Low capacity Diaphragm pump - sizing suction and discharge 1

[Wazobia]

Hi.

I am trying to size a pump for a client, to output chemicals from an injection pot to an overhead tank. The capacity of the pump they want is in the 10gallon per hour range (the chemical injection shot pot is 2 gallons in volume). However, I am getting very small values for suction size based on a velocity of 1f/s at the suction side (less than 0.5 inches), as mentioned in API 14E "typical flow velocities".

Does anyone have experience on sizing the suction / discharge lines for such low capacity pumps? Or to be more precise, anyone sized chemical injection pumps with suction size of 0.5 inches?

 

[Wazobia]

Also, i know it is recommended to have a smaller suction flange size [1 or 2 sizes] than the suction line size. But is it necessary?

 

[rickfer]

You can forget about flanges. Yamada, Graco, and very other manufacturers have 1/4" ,3/8" & 1/2" ø NPT air operated diaphragm pumps, that can handle very low flow, and several fluids.

However, lets remember that this kind of pumps doesn't give you an uniform flow, but pulses of high-low flow. This can be avoided with a surge tank, but this will be an extra equipment you'll need to consider.

You may want to consider a metering pump or a peristaltic, (LMI Milton Roy is a manufacturer, I don't remember another right now).

 

[wimple]

I had a similar cas once for selecting a pump feeding a tank at low flow and relatively moderate pressure (About 110 m if I remember well). The OP was positionned in the lower limit of classical multistage horizontal pumps and we declined. In fact it was needed a eccentric screw pumps to perform such duty.

Are you considering a centrifugal pump? why not look at eccentric screw pumps for such low flow application?
Can you provide details about head ?


Wimple

 

[Wazobia]

Thanks for the information. I kinda feel sheepish for mentioning 1/2 inch pipe sizes and flanges, thanks for the reminder.

Rickfer, for one application the flow is to a heat medium system make up tank, so intermittent flow isnt a problem. The second application is to another pump suction. However, a pulsation dampener should help reduce effects of high low flow, as you call it. The pictures i have seen of pumps suitable for this kind of application are very small pumps, they can actually fit in my hand! I'm not very experienced with pumps for such low flow.

Wimple, we are considering a pneumatic diaphragm pump. Total head needed is no more than 9 feet.

 

[rickfer]

If you don't require an accurate metering flow, and you have plant air near, I'll agree that a pneumatic diaphragm pump may be the best options, as these are very low cost, and very small on size and weight.

Smallest pump I know has 1/4" NPT connections and fits on a 6"x6"x6" box, and can flow up to 3.1 gpm.

Remember that this pumps operate different that centrifugal pump. You can adjust pump flow by adjusting inlet air pressure and flow (with the FRL).
If you need to change the flow head, you need to change the inlet air pressure. If you need to change the flow, you need to change the inlet air flow.

If you need to control the flow, you may need the surge tank and a control valve.

Regards,

 

[BigInch]

In lieu of a suction surge tank, increase the diameter... same effect.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[roydm]

One good way of controlling the output of a diaphragm pump is to specify one with a solenoid control rather than the usual free running variety. One pulse of the solenoid gives one stroke of the pump.
Vary the pulse rate to vary the flow.
Roy

 

[Wazobia]

Thanks guys, i have learned a lot from these posts.

BigInch, I am assuming you mean to increase the suction pipe diameter in lieu of a larger surge tank, right?

With a flow of 10gph [just under 0.2gpm], the flow velocity for a half inch pipe comes to about 0.28ft/s. This is far lower than recommended velocities, but as i was at pains to explain to my colleague yesterday, these recommendations (API 14E, for example) mention the words "if practicable".

In the first case, for the heat medium system, the client expressly asked for a "small, pneumatic pump with a flow of 4 - 6gph". The original design was supposed to be by gravity flow to the make up tank, but the installers installed it at a lower elevation than the make up tank, hence my designing a pump. The pot is only 150mm diameter and 450mm height [2 gallons when full] and is filled manually, hence the low flow needed. I figure a design margin of 10gph pump should be adequate.

The second case, for the cooling water system, is similar (same size pot, same flawed installation with not enough static head) but they want to bypass the makeup tank and instead pump chemicals directly to the suction of the cooling water circulating pumps. The discharge pressure head is the pressure at the suction of these pumps.

(I realise i should have described the system in more detail earlier)

Anyways, my question now is, has anyone worked with such low velocities with such pumps? 0.3 ft/s seems quite slow.

Regards

 

[BigInch]

Or a surge tank. The larger suction line can act as a surge tank itself.

If the fluid has/gives no problems flowing slow, let it. Its going to be equally as hard making the suction line smaller than 1/2 inch. The alternative is to dose the line for a few minutes at higher flow and shut it down until its time for the next dose.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[moltenmetal]

It's a mistake to assume that 1/2" NPS is the smallest linesize available.

OD tubing and compression fittings are commonly used for linesizes lower than 1/2" NPS such as for metering applications. With proper routing and support, they perform extremely well, for far less than the installed cost of pipe.

 

[Wazobia]

moletenmetal
The Client specifications rule out anything less than 1/2" NPS pipes for this service of fluid. But thanks for the input.

BigInch,
the client was happy the size / flow. So i think this issue is solved now. Thanks to all who made an input.

I think the thread can be closed now. Err...who closes it, exactly?

 

[JJPellin]

I have brought this up before. I keep mentioning it because I have been burned on this issue before. If you put in a diaphragm pump, you need to make sure that the length of the suction line is not going to be a problem. This is independent of the line size. As was already pointed out, the flow in this line has to accelerate and decelerate with each stroke. If the line is too long, the pump will cavitate because it cannot accelerate the long column of fluid rapidly enough. This problem can be avoided by installing an expansion bottle close to the pump to affectively de-couple the pump flow from the suction line flow. With the expansion bottle, you can run a smaller line and expect more of a smooth, continuous flow through the line rather than the oscillating flow that the pump will produce. "Cameron Hydraulic Data" has a good write-up on the design of piping for reciprocating pumps.

Johnny Pellin

 

[BigInch]

I'll take a rare step and disagree with JJ just a tiny little bit. (Actually split hairs, if you will.) All the pump needs to do on the suction side is to draw in enough fluid at the suction flange to prevent cavitation. It does not have to accelerate the entire suction column of fluid, although that would be what would be the assumed requirement for a typical conservative design that would not consider the effects of transient pressures in the suction line. The ability to accelerate that specific amount of fluid is facilitated by the inlet pressure at the inlet to the suction pipe, the lowest pressure at the pump's suction inlet, the compressibility of the product and its viscosity, if light. It is restricted by long length, small diameters, heavy density and higher viscosity. As a surge analysis of the suction line would show, and if the above parameters are suitable, a sufficient amount of product to load the pump without cavitation can be delivered without acceleratig the entire fluid column in the suction line, or without a surge bottle. But do it like JJ says, if you can't otherwise verify by a surge analysis of that line that it is designed appropriately to be used without a surge bottle.

**********************
"Pumping accounts for 20% of the world’s energy used by electric motors and 25-50% of the total electrical energy usage in certain industrial facilities."-DOE statistic (Note: Make that 99% for pipeline companies)

http://virtualpipeline.spaces.live.com/

 

[MartinSr00]

I know the question was about small suction sizes, but given such low suction velocities, as long as viscosities aren't super high, these small lines are only small because they hurt our eyes when we look at them.

Just a point for your consideration. I work in the water treatment area, and depending on the chemical you're pumping, you might want to consider small magnetic drive gear pumps. They're available in exotic metals (e.g. titanium, stainless steel, Alloy 20, etc), as well as various high tech polymers, and because the magnetic drives, you don't have a seal to deal with, and the output flow is very nearly pulseless. Flow can then be monitored with a small mag-meter, and you're all set. For flow control, you can either go variable speed, but if your required flow turndown is too great, simply put a relief valve back to suction in the flow loop (it's a positive displacement pump after all), and connect the discharge line to a flow control valve. Use the valve to meter the flow. Yes, there's a bit more power loss, but control wise it's smooth and simple control that you can set "right on the money". For most water treatment chemicals, we've found that any stuff in the as-furnished chemicals contains soft sludges, so filtering is commonly not necessary. At my plant, we've been doing this for a decade or so. Works real slick.

Then you don't have to worry about vibrating piping and hard-to-measure flowrates.

Just a thought.

 

[ccfowler]

MartinSr00,

I agree with your recommendations. There's nothing better than a nice smooth stable flow with moderate velicities.