Double Diaphragm Pumps

[DonyWane]

I posted this question in the Piping and Fluid Mechanics forum and didn't generate any responses - so, I thought that I would give it a try here.

If I have one double diapragm pump on a closed loop, I know that I can simply take the pump curves for the pump, use the air pressure curve that corresponds to my air pressure, and find the flow at any given frictional resistance.

Here's the question (somewhat spurred by previous discussion concerning centripetal pumps in series): What if I add a second air fed double diaphragm pump in series with the first? The resulting pressure obviously isn't purely additive because the outlet pressure can never exceed the set pressure on the air (correct?), even if the pumps are in series. Any suggestions on how one would calculate the resulting flow for a given frictional resistance?

 

[kenvlach]

There was probably no response because AOD pumps aren't operated in series; there would be no improvement in flow and possible diaphragm damage. Wilden recommends that the liquid inlet pressure should be no more than 10 psig (0.7 bar) -- Wilden Pump Guide

http://www.wildenpump.com/pugs/ENGLISH_PUG.pdf

The pulse flow and check ball nature of the pump will create problems for just about any possible combination of pump in series, regardless of sizes and air pressures.
Ken

 

[DonyWane]

Ken,

Thanks for the response! The particular setup I had in mind was a AOD pump followed by a run of piping/filters then a second AOD pump followed by a run of piping/filters and then back into the first AOD pump. This could possibly address the 10 psig inlet issue.

So, even in this instance, is no extra performance offered to the system by the second AOD pump? Is this because as one pump pulses - the other may be on the opposite phase, which could cause some damage. If they were on the same phase consistently (which may not be possible to ensure), would the push/pull nature of both pumps not enhance flow?

These may be stupid questions - I can't quite visualise the operation as well as I can for centripetal pumps.

Thanks!

 

[kenvlach]

1) To run AOD pump output through a filter w/o damaging the filter, you may need a surge dampener (something with an air-filled bladder – see the Wilden site). We use an AOD pump to feed slurry into a plate-and-frame filter press, and it is necessary to do a staged startup, gradually raising the air pressure until the press is filled and develops some back pressure.
2) Trying to match slug flows of 2 AOD pumps in series generally doesn’t work because you need to maintain a total of 4 workable inlet and outlet pressures at a given volumetric flow. Much worse than a 2-stage piston-type air compressor because of the incompressible fluid.
3) Your design may work with a surge dampener and a pressure drop between pumps comparable to the downstream backpressure. Keep the upstream pump throttled down so as to not exceed 10 psig at the upstream pump inlet. Always start the downstream pump first (it should be able to expel air and prime the system by siphon-pulling through the upstream pump), then slowly start the upstream pump. When the downstream pump reaches the desired flow, maintain the air pressure to the upstream pump. Always shut down the upstream pump first.

Another consideration: If you are pumping a hazardous liquid and a diaphragm ruptures, liquid can escape through the air exhaust. A lot worse in a 2-pump system if it happens on the downstream pump.

Good luck. Please give some feedback on your results.
Ken

 

[Compositepro]

The feed pressure to a double diaphragm should never exceed the exhaust air pressure (generally ambient pressure). The pump diaphragm is curved so that the the liquid-side pressure is always less than the air-side pressure. If the liquid-side pressure exceeds the air-side pressure the liquid pressure will drive the pump. This reverses the diaphram curvature on every stroke and will very rapidly cause the diaphram to fail in fatigue ( like breaking a wire by flexing it back and forth).

I don't see how you would benefit by a second pump. What is the problem that you are trying to solve?

 

[DonyWane]

I was actually asked to perform some calculations concerning system performance (i.e., flow rates at various filter DP and other system drops). The conceptual design was performed by another engineer. The pumps will "live" on a cart that will be attached to tanks with cleaning pigs. The system's filters will be located between the two pumps on the cart. Before I could address flow issues, I wanted to better understand the dynamics of two AODD pumps. I am looking for a way to easily set up a model, but the AODD pumps in series are throwing me a bit (i.e., this engineer's way of saying it is throwing me a lot).

 

[kenvlach]

Hi,
Have a typo in my last post. Sentence in 3) should of course read “Keep the upstream pump throttled down so as to not exceed 10 psig at the downstream pump inlet.”

Thank you for the additional design details. AOD pumps are generally only used unless utilizing some of their advantages (good self priming, can pump slurry, high output pressure ~inlet air pressure, non-electrical so no sparks nor burn-out if stall). Your designer must have had these in mind.

Pumping speeds will be difficult to control as the filters become loaded. To maintain flow rates may require regulating air pressures using a flow controller. This only makes sense if starting out at low air pressure. You definitely need pressure relief on the inlet to the downstream pump, perhaps a bypass loop with a spring-loaded check valve. In any event, AOD pumps in series are not suitable for boosting outlet pressure. They may work with a filtration load between pumps, but require surge equalizers on both outlets and pressure relief on the downstream pump inlet.

 

[DonyWane]

Thanks for your help on this Ken. Sounds like running two in series may be more trouble than it is worth.

I have ran a lot of calcs invoving an AODD pump motivating fluid from a reservoir to another reservoir (essentially - open systems). Have you ever ran the calc on a closed loop system? I find myself rather confused by the calculation. With a centripetal pump the calculation is rather simple because the pumps develop head. Here, with an AODD, that is not really the case. The pump is setting discharge pressure (with little or no regard to supply pressure). So, is there a set suction pressure based on air set pressure? If that is the case, then do you treat the system as an open system discharging to an atmosphere equal to the pressure of the suction?

Thanks!

 

[kenvlach]

I only use AODD pumps in open systems – pumping chemicals from drums, transferring proces solutions and precipitated metal hydroxide slurry, cleaning tanks, etc. Main reasons are very good suction, portability (especially with inlet & outlet hoses) and operation from compressed air. But, all of our hard plumbed, closed systems use hard-wired centripetal pumps (except for small electric metering pumps).

A problem for both open & closed systems is that AODD outlet pressure is sinusoidal. We have a policy of double clamping outlet hose ends because of the danger of shaking free and spraying chemicals.

With outlet surge dampeners, several filters of enough flow resistance and volume, and some type of pressure relief prior on the inlet to the 2nd pump, you should be able to get the proposed setup to work. Be aware that too much backpressure on the 2nd output from filter loading will be a problem without fancy inlet air controls and/or pressure relief on the inlet. I don’t have the means to model the system, but by inputting the desired flows and resistances (heads), you should be able to iterate to find the right modulation for the air pressures.

Getting 2 AODD pumps to operate successfully in series is certainly an interesting engineering challenge.

 

[Compositepro]

I've used double diaphagm pumps on a solvent wash cart for resin mixers that included a filter. This is a good application because of the flammable environment and the self limiting of outlet pressure as the filter clogs. For continuous pumping they are not energy efficient and are high maintenance. In practice, pump curves are only useful for selecting the size of pum to buy. The flow rates are often limited by the air supply lines and regulators rather than the liquid side.

When you say pumping in closed loop I guess you mean there is no tank or surge volume in your loop. I still can't see any benefit from a second pump. If you want more pressure across your filter it can only be done with more air pressure. However, more pressure will usually just compact the filter cake and reduce its permeability and not increase flow.

I'm sending this several days ater writing it since I lost my modem connection while drating this and haven't been on-line since. Just in case it seems out dated.