Air-Operated Double Diaphragm Pump Sizing 1

[RJB32482]

Hey,
Whenever sizing an Air-Operated Double Diaphragm pump (or any pump in that matter), do you size for just the discarge head? Or do you size the pump (determine operating point on pump curve) by using:

Head=Discharge head-Suction head

Thanks

 

[StoneCold]

RJB32482
We use a lot of these pumps. Almost all of ours are 1" pumps. We use them in all kinds of intermitant services. We have pressure regulators on the compressed air to the pumps. That controls the total allowable discharce pressure (Discharge=Supply) and we also have a ball valve on the air line that we throttle to slow or speed up the pump. It is pretty crude but effective for what we do. If you actually have a standard application maybe you would be more interested in a vane pump or centrifugal pump but for miscellaneous pumping needs the 1" or 2" diaphragm pumps work well.


Goodluck
StoneCold

 

[RJB32482]

Thanks for the Info,
I have heard two different ways to determine the operating points on pump curves. Do you just use discharge pressure or (Discharge-Suction)?

Thanks

 

[StoneCold]

RJB32482
I only look at the required discharge pressure and the volume I want to pump to pick the pump size. It is not like a centrifugal pump because in most cases you do not have a stable NPHS to work with. Lets say you are just pumping out of the bottom of a full tank to the bottom of an empty tank. In the beginning you have lots of NPHS and low head. That situation slowly reverses as the suction side tank empties until you are pumping air.
So the sizing of the pump is not that critical because it is a transient operating point.
Maybe if you explain to me in detail what you want to do I can help you a lot more.

StoneCold

 

[RJB32482]

Thanks for the reply,
The pump is going to pump antifoam from an tote into an empty reactor with about 5 psig pressure in it. Don't have the piping configuration yet, but we can place the pump pretty closeto the tote (within 5 feet). The maximum level of the tote above the pump suction is about 5 feet. Which would coincide to about 2.1 psig of head from the tank level. This is going to decrease until the tote is about completely empty. So it would decrease. So I guess you would be correct as to just use the discharge pressure head to size it. Correct?

Thanks

 

[MikeHalloran]

If you're using the DD pump to meter the antifoam, i.e., stop pumping when X amount has been pumped, you might be better off counting its cycles than trying to use rate x time to get volume, because the cycle rate is sensitive to temperature and air pressure and liquid head.





Mike Halloran
Pembroke Pines, FL, USA

 

[RJB32482]

When do you mean by cycles? Also please answer my previous post.

Thanks

 

[MikeHalloran]

The output of a DD pump is not continuous, but pulsatile. If you put an undamped pressure gage on the discharge, the gage will indicate wild pressure fluctuations until it breaks, which won't take long. This happens because the diaphragms are alternately pressurized and relieved of air, applied in alternation by a mechanical toggle valve that is triggered by the position of the diaphragms, or the bar linking the two.

I am not aware of any air- operated DD pumps that actually include a cycle output, but a pressure transducer and Schmitt trigger could give you a usable square wave. You would then have to know the discharge volume per cycle; it appears that you might have to measure that for yourself, because the pump manufacturers rate the pumps as if they were continuous discharge devices. You wouldn't need real fast electronics; DD pumps typically cycle much slower than relays.

I've never used 'em as metering pumps myself; they're better adapted as transfer pumps, i.e. pump the entire tote-load into the reactor, and when the discharge pressure drops or an attendant notices the pump speed up (because it's pumping air), shut off the air supply.

There exist high purity, high precision, high cost "diaphragm metering pumps" with mechanically driven diaphragms that are intended for metering, and can include some kind of cycle count or rotation count mechanism.

I.e., if your mix ratio is other than one entire tote of antifoam per reactor charge, you might consider a different pump technology.

The curves provided are for flow vs. total head, not just discharge head.


Mike Halloran
Pembroke Pines, FL, USA

 

[RJB32482]

Well, I'm going to have a flow meter on the discharge where the operators can look at when the flow totalizer hits a certain amount, then they will shut off the air. What u think about that?

 

[MikeHalloran]

A rotameter will not give a stable reading.

A paddlewheel might be inaccurate because of constantly accelerating and decelerating.

A true positive displacment flowmeter might be accurate, but might suffer reduced life.

Check with your flowmeter supplier.



Mike Halloran
Pembroke Pines, FL, USA

 

[StoneCold]

You may want to switch pumps as MikeHalloran suggested. You will be happier with something like an LMI pump for real small flows or a Pulsafeeder pump for larger flows. Then control based on the number of strokes, or time which ever is easier for you. The diaphragm pump is not going to give you what you want. Maybe you could fill an intermediate tank from the tote and then pump the whole intermediate tank into the reactor but that is more operator intensive.

Goodluck
StoneCold

 

[RJB32482]

We wanted an air operated pump for the application. Any ideas for air operated pumps for the application??

 

[MikeHalloran]

The air operated DD pump can work okay, if you use it to fill a charging cylinder from the tote, and then drain or pump the contents into the reactor.

If your reactor dosage is small enough, you could use an air operated single stroke dispenser, e.g., air cylinder coupled to hydraulic cylinder, coupla check valves, etc.



Mike Halloran
Pembroke Pines, FL, USA

 

[checman]

RJB32482

On some pumps like your average centrifugal, total head at the discharge of the pump is the pumps discharge head plus suction head. But on an average PD gear pump the suction pressure just lowers the hp required to do the job. When sizing an AOD pump the suction pressure is a positive. But AOD pumps do not like a pressurized suction. The diaphragms and there drive assembly parts are not designed for the higher pressure to be on the product side. Bent shafts and short diaphragm life is the result of high inlet pressure. You do not have high inlet pressure.

I have AOD pumps that roughly meter chemicals into batch reactors. Some are directly equipped with solenoids and others with stroke counters that tie into a solenoid and allow you electronically control the number of strokes or the output of the pump. Most large AOD pump manufactures have an option for batching electronics.

To improve accuracy and assure that the check valves are seated properly you will need a back pressure valve on the pump discharge. This will also stop the siphon flow. If you want to try and use your flow meter you would want a surge suppressor between the pump and the back pressure valve. Your flow meter will not display accurately but it will be some what repeatable if the surges are suppressed. To regulate the pump through its air supply you can put a valve ahead of the pump to control the air volume which will control the pump speed. If the supply air can exceed the pumps maximum inlet pressure you would want a pressure regulator between the supply and the volume control valve.

Regards, checman

 

[RJB32482]

Checman
So since AOD pumps don't like too much inlet suction pressure, it is better to make the suction line longer since the friction loss will decrease the pressure added to the suction side by the static head (level in the tote)?

 

[checman]

I would not worry about your positive suction pressure, it is not enough to do damage. It is enough to support a siphon flow that will keep the check balls from seating correctly. The back pressure valve on the pump discharge is key for your system, if it does not have very much back pressure on the discharge.

Regards checman

 

[kenvlach]

All of the advice given above is very good. I just want to mention a few more variables.

-- Diaphragm material affects the flow rate. Teflon (+Neoprene backing) is more rigid, hence slower (and a lot more expensive) than Hypalon. Only use Teflon if your chemicals require it.

-- Exhaust air muffler. For less noise or better removal of lubricating oil mist, use a more restrictive muffler (smaller micron pore size), which slightly decreases flow. Also, the small, sintered bronze ones don't get broken off like the larger plastic ones.

--- AOD pump body construction. For emptying 55-gallon drums & totes, I like the lightweight polypropylene body pumps with intake pointing directly down. These sit directly above the bung (opening).

--- AOD check valve design. Maybe doesn't matter if you only pump liquids, but if you have solids, the flapper valves don't jam as much the check balls. I've also heard a rumor that some manufacturer had the clever idea to make the outlet check balls larger diameter than the inlet check balls to reduce jams (whatever enters can get out!).

--- Location. Try to avoid any possibility of siphoning; eventually it will happen. Better to put the pump above the source & below the outlet. Also, keep the inlet hose as short as possible.

--- Surge dampener. The ones I have seen are like a flexible bladder & very expensive. But, as others have mentioned, you may need for a flowmeter to work. Also, you don't want 100 psi [or ~whatever your air pressure is] pulses going down your outlet hose or piping. Hoses get loose, PVC pipe breaks. Not nice to be around when it happens.

--- Air valve construction. I suggest one that needs little or no lubricating oil – presume Teflon-coated. Starting with used equipment, as the air valves wear out, I replace them with the anti-freeze up (maybe anti-seize up is correct phrase) type air valves. Some advertising makes pretty good claims, but I think they all wear out and seize up.

 

[RJB32482]

Will lowering the air pressure going into the pump help to decrease pulsations that shake inlet and outlet piping?

 

[kenvlach]

Yes,
the liquid pressure spikes at the pump outlet are basically equal to the air pressure suuplied, minus a few frictional losses, e.g., inelasticity in the diaphragms (internal friction). The pressure spike dampens somewhat as it travels down the outlet hose or pipe, but the mathematics depend upon the fluid & hose properties.