What kind of pump can do this?

[RCH]

TL;DR Summary

How to pump a fluid with suspended particles, without damaging them?

Hello --

I am working on a project where I need to recirculate a fluid which has a large quantity of glass spheres of say, 1mm diameter suspended in it, but without the pump crushing the spheres. I would appreciate any advice.

To give a simple description of the desired function: it will be a closed system with a vertically-oriented tank about 1m tall by 75cm wide, whose front and back are two sheets of glass spaced a few mm apart. The suspension will enter the tank at either the top or the bottom and will flow to the opposite end of the tank where it will collect and be pumped up (or down) about 1.5m, to start the process again. (That extra 50cm is to account for curved entry / exit tubes, rather than elbows, to use gravity to prevent the spheres from their expected tendency to collect in elbows.) As you might suspect, there is a bit more to the project than that, but that is the gist of it. The real goal is to move the glass spheres; the fluid is just there to act as a circulatory medium.

I don't know yet what fluid would be optimal for this suspension, but the spheres will likely be denser than it, which could affect the spheres' behavior with various pump types.

I have three main concerns: the primary one is to avoid crushing the glass spheres into dust by the mechanical action of the pump on them. The secondary concern is avoiding the spheres causing wear to the pump's parts where they come in contact; I want this to be low-maintenance. And the tertiary is to make it as quiet as possible. It seems to me that impeller pumps might be prone to the first two problems. A diaphragm pump doesn't seem practical here, and I don't want to use anything that reciprocates, due to the noise and vibration. It seems to me that a peristaltic pump would almost certainly cause crushing of the spheres.

Perhaps some version of an Archimedes' screw would be a good choice for avoiding crushing the spheres, and since the suspension has to be moved vertically about 1.5m, their long, cylindrical configuration might be a plus. The "Olds Elevator" is interesting as well, but rotating the cylinder, rather than the augur, might not actually give significant advantages over rotating the augur, and might add complexity ( ). Perhaps an augur and cylinder that are in one piece, the augur of which has "steps" or "buckets" formed into it to help collect the spheres, would be better than the standard screw?

I love the idea of a magnetohydrodynamic pump (no moving parts!), but I doubt that the glass spheres could be convinced to move along with the fluid (but it might be interesting to try anyway).

I'd welcome any constructive opinions!

Thanks --

Robert

 

[jrmichler]

It would be helpful to know:
1) Your desired flow rate?
2) Specific gravity and viscosity of your fluid.
3) Density of the glass spheres.
4) Concentration of the glass spheres. Percentage of total volume that is fluid and what percentage is spheres.
5) Will the flow be up through the tank, down through the tank, or do you want the capability of flowing in both directions?
6) If you connect to the tank with hoses as in your sketch, the flow will be mostly through the middle of the tank, zero in the tank corners, and slow down the sides. Is that acceptable?

 

[DaveE]

IDK, but I would look into mimicking a heart. i.e. a diaphragm pump with leaf valves. Hearts have to pump blood without breaking cell walls.

 

[Lnewqban]

Welcome, @RCH !

I would by-pass the glass spheres from the fluid around the point of mechanical pumping.

The most destructive effects will be the friction and impact interaction among the glass surfaces of the spheres.

 

[Baluncore]

1. Use a centrifugal pump with sufficient clearance to pass the spheres.

2. Use a continuous flow, three piston pump, with head clearance and soft rubber flap valves.

3. Remove the spheres with a cylindrical sieve or filter. Pump the liquid, then reintroduce it to flush the spheres forward.

4. Use a jet injector pump.

5. How do they pump the fluid when "abrasive flow machining"?
https://en.wikipedia.org/wiki/Abrasive_flow_machining

 

[RCH]

It would be helpful to know:
1) Your desired flow rate?
2) Specific gravity and viscosity of your fluid.
3) Density of the glass spheres.
4) Concentration of the glass spheres. Percentage of total volume that is fluid and what percentage is spheres.
5) Will the flow be up through the tank, down through the tank, or do you want the capability of flowing in both directions?
6) If you connect to the tank with hoses as in your sketch, the flow will be mostly through the middle of the tank, zero in the tank corners, and slow down the sides. Is that acceptable?

Thank you for the response!

1. The desired flow rate will be low.

Perhaps it would help if I explained the actual point of the project:

I am making a set of windows for a chapel. I think of them as being sort of "kinetic stained glass", although that's probably more confusing than illuminating. These will be made of fairly thick clear plate glass. I didn't show it in the "diagram", but rather than being just flat, open boxes for the fluid & suspended spheres to flow through, the sheets will be media-etched in patterns on one side of each piece of glass (in mirror images), so that when two sheets are mated together, the resulting open pattern (cavity/channel/whatever you want to call it) will be about 3mm "deep" in total, for the fluid to be pumped through.

The intent is for colored glass spheres to slowly flow through these etched channels, collecting here and there, and periodically being swept down (or up, depending upon what the programmed flow is supposed to produce -- see #5). When viewed from one of the faces, with the light coming from the opposite side, the clear glass will have swirling patterns of colored light.

2. I have not yet chosen a fluid; that is something I will need some input from those more knowledgeable than I. I have kind of been thinking of something like a low-viscosity mineral oil, since it would (in theory) lubricate the pumps, exclude water, be less likely to leak, and would cause the spheres to drift more slowly through it.

3. I'm arbitrarily saying "1mm" diameter for the spheres, but they may be larger or smaller, depending upon what I can source. I want them to be small enough that from the expected viewing distance, they will more or less lose their individual forms and just add color to the suspension. But I would presume they need to be solid rather than hollow, so their density will be whatever glass' density is.

4. For concentration -- I'm presuming I'll have to work that out by trial & error, but I'm guessing something on the order of 15-20% of the solution, by volume. I want the effect to be impressionistic, not saturated.

5. As for the direction of flow: I am planning to have half of them flowing downward, with light blue glass spheres, to give a slow-motion rain effect (with the etched channels being slightly diagonal from vertical, and in a sort of random spray pattern), and the other half of the windows having the suspension flowing upwards, with green spheres, to give a plant growth effect (with more sort of fractal branching). If that doesn't make sense, I can sketch something to give a clearer impression.

6. I hope that my earlier explanation covers this question.

Again, thank you so much for (obviously) giving this some thought!

 

[RCH]

1. Use a centrifugal pump with sufficient clearance to pass the spheres.

2. Use a continuous flow, three piston pump, with head clearance and soft rubber flap valves.

3. Remove the spheres with a cylindrical sieve or filter. Pump the liquid, then reintroduce it to flush the spheres forward.

4. Use a jet injector pump.

5. How do they pump the fluid when "abrasive flow machining"?
https://en.wikipedia.org/wiki/Abrasive_flow_machining

Some very interesting ideas, there. Thanks!

Soft rubber flap valves seem to me to be a very good idea for preventing backflow, regardless of what kind of pump I decide to go with. Thank you for that!

I had toyed with the idea of separating the glass spheres from the fluid for more precise metering at the tops (or bottoms) of their journeys through the windows (see my response to jrmichler for context), but I had set that aside in the interest of not making it more complex than necessary. I'm still going to keep it as a possibility if I have trouble producing the desired "random" effect (I know -- using metering to produce a "random" effect makes no sense).

I'm unfamiliar with jet injector pumps. I'm only passingly familiar with water-jet cutting, which (as I understand it) uses a series of cylinder and piston pumps running out of phase, with water from the output of each being fed into a sort of plenum that maintains even and very high pressure. Although I may have that wrong.

 

[RCH]

Welcome, @RCH !

I would by-pass the glass spheres from the fluid around the point of mechanical pumping.

The most destructive effects will be the friction and impact interaction among the glass surfaces of the spheres.

Thanks for the welcome, Lnewqban!

I'm not averse to separating the glass spheres from the fluid (and I presume that could be fairly easily accomplished with sieves), but I can't figure out a way to move the spheres to the opposite end of the windows to make their journeys through them again (see my explanation to jrmichler). The only that occur to me are things like conveyor belts/buckets, which would likely add considerable complication. Any suggestions?

 

[Lnewqban]

... I can't figure out a way to move the spheres to the opposite end of the windows to make their journeys through them again (see my explanation to jrmichler).

That I don't understand, sorry.
You will have two closed circuits of channels with two pumps, one for the blue balls and one for the green ones.

 

[RCH]

IDK, but I would look into mimicking a heart. i.e. a diaphragm pump with leaf valves. Hearts have to pump blood without breaking cell walls.

Well! Inspired by your suggestion, I just spent a few hours falling down the rabbit hole of the technology behind artificial hearts. I had no idea that so many different kinds of pumps have been tried, including an Archimedes' screw (still searching for a photo or diagram of that one)!
Re. avoiding breaking cell walls (or glass spheres): I wonder if molding my own impellers out of say, silicone or TPE over a stiffer (perhaps 3-D printed) core could reduce friction against the glass spheres to an acceptable level...
Thanks!

 

[RCH]

That I don't understand, sorry.
You will have two closed circuits of channels with two pumps, one for the blue balls and one for the green ones.

I'm sorry if I'm not explaining things clearly.
Both colors won't be used in the same plates of glass, but rather, in different ones.
Just forget about the colors; they won't be a factor.

 

[Baluncore]

4. Use a jet injector pump.

I'm unfamiliar with jet injector pumps.

Google images of "jet injector pump".
They operate by the Bernoulli effect with no moving parts.
Then imagine one as the pump that drives my suggestion 3.

3. Remove the spheres with a cylindrical sieve or filter. Pump the liquid, then reintroduce it to flush the spheres forward.

Part of the fluid is extracted, pumped, then pushed out of a jet to accelerate the fluid remaining in the cylindrical sieve, along with the glass beads.

I'm only passingly familiar with water-jet cutting, which (as I understand it) uses a series of cylinder and piston pumps running out of phase, with water from the output of each being fed into a sort of plenum that maintains even and very high pressure.

That is a triplex type pump with steady flow.

2. Use a continuous flow, three piston pump, with head clearance and soft rubber flap valves.

Three pistons operating at 120° phase, will generate a steady flow rate without huge fluctuations in pressure. They are used for pressure washing and water jet cutting.
https://en.wikipedia.org/wiki/Pump#Triplex-style_plunger_pumps

 

[Rive]

Conveyor belts can be modified to transport liquids, even upward

(Ok, that steel mill transport belt as an example is a bit exaggerated, but anyway )

 

[jrmichler]

Try searching flexible impeller pump. This image of that pump is from the Wikipedia article:

Following some of the links leads to this quote: Solid Handling capability – solids can be handled with ease, but also without damage. Seeds, grapes and fruit in suspension can be handled gently, retaining solid integrity. It is favoured in the brewing industry as not only does it pump gently, but it can handle solids. Solids are also kept in suspension which can be important with textured products, such as; yogurts, drinks, or sauces. The quote is from this site: https://www.northridgepumps.com/article-141_flexible-impeller-pump-guide. If this type of pump can pump grapes, it certainly can pump glass spheres. Since it is a positive displacement pump, it can run as slow as needed to get the effect you want.

 

[Flyboy]

I'm unfamiliar with jet injector pumps.

Google images of "jet injector pump".
They operate by the Bernoulli effect with no moving parts.
Then imagine one as the pump that drives my suggestion 3.
Part of the fluid is extracted, pumped, then pushed out of a jet to accelerate the fluid remaining in the cylindrical sieve, along with the glass beads.

Jet pumps are widely used in aviation for high volume, low pressure applications, especially for fuel pumps. Usually you have a gear or vane pump providing high pressure, low volume fuel flow from the pump to the tank where it's injected into a return pipe and it drags fuel with it.

As Baluncore suggested, it should work well for your application. I'd say you use a mechanical screen to separate out the spheres from the working fluid at the intake, put it through a suitable pump, then I'd suggest at least a filter screen of some sort to catch any large particulate to keep the FOD to a minimum. Screen filters are highly reusable and reliable, and generally just need a cleaning every so often. (Doing it after the pump allows it to work well and keeps any debris from a failing pump to be caught before making into the rest of the system.) The outlet from the filter then goes to your jet nozzle and pumps the fluid.

Given the proposed application, I'd say a couple of jets along the perimeter of the tube is the best option. Think of how the jets in a lazy river at a swim park work. They're flush with the wall and you can't snag or bump into them, but they still move large amounts of water with small input.

 

[DaveC426913]

Why do the spheres have to be glass of all things?

Can't you choose a more durable, resilient yet still colourful substance such as plastic or acrylic?

 

[Flyboy]

Why do the spheres have to be glass of all things?

Can't you choose a more durable, resilient yet still colourful substance such as plastic or acrylic?

I think the glass will hold its color for longer than the plastic would in a window application like he's intending to do.

That said... I bet you could get something of the right color and density with plastic to be mildly buoyant in mineral oil so getting the upwards flowing ones to be easier. I was thinking about that and was scratching my head a little about the pressures and flow rates that might be involved to go uphill through small tubes like he's describing, if I'm understanding the project right.

 

[Baluncore]

I was thinking of silicone oil as the fluid, which is thick and clear, so will suspend glass beads.
A water trap like a diesel sedimenter is needed at the bottom of the system after the pump. That is usually integrated with a filter.

 

[DaveC426913]

I think the glass will hold its color for longer than the plastic would in a window application like he's intending to do.

Great, but the downside of glass is why were all here. Solve the biggest problems first?
<sup>"Some problems are so complex they require magic to solve them. That's called a poof of concept."
- me, just now</sup>

 

[Baluncore]

I think a coaxial sieve, followed immediately by the jet, will do the pumping job. Only the part of the total flow that escapes through the sieve, is then externally pumped and returned to the axial jet. The glass beads move in the main flow, along the centre passage, so remain in the fluid that is then accelerated by the pumped jet.

 

[Flyboy]

I think a coaxial sieve, followed immediately by the jet, will do the pumping job. Only the part of the total flow that escapes through the sieve, is then externally pumped and returned to the axial jet. The glass beads move in the main flow, along the centre passage, so remain in the fluid that is then accelerated by the pumped jet.

Yeah. I think at this point the biggest issue is what type of pressure pump to drive the jets, and how much pressure is needed.

... What kind of power supply do you have to work with, and what kind of reliability do you need?

 

[tech99]

The peristaltic pump is used for pumping blood without harming the corpuscles.

 

[Baluncore]

The peristaltic pump is used for pumping blood without harming the corpuscles.

The big advantage of the peristaltic pump for blood, is the ability to quickly replace the previously contaminated tube, with a sterile tube for the next patient.

Glass beads are much larger than blood cells, so may damage the integrity of the tube if the tube is not changed often. The glass beads will also tend to be pressed against each other in the pinched tube. That will wear the bead surface and contaminate the fluid with a ground-glass flour.

Centrifugal pumps that move water with suspended sand, such as in a suction dredge, are worn out quickly. They must be coated internally with wear resistant materials. If a fast centrifugal pump was used to move glass beads in a fluid, both the pump and the beads would be scoured, resulting in contamination of the fluid, replacement of the pump, and the beads.