Ballast tank fluid dynamics

[Polarity]

Hi all

Could anyone help me on this one please...

We have 750L (200 US gal) ballast tanks to fill on the 40 ft racing sail boat we build. When we designed the ballast system I put the discharge of the electric pump to discharge into the top of the tanks. This means that the pump always works against the full head (about 1.5m/7ft) even when the tank is empty, however it seemed intuitively that this would be less than the resistance of pumping in at the bottom of the tank.
So the question is .. am I correct ? If I am, and by enough to make up for the extra pipe work I am thinking about having the through hull scoops that we also use to fill the tanks discharge into the tops of the tanks instead of the cross over pipe in order to speed up fill time.

What do you think?

Thanks!

Paul

[rwatson]

Off the top of my head - you would probably get a brief benefit intially pumping in from the bottom - but once the tank gets fuller, you have the reverse pressure from the liquid in the tank to compete with.
On that basis, the inward scoops should probably fill from the top also, especially if you are using the boats momentum to provide the pressure.
I wouldnt think it would be that hard to set up a small test rig to verify the concept before the hard work has to be done.
Either that or kidnap some physics student on their way home from uni!

[marshmat]

or kidnap some physics student on their way home from uni!

I resent that!

Paul, in the first case where you are using the electric pump, there are a few more factors that will affect the pump's performance as the back pressure increases (type of pump, for instance). Assuming, though, that the water ballast is essentially free to flow within the tank (not too heavily baffled, etc.) I would look at it this way:

By putting the pump in the bilge and the discharge at the top of the tank, your pump is working against constant back pressure (assume 1.5 m head in your case). Remember, the pump doesn't care what volume of water is on the other side, it only cares what pressure it sees at its outlet. Assuming there's no pressure loss in the pipe due to shear on the pipe walls (a reasonable assumption for the present purposes), the pump will move water at the same rate up from the bottom of a 2' wide square tank as it will through a vertical 2" hose with its top at the same level as the surface of the water in the tank.

By putting the pump in the bilge and the discharge at the bottom of the tank, your back pressure will increase linearly with the water level in the tank, and your flow rate may (depending on the type of pump) drop as the tank fills. The back pressure starts at zero when the tank is empty, and reaches 1.5 m water column when the tank is full.

If your pump is a positive displacement type- a gear or piston pump, for instance- your flow rate is very nearly independent of back pressure throughout the normal operating range. Gear pumps are popular for hydraulic power systems precisely because they can maintain a constant flow rate at very high back pressures, so long as you have enough torque to keep the pump turning at the same speed. In this case, there would be no difference between the two possible discharge points.

If your pump is a turbine or dynamic type, the flow rate will drop off somewhat as the pressure increases. As the back pressure goes up, it is harder for water to be forced into the high-pressure side, and so less water is pumped. In this case, the flow rate at full 1.5 m head would be equal for either discharge point, but while the flow rate would remain constant at any ballast level with top-discharge, the bottom-discharge system would show a higher flow rate when the tank is empty. Thus, it would seem to be the faster option if you have a dynamic or turbine style pump.

When filling with the scoops, you have a constant pressure source. In this case, the difference in pressure between the surface of the water in the tank and that of the water entering the scoops is what matters, and the arguments applied to the dynamic pump case would seem to apply.

[Polarity]

consider yourself kidnapped!

Your reasoning was how far I have got.
It's a centrifugal pump (other types wont give me the high flow rate) so flow rate drops off considerably with back pressure/head/pipe friction. As you say a calculation of the head is no different in a pipe or in a tank. So that states that it is better to put all entry's at the bottom of the tank so as to only work against full head when the tank is full.
BUT... experience and observation seems to indicate that there is another principle at work other than just the head of the water when you pump into the bottom of a tank, thats what I am after - is it just my imagination or is there a difference?

Thanks !

Paul

[rwatson]

[quote= As you say a calculation of the head is no different in a pipe or in a tank. So that states that it is better to put all entry's at the bottom of the tank so as to only work against full head when the tank is full.
[/QUOTE]

I cant follow the logic in that section. I dont think the first sentence leads to the conclusion in the second sentence at all.

The head height inside or outside the tank makes no difference, true, but the back pressure starts long before the tank is full, it fact it starts a soon as the inlet pipe is covered - as you have observed. If you extend the inlet pipe so it almost reaches the top of the tank, you do of course, create much the same head height to pump to as a top loading pipe.

The only other effect that comes into play beside friction on the entry pipe, is the pressure create by the turbulence where the inlet pipe ends.

Whatever the reason though, you have observed that bottom filling tanks slow down a lot.
I suppose all you can do is get a more effective pump, or go back to top filling.

[jeff spinney]

i think the top filling option would only fill as fast as the bottom filling tank when almost full(slower as observed),but if you had a gear,piston,and turbine pump all geared to push the same volume of water with the same power driving them it's hard to think the piston or gear pump wouldn't be bothered by back preasure as much as the turbine.

[Polarity]

Originally Posted by rwatson

I cant follow the logic in that section. I dont think the first sentence leads to the conclusion in the second sentence at all.

The only other effect that comes into play beside friction on the entry pipe, is the pressure create by the turbulence where the inlet pipe ends.

Thanks for the reply, assuming pump level with bottom of tank, and top fill pipe runs up the outside of the tank and enters just at the top, am I correct in assuming the following regards the head?

Empty tank (1.5 m high)
Bottom fill head=0
Top fill head = 1.5m (with a full pipe but none in the tank)

Half full tank
bottom fill head =.75m
top fill head= 1.5m

full tank
bottom fill head = 1.5m
top fill head =1.5m

If the above is true I can draw the conclusion that "it is better to put all entry's at the bottom of the tank so as to only work against full head when the tank is full."

This seems the logical position then, however the general observation seems contrary - hence the post! You mentioned the "pressure create by the turbulence where the inlet pipe ends." Is this a high enough value in a bottom fill situation to warrant a top fill despite the additional head? Or am I missing something else altogether?

The pump has to be centrifugal/turbine type for flow rate to be anywhere close, and the scoops are behaving in much the same way as this type of pump I believe.

Paul

[jehardiman]

Originally Posted by rwatson

The only other effect that comes into play beside friction on the entry pipe, is the pressure create by the turbulence where the inlet pipe ends.

Polarity, pay very careful attention to this excelent statement made by rwatson.

Normally, pumps are sized based upon expected head need and losses. This is based upon not only the true fluid head, but losses in both the suction and discharge piping and valving geometry ....which includes the inlet and outlet...(the infamous "L/D" and "K" factors), which increase with increasing flow rate. Selection of a poor inlet and outlet geometry and valving can result in easily doubling or trippling the head losses which could then easily excede the required head. A single gate valve in 4" piping is the equalivent of 4 feet of pipe, an elbow is ~6 feet, and a sharp edged entrant and exit can double the head loss in a system.

If you sized a low pressure, high volume centrifical pump for just the head difference, I would expect you got a rude awakening if you paid no attention to the piping layout or inlets and discharges.

Get a copy of Crane TP-410,The Piping Handbook, or any basic piping flow text.

[Polarity]

Other constraints drove the pump specification and now I am looking for the best performance possible from what we have (it is already adequate, I want better!), ditto the scoops.

I am familiar with the issue of valves and bends adding additional back pressure - (we use Valterra type slide valves in all the pipes) , though your revelation of 6ft for a 90deg bend means I will be changing the one 90 we have to a smooth radius!

My question relates specifically to the head loss of an exit pipe into air (top fill) V's the head loss in an identical pipe into a (say) half full tank of water - bottom fill

Thanks

Paul

[marshmat]

Selection of a poor inlet and outlet geometry and valving can result in easily doubling or trippling the head losses....

I am familiar with the issue of valves and bends adding additional back pressure....

If I'm not mistaken, rwatson and jehardiman have pointed out a very significant factor that will affect your results.
The commentary I posted earlier is valid for the case where you have negligible losses in the pipe and in the outlet to the tank. (I was under the impression you were going to start with the theoretical, lossless scenario and work from there.) Is this the case in your boat? Are you sure?
In practical applications, there are losses to consider. Not only the pipe and fittings are at fault, but there can still be resistance to the flow after it enters the tank, due to whatever medium is already in there.
When you fill from the top, your water is exiting the pipe into... air. At ~600 times less dense than water, this presents little resistance, thus the flow loses very little momentum as it exits the nozzle. (Ever notice how jetboats almost always have the pump nozzle above the operating waterline?)
When you fill from the bottom, the static head you have to contend with is much less than when you fill from the top. At least, it is at first. Once you are discharging into water, though- and with a free surface, no less- things get unpredictable. If the velocity of the outflow is high, thus a fair bit of momentum and a fair bit of kinetic energy tied up in the moving water, the back pressure on the outflow stream can (note I say can, not will) increase significantly. I doubt your tanks are designed to maintain clean, laminar flow within the tank? Because there's no way in hell that they can. So you get turbulence, lots of it, and you get an outlet stream that is decelerated very rapidly upon exiting the nozzle. My understanding is that this translates to a higher pressure in the nozzle and its supply line, thus a higher pressure at the pump outlet- possibly as high, or higher than, the pressure on the pump outlet in the top-filling case where there is no additional resistance after the flow exits the nozzle.

[jeff spinney]

so if all piping is the same in both methods of filling the tank except the extra length of vertical in the top filling method,the friction of forcing water into water by the bottom filling method will or won't be greater than the always higher head of the top filling?maybe if you shoot the water in from a side you could create a vortex that should get rid of some turbulance.
i'm just trying to keep up with you guys for fun,but i'm learning few things here that might come in handy,we use live wells that we have to drain and fill to put away crates of fish.Let me know if i'm buggin you too much.
also rethinking my statement of all pumps should pump the same presure if loaded with the same power driving them,i should have thought more about your example of the hydraulic pumps,thinking of the higher presure a dinamic pump wouldn't be able to make seems less complicated to me,anyhow i'll keep waiting for a verdict on the fastest way to fill the tank

[jeff spinney]

matt i guess your comment wasnt on my computer before i sent mine, but how come creating flow on the inside is impossible,,,because of baffling?

[marshmat]

Once in the tank, the flow is going to be completely irregular and very turbulent, unlike in the pipe where, although there is turbulence, the flow has a reasonably uniform overall momentum.

On the different types of pumps, positive-displacement vs. dynamic: Drive both at a constant shaft speed, and think of it this way (reasonably valid within normal operating regimes): the former will have a roughly constant flow rate, the latter will have a roughly constant pressure rise across the pump.

[Willallison]

It's anecdotal and only of passing interest, but I supervise the discharge of bulk liquid chemicals from ships into a 6000 ton shoreside tank. As ours is top-filling, the discharge rate(s) predicted by the various c/o's is around 50% higher than if it were bottom filled....

[jeff spinney]

i see, well it's hard to believe how many differant forces apply to pumping water,i knew of pipe friction before but didn't consider how much it applied,and didn't even consider pressure, liquid would cause when pumping into it.I will be able to use some of this info the next time i have to install a bilge pump at the least, defenatly want them to pump as fast as possible. thanks jeff