• Publication Date: 10/01/2020
  • Author(s):
    Cocco, Ray
  • Organization(s):
    Particulate Solid Research Inc.
  • Article Type: Technical Articles
  • Subjects: Dust collection and dust control, Solids flow

Ray Cocco, president | Particulate Solid Research Inc.

The first patent of a cyclone was issued in 1885 to John M. Finch of the Knickerbocker Company.1 Instead of using large, stagnant settling chambers, as was the case with the gunpowder industry, dust-laden air could now be sent to a small cylindrical housing with a tangential inlet. The tangential inlet caused a swirling flow in the cyclone, where centrifugal force instead of gravity provided the gas-solids separation. By adjusting the inlet velocity and cylindrical housing (barrel) diameter, a significant centrifugal force developed. This force could be 100 times greater than the force of gravity.

The fundamental physics describing a cyclone’s performance were laid out more than 150 years ago by Isaac Newton and George Stokes. Reasons why the reverse-flow cyclone worked so well, however, weren’t totally understood until the 1930s to 1960s, when contributions from individuals such as Eugen Feifel, J. Stairmand, Walter Barth, and Edgar Muschelknautz provided that understanding. Specifically, these scientists showed that the cyclone had outer and inner vortexes that provided two-stage solids separation for dust collection.

Cyclone pros and cons

Indeed, cyclones offer many advantages in particle processing. They’re relatively compact and quiet as they have no moving parts. A cyclone requires low capital investment and, for a properly designed cyclone, relatively low maintenance costs (for most applications). Cyclones can be constructed of almost any material and lined with refractory, which makes them suitable for cost-effective, high-temperature applications. Cyclones also are forgiving and can continue to operate when damaged, even with holes, but at reduced collection efficiencies.

Cyclones aren’t without their issues and limitations, however. Cyclones can have a low collection efficiency for particles smaller than 10 microns; low-loading cyclones can be crippled by erosion in the cone region; and particle attrition in a cyclone can be significant.

Understanding particle hydrodynamics is crucial when designing and optimizing a cyclone system. Particle flow is more than the centrifugal force on the cyclone’s wall, and deviations in this simple concept are at the root of the cyclone’s issues and limitations. In an earlier PBE column, we discussed cyclone design and operation,2 with an emphasis on the macroscopic flows. In this column, we’ll discuss microscopic physics and how these features translate to cyclone failure and limitation.

Feeding a cyclone

The basic design of a cyclone is simple, as shown in Figure 1. To review, a tangential inlet feeds the gas-­solids flow into a cylindrical barrel, creating a swirling gas flow. Centrifugal force causes the solids to migrate to the barrel wall. A conical section below the barrel confines this swirling flow and directs the solids down to the dipleg, which extends from the bottom of the cyclone, to be either discharged or recycled. The cleaned gas reverses its flow (typically in the conical region) and flows upward and out through the gas outlet tube (sometimes called the vortex finder).

Parts of a reverse-flow cyclone. Particle flow in cyclones.

While the cyclone’s basic design is simple, the resulting hydrodynamics are not. As gas and particles enter the cyclone through the tangential inlet, the swirling flow produces a centrifugal force, which affects the more massive particles more than the much lighter gas. As a result, particles concentrate at the barrel wall and fall into the conical region, or cone, as “ropes.” Figure 2 shows this roping for a Geldart-­classified group A powder in a 17-­inch-­barrel-­diameter cyclone.

Inducing swirling flow

For high-loading cyclones, such as primary units (first in a series of more than one), only one or two swirls or ropes of particles develop, whereas low-loading units, such as secondary cyclones, can have five to eight swirls, as seen in Figure 2. These swirls are important to the cyclone’s collection efficiency. Particles that are clustered together in ropes are less susceptible to wall friction and gas-phase turbulence. Having a good inlet design is essential in the development of these swirls. Too short of an inlet or a biasing of solids in the inlet toward the inside can reduce these swirls in number and stability, which results in lower collection efficiency.

Ropes or particle swirls in a 17-inchdiameter cyclone with Geldartclassified group A particles.

Some cyclone designs maximize this swirling flow. In fluidized beds, an asymmetric horn3,4 (nonsymmetric expansion at the cyclone inlet) is added to the inlet to compress the solids to the outer wall and, thereby, further promote solids moving to the wall and roping. Other cyclones use a helical (corkscrew-shaped) roof to preset the swirling flow.5

An unobstructed path

For high-loading cyclones, the path of the incoming particles is also a consideration. As shown in Figure 3a., particles and gas entering a cyclone tend to expand, and if the outlet tube is in the way, secondary flows are generated. Research by Muschelknautz and Muschelknautz6 showed that moving the outlet tube away from this expansion may help cyclone collection efficiency. This was successfully put into practice in several combustors in Germany.

Computational fluid dynamics (CFD) studies7 suggest that the real issue with an outlet tube interfering with an incoming cyclone gas-solids stream is a secondary eddy that develops. Figures 3a. and 3b. (generated by CFD software) show the result of this secondary flow for a concentric and eccentric outlet tube, respectively. The intrusive concentric outlet tube causes a recirculating eddy that promotes particles leaking directly into the outlet tube. The effect of this leakage can be seen in the outlet particle velocity profile (black and grey vectors on the outlet tube exits). With a concentric outlet tube, particle leakage is significant enough to skew, or change direction of, the particle flow in the outlet tube. In other words, some of the particles never make it into the swirling/roping flow in the cyclone’s barrel region, which is the heart of the cyclone’s dust collection capability. For the case shown in Figure 3, a small ¾-inch shift of the outlet tube position away from the incoming flow results in a reduction of particles leaking into the outlet tube from the inlet.

Computational fluid dynamics simulations for high solids loadings in a 17-inch-diameter cyclone with concentric and eccentric outlet tube shifted
¾-inch away from inlet

Furthermore, if the outlet tube is in the path of the particles entering the cyclone, the outlet tube will be prone to erosion — even when refractory-lined. Often, a volute cyclone, where the inlet scrolls the outside of the barrel upon entry, allowing more room for solids to enter the cyclone and enable better flow, will be used as the primary cyclone to prevent this erosion problem. Volute cyclones are more expensive, but they have better collection efficiency at high loadings and better reliability in the long term.

Vortex in a vortex

What makes reverse-flow cyclones fascinating is that the gas-solids flow does reverse in the cone region of the cyclone in a relatively small space. Large particles and clusters of particles at the wall fall into the dipleg, but the gas and smaller particles reverse their flow to exit via the outlet tube. If the gas velocity (with particles) in the outlet tube is high enough, an inner vortex develops so the cyclone contains a vortex within a vortex, which operates as a double-­effect classifier. Particles that aren’t captured in the outer vortex at the wall have another chance of being captured in the even-faster-­moving inner vortex at the cyclone’s center. Figure 4 provides an illustration of this inner vortex.

Diagram of the inner and outer vortexes that exist in reverse-flow cyclones. Particle flow in cyclones.

As with the swirls or ropes forming in the outer vortex, the stability of the inner vortex is equally critical in developing a good collection efficiency for the smaller particles in the cyclone. This stability is dependent on the outlet velocity, the cone depth, and the dipleg seal.

Effects of outlet velocity

Up to a point, the higher the outlet velocity and the more stable and longer the inner vortex, the better the particle separation. For particles smaller than 20 microns, this is crucial. For high-loading cyclones, the high concentration of particles typically limits the inner vortex to reach only into the barrel region. In contrast, low-loading cyclones can have very long inner vortexes that even penetrate into the dipleg. We’ve seen diplegs with eroded swirls, similar to rifle barrel grooves that increase bullet trajectory, due to too long of an inner vortex. At this point, the outlet velocity flow is too high, and solids entering the dipleg from the cone region can now be in contact with the inner vortex. This nullifies all the particle collection work done by the outer vortex and forces the inner vortex to do most of the particle collection.

Another issue with too long of an inner vortex and low-­loading cyclones is erosion in the cone region. Cone erosion is mostly an issue with low-loading or secondary cyclones. Primary cyclones tend to have inner vortexes that are well short of the cone region. As the inner vortex reaches the cone and possibly the dipleg region, the swirling flow is intensified and erosion can become significant. Thus, the outlet velocity generally should be less than 130 ft/s for typical applications. Higher velocities could result in a significantly reduced cyclone service life unless a vortex stabilizer has been incorporated.8 (A topic for our next “Particle Professor” column.)

Effect of cyclone length

Ideally, the inner vortex should terminate about one-third of the distance up the cyclone cone. This allows the room needed for the flow reversal. The length of the inner vortex needs to be a key parameter in the cyclone’s design. If the cyclone cone is too short, the inner vortex becomes compromised, and erosion in the cone becomes significant. If the cyclone cone is too long, the flow reversal that develops into the inner vortex becomes less stable and, thereby, less efficient in collecting small particles. Too long a cyclone cone also makes the cyclone’s capital cost higher. As a rule of thumb, a cyclone length of 3 to 4 times the barrel diameter is a good minimum length.

Dipleg design and performance

Dipleg performance trumps all other cyclone design considerations. If the dipleg isn’t long enough, collected solids will reach the inner vortex in the cone region. This backing up of solids is called flooding. Such a high concentration of solids at the base of the inner vortex will compromise the cyclone’s collection efficiency. This renders the outer vortex useless, and the high solids concentration the inner vortex is seeing will make it less stable.

Cyclone flooding also can create another issue and that’s plugging. If solids flood into the cone region from the plugged dipleg, the particles could compact to the point that bridging occurs at the dipleg entrance. Now all the particles have nowhere to go but out of the outlet tube.

Another problem with diplegs is the seal. For high-loading cyclones, the solids flowrate down into the dipleg is sufficient to keep the gas from flowing back up — even for negative-pressure cyclones (cyclones where the pressure inside the cyclone is less than outside). However, for low-loading cyclones, such as secondary units, diplegs don’t have that advantage. As a result, gas can flow up the dipleg while particles are flowing down the wall region. Yes, you can have counter-­current flow in a dipleg, and that upward gas flow is bringing already collected small particles back up to the base of the inner vortex, as shown in Figure 5. This gas leakage in the dipleg can have a significant impact on collection efficiency.

Sliced view of a CFD simulation for
a low-loading cyclone revealing
the dipleg gas leakage up the
dipleg into the cyclone

To avoid this issue, many secondary cyclones’ diplegs have termination devices such as trickle valves installed. Originally, trickle valves were added to help with startup and prevent gas from bypassing the cyclone inlets. However, such termination is also needed to reduce this countercurrent flow. Many of these termination devices have their own problems, however. Trickle valves can experience a warped flapper plate or erosion around the seal, both of which result in the countercurrent flow up the dipleg. A good and well-designed dipleg termination device operating correctly is well worth the expense.

In summary

To summarize, reverse-flow cyclones develop a vortex within a vortex, which acts as a double-effect classifier. Any particles that aren’t collected in the outer vortex may accumulate in the faster-spinning inner vortex. However, smaller hydrodynamic features in cyclones can compromise this effect. The incoming gas and solids need room to develop the outer swirling vortex. Otherwise, a small eddy flow can develop that provides a shortcut for some of the particles to leak into the outlet tube.

The inner vortex needs room (i.e., length) to develop and terminate without being in contact with the cyclone or, more specifically, the cone walls. If this arrangement is compromised, the benefits of having an outer vortex become diminished. All those collected particles are now being delivered directly into the inner vortex. A less severe problem is having a cyclone that’s too long. The inner vortex develops but may be less stable and have an effect on collection efficiency — but the inner vortex isn’t pulling particles up from the dipleg.

Diplegs need to be long enough so that the dipleg bed height stays in the dipleg. This prevents solids from flooding or backing up into the cyclone. Fortunately, we can calculate what the dipleg height needs to be. This was discussed in an earlier article.9

Another issue is that the dipleg of a low-loading cyclone can develop its own hydrodynamic feature and that’s countercurrent flow. This isn’t a problem for high-loading or primary cyclones. It’s a problem for low-loading or secondary cyclones.

However, if you make sure that: 1) the solids’ path from the inlet can enter the cyclone unobstructed; 2) the cyclone length is such that the inner vortex is stable and not in contact with the solids from the outer vortex; and 3) the dipleg is long enough with a good termination device to provide a reliable seal, you’ll have a cyclone you can brag about at the next company picnic — a socially-­distanced picnic.



  1. J.M. Finch, “Dust collector,” US Patent 325,521, Sept. 1, 1885.
  2. R. Cocco, “The ‘simplicity’ of reverse-flow cyclones,” Powder and Bulk Engineering, (2013) 1–4.
  3. E. Tenney, “FCC Cyclone Problems and How They Can Be Overcome With Current Designs.” Paper presented at Grace-Davison FCC Technology Conference, Toledo, Spain, June 3-5, 1992.
  4. E. Tenney, Cyclone Separator with Compact Inlet, US Patent 6,926,749, Aug 9, 2005.
  5. A.C. Hoffman, L.E. Stein, “Gas Cyclones and Swirl Tubes: Principles, Design, and Operation,” 2nd Ed., Springer-Verlag, Berlin, 2008, p. 264.
  6. U. Muschelknautz, E. Muschelknautz, “Separation efficiency of recirculating cyclones in circulating fluidized bed combustors,” VGB PowerTech. (1999).
  7. R. Cocco, U. Muschelknautz, B. Freireich, S.B.R. Karri, T. Knowlton, “Attributes of an eccentrically positioned vortex finder on primary cyclones,” Circulating Fluidized Bed XIII, Vancouver, May 10-14, 2021.
  8. R. Cocco, S.B.R. Karri, T. Knowlton, “Avoid Fluidization Pitfalls,” Chemical Engineering Progress. (2014) 1–6.
  9. R.A. Cocco, “Pressure’s role in fluidized-­bed performance,” Powder and Bulk Engineering. (2011) 1–4.

For further reading

Find more information on this topic in articles listed under “Solids flow” and “Dust collection and dust control” in the article archive.

Ray Cocco (773-523-7227) PhD, is president and CEO of Particulate Solid Research Inc. and holds a PhD in chemical engineering from Auburn University in Auburn, AL. He has more than 25 years of experience in particle technology, holds several patents, and has published numerous technical articles on particle technology topics.

Particulate Solid Research Inc. • Chicago, IL
773-523-7227 • psri.org

Copyright CSC Publishing Inc.

Download PDF