Transition from pressure to gravitional flow.Possible? 4

[carletes]

Hello all!!

I'll try to explain a situation that during the last weeks is making me go crazy, in case any of you have ever faced to it (I'm sure you have).

I have a system with a pump (elevation of the pump 0 m) that pumps water towards an atmospherical tank (nothing strange so far). After the pump, the water goes up (up to a height of 15 m aprox.) and then it goes down by means of a vertical pipe (DN 300) into the atmospherical tank's inlet (elevation of the inlet is 6 m aprox).
My question is the following: is it possible to size the pump just to make the water reach the highest point of the system (15 m heigth) with atmospherical pressure and then, let the flow to be gravitional up to the discharge to the tank? Which are the conditions for change the pressure flow into a gravitional flow?I have problems when using the Manning equation to study the flow of that gravitional vertical pipe that goes from the highest point to the tank, because as the slope is infinite, I can't calculate the maximum flow that can be discharged just by gravity.

Perhaps I am completely wrong and it is not possibble at all to change from pressure to gravitional flow, and what I have to study is a flow under pressure (Bernouilli equation) from the pump up to the discharge to the tank.

I hope having explained my problem well. If not I can give any explanation you need.
Thank you very much.

Carletes

 

[BobPE]

sure you can do it, you have to design a break to atmosphere at the point you want to transition to gravity, then design the gravity pipe to handle your delivered flow from the pump system. Seems to me that you would want to run it right into the pump. The cost of the gravity system vrs the cost savings form the few inches of headloss saved are not worth the trouble.



BobPE

 

[carletes]

BobPe,

Thanks (again) for your help. One question I still have is how can I calculate the flow by gravity that is possible to discharge with a vertical pipe, because when I use the Manning equation, if I take an infinite gradient (because the pipe is vertical) then the velocity becomes infinite as well, so the discharged maximum flow could be any without limitation. Am I wrong? Has Manning equation, let's say, a limit of gradient?

Finally, what do you mean with a "break" to the atmosphere? Some kind of venting? My problem (perhaps I haven't explain myself very well) is not that I want a gravitional flow from the highest point to the tank's inlet, but on the contrary, I am afraid that if I size the pump to reach the higest point of the system with atmospherical pressure, then, a gravitional flow will "spontaneously" happen in that vertical pipe. Is not possible to happen without that "break" to the atmosphere so all the flow will be under pressure (Bernouilli equation valid so)? I can't catch it...

I hope that with these 2 questions I will completely understand the water behavior in this kind of systems.

Thanks a lot.

Carletes

 

[DLANDISSR]

Carletes:

You should not use the manning formula for this type of flow. You should use the Darcy-Weisbach/Colebrook method or the Hazen and Williams empirical formula (The Hazen and Williams formula is : Hf (friction head in feet) = .002083 x Length(ft)x(100/C)^1.85 x (gpm ^1.85/ d(pipe inside diameter in inches)^4.8655)). (^ means exponent so: 3^2 would mean 3 squared). ) C for clean steel pipe is 130

The pump must be sized to overcome the head difference between the suction and discharge.

The suction head will be the static head plus any pressure above atmotpheric minus velocity pressure and the friction loss in the pipe, fittings, etc in the suciton side piping. This will give you the head at the pump suction.

Subtract the vapor pressure from this number and you have the Net Positive Suction Head which is the pressure of water at the entrance of the pump above the vapor pressure of the water.

The NPSH available must be greater than the HPSH required by the pump or the pump will cavitate (The water at the pump suciton will vaporize and cause reduced pumping capacity. When the vapor recondenses in the pump the collapse of the vapor will cause hammering on the pump impeller and can cause extensive damage to the pump.

The head at the discharge is the sum of the static pressure (from the pump up to the highest point) and the friction loss in all the piping and components, including velocity pressure.

The static pressure during start up (that is, with an empty pipe) will be the the height up to the highest point in the piping (15M in your case). After the pipe is full the static head will be reduced because the flow on the down side will be helped by gravity. The static head would be reduced by (6M) and would therefore be 9M. If you sized the pump for only 9M of static (fully operating conditions) you would not be able to start the system. The flow would never reach the top and begin to flow down to get the benefit of the down flow.

If the head loss is less than the height on the down side there will be negative pressure in the pipe (head loss minus static head). This is OK unless the negative pressure is more than the vapor pressure. If the negative pressure is lower than the vapor pressure of the water (at what ever temperature the water is at) the water will vaporize and create water hammer and reduced pumping rate.

To eliminate the negative pressure the pipe could be vented at the high point (break to atmosphere). This would introduce air into the flow, which may be undesirable. After the vent the flow would be due to gravity. Since the flow would be mixed (air and water) is would be very difficult to accurate predict the friction loss and the flow rate.

Another subject.
In a vertical full pipe the flow will increase until the head loss in the pipe equals the head available. If you had an atmospheric tank and a pipe running vertically down out of the tank and the total depth of the water from the surface to the end of the pipe was 10 feet. If the pipe were a 1" pipe the flow would be about 30GPM. There is 10ft of head avaible and about 10 equivalent feet of pipe, fittings, etc., at 40GPM the head drop is about 10FT.


Probably more that you wanted to know but Hope this helps!

 

[carletes]

Yes Dlandissr!! It has helped me a lot. So, if I have understood your explanation, if I want to get a flow of (for example) 50 gpm, I have to calculate the head loss with that flow from the pump up to the heighest point (15 m), and adding it to that height I obtain the discharge head, OK? Then, I have to calculate the difference between discharge and suction head to obtain the required head of the pump, and I have to buy a pump capable of pumping 50 gpm with that head, don't I?
But, in operation, the required head of the pump will be smaller (6 m smaller), so, because of the characteristic curve of the pump (if it is centrifugal, for exmaple) the pumped flow will be bigger, so if I want a flow of 50 gpm and not bigger, I will have to install a valve or similar to generate new head loss, I suppose.Am I right?
Finally, if I vented the highest point of the system I would be able to guarantee atmospherical pressure in that point (and not negative pressure), but the new mixed flow, as you say, is difficult to predict, isn't it? It is not possible to know the discharge velocity, for example? Isn't it like a flow in a channel? If instead of vertical it had another slope would it be like an open channel?

Thanks a lot.

 

[25362]

I can recall a transformation of pressure flow to gravitational on a reflux line to the top of an atmospheric tower. On reaching the needed level the vertical line turned horizontal before connecting with the tower. A thermometer was installed at the elbow in a horizontal position. Its readings were erratic because the gravitational liquid flow (in an open channel) didn't fill the horizontal pipe section, barely touching the sensing element.

The problem was solved by changing the TI to a vertical position thus being fully wetted by the pumped upflowing liquid.

 

[carletes]

Hello 25362,

The question is why did your pressure flow turn into gravitional flow? Perhaps because there was some kind of air infiltration in the pipe that acted like a venting? I think that if the pipe is perfectly sealed (no air infiltration) thre can't be that transition, but I posted this question to confirm that "hypothesis" with your experienced opinions in case there is something I have not taken into account.

Thanks a lot,


carletes

 

[BobPE]

carletes:

You are exactly right!!!!IT is a piped pressure system, the only way it transition to gravity is to make it do so.

All system calculations should be done on the entire pressure system or it will be wrong.

BobPE

 

[25362]

The pipe was vented by pouring freely the reflux into the atmospheric pressure column top tray.

 

[greg87]

I responded to this same question in a different forum, but since there are different participants (and information) here, I would like to respond again.

The answer to your original question is: Yes, you can just size the pump based on the total frictional loss plus the elevation up to the top of the discharge pipe(and of course any suction head). The contribution to the discharge head due to the downward section of pipe will not add to but would only decrease the total head, regardless of the flow regime in that line, or what word is used to describe it. (One exception would be if the flow rates are extremely high, see below) The pump will need to handle that total head during startup. It will probably run at a somewhat faster flow rate once the line is full. If you must have 50.0 gpm, make sure you have a manual valve you can adjust (throttle) the flow down to exactly what you want.

However, it appears that, besides sizing the pump, you also want to understand what happens in that downward section of pipe, and on how you could calculate the head contributed by that vertical section. I don’t believe that question has been adequately answered.

For the case of very low flow and a very large pipe, you would have true gravity flow in that downward section of pipe. For the case of extremely high flow rate, you could have a full pipe, and you could use a standard formula for pressure loss for full pipes. However, in order to have a downward vertical line full, the frictional pressure loss in the line would need to be equal or greater than 0.43 psi (2.31 ft) per foot of pipe. If the frictional loss is not that high, the water will be “falling” faster than it will be supplied, and the line will not be full. As DLANDER pointed out as “Another Subject” (which is actually the subject ): the flow will increase until the head loss in the pipe equals the head available. He pointed out that in a 1 inch pipe this would be 40 gpm (almost 15 fps). Note for a 2 inch line the flow would be over 200 gpm, over 20 fps (the pump would need to be pumping that rate for the line to be full) Your original question indicated a DN300 pipe, meaning 12 inch, is that correct? For a 12 inch you would need to be pumping about 12,000 gpm…

Under flows less than these the pipe will not be full. There will be air in the line initially, which will not be expelled. Air will also tend to be drawn upward as vacuum starts to form. Therefore, in order to predict the actual vacuum level in that vertical section (to satisfy your intellectual curiosity), you would need to calculate the friction drop due to a type of 2 phase flow (not at all straightforward). However, as per above, you do not need to predict the pressure in that portion of the line to size the pump, or you do not need to try to “make” it a gravity flow system. If you are really in true gravity flow the pressure will be 0 psig at the top of the pipe. If you are anything between that and the velocity that exceeds those above, the pressure at the top of the line will be in vacuum, less than 0 psig.

 

[BobPE]

Gregg87:

How will the system change from pressure to gravity? The downward section of pipe will add quite dramitically to headloss, that is why most people get this design wrong.

You cant have a pipe that has not referenced to atmoshpere going to gravity. If the atmosphere goes in the discharge end, then it will serve to decrease the flow area cross section area, increasing flow velocities, and increasing headloss. The goal in a system like this owuld be to eliminate the atmoshpere penetration into the discharge end. This will reduce frictional losses and make for the most efficient pumping system design.

Be careful when you describe answers, people jobs may be dependent upon them.

BobPE

 

[carletes]

Sorry BobPe and Greg87,

Now I don´t know if you both say the same or not (although I find both answers quite interesting). I understand from your answers that if there is not air infiltration into the pipe, it is not possible a flow by gravity and that is exactly what is desirable, because a gravity flow involves a bigger head loss so a a bigger pump. Is it right? You both agree with that?

Thanks again,


Carletes

 

[BobPE]

carletes:

Now you threw something new at us LOL!!! The cost you will have to look at. Breaking to gravity may or may not be a cost savings. The gravity flow would not be included in the headloss of the pump system, but like we all said beofre, you need to design the transition to graviry. Just remember, gravity flow will occur where you design it to occur by breaking the pipe to atmosphere from the piped system, not from atmosphere coming in from the end of the pipe. The gravity pipe will be larger in diameter that the pressure pipe, so more capital cost. Compare that to cost to pump over life of pump system without making the transition to gravity (including headloss and capital costs for increased pump size for what you could turn to gravity).

Take care

BobPE

 

[carletes]

Bobpe,

You say that if I size the pump for a pressure flow (let's call it case A) in the whole system the pump will be bigger than if I manage to get a gravitional flow in the discharge vertical pipe (case B). But if I am not wrong, in both cases the total head of the pump at the required flow must be sized with the difference between the suction and discharge static heads,dynamic heads, pressures and head loss of the system between inlet and outlet with such normal flow. Then, taking into account that the static head in case B is 9 m bigger than in case A (the discharge point is the highest point of the system), and that the head loss in that 9 m. of vertical pipe is quite small (0.8 w.c.m approx.), the pump will be smaller in case A, as long as I guarantee that shut-off head of the pump is bigger than the static head of the highest point of the system, to make water reach the highest point during start up. Do you agree? Perhaps is not enough for the start up of the pump to have a shut off head bigger than the maximum height of the system and I need a bigger pump?

I hope you don't get tired of "resolving" my problems.

Carletes

 

[BobPE]

carletes:

I think you are getting it. The high point in your system is the point you want to design for in your calculations. If your pump can't make it to that point then fluid will not flow either. I am not sure about what you say with shutoff head being higher that the high point head. We dont know enough about your system and the curves to say if that is the way to design it or not. I would strongly recommend once you design a sloution to get an engineer that is experienced in hydraulics to check it for you. Once the system is built it is tought to run and hide from it and if it doesnt work right you will be remembered for that....I have worked on several "pump-siphon" projects where the pumps system did not work for reasons similar to what you were asking questions of, it is not an easy design, so you are right in asking questions!!!

I don't ever get tired of talking intelegently with someone!!!!

BobPE

 

[carletes]

BobPe,

What I wanted to say with "shutoff head being higher that the high point head", is that to make fluid run along the whole system and fill it during the start up, is enough making sure that the shut-off head of the pump is gretaer than the static head required to reach the high point (no matter head loss up to that point or other facts). Once the pipe is full of liquid, the discharge will be at a lower elevation (the tank's roof) and the flow in the system will be the calculated one, considering the discharge at the tank's roof and the head loss up to that point.
I am sure you know it, but I have read of that in a very good web page (

http://www.fluidedesign.com).

I am getting hooked by your explanations. Thanks.

 

[greg87]

BobPE:

I agree with your conclusion that "The high point in your system is the point you want to design for in your calculations.", which is exactly what I stated in my response. I do not think Carlets will lose his job if he follows that advice. I also agree that the best design would be to have the pipe subsurface in the tank, however, that is not the system he has.
I'm not sure of why you asked me "how does the flow change from pressure to gravity?". In order to frame the issue (true gravity flow as one extreme, and full pipe flow on the other, which seemed to be causing some confusion), I pointed out that at some very low flow, the flow would be gravity, meaning just cascading into the pipe. (gravity flow defined as referenced to atmosphere from the bottom of the pipe) Since we don't know his flows, we (meaning you, or I) can't talk in absolutes, but can provide information. I then pointed out the full pipe flow case. At each of these “extremes”, the pressure loss can be calculated using the std. formulas that had been referred to. However, in between, this is a not the case, and I pointed out was equivalent to a two phase flow regime, not an easy or straightforward calc. You use the term “transitional”, that will “add dramatically to the head”. (sounds like we’re saying the same thing). For flow rates up to that rate where these friction loss equal the elevation difference (here 30 ft), the net impact will be a decrease in head. (now looking over the figures that DLANDISSR gave for a 1 inch line, and that I gave for a 2 " line, both were based on full line flow, vs. a line with air, so those figures were not correct for this application. The flows and velocities before the friction would exceed the elevation head would be less. So, Carlettes, please don't quote them so you don't lose your job!) However, the point I was trying to make, was that you need a pretty high flow to be equal the increase in head from the elevation change, even for a small line. Carlettes (assuming he still has a job) has a 12 inch line! For the application we are discussing, there will need to be one heck of a flow rate before that water will be become so bound up with air that there will be the friction loss will be more than the elevation head. Going to true gravity flow would obviate the need to do the calc, but I agree with you (as I said) that it not necessary.
So Bob, since you felt you needed to give me some (non-technical) advice, I think its OK for me to give you some. If you disagree with someone (like, for instance, me), point out the issue and state your case. Avoid making oblique (and silly) remarks that imply that someone’s statements are so erroneous that they could torpedo careers, especially if you haven’t read them very carefully. This strength of this forum (as well as any organization) is open communication of ideas, opinions, and information. That means listening as well as talking.
Engineers, unfortunately, make a lot of honest mistakes, but they don’t (and I agree they shouldn’t) lose their jobs because of them. However, if anyone is designing systems based exclusively, or even primarily, on what they read on a web site, then they should lose their job. The other behavior that would merit dismissal would be someone who would gratuitously undermine the opinions of other members of the group. So the choice of words, and the job, that you should be concerned about, is your own!
Greg87

 

[BobPE]

greg87

This is a technical forum, so lets stick with it, not emty threats there champ. My advice would be to re-read your advanced fluid mechanics text book from college and you will see what I am talking about.

I do take seriously that no one here should use this advice for anything, but you know what, people do, and misinformed advice can hurt people and their careers.

Off my soap box now lol, cartles does have a 12 inch line and that is all the more important for him to understand correctly what the system is doing, this is not a 1 inch line or interior plumbing system. Flow is conserved in the 12 inch line and atmospheric reference to the discharge end will cause a severe headloss to the pressure system, not gravity flow unless there is a break somewhere else in the line at the point he wants to have the line transition to gravity. I work with this 10 to 12 hours a day 5 days a week so I am confident of what I am saying since I have my nose in my college texts every day. I have also had the unfortunate opportunity to fix problems like this where these assumptions have been incorrectly made.

So sorry if you felt slighted and I hope you still contribute, but I weigh the person receiving the advice more heavily than the person dispensing the advice and that includes me.

BobPE