Air cooled heat exchanger pdf

Great attention should be given to the alternate cooling methods - evaporative or dry cooling towers. Should the use of ACHEs be inevitable, it is difficult to recommend any general rules, for each case will be different. A noise expert and an ACHE expert should be consulted from the earliest possible stage, and a flexible attitude to fan requirements and to ACHE siting taken. Planning authorities sometimes impose a more stringent noise specification at night time than during daytime.

As ambient air temperatures are usually lower at night, it may be possible to run the fans at slower speed during the night time. As noise increases with the fifth or sixth power of the tip speed, this can give a marked reduction in noise.

For overseas locations, significantly higher figures may be regularly attained. In contrast, the wet bulb temperature, which controls the re-cool temperature of a wet cooling tower, does not vary so much, as the relative humidity is generally lower in warmer weather.

In selecting the maximum design inlet air temperature, it is the engineer's responsibility to consider the frequency with which the chosen temperature may be exceeded, and to assess the level of risk involved in under-designing against the cost of a too conservative design. This is discussed in more detail in sub clause 5. The minimum design temperature is important in considering control and winterization requirements see below.

The process fluid flows through the tubeside of one or more bundles of tubes; the air flows in cross flow over the outside of the tubes, assisted by a fan or fans. An example familiar to everyone is the motor car radiator.

In principle, there are many ways in which an ACHE could be arranged; this Guide in general is confined to the sorts of design that are found in the chemical and petrochemical industry. Figure 10 shows the major parts of a typical air cooled heat exchanger.

Although plain tubes could be, and in certain rare circumstances are, used, in almost all cases the tubes are finned on the outside. This is to counter the relatively poor film heat transfer coefficient that occurs on the air side. Sub clause 5. Tubes are typically from 2 to 12 m long. Within the bundle, the tubes are arranged in horizontal rows, with a tube spacing marginally greater than the fin o.

A bundle will usually contain between 3 and 6 rows of tubes, with successive rows staggered to give a triangular tube pitch. Different forms of header are used, depending on the duty. See sub clause 5. An ACHE bundle can have either single pass process flow, with the process fluid inlet connected to the header at one end and the outlet to the other, or a multi-pass arrangement, with pass partition plates dividing up the header s.

Unlike shell and tube exchangers, it is common for the different passes to have significantly differing numbers of tubes. A typical arrangement for an air cooled condenser where sub cooling is required, for example, is to have several rows of tubes in parallel performing the condensing part of the duty, followed by a single row of tubes for the sub cooling duty, resulting in an increased liquid velocity in this stage.

Not all the tubes in one row need be in the same pass. Bundles are usually mounted horizontally, but for condensers there may be a slight slope to assist in drainage. A large ACHE will require several bundles to provide the surface. Bundles are grouped into BAYS, each bay containing one or more typically bundles in parallel. The complete UNIT may contain several bays. Air for cooling is assisted through the bundle by FANS. Axial flow fans, giving a large volumetric flow for a very low pressure drop of the order of inches water gauge are used.

On large units these fans are often m in diameter; diameters of 7 m are not unknown. The width of a bay, the chosen tube length and the fan diameter are loosely interrelated. In order to ensure reasonable air distribution across the unit, it is desirable to divide each bay up into roughly square sections between the headers, each section being served by one fan see Figure It is normal to have between one and three fans for each bay.

On small units the fans may be driven by a directly coupled electric motor, but it is more usual for them to be driven through a gearbox or belt drive. This may be a simple rectilinear box, as shown in Figure 10, or may be shaped to reduce the pressure drop associated with the change in flow from the circular fan ring to the rectangular bundle. It is also possible to arrange pairs of bundles in an "A" or "V" formation see Figure 9 e , but this is not common in the process industries. It is more usual, however, to control air flow, if desired, either by using variable pitch fan blades, variable speed drives or switching off some fans see Clause 6.

In certain cases, especially in locations with extremely cold winters, STEAM COILS may be mounted below the bundle, warming the inlet air somewhat, to prevent over-cooling of the process fluid. The inlet and exit headers on each bundle will have at least one connection for the process fluid; on wide bundles there may be several, to aid flow distribution.

See sub clause 8. The complete ACHE installation will include a support framework to mount it clear of other equipment, to avoid restricting the air flow, and walkways, stairs etc. The normal approach is to specify the required duty, and place the thermal and mechanical design out to tender with selected ACHE manufacturers. In order to obtain an acceptable design, the manufacturer needs to know not only the process conditions, but also any constraints that GBH Enterprises wish to place on the design.

These will include layout constraints, noise specifications, preferred fans and drive systems, control requirements and economic factors. The thermal duty will usually be specified by the process engineer, who should also be responsible for deciding on an appropriate design margin over the flowsheet duty.

A design margin may be specified for several reasons: a The section of plant may be required to run at instantaneous rates above the normal plant throughput as part of the normal plant operation. Designing for this condition does not represent a true design margin, as the higher rate represents normal conditions. If it is probable that the plant will be uprated at some future date, there may be a case for increasing the design throughput, with a corresponding increase in heat load.

However, the heat transfer coefficient under the initial operating conditions will be lower than the design figure because of the lower velocities; the performance under the initial operating conditions should be checked to determine the expected design margin. It may be preferable to make provision for increasing the size of the ACHE at some later date, by adding further bundles in parallel with the original ones. The physical properties of the mixture may be uncertain, and plant measurements of actual flowrates and compositions may be unreliable.

Hence, the possibility of enforcing any thermal guarantee is remote. The manufacturer is under great pressure to design as cheap a unit as possible. Further, the heat transfer data used by the manufacturer to design the cooler are, at best, subject to some uncertainty. It is generally advisable, for a critical duty, to provide some form of safety margin to allow for uncertainties in the design methods. A thermal design margin safety factor may be provided in several different ways, which have their own advantages and disadvantages.

It is important that the engineer understands the implications of these. The engineer should be wary of disclosing design margins to a manufacturer, as the latter may be tempted to design with negative margins himself, knowing that in many cases, actual performance checks under design conditions may be difficult or impossible.

This sheet should be included, suitably annotated, in the plant manual, along with the correct data sheets, so that the true situation is recorded: 1 The provision of excess surface: If the extra surface is provided by increasing the number of tubes per pass, this may prove unsatisfactory. It will result in a more expensive unit but because of the lower process side velocity, and hence coefficient, there may be little effective increase in performance.

It is better to provide the extra area by increasing the exchanger length. It is not possible to use this approach without declaring it to the manufacturer. This suffers from the disadvantage that the actual margin on performance at normal air temperatures will depend on the product temperature. The specification of design ambient temperature is discussed in sub clause 5.

It should be used to ensure that a critical unit is designed to meet its duty on warm days, but it is not recommended to use this parameter to control design margins at other ambient conditions. If this approach is used, and the higher throughput is not actually likely to occur, the allowable pressure drop supplied to the manufacturer should be increased above the actual value by the square law, in order to avoid undue constraints.

As the unit will end up being designed for a flowrate above that at which the plant will run, it will not be possible to do performance checks at design conditions. The manufacturer will seek to minimize the area, within the constraints of allowable pressure drop; the film coefficients used by the manufacturer will not be affected by the "safety margin" as is the case for using an increased throughput.

The approach is useful when dealing with a manufacturer, as it means that the safety margin does not have to be revealed. However, it suffers from the same drawback as does raising the design air temperature, in that the margin will appear greater for units with a low outlet temperature. In some cases the section of the plant upstream of the ACHE is required, for process reasons, to run at a higher pressure than the downstream, and any pressure drop not absorbed by the exchanger will be taken by a control valve.

An example of this might be where the product from a pressure reactor is to be cooled before storage at atmospheric pressure.

In these cases the pressure drop can be regarded as "free" and it will usually pay the engineer to design the unit to absorb as much of the available pressure drop as possible, consistent with the requirements for control. However, in general, pressure drop has to be provided by a pump or compressor. The cost of pressure drop may be considerable, especially with less dense fluids, as the power absorbed is proportional to the volumetric throughput times the pressure drop.

However, a large pressure drop with viscous fluids, by improving the process side heat transfer coefficient and hence reducing the exchanger capital cost, may more than outweigh the cost of the pressure drop. For low pressure condensation duties, particularly vacuum condensers, it is usually necessary to limit the pressure drop, as the condensing temperature, and hence the driving force, falls with reducing pressure. Fouling resistances specified frequently take no account of the effect of fouling layer thickness on pressure drop.

As the pressure drop for a single phase fluid through a pipe varies inversely with the fifth power of the diameter, any significant fouling layer can have a noticeable effect.

Comparison of the estimated exchanger and pressure drop costs, together with common sense, should show if there is a serious problem. If so, the only solution is to make several designs at varying pressure drop, with a computer, and compare the resultant overall costs.

A rigorous examination of the effect of ambient design temperature on plant economics will be so expensive and time consuming as to be impracticable. The best that can be hoped for is a crude optimization of the largest units, perhaps so inaccurate as to be misleading. In general, the effect of too low a design air temperature will be a turndown of the plant on hot days. The true cost of such turndown depends on market conditions at that time and hence is almost impossible to forecast.

The engineer will, therefore, have to make a judgment, based on no sound data. The following data are given as a guide: a Lenient Design Non-critical duties : The chosen temperature is exceeded for approximately hours per year.

Ideally, temperature frequency data should be obtained for the works where the exchanger is to be installed. Failing this, the Meteorological Offices maintain records for a number of locations throughout your geographic region, but it should be remembered that weather conditions can vary significantly over small distances, so these data may not be representative. The proximity of potential sources of warm air e.

The minimum expected air temperature should be specified, as this not only determines the performance of the unit on cold days, and shows up any tendency for process freezing etc. The performance of a given air cooled heat exchanger will be less at higher altitude due to the fall-off in air density, and hence volumetric heat capacity.

The best way of supplying these data, especially for multi-component condensation, is in the form of a "Physical Properties Profile", where the properties of the vapor and liquid phases together with the heat load and weight fraction vapor are given for a range of temperature values spanning the expected operating conditions.

Such data can be generated for most cases. The Process Engineer should discuss this with the manufacturer, based on the use of "normal" ACHE bundles, with welded steel headers, and round steel tubes and aluminium fins.

For many duties, especially with low pressure and clean fluids, other forms of ACHE are more efficient. If offers for "different" ACHEs are required. Three fin materials are commonly used in fin tubes - aluminium, steel and copper. The virtues and disadvantages of the three metals can be summarized: a Aluminium: is the most cost effective of the three, having good thermal conductivity and reasonable cost per square meter.

The cost of heat transfer surface is "per square meter", not "per ton". Aluminium has adequate corrosion resistance for most ACHE applications, though it is reasonable to have some reservations on this question. The almost universal choice of aluminium fins in process ACHEs involves the use of helical fins on round tubes.

The performance of aluminium fins is much better than that of steel fins, and they are much cheaper than copper helical fins. Steel, galvanized, is much the same cost "per square meter" as aluminium. However, it is rather a poor conductor, resulting in low fin efficiencies.

The result is that steel finned exchangers are much more expensive than are aluminium finned. They are, in some atmospheres, more resistant to corrosion. They are also much stronger than are aluminium fins, but cost has limited their use to some particularly corrosive services. The efforts made to improve air quality at these sites has been such that aluminium finned tubes are now acceptable, and there now seems hardly any market for steel finned ACHEs on process plants.

It is thus more expensive "per square meter" and little advantage can be taken of its superior thermal conductivity in round helical fin tubes, where fin thickness is dictated by manufacturing considerations, resulting in very high fin efficiencies for aluminium fins. Especially when tinned, copper offers superior corrosion resistance to either of the other metals.

Such ACHEs will be present on all sites, probably as diesel or transformer coolers. However, copper finned ACHEs have scarcely been used for process units. Thus aluminium finning is the almost invariable choice for process ACHEs. This has an electrolytic potential lower than that of carbon steel, or any other tube metal commonly used in process plants. The aluminium therefore acts as a sacrificial protection to the steel.

The result is that external corrosion of the tube is virtually unknown over the finned portion of fintubes. Some manufacturers leave an unfinned part near the tubesheet. This will be subject to corrosion if it is longer than 10 mm, and the provision of protection of these parts by e. If aluminium tubes are used with aluminium fins, it is necessary to check that the tube is electropositive to the fins at the temper used for both.

If not, preferential pitting and failure of the tube may occur. The corrosion to be avoided is a general corrosion of the fins. Unprotected fins would have corroded rapidly in the atmosphere in certain plant locations; certainly, with the lower rows of fintube protected, life of aluminium surfaces will be similar to that of the plant. At less aggressive site locations, including coastal sites with chlorine in the air, atmospheric corrosion of the general finned surface is rarely important.

As explained in 5. Should atmospheric corrosion occur, the corrosion product is bulky and adherent, and very difficult to remove. It will cause an increased resistance to airflow, and hence loss of performance. Generally, there will be preferential corrosion close to the tube, which will cause further loss of performance due to decreased fin thermal efficiency, and the fin may be seriously weakened. See, for example, sheet AE2 of Ref.

These finned tubes are manufactured by opening up a groove in the base tube, tension winding a strip of fin material into the grove, and then peening the base tube so that the fin is securely held. The resultant tubing is robust, with little likelihood of the fin coming away from the base tube. It is sometimes suggested that water can enter the crack between tube and fin and cause a thermal resistance at this point. Some tubes submitted to the British Non-ferrous Research Association for long term marine and industrial corrosion tests indeed show corrosion at this point; however, when tested for heat transfer, they showed a small increase in heat transfer coefficient compared to new tubes.

There is a suspicion that preferential corrosion may occur near the base of "G" fins, which would lead to a weakening of the fins and a loss of performance. There is no known evidence to support this suggestion, but it remains a nagging doubt. The remaining types of finned tubing are not generally recommended for process duties, but are described below for completeness.

This fin-tube interface is not recommended, and will rarely be found on process ACHEs, although such tubing may be found on steam heated process air heaters, which can be considered to be a type of ACHE. As the fins are not positively located onto the base tube, relative movement of fins tends to occur, and continuous contact between the fin base and the tube cannot be guaranteed.

In the extreme, if the fin should break or become detached at one end, the complete fin spiral can end up at one end of the tube, leaving a bare tube. Although this construction does give an improved heat transfer area between the tube and the fin and more positive location, its use is not recommended. A particularly damaging form of corrosion occurs when a bundle is wetted, possibly during construction or shut-down. Water between the fins can infiltrate the space between tube and fin by capillary action.

A galvanic cell is set up between aluminium and steel, and aluminium oxide corrosion product is formed. This makes an effective insulating blanket between tube and fin. Although only indirectly concerned with corrosion, there is another point to avoid with "L"! The base of the fin will not be truly flat, and there will only be a relatively small proportion of the base of the fin in contact with the tube.

The result is that any interface thermal resistance will be multiplied by this ratio, when related to the whole outside surface of the tube. Such a resistance will be present if mill scale is not removed from the tube before finning, and can be appreciable.

Values as high as 0. K based on bare tube area have been measured with "L" fin tubes in new condition. If the mill scale is removed, then the tube is very liable to corrosion before the finning is applied. Some McElroy machines have a sand blast incorporated, thus avoiding these troubles.

Careful inspection of tubing is necessary before "L" fins are applied. These are intended to give better cover of the base tube with the aluminium. Since there is no risk of corrosion of the base tube, there seems little point in paying extra for this type of tube.

They have the disadvantages of simple "L" fins. If an aluminium tube is threaded over a steel tube, and fins formed on the aluminium, then a "muff" fin or "E" fin is formed. Many advantages are claimed for these fintubes, especially that the continuous cover of the steel prevents corrosion. However, no external corrosion of the steel will occur in any case, because of the galvanic protection afforded by ordinary aluminium fins, so this advantage can be dismissed.

GEA claimed as the advantage for this type of tubing that the airside pressure drop characteristics are superior to those of round tube. Recent experience of trying to re-tube exchangers dating from that period has shown that the elliptical tubing is expensive and hard to obtain. Moreover, the manufacturing process for the finned tube, which involved rolling round tube to an elliptical cross section, threading the fins on and re-rolling the ends to a circular cross section for welding into the tubesheet, was prone to cause cracking of the tube ends.

GEA appear no longer to offer it as their standard. Fin heights are usually either 0. However, in particularly dirty environments it may be advisable to reduce this to 8!

The temperature limits for the various types of fin should preferably refer to the tube metal temperature, rather than the fluid temperature. However, any proposal which is based on metal, rather than fluid temperature, may be considered carefully.

Work by CMB Russell showed that obstructions in the fan discharge are more damaging than those in the inlet, so the provisions of S. The specification of noise levels near the ACHE to protect operator hearing is straightforward. In practice, the noise level due to the fans away from the near field of an ACHE bank is often below the background noise level.

Insist that the specialist translate the allowable sound power into a sound pressure level SPL near the fan. The allowable SPL near the fan can be calculated from the PWL with a loss of accuracy of only a decibel or so, and can be guaranteed and measured. It is probable that a lower noise level will be required at night than during the day. As ambient temperatures drop at night, the fan speed can be reduced with a reduction in noise level, provided that variable speed fan control is used.

This advantage does not apply to variable pitch control, the noise being almost independent of blade pitch. The reduction in noise can be very dramatic: the sound power level for a given fan varies typically with the speed raised to the power 5 or 6. Fans meeting these requirements will be operating at a very poor efficiency when at the design point with clean fin surface. The requirement may affect the thermal design adversely, especially if there are severe noise limitations.

The effect of stall is much more severe with broad chord fans, than is the case with the narrow chord. Toothed "timing" belts, however, although they are specifically excluded, have shown good performance on many duties. It seems reasonable to recommend them for drive motors up to 30! Should it be essential to avoid tubeside leakage of an ACHE, then a manifold type of header may be used.

This permits radiography of tube and manifold welds; the tube may be left unfinned to permit ultrasonic inspection to the first, say, mm of the tube from the manifold, to check against erosion but see 5. Headers between passes may be avoided by the use of U-bends.

Tube fixing will be by welding when leakage is feared, and, although welding and inspection are possible when plug headers are used, both are more difficult than is the case when cover plate or "D" type headers are used. Equally, inspection of tubes and tube ends for damage, corrosion or erosion is more difficult with plug headers. Although plugs resist leakage better than will rectangular joints, cover plate or "D" type headers will normally be the choice when manifold headers are unacceptable, and precautions against leakage are necessary.

A dummy tubesheet may be used to prevent the spread to atmosphere of any leakage that might occur at the tube ends. In steam condensing applications, the A-frame type Fig. Headers of air coolers are the boxes at the ends of the tubes which distribute the fluid from the piping to the tubes.

The cover plate header design shall permit the removal of the cover without disturbing header piping connections. The bonnet header design shall permit the removal of the bonnet with the minimum dismantling of header piping connections. Threaded plug holes shall be provided opposite the ends of each tube for access. Holes shall be threaded to the full depth of the plug sheet or 50 mm 2 inches , whichever is less.

In addition to the above, The major differences between an Induced Draft and a Forced Draft Air Cooled Heat exchanger are listed below with respect to some design parameters:. As one might expect the best kind of control scheme depends on the application. Does the process require a very tight control on the process outlet temperature or is it better to allow the process temperature to go down with the ambient air temp.

Following is a list of some of the commonly used control devices for air coolers, but in no particular order. VFD if used shall suit the motor nameplate rating and similarly for the Belt Pulley transmission it shall suit the motor nameplate rating.

For extremely cold regions like Canada, Siberia, etc an air-cooled heat exchanger with an internal recirculation system is used. Such systems can control the cooling air temperature regardless of ambient temperature. Positive or Negative Step Auto variable fans are used for such systems. By using one fixed-pitch fan blowing upward and one Auto-variable pitch fan, which is capable of negative pitch and thus of blowing air downward, it is possible to temper the air to the coldest portion of the tubes and thus prevents 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.

The performance and reliability of ACHE depend on proper inspection and maintenance of components. Regular maintenance and inspection of Air Cooler systems include. Refer to the below attached animated video for proper understanding and visualization of each part of a forced draft type air fin fan cooler.

The assembling sequence of one single bay is shown in the video. I am a Mechanical Engineer turned into a Piping Engineer. This includes special high-performance finned tubes with optimized material and design, Depending on your specific requirement: GEA air coolers are oriented towards the respective process requirements and the particular features of the location.

Regardless of design characteristics, unhindered air feeding and discharge must be ensured. Where space is at a premium, roof structures are used. The economic profitability of an air-cooled heat exchanger depends on the capacity of the finned tubes.

The capacity results from the transfer of the maximum heat exchanged in the smallest possible space while keeping thermodynamic losses on the air side low. Cover plate header with through bolts Cover plate header with stud bolts Welded header D type This header version has a removable cover In this version, the nozzles are arranged at The inexpensive header design is mainly to enable easy inspection and cleaning of used for clean products or a high vacuum.

For inspections it is can remove the cover plate without The tubes are welded into tube sheets, and generally sufficient to remove only a dismantling any tubes. The cover plates are D type headers with the required manifold because this already offers a also available The specific feature: GEA headers are available in all material types.

They can be used in all pressure and temperature ranges and therefore predestined for various media and marginal conditions. Plug-type header High-pressure header with return segments High-pressure header with welded return bends Here the tubes are usually rolled into the A threaded plug with a soft iron seal is This high-pressure header, just as the variant plug headers; in special cases, they are provided opposite every tube.

This plug can with the return segments, is perfectly suited directly welded into the tube Direct drive Belt drive for forced draft configuration The fan is fitted directly to the motor shaft, using a flange. Fan This version uses pulleys for reducing the motor speed to the required speed and motor speed are identical. Belt drive for induced draft configuration Gearbox drive When waste air temperatures are low, the motor can also be This drive includes a fan joined directly to the output shaft of the installed above the bundle.

Different versions, one common feature: GEA air cooler fan drives are convincing with their high degree of reliability and their low Your most recent searches Delete. Frequent searches. Other sections. Air-cooled heat exchangers 20 Pages Add to favorites. Catalog excerpts.