Wing function technology - a new rotor technology for the Farrel Banbury mixer.

Batch mixers are widely used within the rubber and plastics industry to mix a great variety of polymeric compounds. They are classified into two distinct categories according to their geometrical configuration and operational characteristics, i.e. tangential or non-intermeshing versus intermeshing. As the names imply, the rotor wings of an intermeshing rotor design interlock with each other in the apex region or window of interaction between the two chamber halves, whereas in tangential designs the rotor wings are tangent to each other (refs. 1-4).

In tangential mixer designs, dispersive mixing occurs primarily in the high shear region between the rotor tip and the housing wall, whereas extensive mixing occurs in rolling pools in front of the rotor wings. Distributive mixing is further enhanced by the continuous splitting and recombination of material from one rotor wing to the other and from one chamber side to the other. In intermeshing mixer designs, dispersive mixing occurs between the two rotors and in the region between the rotor "nogs" and the inner housing wall, whereas extensive mixing occurs in the region between the two rotors and in rolling pools formed in front of the rotor nogs. As in tangential mixer designs, continuous splitting and recombination of material from one chamber side to the other further enhance distributive mixing (refs. 5 and 6).

Tangential mixer designs offer excellent dispersive mixing characteristics, but frequently can not match the composition and temperature uniformity observed with the intermeshing rotor mixer designs. All batch mixer manufacturers realized the need for an improved version of tangential mixer design, one that offers improved levels of extensive mixing over existing technologies while retaining the high productivity and dispersive mixing capabilities typical of tangential mixer designs. The general approach by each manufacturer was to add additional wings or rearrange the existing rotor wings (ref. 7). The basic wing geometry in all cases appeared to remain practically unchanged.

Farrel's approach to rotor design was to combine the best features offered by the intermeshing and tangential rotor designs in a single rotor geometry. To achieve these objectives, a specific process function was assigned to one or more rotor wings. Each wing geometry was subsequently optimized for that process function. Under this design philosophy, one or more of the rotor wings were designed to promote distributive mixing, (the pushing rotor wing), whereas the other(s) is/are designed to promote dispersive mixing, (the shearing rotor wing). These unique features of the new rotor design will be presented and discussed in this article.

Wing function technology (WFT) rotor design

The guiding principles in the development of the new rotor for the Banbury family of batch mixers were:

* Provide high levels of productivity while retaining the dispersive mixing characteristics commonly observed with tangential mixers.

* Improve the extensive mixing characteristics and provide enforced material flow within each processing chamber side and from one side to the other:

* enhance product uniformity within the batch;

* ensure good batch ingestion and discharge characteristics;

* provide good material temperature control;

* avoid any areas of poor material flow; and

* ensure that WFT rotors can be retrofitted to existing mixer designs.

These objectives were achieved by assigning a specific process function to one or more rotor wings and by designing the rotor wing configuration, placement and geometry according to the process function they were assigned to perform. Under these design guidelines, one or more rotor wings are designed to promote extensive mixing, whereas the other rotor wing(s) is/are designed to promote dispersive mixing. Since each wing is designed to perform a specific process function, different design considerations applied to each rotor wing, and the rotor wing geometry was accordingly optimized for the process function assigned to that rotor wing. Additionally, the wing arrangement and configuration was selected to ensure effective surface renewal exposure of mixed material elements to temperature controlled surfaces.

Figure 1 embodies the general design principles used as guidelines in establishing the basic rotor wing configuration. It shows the calculated fraction of the maximum flow rate over the rotor wing per unit length, as a function of rotor wing helix angle for a fixed rotor speed, wing length, lead angle and land width. By varying the helix angle, the fraction of material that is allowed to flow over the rotor wing, for a given rotor wing geometry and given operating conditions, can be controlled. For example, the maximum flow rate for any of the above listed design parameters corresponds to a "0 [degrees]" rotor wing helix angle, simply because the wing under these conditions is perpendicular to the velocity vector. As the rotor helix wing angle increases, the flow rate over the rotor wing decreases, approaching a zero value at 90 [degrees] helix angle. At this condition, the rotor wing is in line, parallel, with the velocity vector and thus no material flows over the rotor wing.

[FIGURE 1 OMITTED]

At a typical rotor wing helix angle of 30-33 [degrees], approximately 85% of the material accumulated in front of the rotor wing is forced through the rotor wing tip/housing clearance. This fraction is reduced to approximately 60% for a helix angle in the range of 50-60 [degrees]. The dispersive and distributive rotor wing helix angles were thus selected according to the above considerations. For the dispersive rotor wing, a low helix angle was selected to maximize material flow per unit rotor wing length over the high shear region formed between the rotor wing tip and the inner housing wall. For the distributive wing, a high wing helix angle was selected to provide increased material flow along the rotor axis. Figure 2 shows schematically the wing configuration for a four-wing WFT rotor design. For reference, the wing configuration of the Farrel ST rotor design is also shown.

[FIGURE 2 OMITTED]

In addition to the helix angles, the specific rotor wing geometries in the wing tip region were selected to further promote the above described flow patterns. Figure 3 shows a schematic view of the rotor wing geometry in the tip region and identifies the key design parameters for the rotor wing. Each of the design parameters associated with the wing geometry, i.e., wing approach and trail angles [alpha], [beta], tip clearance [delta] and land width e, were optimized via mathematical modeling of the localized flow patterns in that region of the mixer according to their process function (ref. 8). Details of the rotor wing tip regions for the two rotor wings are shown schematically in figure 4. A low approach angle ct, for the dispersive rotor wing, is expected to yield high fill levels, high localized pressures in the wedge region, thus maximizing the flow of material over the rotor tip. The high approach angle for the distributive wing is expected to yield a smaller pool size, reduced pressure generation and higher axial material flow.

[FIGURES 3-4 OMITTED]

The helical length of each rotor wing, its respective placement within the mixing chamber, and the orientation with respect to each other, as the rotor wings cross the window of interaction between the two rotors, were all selected to ensure effective material transfer from one chamber side into the other. To maximize temperature control of the processed material, temperature control cavities were placed in each rotor wing, with their placement and geometry optimized for maximum heat transfer and rotor strength.

Two-wing and four-wing rotor designs were fabricated along the above principles. The two-wing rotor design was installed in a laboratory size mixer equipped with special instrumentation and a high-speed data acquisition system. The initial testing was to verify the design principles and the basic design concept. The results from these studies are reported in this presentation. The four-wing rotor was installed in a F270 mixer for field evaluation. The results from that study will be reported elsewhere (ref. 9).

Experimental setup 

To verify the rotor design concepts presented above, two-ring rotors were designed and fabricated for a Farrel laboratory size mixer. The rotors were installed in a mixer with swing sides and a fixed door section. Additionally, two pressure transducers were installed in one of the sides to record pressure profiles and fill levels for the dispersive and distributive rotor wings. Additionally, the two rotor wings were equipped with an indexer, which allows determination of the location of the two rotor wings during each revolution.

The rotors were driven by a 40 hp variable frequency drive with a 225 rpm maximum rotor speed. The body sides and rotor body were both temperature controlled by means of a closed loop water circulating system. The mixer was also equipped with a four-inch diameter air cylinder with integral (internal) ram position indicator (RPI). Net chamber volume was calculated, based on rotor geometry, to be 1.55 liters. Experimentally, the net chamber volume was found to be of the same order. The mixer was equipped with even speed gearing.

The flow characteristics in the mixing chamber were established by measuring the pressure profiles associated with each rotor wing at different mixer fill factors and rotor speeds. The general procedure employed in these experiments was to load the mixer with a pre-determined batch weight, in the range of 1,000 g to 1,400 g, and mix it until it reached a processing temperature of 350 [degrees] F. From that time on, the rotor position and pressure profiles were recorded until termination of the batch cycle. The material used in the investigation was an EPDM with a Mooney viscosity of 76 ML. The mixer body was set at 300 [degrees] F, the rotors neutral and the ram at 40 psi air cylinder pressure.

In another set of experiments, flow patterns and mixing processes were established by tracer techniques. The mixer was first loaded with a white compound, which was masticated to a set temperature of about 200 [degrees] F. The mixer was then stopped, and a black tracer was inserted through the hopper opening in the region between the two rotors at the rotor end plate. The mixer was subsequently operated at a fixed rotor speed of 25 rpm and for mixing times corresponding to 10, 20 and 30, 40 and 60 revolutions. At the completion of each test run, the mixer was stopped and the carcass was removed as a whole piece for visual observation of the prevailing flow patterns. For comparison the same tests were also performed with standard Banbury type rotors.

Results and observations

First the flow characteristics in the mixing chamber were established by measuring the pressure profiles associated with each rotor wing at different mixer fill levels. Figure 5 shows the recorded pressure profiles for the two rotor wings at a batch weight of 1,000 gm, and a rotor speed of 10 rpm. Rotor speeds were purposely maintained at low levels to ensure quasi-isothermal conditions during data collection and to ensure that sufficient data points can be collected during each rotor revolution. Wing fill factors, which are different from mixer fill levels, are easily determined by the onset of the pressure build up in front of the rotor wing during each rotor revolution. Figures 8 to 10 show the pressure profiles for batch weights of 1,200 gm, 1,300 gm and 1,400 gm.

[FIGURE 5 OMITTED]

From figure 5, it can be easily concluded that the amount of material accumulated in front of the shearing wing is significantly higher than that of the pushing wing. It simply indicates that a larger pool of material forms in front of the dispersive rotor wing as compared to the distributive wing. Additionally, and as expected, the maximum pressure generated by the shearing wing is also significantly higher than the pressure generated by the pushing wing.

Both observations are direct verifications of the basic design principles previously outlined in this presentation. The higher pressures and higher fill factors associated with the shearing wing are both consistent with the smaller rotor helix and wing approach angles used in the design. Similarly, the low fill factors associated with the distributive rotor wing can be attributed to high helix and approach angles. For the experimental conditions of figure 5, calculated maximum pressures, based on a non-Newtonian isothermal flow in a linear wedge (ref. 8), were 34 psi and 67 psi, respectively, with average shear rates for both wings on the order of 85 1/sec. A close agreement between the experimentally measured and the calculated values is observed, in spite of the simplified assumptions used in the calculations. However, considering the low rotor speeds used in the experimentation, the isothermal assumption used in the mathematical model is quite reasonable. Figures 6 and 7 show calculated velocity profiles in the wing tip regions for the two rotor wings. The velocity profiles of figure 6 were normalized with respect to the maximum fluid velocity for each rotor wing. No significant difference in velocity profiles is observed between the two rotor wings. In both cases, and as expected, the general velocity profiles are consistent with those where flow in the tip region is combined drag and pressure flow. However, when the velocity profiles for the dispersive rotor wing were normalized using the maximum local velocity calculated for the distributive rotor wing (figure 7), the difference in the flow patterns between the two rotor wings is apparent and consistent with the basic rotor design.

[FIGURES 6-7 OMITTED]

Increasing the batch size to 1,100 gm while retaining the same rotor speed (figure 8) results in an increase in the pushing wing fill level, whereas the fill level for the shearing wing remains practically unchanged. The maximum pressure generated by the shearing wing is of the same order of magnitude as observed with the lower mixer batch size, hut the maximum pressure generation has now shifted away from the wing tip. The general pressure profile for the pushing wing remained practically unchanged, although a higher fill level and the associated higher maximum pressure are now observed. The data of figure 8 clearly indicate that the additional material introduced to the mixer, as a result of the higher mixer fill level, accumulates in front of the pushing rotor wing, thus increasing the rotor wing fill level.

[FIGURE 8 OMITTED]

Figures 9 and 10 are consistent with higher mixer fill levels. In both cases, a significant change in the fill level of the pushing rotor wing is observed with increased mixer batch size.

[FIGURES 9-10 OMITTED]

At a batch weight of 1,400 gm, both the pushing and shearing wings exhibit similar flow characteristics, and the maximum pressure generated by either wing is of the same order. In both cases, the location of maximum pressure is shifted away from the narrow tip clearance region.

From the general pressure profiles at low batch sizes, one can conclude that the amount of material forced over the dispersive rotor wing per unit length is higher than that of the pushing wing as a result of the higher wing fill levels and the associated higher pressures. At high mixer fill levels, at approximately 80%, the difference in the fill levels between the two rotor wings diminishes. At these conditions, significant material flow is expected to prevail over both rotor wings. Although the localized flows are different for the two rotor wings, wing length can compensate for the lower flow rates per the unit wing length observed with the distributive rotor wing. Of the four batch size cases considered in this investigation, mixer fill levels on the order of 80%, i.e., batch size on the order of 1,300 gm, are therefore, the preferred mode of operation of the new rotors. At higher mixer fill levels (figure 10), the initial starting pressure for the dispersive rotor wing is somewhat higher than that of the distributive wing. This clearly indicates that the mixer in the transition region between the two wings is fully pressurized and may hinder material flow and exchange from one wing to the other.

In another set of experiments, tracer techniques were used to identify the flow patterns and establish regions of possible poor flow within the mixing chamber. Figure 11 shows the mixed materials, top and bottom views, at different total number of revolutions. Areas of poor flow could possibly be regions close to the ends of the rotor, the bottom of the ram and the top of the drop door. In these experiments, a tracer, a small piece of black compound, was placed between the two rotors on the waterside rotor end.

[FIGURE 11 OMITTED]

Close examination of the carcasses removed from the mixer clearly indicates that after a total of 10 revolutions, 24 seconds of mixing time, the tracer is already spread along the wing length, but the major concentration is still within one of the chamber halves. However, after an additional 10 revolutions, the tracer is now well distributed between the two chamber halves. Black striations are still evident, and it took another 10 revolutions to a total of 30 to generate a completely uniformly colored batch. No distinguishable differences in the mixed samples were observed between 30 and 40 revolutions of mixing. In both cases, the tracer is well distributed within the batch and no areas of stagnation are evident anywhere.

To demonstrate the efficiency of mixing of the new rotor design, test specimens were also generated using a BR1600 laboratory mixer equipped with standard two-wing rotors. The results are shown in figure 12. The difference in the mixing characteristics between the two samples is significant. For a total of 30 revolutions, the sample generated by the new rotor is completely mixed, whereas that generated by the standard Banbury rotor design exhibits unmixed regions. Clearly, longer mixing times are required for the mixer equipped with the standard rotor design to reach the mixing levels achieved with the new WFT rotors.

[FIGURE 12 OMITTED]

Conclusions

The process functions of the distributive and dispersive rotor wings of the new Wing Function Technology rotors for the Banbury mixer were verified by means of experimental pressure profiles for each rotor wing. These profiles were established at different mixer fill levels and showed that the fill level and the pressures generated by the dispersive rotor wing are practically unaffected by the mixer fill level. For the distributive wing, the wing fill level and the maximum pressure generated increased with mixer batch size. Tracer profiles confirmed the effective utilization of the entire mixing chamber, with no areas of material of stagnation and, compared to standard type Banbury rotors, reduced mixing times for a comparable compound quality.

Finally, based on the recorded pressure profiles for the two rotor wings, a mixer batch size which provides a mixer fill level on the order of 80% provides best overall mixer performance. At this fill level, both rotor wings are effectively used in performing their designated process functions.

References

(1.) Schmid, H.M., Rubber World, February 1984, p. 33.

(2.) Melloto, M.A., Rubber World, February 1989, p. 34.

(3.) White, U., Rubber Processing: Technology - Materials - Principles, Carl Hanser: Munich (1995).

(4.) Valsamis, L.N., Canedo, E.L., Donoian, G.S., The Mixing of Rubber, R.F. Grossman, Ed., Chapman & Hall (1997).

(5.) Inoue, K., "Internal mixers," in Mixing and Compounding of Polymers: Theory and Practice, Ica Manas-Zloczower and Z. Tadmor, Eds., Carl Hanser Munich (1993).

(6.) Ghafouri, S.N., Rubber World, March 2000, p.31.

(7.) Wood, P.R., Rubber Technology International 1997, p. 172.

(8.) Canedo, E.L., Valsamis, L.N, "Mixing in the Farrel continuous mixer" in Mixing and Compounding of Polymers: Theory and Practice, Ica Manas-Zloczower and Z. Tadmor, Eds., Carl Hanser: Munich (1993).

(9.) Borzenski, F.J., Valsamis, L.N., "Optimizing mixing in the Farrel Banbury mixer with wing function technology (WFT) rotors," presented at ACS Rubber Division Meeting, Cleveland, OH, October 2001.