43

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43.1 Scaleup/Scaledown 43.2 Effect of the Circulation Time Spectrum and the Spectrum of Shear Rates on Ten Different Mixing Technologies

James Y. Oldshue
Oldshue Technologies International, Inc.

GasLiquid Dispersion GasLiquid Mass Transfer Solids Suspension and Dispersion SolidLiquid Mass Transfer LiquidLiquid Emulsions LiquidLiquid Extraction Blending Chemical Reactions Fluid Motion Heat Transfer

43.3 Computational Fluid Dynamics

The uid mixing process involves three different areas of viscosity which affect ow patterns and scaleup and two different scales within the uid itself: macro scale and micro scale. Design questions come up with regard to the performance of mixing processes in a given volume. Consideration must be given to proper impeller and tank geometry as well as to the proper speed and power for the impeller. Similar considerations arise when it is desired to scale up or scale down, and this involves another set of mixing considerations. If the uid discharge from an impeller is measured with a device that has a high frequency response, one can track the velocity of the uid as a function of time. The velocity at a given point in time can then be expressed as an average velocity ( v ) plus a uctuating component (v). Average velocities can be integrated across the discharge of the impeller, and the pumping capacity normal to an arbitrary discharge plane can be calculated. This arbitrary discharge plane is often dened by the boundaries of the impeller blade diameter and height. Because there is no casing, however, an additional 10 to 20% of ow typically can be considered as the primary ow of an impeller. The velocity gradients between the average velocities operate only on larger particles. Typically, these particles are greater than 1000 mm in size. This is not a precise denition, but it does give a feel for the magnitudes involved. This denes macro-scale mixing. In the turbulent region, these macro-scale uctuations can also arise from the nite number of impeller blades passing a nite number of bafes. These set up velocity uctuations that can also operate on the macro scale. Smaller particles primarily see only the uctuating velocity component. When the particle size is much less than 100 mm, the turbulent properties of the uid become important. This is the denition of the boundary size for micro-scale mixing.

2005 by CRC Press LLC

43-2

The Engineering Handbook, Second Edition

OUTLET VELOCITY vs TIME A310 VR 3 2 1 m/s 0 1 0 1 s 2 3 2 1 0 1 0 1 s 2 A200 VR 3 2 1 0 1 0 1 s 2 R100 VZ

FIGURE 43.1 Typical velocity as a function of time for three different impellers, all at the same total pumping capacity. (Courtesy of LIGHTNIN.)

All of the power applied by a mixer to a uid through the impeller appears as heat. The conversion of power to heat is through viscous shear and is 2542 Btu/h/hp. Viscous shear is present in turbulent ow only at the micro-scale level. As a result, the power per unit volume is a major component of the phenomenon of micro-scale mixing. At a 1 mm level, in fact, it does not matter what specic impeller design is used to apply the power. Numerous experiments show that the power per unit volume in the zone of the impeller (which is about 5% of the total tank volume) is about 100 times higher than the power per unit volume in the rest of the vessel. Based on some reasonable assumptions about the uid mechanics parameters, the rootmean-square (rms) velocity uctuation in the zone of the impeller appears to be approximately 5 to 10 times higher than in the rest of the vessel. This conclusion has been veried by experimental measurements. The ratio of the rms velocity uctuation to the average velocity in the impeller zone is about 50% for many open impellers. If the rms velocity uctuation is divided by the average velocity in the rest of the vessel, however, the ratio is on the order of 5 to 10%. This is also the ratio of rms velocity uctuation to the mean velocity in pipeline ow. In micro-scale mixing, phenomena can occur in mixing tanks that do not occur in pipeline reactors. Whether this is good or bad depends upon the process requirements. Figure 43.1 shows velocity versus time for three different impellers. The differences between the impellers are quite signicant and can be important for mixing processes. All three impeller velocities are calculated for the same impeller ow, Q, and same diameter. The A310 (Figure 43.2) draws the least power and has the lowest velocity uctuations. This gives the lowest micro-scale turbulence and shear rate. The A200 (Figure 43.3) displays increased velocity uctuations and draws more power. The R100

FIGURE 43.2 Fluidfoil impeller (A310). (Courtesy of LIGHTNIN.)

2005 by CRC Press LLC

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43-3

FIGURE 43.3 Typical axial-ow turbine (A200). (Courtesy of LIGHTNIN.)

FIGURE 43.4 Radial-ow Rushton turbine (R100). (Courtesy of LIGHTNIN.)

(Figure 43.4) draws the most power and has the highest micro-scale shear rate. The proper impeller should be used for each individual process requirement. The velocity spectra in the axial direction for the axial-ow impeller A200 are shown in Figure 43.5. A decibel correlation has been used in this gure because of its well-known applicability in mathematical modeling as well as the practicality of putting many orders of magnitude of data in a reasonably sized chart. Other spectra of importance are the power spectra (the square of the velocity) and the Reynolds stress (the product of the R and Z velocity components), which is a measure of the momentum at a point. The ultimate question is this: How do all these phenomena apply to process design in mixing vessels? No one today is specifying mixers for industrial processes based on meeting criteria of this type. This is largely because processes are so complex that it is not possible to dene the process requirements in terms of these uid mechanics parameters. If the process results could be dened in terms of these parameters, sufcient information probably exists to permit the calculation of an approximate mixer design. It is important to continue studying uid mechanics parameters in both mixing and pipeline reactors to establish what is required by different processes in fundamental terms. One of the most practical recent results of these studies has been the ability to design pilot plant experiments (and, in many cases, plant-scale experiments) that can establish the sensitivity of a process to macro-scale mixing variables (as a function of power, pumping capacity, impeller diameter, impeller tip
VZ & VR SPECTRUMS FOR 15.9 IN. A200 (PBT) 10 VZ (dB) 30 50 70 10 VR (dB) 30 50 70 0 5 10 HZ 15 20 .37 FT/SEC N = 2.0 RPS C = 16 IN. ZC = 12.8 IN. RC = 5.6 IN. RUN 23 JUL 87-4 .21 FT/SEC 8.0 HZ

FIGURE 43.5 Typical velocity spectrum as a function of uctuation frequency. (Courtesy of LIGHTNIN.)
2005 by CRC Press LLC

43-4

The Engineering Handbook, Second Edition

speeds, and macro-scale shear rates) in contrast to micro-scale mixing variables (which are relative to power per unit volume, rms velocity uctuations, and some estimation of the size of the micro-scale eddies). Another useful and interesting concept is the size of the eddies at which the power of an impeller is eventually dissipated. This concept utilizes the principles of isotropic turbulence developed by Komolgoroff. The calculations assume some reasonable approach to the degree of isotropic turbulence, and the estimates do give some idea as to how far down in the micro scale the power per unit volume can effectively reach. The equation is L = (v 3 / e)1/4

43.1 Scaleup/Scaledown
Two applications of scaleup frequently arise. One is building a model for pilot plant studies to develop an understanding of the process variables for an existing full-scale mixing installation. The other is taking a new process and studying it in the pilot plant to work out pertinent scaleup variables for a new mixing installation. Because there are thousands of specic processes each year that involve mixing, there will be at least hundreds of different situations requiring a somewhat different pilot plant approach. Unfortunately, no set of rules states how to carry out studies for any specic program, but here are a few guidelines that can help one carry out a pilot plant program: For any given process, take a qualitative look at the possible role of uid shear stresses. Try to consider pathways related to uid shear stress that may affect the process. If none exist, then this extremely complex phenomenon can be dismissed and the process design can be based on such things as uniformity, circulation time, blend time, or velocity specications. This is often the case in the blending of miscible uids and the suspension of solids. If uid shear stresses are likely to be involved in obtaining a process result, then one must qualitatively look at the scale at which the shear stresses inuence the result. If the particles, bubbles, droplets, or uid clumps are on the order of 1000 mm or larger, the variables are macroscale, and average velocity at a point is the predominant variable. When macro-scale variables are involved, every geometric design variable can affect the role of shear stresses. These variables can include power, impeller speed, impeller diameter, impeller blade shape, impeller blade width or height, thickness of the material used to make the impeller, number of blades, impeller location, bafe location, and number of impellers. Micro-scale variables are involved when the particles, droplets, bafes, or uid clumps are on the order of 100 mm or less. In this case, the critical parameters usually are power per unit volume, distribution of power per unit volume between the impeller and the rest of the tank, rms velocity uctuation, energy spectra, dissipation length, the smallest micro-scale eddy size for the particular power level, and viscosity of the uid. The overall circulating pattern, including the circulation time and the deviation of the circulation times, can never be neglected. No matter what else a mixer does, it must be able to circulate uid throughout an entire vessel appropriately. If it cannot, then that mixer is not suited for the tank being considered.

43.2 Effect of the Circulation Time Spectrum and the Spectrum of Shear Rates on Ten Different Mixing Technologies
GasLiquid Dispersion
The macro-scale rate shear change affects the bubble size distribution in tanks of various sizes. As processes are scaled up, the linear, supercial gas velocity tends to be higher in the larger tank. This is
2005 by CRC Press LLC

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43-5

the major contributor to the energy input of the gas stream. If the power per unit volume put in by the mixer remains relatively constant, then small tanks have a different ratio of mixing energy to gas expansion energy, which affects the ow pattern and a variety of other uid mechanics parameters. The large tank will tend to have a larger variation of the size distribution of bubbles than will the small tank. This phenomenon is affected by the fact that the surface tension and viscosity vary all the way from a relatively pure liquid phase to all types of situations with dissolved chemicals, either electrolytes or nonelectrolytes, and other types of surface-active agents.

GasLiquid Mass Transfer

If we are concerned only with the total volumetric mass transfer rate, then we can achieve very similar KGa values in large tanks and in small tanks. Blend time enters the picture primarily for other process steps immediately preceding or following the gas-liquid mass transfer step. Blending can play an important role in the total process, of which gasliquid mass transfer is only one components.

Solids Suspension and Dispersion

Solids suspension is not usually affected by blend time or shear rate changes in the relatively low to medium solids concentration in the range from 0 to 40% by weight. However, as solids become more concentrated, the effect of solids concentration on the power required changes the criteria from the settling velocity of the individual particles in the mixture to the apparent viscosity of the more concentrated slurry. This means that we enter an area where the blending of non-Newtonian uid regions, the shear rates, and circulation patterns play marked roles (see Figure 43.6). The suspension of a single solid particle should depend primarily on the upward velocity at a given point and also should be affected by the uniformity of this velocity prole across the entire tank crosssection. There are upward velocities in the tank, and there also must be corresponding downward velocities. In addition to the effect of the upward velocity on a settling particle, there is also the random motion of the micro-scale environment, which does not affect large particles very much but is a major factor in the concentration and uniformity of particles in the transition and micro-scale range. Using a draft tube in the tank for solids suspension introduces another, different set of variables. There are other relationships that are very much affected by scaleup in this type of process. Different scaleup problems exist depending on whether the impeller is pumping up or down within the draft tube. If the process involves the dispersion of solids in a liquid, then we may either be concerned with breaking up agglomerates or possibly physically breaking or shattering particles that have a low cohesive

H.P.

20

40

60

WEIGHT PERCENT SOLIDS

FIGURE 43.6 Effect of percent solids by weight and power required for uniformity and uid motion. (Courtesy of LIGHTNIN.)

2005 by CRC Press LLC

43-6

The Engineering Handbook, Second Edition

force between their components. Normally, we do not think of breaking ionic bonds with the shear rates available in mixing machinery. If we know the shear stress required to break up a particle, we can determine the shear rate required of the machinery by various viscosities with the equation Shear stress = Viscosity Shear rate The shear rate available from various types of mixing and dispersion devices is known approximately, as is the range of viscosities in which they can operate. This makes the selection of the mixing equipment subject to calculation of the shear stress required for the viscosity to be used. In the equation above, it is assumed that there is 100% transmission of the shear rate in the shear stress. However, with the slurry viscosity determined essentially by the properties of the slurry, at high slurry concentrations there is a slippage factor in which internal motion of particles in the uids over and around each other can reduce the effective transmission of viscosity efciencies from 100% to as low as 30%. Animal cells in biotechnology do not normally have a tough skin as fungal cells do and are very sensitive to mixing effects. Many approaches have been tried to minimize the effect of increased shear rates on scaleup, and these include encapsulating the organism in or on micro particles and conditioning cells selectively to shear rates. In addition, traditional fermentation processes have maximum shear rate requirements in which cells become progressively more and more damaged until they become motile.

SolidLiquid Mass Transfer

There is potentially a major effect of both shear rate and circulation time in these processes. The solids may be inorganic, in which case we are looking at the slip velocity of the particle and also whether we can break up agglomerates of particles, which may enhance mass transfer. When the particles become small enough, they tend to follow the ow pattern, so the slip velocity necessary to affect the mass transfer becomes less and available. This shows that from the denition of off-bottom motion to complete uniformity, the effect of mixer power is much less than from going to on-bottom motion to off-bottom suspension. The initial increase in power causes more and more solids to become in active communication with the liquid and has a much greater mass transfer rate than that occurring above the power level for off-bottom suspension, in which slip velocity between the particles of uid is the major contributor. Since there may well be chemical or biological reactions happening on or in the solid phase, depending upon the size of the process participants, it may or may not be appropriate to consider macro- or microscale effects. In the case of living organisms, their access to dissolved oxygen throughout the tank is of great concern. Large tanks in the fermentation industry often have a Z/T ratio of 2:1 to 4:1; thus, top to bottom blending can be a major factor. Some biological particles are facultative and can adapt and reestablish their metabolism at different dissolved oxygen levels. Other organisms are irreversibly destroyed by sufcient exposure to low dissolved oxygen levels.

LiquidLiquid Emulsions
Almost every shear rate parameter we have affects liquidliquid emulsion formation. Some of the effects are dependent upon whether the emulsion is both dispersing and coalescing in the tank, or whether there are sufcient stabilizers present to maintain the smallest droplet size produced for long periods of time. Blend time and the standard deviation of circulation times affect the length of time it takes for a particle to be exposed to the various levels of shear work and thus the time it takes to achieve the ultimate small particle size desired. As an aside, when a large liquid droplet is broken up by shear stress, it tends to initially elongate into a dumbbell type of shape, which determines the particle size of the two large droplets formed. Then the neck in the center of the dumbbell may explode or shatter. This would give a debris of particle sizes that can be quite different from the two major particles produced.
2005 by CRC Press LLC

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43-7

LiquidLiquid Extraction
If our main interest is in the total volumetric mass transfer between the liquids, the role of shear rate and blend time is relatively minor. However, if we are interested in the bubble size distribution and we often are because that affects the settling time of an emulsion in a multistage cocurrent or countercurrent extraction process then the change in macro and micro rates on scaleup is a major factor. Blend time and circulation time are usually not major factors on scaleup.

Blending
If the blending process occurs between two or more uids with relatively low viscosity such that the blending is not affected by uid shear rates, then the difference in blend time and circulation between small and large tanks is the only factor involved. However, if the blending involves wide disparities in the density of viscosity and surface tension between the various phases, a certain level or shear rate may be required before blending can proceed to its ultimate degree of uniformity. The role of viscosity is a major factor in going from the turbulent regime, through the transition region, into the viscous regime, and there is a change in the rate of energy dissipation discussed previously. The role of non-Newtonian viscosity is very strong since that tends to markedly change the inuence of impellers and determines the appropriate geometry. Another factor here is the relative increase in Reynolds number on scaleup. This means that we could have pilot plants, as well as the plant, running in the turbulent region. We could have the pilot plant running in the transition region and the plant in the turbulent, or the pilot plant could be in the viscous region while the plant is in the transition region. There is no apparent way to prevent this Reynolds number change upon scaleup. In reviewing the qualitative ow pattern in a pilot scale system, it should be realized that the ow pattern in the large tank will be at an apparently much lower viscosity and therefore at a much higher Reynolds number than is observed in the pilot plant. This means that the roles of tank shape, D/T ratio, bafes, and impeller locations can be based on different criteria in the plant size unit than in the pilot size unit under observation.

Chemical Reactions
Chemical reactions are inuenced by the uniformity of concentration both at the feed point and in the rest of the tank and can be markedly affected by changes in overall blend time and circulation time as well as the micro-scale environment. It is possible to keep the ratio between the power per unit volume at the impeller and that in the rest of the tank relatively similar on scaleup, but much detail needs to be considered regarding the reaction conditions, particularly where selectivity is involved. This means that reactions can take different paths depending upon chemistry and uid mechanics, and this is a major consideration in what should be examined. The method of introducing the reagent stream can be projected in several different ways depending upon the geometry of the impeller and the feed system.

Fluid Motion
Sometimes the specication is purely in terms of pumping capacity. Obviously, the change in volume and velocity relationships depends upon the size of the two- and three-dimensional area or volume involved. The impeller ow is treated in a head/ow concept, and the head required for various types of mixing systems can be calculated or estimated.

Heat Transfer
In general, the uid mechanics of the lm on the mixer side of the heat transfer surface is a function of what happens at that surface rather than the uid mechanics around the impeller zone. The impeller provides ow largely across and adjacent to the heat transfer surface, and that is the major consideration of the heat transfer result obtained. Many of the correlations are in terms of traditional dimensionless groups in heat transfer, while the impeller performance is often expressed as the impeller Reynolds number.
2005 by CRC Press LLC

43-8

The Engineering Handbook, Second Edition

43.3 Computational Fluid Dynamics

Several software programs available to model ow patterns of mixing tanks are available. They allow the prediction of ow patterns based on certain boundary conditions. The most reliable models use accurate uid mechanics data generated for the impellers in question and a reasonable number of modeling cells to give the overall tank ow pattern. These ow patterns can give velocities, streamlines, and localized kinetic energy values for the system. Their main use at the present time is in examining the effect of changes in mixing variables based on adjustments to the mixing process. These programs can model velocity, shear rates, and kinetic energy but probably cannot adapt to the chemistry of diffusion or mass transfer kinetics of actual industrial processes at the present time. Relatively uncomplicated transparent tank studies using tracer uids or particles can also give a feel for the overall ow pattern. The time and expense of calculating these ow patterns with computational uid dynamics should be considered in relation to their applicability to an actual industrial process. The future of computational uid dynamics looks very encouraging, and a reasonable amount of time and effort placed in this regard can yield immediate results as well as the potential for future process evaluation. Figure 43.7 through Figure 43.9 show some approaches. Figure 43.7 shows velocity vectors for an A310 impeller. Figure 43.8 shows contours of kinetic energy of turbulence. Figure 43.9 uses a particle trajectory approach with neutral buoyancy particles. Numerical uid mechanics can dene many of the uid mechanics parameters for an overall reactor system. Many of the models break the mixing tank up into small microcells. Suitable material and mass transfer balances between these cells throughout the reactor are then made. This can involve massive computational requirements. Programs are available that can give reasonably acceptable models of experimental data in mixing vessels. Modeling the three-dimensional aspect of a ow pattern in a mixing tank can require a large amount of computing power.

FIGURE 43.7 Typical velocity pattern for a three-dimensional model using computational uid dynamics for an axial ow impeller (A310). (Courtesy of LIGHTNIN.)

2005 by CRC Press LLC

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43-9

FIGURE 43.8 Typical contours of kinetic energy of turbulence using a three-dimensional model with computational uid dynamics for an axial ow impeller (A310). (Courtesy of LIGHTNIN.)

FIGURE 43.9 Typical particle trajectory using an axial ow impeller (A310) with a 100-micron particle using computational uid dynamics. (Courtesy of LIGHTNIN.)

2005 by CRC Press LLC

43-10

The Engineering Handbook, Second Edition

Dening Terms
Isotropic turbulence Fluid shear rate is a velocity gradient that results in shear stress, which can break up, disperse, or otherwise affect particles. Macro scale Any process governed by large particles on the order of 1000 mm or more. Micro scale Any process governed by small particles on the order of less than 100 mm.

Nomenclature
N D T Z P/V SR impeller speed impeller diameter tank diameter liquid level power per unit volume solidity ratio, obtained by dividing the projected area of the impeller blades by the area of a disk circumscribing the impeller blades power number velocity head, v 2 / 2g power L v v v KGa KLa ks L e u u length scale uid velocity uid velocity uctuation average uid velocity gasliquid mass transfer coefcient liquidliquid mass transfer coefcient liquidsolid mass transfer coefcient size of microscale eddy energy dissipation rate kinematic viscosity dynamic viscosity

NP H P

References
Levich, V. 1962. Physico-Chemical Hydrodynamics, Prentice Hall, Englewood Cliffs, NJ. Middleton, J. C. 1989. Proceedings of the Third European Conference on Mixing, BHRA, Craneld, England, pp. 1536. Neinow, A. W., Buckland, B., and Weetman, R. J. 1989. Mixing XII Research Conference, Potosi, MO. Oldshue, J. Y. 1989. Mixing 89. Chem. Eng. Prog. pp. 3342. Oldshue, J. Y., Post, T. A., and Weetman, R. J. 1988. Comparison of mass transfer characteristics of radial and axial ow impellers. Proceedings of the Sixth European Conference on Mixing, BHRA, Craneld, England.

Further Information
Harnby, N., Edwards, M. F., and Neinow, A. W., Eds. 1986. Mixing in the Process Industries, Butterworth, Stoneham, MA. Lo, T. C., Baird, M. H. I., and Hanson, C. 1983. Handbook of Solvent Extraction, John Wiley & Sons, New York. McDonough, R. J. 1992. Mixing for the Process Industries, Van Nostrand Reinhold, New York. Nagata, S. 1975. Mixing: Principles and Applications, Kodansha Ltd., Tokyo. Oldshue, J. Y. 1983. Fluid Mixing Technology, McGraw-Hill, New York. Tatterson, G. B. 1991. Fluid Mixing and Gas Dispersion in Agitated Tanks, McGraw-Hill, New York. Uhl, V. W. and Gray, J. B., Eds. 1966, 1986. Mixing, Vols. I and II, Academic Press, New York; Vol. III, Academic Press, Orlando, FL. Ulbrecht, J. J. and Paterson, G. K., Eds. 1985. Mixing of Liquids by Mechanical Agitation, Gordon and Breach Science Publishers, New York.

Proceedings
Fluid Mechanics of Mixing, ed. R. King. Kluwer Academic Publishers, Dordecht, Netherlands, 1992.

2005 by CRC Press LLC

Basic Mixing Principles for Various Types of Fluid Mixing Applications

43-11

Fluid Mixing, Vol. I. Inst. Chem. Eng. Symp., Ser. No. 64 (Bradford, England). Institute of Chemical Engineers, Rugby, England, 1984. Mixing Theory Related to Practice, AIChE, Inst. Chem. Eng. Symp. Ser. No. 10 (London). AIChE and Institute of Chemical Engineers, London, 1965. Proceedings of the First European Conf. on Mixing, Ed. N. G. Coles. BHRA Fluid Eng., Craneld, England, 1974. Proceedings of the Second European Conference on Mixing, Ed. H. S. Stephens and J. A. Clark. BHRA Fluid Eng., Craneld, England, 1977. Proceedings of the Third European Conference on Mixing, Ed. H. S. Stephens and C. A. Stapleton. BHRA Fluid Eng., Craneld, England, 1979. Proceedings of the Fourth European Conference on Mixing, Ed. H. S. Stephens and D. H. Goodes. BHRA Fluid Eng., Craneld, England, 1982. Proceedings of the Fifth European Conference on Mixing, Ed. S. Stanbury. BHRA Fluid Eng., Craneld, England, 1985. Proceedings of the Sixth European Conference on Mixing, BHRA Fluid Eng., Craneld, England, 1988. Process Mixing: Chemical and Biochemical Applications, Ed. G. B. Tatterson and R. V. Calabrese. AIChE Symp. Ser. No. 286, 1992.

2005 by CRC Press LLC