I am designing a simple centrifuge to separate soil by particle size (as well as decanting water). My goal is to be able to continuously funnel in 500 $cm^3$ of soil and 1.25 liters of water and have the components of the soil (clay, silt, sand) separated from each other and funneled into individual containers for each component. Water will be easily decanted and I know how to do so.

The problem is that I am having difficulties trying to figure out how to easily and continuously extract the soil layers without remixing them together. It is important that the machine is able to sort the soil layers while the centrifuge is still spinning. This is due to waiting for the centrifuge to stop spinning/restart spinning adding large time wastes over many kilograms of soil processed.

I was thinking about adding opening/closing dividers to attempt to keep the layers separate while funneled out, however, there are two problems with this:

  1. The ratio/amount of soil components (again, clay, silt, and sand) is unknown. Installing opening/closing dividers will have to be at fixed positions and as the soil components are unknown, dividing along these fixed positions will cause some remixing of the soil.
  2. I am hesitant to introduce moving parts on a spinning part spinning at 2,500 RPM (the inner drum of the centrifuge). This seems like an unnecessary safety risk that I would like to avoid, if possible.

Here is a diagram of a simple example of what the inner drum looks like: Simple soil centrifuge

What is supposed to be happening in the centrifuge is similar to if you grabbed some soil, mixed it with water, and let it set, but exceptionally faster. If I wait for it to settle naturally, I have to wait 4-5 days for a small 500g sample to fully settle, whereas with a continuously flowing/separating centrifuge (and additional mechanisms), I can expect to sort hundreds of kilograms within a day. This, along with it automatically sorting soil components, drastically improves my workflow.

Short explanation of expected sorting behavior:

  • Sand particles (defined as particles with diameters from 0.05 mm to 2 mm) should separate closest to the wall of the centrifuge due to being the densest particles in the soil.
  • Silt particles (defined as particles with diameters from 0.002 mm to 0.05 mm) should separate and collect in the "middle" of the soil solids due to it being the moderately dense particles in the soil.
  • Clay particles (defined as particles with diameters less than 0.002 mm) should separate closest to the water layer due to it being the least dense particles in the soil. I also expect organic matter to collect close to this layer, but I know of some ways of eliminating the organic matter.
  • Water, of course, would be the furthest away from the wall due to it being much less dense than any of the soil solids.

It is not required for the sorting to be perfect. For instance, I do not expect perfectly sized particles that are perfectly sorted. I do expect some silt to be mixed with clay, sand with silt, etc. However, I would prefer this to be minimal with a reasonable setup.

I greatly look forward to your suggestions. Additionally, if I have missed any oversights, I appreciate pointing them out as well. Thank you very much.

  • $\begingroup$ You seem to be asking for a Trommel sieve. $\endgroup$
    – Solar Mike
    May 2, 2022 at 5:33

2 Answers 2


Separating solids of different densities by centrifuge would be extremely difficult as you have identified.

I would recommend building a multi stage classifier instead. These are used in milling processes to extract the desired material size. They generally use air, but could also be built using water as the working fluid. It is based on the principle of Elutriation, and essentially uses the terminal velocity of each individual particle as the separation principle. With the streams separated you can then remove the water via gravity seperation, centrifuge, screening, filter press, your choice. Depending on the geometric shape and density of the sand, you may have to add a wet or dry screening process to separate it from denser smaller particles that had similar terminal velocities.

I recommend finding some mining engineering books to review other options and to locate the calculations for sizing your system.

  • 1
    $\begingroup$ I would suggest the standard handbook for mineral processing (which is a small subset of mining engineering at best and a separate field in its own right). amazon.com/… $\endgroup$ May 2, 2022 at 6:23

I would suggest a horizontal classifier - fluid with soil mixed in flows through, solids settle out, where solids settle out will depend on particle size, density and flow regime.

enter image description here

These are more appropriate for a continuous application than you batch process. I suggest to build a setup where water flowrate is adjustable, but constant throughout one run, and then slowly mix in your slurry upstream of the pump (peristaltic pumps handle solids reasonably well).

an alternative, probably better, would be to simply sieve your slurry (several stacked sieves). Sieves for grain sizes 0.002 mm are available.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.