Introduction

Selected passive and active augmentation techniques have been shown to be effective for pool boiling and flow boiling/evaporation. Most techniques apply to nucleate boiling; however, some techniques are applicable to transition and film boiling.

It should be noted that phase-change heat transfer coefficients are relatively high. The main thermal resistance in a two-fluid heat exchanger often lies on the non-phase-change side. Fouling of either side can, of course, represent the dominant thermal resistance. For this reason, the emphasis is often on augmentation of single-phase flow. On the other hand, the overall thermal resistance may then be reduced to the point whereas significant improvement in the overall performance can be achieved by augmenting the two-phase flow. Two-phase augmentation would also be important in double phase-change (boiling/condensing) heat exchangers.

As discussed elsewhere, surface material and finish have a strong effect on nucleate and transition pool boiling. However, reliable control of nucleation on plain surfaces is not easily accomplished. Accordingly, since the earliest days of boiling research, there have been attempts to relocate the boiling curve through use of relatively gross modification of the surface. For many years, this was accomplished simply by area increase in the form of low helical fins. The subsequent tendency was to structure surfaces to improve the nucleate boiling characteristics by a fundamental change in the boiling process [Webb (1981)]. Many of these advanced surfaces are being used in commercial shell-and-tube boilers. A book on the subject was prepared by Thome (1990). The application to high-heat-flux systems is discussed by Bergles (1992).

Pool Boiling

Numerous types of surface treatment and structuring have been utilized to intentionally reduce the wall-minus-saturation temperature difference ΔTs. For water, the placement of small nonwetting spots (teflon or epoxy), either splattered on the surface or placed in pits, reduces ΔTs at constant q by a factor of 3-4. This comparison, as well as others in this section, is based on heat flux evaluated for the surface area of the plain tube.

The usual surface modifications are of a rather fine scale, in keeping with the nature of the nucleate boiling process. In the present classification, these surfaces mostly qualify as treated surfaces; however, some of them could be considered rough. Several manufacturing processes have been employed: machining, forming, layering, and coating. Figure 1 indicates seven of the presently commercialized surfaces.

In Figure 1a, standard low-fin tubing is shown. Figure 1c depicts a tunnel-and-pore arrangement produced by rolling, upsetting, and brushing. An alternative modification of the low fins is shown in Fig. 1d, where the rolled fins have been split and rolled to a T shape. Knurling and rolling are involved in producing the surface shown in Fig. 1g. The earliest example of a commercial structured surface, shown in Fig. 1b is the porous metallic matrix produced by sintering or brazing small particles.

Examples of commercial structured boiling surfaces.

Figure 1. Examples of commercial structured boiling surfaces.

The relative performance of three of these surfaces, tested as single tubes, is shown in Figure 2. Wall superheat reductions of up to a factor often are common with these surfaces. The advantage is seen to be not only a high nucleate boiling heat transfer coefficient, but the fact that boiling can take place at very low temperature differences.

Pool boiling from smooth and structured surfaces on the same apparatus [Yilmaz et al. (1980)]. (a) Sketch of cross-sections of the three enhanced heat transfer surfaces tested. (b) Boiling curves for the enhanced tubes and smooth tube.

Figure 2. Pool boiling from smooth and structured surfaces on the same apparatus [Yilmaz et al. (1980)]. (a) Sketch of cross-sections of the three enhanced heat transfer surfaces tested. (b) Boiling curves for the enhanced tubes and smooth tube.

In all cases, a complex liquid-vapor exchange is involved. The liquid flows at random locations around the helical aperture or through selected pores to the interior of the structure, where thin-film evaporation occurs over a large surface area. The vapor is then ejected through other paths by bubbling. The latent heat transport is complemented by agitated free convection from the exposed surfaces. Data indicate that these surfaces are effective for pure liquids and mixtures with a wide range of boiling points.

The behavior of structured surfaces is not yet understood to the point where correlations are available to guide custom production of the surfaces for a particular fluid and pressure level. For example, a model for nucleate boiling from Thermoexcel-E surfaces requires eight empirically determined constants. Such models are useful in confirming the probable physics of the boiling process, but are of no use in engineering design. Some manufacturers have accumulated sufficient experience to provide optimized surfaces for some important applications, such as flooded evaporators for refrigerant dry-expansion chillers and thermosiphon reboilers for hydrocarbon distillation columns. The structured boiling surfaces developed for refrigeration and process applications have been used as “heat sinks” for microelectronic chips [Bergles (1990)].

The behavior of tube bundles is often different with structured-surface tubes. The nucleate boiling augmentation dominates, and the convective boiling enhancement, found in plain tube bundles, does not occur.

Structured surfaces are not exempt from temperature overshoots and resultant boiling curve hysteresis. However, the superheats required for incipient boiling with the highly wetting liquids are generally lower than for plain surfaces, due to the greater probability of finding active sites over the larger surface area. The practical problem is to provide, at least on a transient basis, the necessary superheat to initiate boiling. In some cases, the stimulus can be provided by injected vapor, either naturally as with a dry-expansion chiller, or intentionally through sparging.

Structured surfaces have recently been applied to thin film evaporation. Here, in contrast to the pool experiments noted above, the liquid to be vaporized is sprayed on or dripped on heated horizontal tubes to form a thin film. Structured surfaces promote nucleate boiling in the film at modest temperature differences.

Little attention has been given to the use of surfaces with discrete roughness elements in pool boiling. Improvement is not expected due to the microscale of the nucleate boiling. On the other hand, knurled surfaces are available commercially, and they apparently improve nucleate boiling. This would be expected, however, from the area increase. More commonly, rough surfaces have been applied to horizontal tube spray film evaporators. Longitudinal ribs or grooves may promote turbulence in the film; however, there is the possibility that film drainage is impeded. Increases of 100 percent in heat transfer coefficient have been referred with knurling. This is apparently a convective enhancement due to turbulence promotion within the film, as well as a substantial increase in surface area.

Low fin tubes (Figure 1) have been the standard enhancement technique for pool boiling of refrigerants and organics. Heat transfer coefficients (based on total area) are typically greater than those for the reference plain tube.

Circumferential or helical fins with various profiles have been considered for augmentation of horizontal-tube, spray film evaporation. Area increases are typically 100 percent. Surface tension causes redistribution of the liquid so that thin films are present at the fin peaks; accordingly, heat transfer coefficients (based on projected area) can increase by more than the area increase. For example, V-shaped grooves or threads increase heat transfer coefficients up to 200 percent. It was found that heat transfer coefficients for falling film evaporation inside vertical tubes could be increased by more than a factor often with loosely attached internal fins.

Active augmentation techniques include heated surface rotation, surface wiping, surface vibration, fluid vibration, electrostatic fields, and suction at the heated surface. Although active techniques are effective in reducing ΔTs and/or increasing , the practical applications are very limited, largely because of the difficulty of reliably providing the mechanical or electrical effect. Perhaps the major contribution of the many studies in this area is the information that is available to predict changes in pool boiling when these effects occur naturally in heat transfer equipment

Compound augmentation, which involves two or more techniques applied simultaneously, has also been studied. The addition of surface roughness to the evaporator side of a rotating evaporator-condenser increased the overall coefficient by 10 percent. Electrohydrodynamic augmentation was applied to a finned tube bundle, resulting in nearly a 200 percent increase in the average boiling heat transfer coefficient of the bundle [Cheung et al. (1995)]. The working fluid was an alternate refrigerant R-134a, and the maximum EHD power consumption was 1.2 percent of the bundle heat transfer rate.

Convective Boiling/Evaporation

The situation of main interest is boiling or vaporization inside tubes. Forced flow normal or parallel to enhanced tubes has received little attention, in spite of the fact that it might be desirable to augment heat transfer on the shell side of a shell-and-tube heat exchanger. Both process and power applications benefit directly from single-tube testing. In other words, there is no difference between single and multiple channels, assuming that the flow is known and steady.

The structured surfaces described in the previous section are generally not used for intube vaporization, due to the difficulty of manufacture. One notable exception is the high flux surface in a vertical thermosiphon reboiler. The considerable increase in the low quality, nucleate boiling coefficient is desirable, but it is also important that more vapor is generated to promote circulation.

Helical repeated rib tubes have been found to increase local heat transfer coefficients for vaporization of R-12 by up to 100 percent. The dryout heat flux was increased by 200 percent.

The use of helically coiled wire inserts to increase critical heat flux was reported. Increases in critical heat flux up to 50 percent were reported, with a pressure drop increase of only 25 percent.

In an important power application, widely spaced helical ribs with a gentle twist considerably increase the dryout heat flux in a once-through boiler. The probable mechanism is stabilization of the annular liquid film through the partially confined rotation. Pseudo-film boiling of supercritical water is a similar process; hence, it is not surprising that the ribbed tubes suppress this condition. Heat transfer coefficients in the post-dryout or mist-flow film boiling region are increased with roughness.

Numerous tubes with internal fins, either integral or attached, are available for refrigerant evaporators. Original configurations were tightly packed, copper, offset strip fin inserts soldered to the copper tube or aluminum, star-shaped inserts secured by drawing the tube over the insert. Examples are shown in Figure 3. Average heat transfer coefficients (based on surface area of smooth tube of the same diameter) for typical evaporator conditions are increased by as mush as 200 percent. Although the integral fins represented in this figure are small, the trend has been toward more numerous and even smaller fins.

A cross-sectional view of a typical “micro-fin” tube is included in Figure 3. These tubes are 9.5 mm outside diameter and have 60 to 70 spiral fins ranging from 0.10 to 0.19 mm in height. The average evaporation boiling coefficient is increased 30-80 percent. The pressure drop penalties are less; that is, lower percentage increases in pressure drop are frequently observed. The outstanding thermal-hydraulic performance is related to increased surface area, increase in the film turbulence level, delay in film dryout, and alteration of the flow pattern. These tubes also have promise for micro-gravity situations, where the capillary forces associated with the small fins promote a favorable flow pattern, even at reduced levels of local acceleration.

Inner-fin tubes for refrigerant evaporators. (a) strip-fin inserts, (b) star-shaped inserts, (c) micro-fin.

Figure 3. Inner-fin tubes for refrigerant evaporators. (a) strip-fin inserts, (b) star-shaped inserts, (c) micro-fin.

Twisted-tape inserts are generally used to increase the burnout heat flux for subcooled boiling at high imposed heat fluxes (107-108 W/m2), as might be encountered in the cooling of fusion reactor components. Increases in burnout heat flux of up to 200 percent were obtained. Unfortunately, the augmentation of the twisted tape drops off rapidly as pressure is increased above 2 MPa. A further problem is promotion of the burnout condition by virtual insulation of the heated wall at the wall-tape interface.

Local heat transfer coefficients for the evaporation of R-113 are increased by as much as 90 percent with twisted-tape inserts. An increase in the dryout heat flux is generally observed, because the rotating flow centrifuges the core droplets to the tube wall.

The effect of the centrifugal force on the droplets is dramatic in the case of mist flow (or dispersed flow) film boiling. There are reductions in wall temperature for both the traditional dryout, at some point along the tube, and for dryout at the tube inlet.

Coiled-tube generators have advantages of higher heat transfer performance as well as compactness. (See also Coiled Tubes, Heat Transfer in.) Modest improvements in axially local, but circumferential average, heat transfer coefficients are obtained, especially when the coil diameter is small. The dryout heat fluxes are substantially higher than the straight-tube values at interrupted at the leading edge of a strip, and the film is thinner when it reforms past the training edge.

Rough surfaces augment condensation primarily by introducing turbulence in the film. The average heat transfer coefficients for steam condensing on vertical tubes can be nearly doubled by knurling the surface. Condensing coefficients for R-11 on horizontal tubes have been found to be four to five times the smooth-tube values. Part of the improvement, of course, is due to the increase in surface area.

Surface extensions are employed for augmentation of condensation. The integral low fin tubing (Figure 1), used for kettle boilers, is also used for horizontal tube condensers. With proper spacing of the fins to provide adequate condensate drainage, the average coefficients can be several times those of a plain tube with the same base diameter. These fins are normally used with refrigerants and other organic fluids that have low condensing coefficients, but which drain effectively because of low surface tension.

The fin profile can be altered according to mathematical analysis to take full advantage of the Gregorig effect, whereby condensation occurs mainly at the tops of convex ridges. Surface-tension forces then pull the condensate into concave grooves where it runs off. The average heat transfer coefficient is greater than that for an axially uniform film thickness. The initial application was for condensation of steam on vertical tubes used for reboilers and in desalination. The analysis of Mori et al. (1981) is typical of those that outlet qualifies above 0.2. Post-dryout heat transfer coefficients are also higher with helical coils.

Vapor Space Condensation

The horizontal and vertical condensers found in both the power and process industries involve condensation on the outside of tubes. Of course, due to the high vapor velocities normally encountered, the condensation process in an actual heat exchanger is different from the classical laboratory experiment using a single tube in a large vapor space.

As discussed elsewhere, condensation can be either filmwise or dropwise. In a sense, dropwise condensation is augmentation of the normally occurring film condensation by surface treatment. The only real application is for steam condensers because nonwetting coatings are not available for most other working fluids. Even after much study, little progress has been made in developing permanent hydrophobic coatings for practical steam condensers. The augmentation of dropwise condensation is pointless, because the heat transfer coefficients are already so high.

It has been found that average coefficients for film condensation of steam on horizontal tubes can be improved up to 20 percent by strategically placing strips of teflon or other nonwetting material around the tube circumference. The condensate film is interrupted at the leading edge of a strip, and the film is thinner when it reforms past the trailing edge.

Rough surfaces augment condensation primarily by introducing turbulence in the film. The average heat transfer coefficients for steam condensing on vertical tubes can be nearly doubled by knurling the surface. Condensing coefficients for R-11 on horizontal tubes have been found to be four to five times the smooth-tube values. Part of the improvement, of course, is due to the increase in surface area.

Surface extensions are widely employed for augmentation of condensation. The integral low fin tubing (Figure 1), used for kettle boilers, is also used for horizontal tube condensers. With proper spacing of the fins to provide adequate condensate drainage, the average coefficients can be several times those of a plain tube with the same base diameter. These fins are normally used with refrigerants and other organic fluids that have low condensing coefficients, but which drain effectively because of low surface tension.

The fin profile can be altered according to mathematical analysis to take full advantage of the Gregorig effect, whereby condensation occurs mainly at the tops of convex ridges. Surface-tension forces then pull the condensate into concave grooves where it runs off. The average heat transfer coefficient is greater than that for an axially uniform film thickness. The initial application was for condensation of steam on vertical tubes used for reboilers and in desalination. The analysis of Mori et al. (1981) is typical of those that can be performed to obtain the optimum geometry. According to their numerical solutions, the optimum geometry is characterized by a sharp fin tip, gradually changing curvature of the fin surface from tip to root, wide grooves between fins to collect condensate, and periodic condensate strippers. Figure 4schematically presents the configuration. Optimum fin pitches and stripper spacing were calculated for water and R-113.

IRecommended flute profile and schematic of condensate strippers, according to Mori et al. (1981).

Figure 4. IRecommended flute profile and schematic of condensate strippers, according to Mori et al. (1981).

Experience has shown that vertical tube condenser performance can be improved by a factor of five to ten by a variety of axial flutes, not just those that have the mathematically correct shape. For example, the effectiveness of simply tying small, vertically oriented wires around the circumference has been demonstrated.

The same concepts for essentially two-dimensional-shaped surfaces can be applied to horizontal tubes; however, the improvements are usually less than a factor of five. This is important in steam surface condenser technology.

Recent interest has centered on three-dimensional surfaces for horizontal-tube condensers. A rolling and upsetting process is used to produce the surface shown in Figure 5. (In fact, the Thermoexcel-C condensing surface is the Thermoexcel-E boiling surface without the final high-speed brushing operation). It is seen that the coefficients for R-113 are as mush as seven times the smooth-tube values. The considerable improvement relative to low fins or other two-dimensional profiles is apparently due to multidimensional drainage at the fin tips. Other three-dimensional strips include circular pin fins, square pins, and small metal particles that are bonded randomly to the surface.

Get Adobe Flash player

Figure 5. Performance of enhanced tubes compared to a smooth tube for condensation of R-113.

Convective Condensation

This final section on augmentation of the various modes of heat transfer focuses in in-tube condensation. The applications include horizontal kettle-type reboilers, moisture separator reheaters for nuclear power plants, and air-conditioner condensers. In the latter application, the tubes must also perform well during evaporation, if a heat pump system is involved.

Internally grooved and knurled tubes have been studied, and average heat transfer coefficients were increased for several of the configurations; however, it appears that further improvements can be realized by optimizing the geometry.

It has been found that average coefficients for complete condensation are increased 80 percent above smooth-tube values when helical repeated ribs, were used. For an extreme case of roughness, deep spirally fluted tubes were found to have coefficients 50 percent above those of smooth tubes of the same maximum inside diameter.

Random roughness has been applied by means of metallic particles with about 50 percent area density. With R-12, the condensing coefficient was increased 300 percent for average qualities greater than 0.6, and 140 percent for lower qualities.

Conventional inner-fin tubes have been used for condensation of steam and other fluids. The fins are relatively long but few in number, and may be straight or spiraled. Up to 150 percent increases in average condensing coefficients were observed for complete condensation. Design correlations for both heat transfer and pressure drop were developed. With similar tubes with R-113, coefficient increases up to 120 percent were observed.

The micro-fin tubes mentioned earlier have also been applied successfully to in-tube condensing. As in the case of evaporation, the substantial heat transfer improvement is achieved at the expense of a lesser percentage increase in pressure drop. By testing a wide variety of tubes, it has been possible to suggest some guidelines for the geometry, e.g., more fins, longer fins, and sharper tips; however, general correlations are not yet available. Fortunately for heat-pump operation, the tube that performs best for evaporation also performs best for condensation.

Data have been reported for condensation of R-113 in tubes with Kenics static mixer inserts. The average heat transfer coefficients were increased considerably; however, the increases in pressure drop were large. Surface renewal/penetration theory (with one empirical constant) is in good agreement with the experimental data.

Twisted-tape inserts result in rather modest increases in heat transfer coefficients for complete condensation of either steam or R-113. The pressure drop increases are large due to the large wetted surface. Coiled tubular condensers provide a modest improvement in average heat transfer coefficient.

Concluding Remarks

These two entries give some indication as to why heat transfer augmentation is one of the fastest growing areas of heat transfer. Many techniques are available for improvement of the various modes of heat transfer. Fundamental understanding of the transport mechanism is growing, but more importantly, design correlations are being established. Heat transfer augmentation has indeed become a second-generation heat transfer technology that is becoming widely used in industrial heat exchangers, particularly those that involve boiling. New journals, e.g., Enhanced Heat Transfer and International Journal of Heating, Ventilating, Air-Conditioning and Refrigerating Research feature this technology.

REFERENCES

Bergles, A. E. (1988) Heat Transfer Augmentation, Two-Phase Heat Exchangers, Thermal-Hydraulic Fundamentals and Design, (Ed. S. Kakac, A. E. Bergles, and E. O. Fernandes) Kluwer, Dordrecht, The Netherlands, pp. 343-373.

Bergles, A. E., Ed. (1990) Heat Transfer in Electronic and Microelectronic Equipment , Hemisphere, New York, NY.

Bergles, A. E. (1992) Enhanced Heat Transfer Techniques for High-Heat-Flux Boiling, High Heat Flux Engineering (Ed. A. M. Khounsary), Vol. 1739, SPIE, Bellingham, WA, pp. 2-16.

Cheung, K. H., Ohadi, M. M., and Dessiatoun, S. (1995) Compound Enhancement of Boiling Heat Transfer of R-134a in a Tube Bundle, to be published in ASHRAE Transactions, Part 1.

Pate, M. B., Ayub, Z. H., and Kohler, J. (1990) “Heat Exchangers for the Air-Conditioning and Refrigeration Industry”: State-of-the-Art Design and Technology, Compact Heat Exchangers, (Ed. R. K. Shah, A. D. Kraus, and D. Metzger) Hemisphere, New York, NY, pp. 567-590.

Thome, J. R. (1990) Enhanced Boiling Heat Transfer, Hemisphere, New York.

Webb, R. L. (1981) The Evolution of Enhanced Surface Geometries for Nucleate Boiling, Heat Transfer Engineering, Vol. 2, (Nos. 3-4), pp. 46-69.

Yilmaz, S., Hwalck, J. J., and Westwater, J. W. (1980) Pool Boiling Heat Transfer Performance for Commercial Enhanced Tube Surfaces, ASME Paper 90-HT-41.

Les références

  1. Bergles, A. E. (1988) Heat Transfer Augmentation, Two-Phase Heat Exchangers, Thermal-Hydraulic Fundamentals and Design, (Ed. S. Kakac, A. E. Bergles, and E. O. Fernandes) Kluwer, Dordrecht, The Netherlands, pp. 343-373.
  2. Bergles, A. E., Ed. (1990) Heat Transfer in Electronic and Microelectronic Equipment , Hemisphere, New York, NY.
  3. Bergles, A. E. (1992) Enhanced Heat Transfer Techniques for High-Heat-Flux Boiling, High Heat Flux Engineering (Ed. A. M. Khounsary), Vol. 1739, SPIE, Bellingham, WA, pp. 2-16.
  4. Cheung, K. H., Ohadi, M. M., and Dessiatoun, S. (1995) Compound Enhancement of Boiling Heat Transfer of R-134a in a Tube Bundle, to be published in ASHRAE Transactions, Part 1.
  5. Pate, M. B., Ayub, Z. H., and Kohler, J. (1990) “Heat Exchangers for the Air-Conditioning and Refrigeration Industry”: State-of-the-Art Design and Technology, Compact Heat Exchangers, (Ed. R. K. Shah, A. D. Kraus, and D. Metzger) Hemisphere, New York, NY, pp. 567-590.
  6. Thome, J. R. (1990) Enhanced Boiling Heat Transfer, Hemisphere, New York.
  7. Webb, R. L. (1981) The Evolution of Enhanced Surface Geometries for Nucleate Boiling, Heat Transfer Engineering, Vol. 2, (Nos. 3-4), pp. 46-69. DOI: 10.1080/01457638108962760
  8. Yilmaz, S., Hwalck, J. J., and Westwater, J. W. (1980) Pool Boiling Heat Transfer Performance for Commercial Enhanced Tube Surfaces, ASME Paper 90-HT-41.
Retour en haut de page © Copyright 2008-2024