometime 6 meter ) . before employing anything new, a propper justification will be needed..

And one last question: does this situation is consider gravity flow ,or is it not ( because there is a pump at the end of the line...)

Thanks again

 

[katmar]

One large pipe will usually be cheaper than two smaller ones. Also, as the pipe size increases the allowable velocity for self-venting flow increases. This means that one pipe of a given cross sectional area will have a higher capacity than two pipes each of half that cross sectional area.

The main argument against allowing the pump to suck air is that it can cause vibration and lead to high maintenance. If your company has experience in doing this, and can select the right pump, then there is no real reason to switch to tanks and level controllers. As I said before, I have also seen companies doing it that way and they were happy with that.

Gravity flow is simply flow where the cause of the flow is only a difference in level and there is no other source of pressure. I would classify the flow in your header and into the vertical pipe as gravity flow, but the flow in the bottom of the pipe will be influenced by the decreased pressure caused by the pump suction - so perhaps that is not gravity flow. The reason gravity flow causes complications is because the pipes often do not run full, and the resulting two-phase flow is more difficult to design for. If you can ensure that the pipe runs full then there is no difference between gravity flow and pumped flow.

The suggestions by Hills are based on specific velocities for given pipe sizes, and if you have a pump that is influencing the velocity and changing it from Hills' recommendation then you cannot guarantee there will be no bubbles. If the pumped flow is carefully controlled to ensure that the velocities remain self-venting then that should work, but it sounds like your traditional way of doing it was to have no controller and simply allow the pump to suck some air. I think if I were in your shoes I would stay with the system that I know works, unless the pump maintenance has been expensive and you would save money in the long run by installing the extra equipment.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Mike4chemic]

Thanks Katmar

Again thank u for all your answers so far.It is starting to look more clear to me

First I want to correct my data from before and to add that the gas bubbles are not air.
This is a closed system which have a purge arrangment to get out all the non-condensible gases in the system, so the gas is actually the hydrocarbon gas itself ( I don't know if this has any affect on the answers or not, but just for common knowledge.. )

I hope u bear with me a bit more, since I am still struggling to understand this fully :

1) Can these gas bubbles cause cavitation, or it is just a few vibrations issue ( as u mentioned ) ? Do they actually reduce the NPSHa to a lower effective NPSHa than just the head from the condenser? how can the real NPSHa be estimated correctly then ?
2)If the first section of the vertical pipe is designed to have a self-venting small section and then designed to be fully-flooded ( according to Hills ), then why should the pump at the bottom end affect in a way that will entraine buubles? Aren't they "blocked" in the high spot? why do they reappear at the bottom ?
3) again- why is it so important to have a self-venting section at all? what will happen if I will desing the vertical pipe to fully-flooded from the collector and straight down? ( and by that also make it smaller )?
4) How do u think about a horizontal pipe section before the pump?
Can it actually serve as some kind of a vessel replacement? do u know or can refer to some refference as to guidlines as for how long should a horizontal pipe be before it assures fully-flooded flow after a vertical flow?

Thanks again in advance

 

[katmar]

The fact that you are not working with water and air should not make any difference, as long as you use the volumetric carrying capacity of the pipe as being the same for the liquid as for water. This means you will get less capacity in mass terms of course. You never said that it was a water-air sytem - I just assumed that. Sorry.

Cavitation is caused by collapsing bubbles. If the vapor in the bubbles is not going to condense and cause the bubble to collapse then you will not get cavitation. You could still get a noise and vibration from the vapor being carried through the pump, but the impeller will not be eroded the way it happens when cavitating. You should check with the pump supplier for your NPSH questions - there is no general answer.

The problem with designing a vertical pipe the way you describe with the top part being self venting and the bottom part running full is that the design would be for a very specific flow rate. Any variation in your flow rate will take you away from your design point, and then the behavior could be different. It's more of a control problem than a hydraulics problem. If you do the vapor-liquid separation in the horizontal pipe before the vertical section then yes, the vertical pipe can run full. But again it is a design for a specific flow rate and you get back to the control problem.

I would not rely on the bottom horizontal section for ensuring fully flooded flow. Once the vapor is in that line it will dragged along by the liquid, unless the pipe is very large and the velocity is very low. I have never seen it done this way, but that is no guarantee that it can't be done ;-)

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"