An ACHE is a device for rejecting heat from a fluid directly to ambient air. This is in contrast to rejecting heat to water and then rejecting it to air, as with a shell and tube heat exchanger and a wet cooling tower system.
The obvious advantage of an ACHE is that it does not require water, which means that plants requiring large cooling capacities need not be located near a supply of cooling water.
An ACHE may be as small as an automobile radiator or large enough to reject the heat of turbine exhaust steam condensation from a 1,200 MW power plant which would require 42 modules, each 90 feet wide by 180 feet long and served by two 60-foot diameter fans driven by 500-horsepower motors.
The prime tube is usually round and of any metal suitable for the process, due consideration being given to corrosion, pressure, and temperature limitations. Fins are helical or plate type, and are usually of aluminum for reasons of good thermal conductivity and economy of fabrication. Steel fins are used for very high temperature applications.
Fins are attached to the tubes in a number of ways:
Sometimes serrations are
cut in the fins. This causes an interruption of the air boundary layer, which increases
turbulence which in turn increases the airside heat transfer coefficient with a modest
increase in the air-side pressure drop and the fan horsepower.
The choice of fin types is critical. This choice is influenced by cost, operating temperatures, and the atmospheric conditions. Each type has different heat transfer and pressure drop characteristics. The extruded finned tube affords the best protection of the liner tube from atmospheric corrosion as well as consistent heat transfer from the initial installation and throughout the life of the cooler. This is the preferred tube for operating temperatures up to 600°F. The embedded fin also affords a continued predictable heat transfer and should be used for all coolers operating above 600°F and below 750°F. The wrap-on footed fin tube can be used below 250°F; however, the bond between the fin and the tube will loosen in time and the heat transfer is not predictable with certainty over the life of the cooler. It is advisable to derate the effectiveness of the wrap-on tube to allow for this probability.
There are many configurations of finned tubes, but manufacturers find it economically practical to limit production to a few standard designs. Tubes are manufactured in lengths from 6 to 60 feet and in diameters ranging from 5/8 inch to 6 inches, the most common being I inch. Fins are commonly helical, 7 to 11 fins per inch, 5/16 to I inch high, and 0.010 to 0.035 inch thick. The ratio of extended to prime surface varies from 7:1 to 25:1. Bundles are rectangular and typically consist of 2 to 10 rows of finned tubes arranged on triangular pitch. Bundles may be stacked in depths of up to 30 rows to suit unusual services. The tube pitch is usually between 2 and 2.5 tube diameters. Net free area for air flow through bundles is about 50% of face area. Tubes are rolled or welded into the tube sheets of a pair of box headers.
The box header consists of tube sheet, top, bottom, and end plates, and a cover plate that may be welded or bolted on. If the cover is welded on, holes must be drilled and threaded opposite each tube for maintenance of the tubes. A plug is screwed into each hole, and the cover is called the plug sheet. Bolted removable cover plates are used for improved access to headers in severe fouling services. Partitions are welded in the headers to establish the tube-side flow pattern, which generates suitable velocities in as near countercurrent flow as possible for maximum mean temperature difference. Partitions and stiffeners (partitions with flow openings) also act as structural stays. Horizontally split headers may be required to accommodate differential tube expansion in services having high fluid temperature differences per pass. The figure below illustrates common head types.
Bundles are usually arranged horizontally with the air entering below and discharging vertically. Occasionally bundles are arranged vertically with the air passing across horizontally, such as in a natural draft tower where the bundles are arranged vertically at the periphery of the tower base. Bundles can also be arranged in an "A" or "V" configuration, the principal advantage of this being a saving of plot area. The disadvantages are higher horsepower requirements for a given capacity and decreased performance when winds on exposed sides inhibit air flow.
Within practical limits, the longer the tubes and the greater the number of rows, the less the heat transfer surface costs per square foot. One or more bundles of the same or differing service may be combined in one unit (bay) with one set of fans. All bundles combined in a single unit will have the same air-side static pressure loss. Consequently, combined bundles having different numbers of rows must be designed for different face velocities.
Axial Flow Fans
Even distribution of the air across the tube bundle is critical for predictable, uniform heat transfer. This is achieved by adequate fan coverage and static pressure loss across the bundle. Good practice is to keep the fan projected area to a minimum of 40% of the projected face area of the tube bundle and the bundle static pressure loss at least 3.5 times the velocity pressure loss through the fan ring. For a two fan unit this is generally assured if the ratio of tube length to bundle width is in the range of 3 to 3.5 and the number of tube rows is held to 4 rows minimum with the net free area for air flow at about 50% of the face area of the bundle.
Fans can vary in size from 3 to 60 feet in diameter, and can have from 2 to 20 blades. Blades can be made of wood, steel, aluminum, or fiberglass-reinforced plastic, and can be solid or hollow. Hollow plastic blades are by far the most popular. Blades can have straight sides or be contoured. The more efficient type has a wide chord near the center and tapers to a narrow chord at the tip, with a slight twist. The twist and taper compensate for the slower velocity of the blade nearer the center to produce a uniform, efficient air velocity profile.
Fans may have fixed or adjustable pitch blades. Except for small diameters (less than 5 feet) most ACHEs have adjustable pitch blades. Adjustable pitch fans are manufactured in two types. One is manually adjustable (with the fans off) and the other is automatically adjustable (while running). Most automatically adjustable pitch fans change their pitch by means of a pneumatically actuated diaphragm working against large springs inside the hub.
The most popular speed reducer is the high-torque positive type belt drive, which uses sprockets that mesh with the timing belt cogs. They are used with motors up to 50 or 60 horsepower, and with fans up to about 18 feet in diameter. Banded V-belts are still often used in small to medium sized fans, and gear drives are used with very large motors and fan diameters. Fan speed is set by using a proper combination of sprocket or sheave sizes with timing belts or V-belts, and by selecting a proper reduction ratio with gears. Fan tip speed should not be above 12,000 feet per minute for mechanical reasons, and may be reduced to obtain lower noise levels. Motor and fan speed are sometimes controlled with variable frequency drives. The figure below provides a breakdown of the mechanical equipment.
Comparison of Induced and Forced Draft Units
Disadvantages and limitations
In most cases the advantages of induced draft design outweigh the disadvantages.
Because the number of tube rows, the face area, the air face velocity, and the geometry of the surface can all be varied, it is possible to generate many solutions to a given thermal problem. However, there is obviously an optimum solution in terms of capital and operating costs.
The basic heat transfer relationships that apply to shell and tube exchangers also apply to ACHEs. The fundamental relation is the Fourier equation:
F is a factor that corrects the log mean temperature difference for any deviation from true counter-current flow. In ACHEs the air flows substantially unmixed upward across the bundles and the process fluid can flow back and forth and downward as directed by the pass arrangement. With four or more downward passes, the flow is considered counter-current and so the factor "F" is 1.0.
As is apparent, initially neither the area nor the overall heat transfer rate nor the effluent air temperatures are known. The traditional approach in the design of ACHEs entailed an iterative trial and error procedure both on the CMTD and the transfer rate until the area satisfied both. Specifically, an air rise was assumed, the CMTD was calculated, an overall heat transfer coefficient was assumed, and an exchanger size was selected with the expected necessary area. An appropriate face velocity was then used to calculate an effluent air temperature, and the process was repeated until the assumed effluent air temperature matched the calculated value. The individual coefficients and the overall coefficient were then calculated, and the whole process was repeated until the calculated "U" and CMTD were sufficiently close to the assumed values.
However, there is another method that eliminates trial and error on the CMTD, and leaves only the trial and error on the tube-side film coefficient. The following discussion presents the Ntu Method described by Kays and London in Compact Heat Exchangers, as applied to ACHEs.
The following are definitions based on Compact Heat Exchangers:
Coefficients are based on outside bare tube surface for 1-inch OD tubes with 10 plain extruded aluminum fins per inch, 5/8 inch high, 21.2:1 surface ratio.
Fan Selection - Horsepower Requirements
To calculate the required horsepower for the fan driver:
Motor Shaft Horsepower =
The actual volume at the fan is calculated by multiplying the standard volume of air (scfm) by the density of standard air (0.075 lb/ft) divided by the density of air at the fan. From this relationship it can be seen that the ratio of the fan horsepower required for a forced draft unit to that required for an induced draft unit is approximately equal to the ratio of the exit air density to the inlet air density, which is in turn equal to the ratio of absolute air temperatures (t1 + 460) / t2 + 460). The total pressure difference across the fan is equal to the sum of the velocity pressure for the selected fan diameter, the static pressure loss through the bundle, (which is deten-nined from the equipment manufacturer's test data for a given fin type and tube spacing), and other losses in the aerodynamic system. Fan diameters are selected to give good air distribution and usually result in velocity pressures of approximately 0.1 inch of water.
The design of the fan, the air plenum chamber, and the fan housing, (in particular fan tip clearance), can materially affect system efficiency, which is always lower than shown on fan curves based on idealized wind tunnel tests. Industrial axial flow fans in properly designed ACHEs have fan (system) efficiencies of approximately 75%, based on total pressure. Poorly designed ACHEs may have system efficiencies as low as 40%. Speed reducers usually have about 95% mechanical efficiency. The value of driver output horsepower from the equation above must be divided by the motor efficiency to determine input power.
For estimating purposes refer to the figure below to approximate the horsepower requirement. This chart plots bare tube surface divided by horsepower versus tube bundle depth for the normal range of velocities. Applying the above criteria to our sample problem, we detennine that we must use two 10-foot diameter fans to have 40% of the bundle face area. We find that for a 6-row bundle, the area/horsepower is between 68 and 92 square feet of bare tube surface. If we use an average value of 80, the horsepower requirement for each fan is (336 ·.2618 · 32) (2 · 80) = 17.5 horsepower at maximum design ambient temperature. Power consumption must be calculated for the coldest expected ambient temperature, since at a fixed fan blade angle, fan horsepower consumption is inversely proportional to the absolute temperature. The power required for this minimum ambient temperature will set the required motor size.
Performance Control of ACHEs
Varying Air Flow
Louvers operate by creating an adjustable restriction to air flow and therefore do not save energy when air flow is reduced. In fact, louvers impose a permanent energy loss, even in the open position.
Two-speed motors, AUTO-VARIABLE fans, and variable frequency fan motor control do save power when air flow is reduced. In temperate climates, as much as 67% of the design power may be saved over the course of a year with AUTO-VARIABLE pitch fans. AUTO-VARIABLE hubs will thus pay back their additional cost in about one year or less.
Both louvers and AUTO-VARIABLE fans may be operated automatically through an instrument that senses temperature or pressure in the outlet header. For extreme cases of temperature control, such as prevention of freezing in cold climates in winter, or prevention of solidification of high pour-point or high melting point materials, more sophisticated designs are available.
Extreme Case Controls
Internal Recirculation. By using one fixed-pitch fan blowing upward and one AUTO-VARIABLE pitch fan, which is capable of negative pitch and thus of blowing the air downward, it is possible to temper the air to the coldest portion of the tubes and thus prevent freezing. Normally forced draft units have the negative pitch fan at the outlet end, while induced draft units have the positive pitch fan at the outlet end. In hot weather both fans can blow upward.
External Recirculation. This is a more positive way of tempering coolant air, but is practical only with forced draft units. Hot exhaust air exits the bundle, and enters a top plenum covered by a louver. When no recirculation is required, the top louver is wide open, and the heated air exits through it. When the top louver is partially closed, some of the hot air is diverted to a duct, through which it flows downward and back into the fan intake, mixing with some cold ambient air. An averaging air temperature sensor below the bundle controls the amount of recirculated air, and thus the average air intake temperature, by varying the louver opening.
Co-current Flow. For high pour-point streams it is often advisable to ensure a high tube wall temperature by arranging the flow co-currently, so that the high inlet temperature process fluid is in contact with the coldest air and the low temperature outlet process fluid is in contact with wan-ned air.
Auxiliary Heating Coils - Steam or Glycol. Heating coils are placed directly under bundles. Closing a louver on top of a bundle will allow the heating coil to warm the bundle or keep it warm in freezing weather, so that on start-up or shut-down the material in the bundle will not freeze or solidify. Heating coils are also occasionally used to temper very cold air to the bundles while the fan is operating and the exhaust louver is open.
ACHE noise is mostly generated by fan blade vortex shedding and air turbulence. Other contributors are the speed reducer (high torque drives or gears) and the motor. The noise is generally broad band, except for occasional narrow band noise produced by the motor or speed reducer, or by interaction between these sources and the ACHE structure.
The evidence is that for efficient fans at moderate fan tip speeds, this noise is proportional to the third power of the fan blade tip speed, and to the first power of the consumed fan horsepower. It is at present quite practical and usually economical to reduce the sound pressure level at 3 feet below an ACHE to 85 dB(A), but below 80 dB(A), noise from the drives predominates and special measures must be taken.
Design of ACHEs for Viscous Liquids
For process fluids with outlet viscosities up to 20 centipoises, it is possible by using large diameter tubes and high velocities (up to 10 ft/sec) to achieve a Reynolds number at the outlet above the 2,000 critical Reynolds number, and to keep the flow in the transition region. However, this usually results in pressure drops of 30 to 100 psi. In view of the disadvantages of designing for laminar flow, this increased pressure drop is normally economically justifiable because the increase in the operating and capital cost of the pump is small compared with the decrease in the cost of the turbulent exchanger.
The biggest problem with laminar flow in tubes is that the flow in inherently unstable. The reasons for this can be demonstrated by a comparison of pressure drop and heat transfer coefficient for turbulent versus laminar flow, as functions of viscosity (m) and mass velocity (G):
In an air-cooled heat exchanger, because of imperfect air-side flow distribution due to wind, or because of multiple tube rows per pass, it is likely that the flow through some of the tubes in a given pass is cooled more than that through other tubes.
With turbulent flow, pressure drop is such a weak function of viscosity (0.2 power) and such a strong function of mass velocity (1.8 power), that the flow in the colder tubes must decrease only slightly in order for the pressure drop to be the same as that in the hotter tubes. Also, as the flow slows and the viscosity increases, the heat transfer coefficient drops significantly, (-0.47 power of viscosity, 0.8 power of G), so the over-cooling is self-correcting.
With laminar flow, pressure drop is a much stronger function of viscosity (1.0 power) and a much weaker function of mass velocity (1.0 power), so the flow in the colder tubes must decrease much more to compensate for the higher viscosity. Viscosity of heavy hydrocarbons is usually a very strong function of temperature, but with laminar flow, the heat transfer coefficient is independent of viscosity, and only a weak function of mass velocity (0.33 power), so the selfcorrection of turbulent flow is absent.
The result is that many of the tubes become virtually plugged, and a few tubes carry most of the flow. Stability is ultimately achieved in the high flow tubes as a result of high mass velocity and increased turbulence, but because so many tubes carry little flow and contribute little cooling, a concurrent result is high pressure drop and low performance. The point at which stability is reached depends on the steepness of the viscosity versus temperature curve. Fluids with high pour points may completely plug most of an exchanger.
This problem can sometimes be avoided by designing deep bundles to improve air flow distribution. Bundles should have no more than one row per pass, and should preferably have at least two passes per row, so that the fluid will be mixed between passes.
When a fluid has both a high viscosity and a high pour point, long cooling ranges should be separated into stages. The first exchanger should be designed for turbulent flow, with the outlet temperature high enough to ensure an outlet Reynolds number above 2,000 even with reduced flow. The lower cooling range can be accomplished in a serpentine coil (a coil consisting of tubes or pipes connected by 180' return bends, with a single tube per pass). The low temperature serpentine coil should, of course, be protected from freezing by external warm air recirculation ducts.
Closed loop tempered water systems are often more economical, and are just as effective as a serpentine coil. A shell and tube heat exchanger cools the viscous liquid over its low temperature range on the shell side. Inhibited water is recirculated between the tube side of the shell and tube and an ACHE, where the heat is exhausted to the atmosphere.
For viscous fluids which are reasonably clean, such as lube oil, it is possible to increase the tube side coefficient between four- and tenfold, with no increase in pressure drop, by inserting turbulence promoters, and designing for a lower velocity. It is then advantageous to use external fins to increase the airside coefficient also. In addition to the increase in heat transfer coefficient, turbulence promoters have the great advantage that the pressure drop is proportional to the 1.3 power of mass velocity, and only to the 0.5 power of viscosity, so that non-isothermal flows are much more stable. The simplest and probably the most cost-effective promoters are the swirl strips, a flat strip twisted into a helix.
| About Us | Products / Services
Contacts | News | Links