Choking a Vertical Thermosyphon

Choking down on the channel outlet nozzle and piping reduces the circulation rate through a heat exchanger. Since the tubeside heat transfer rate depends on velocity, the heat transfer is lower at reduced recirculation rates. A rule of thumb says that the inside flow area of the channel outlet nozzle and piping should be the same as the flow area inside the tubing. Shell Oil in an experimental study showed that a ratio of 0.7 in nozzle flow area/tube flow area reduced the heat flux by 10%. A ratio of 0.4 cut the heat flux almost in half.

An approximate equation for the amount of heat flux reduction is:
Reduction = 3.06X -1.63X2 -0.43
Where X = area ratio


When To Limit Tube Passes In Aircooler

For tube side streams that have a high heat transfer coefficient, it is probably not advantageous to use more than two tube passes. This would be for condensing streams like ammonia and steam. This could also be true for high thermal conductivity liquid streams if the LMTD is high. The velocity on these type of streams will have a minor effect on the overall heat transfer coefficient in the typical aircooler. The major thermal resistance is the air side heat transfer coefficient.


Designing Better Use Of Tube Pressure Drop

When the calculated pressure drop inside the tubes is under-utilized, the estimated pressure drop with increased number of tube passes is:

         New tube DP = DP x (NPASS/OPASS)3

         DP =        Pressure drop

         NPASS = New number of tube passes.      

         OPASS = Old number of tube passes

This would be a good estimate if advantage is not taken of the increase in heat transfer. Since the increased number of tube passes gives a higher velocity and increases the calculated heat transfer coefficient, the number of tubes to be used will decrease. Fewer tubes increases the new pressure drop. For a better estimate of the new pressure drop, add 25% if the heat transfer is all sensible heat.


How to Increase Heat Transfer For Low Reynolds Numbers

If pressure drop is available and if the tube side Reynolds number is less than 5000 and more than 1000, you can probably increase the heat transfer considerably by increasing the number of tube passes and using shorter tubes. This will not only increase the tube velocity but there will be a lower L/D correction. Both of these factors will increase the heat transfer.


Calculating Fouled Pressure Drop

There are various ways to account for fouling when calculating pressure drop. One way would be to add a small amount to the tube diameter. This has a complex effect that is not linear in nature. A simple method is to add 10% for each .001 increase in fouling factor. Then multiply this factor by the clean pressure drop. You would use a pressure drop factor of 1.2 for a fouling factor of .002.


Liquid Thermal Conductivity for Light Hydrocarbons

You can make an estimate for the liquid thermal conductivity of light hydrocarbons if you know their specific heat. It is good for propane and heavier.
K = 0.025 / (specific heat)1.5


Cooling Water Flowing Inside 304SS U-tubes

Normally it is OK to use 304SS when cooling water with low chloride content is flowing inside u-tubes. But if for some reason the operating pressure drops to saturation there can be corrosion problems. The tube vibration that results from the flashing of steam amplifies the stress that causes stress corrosion cracking.


Flange Gasket Location

There is an optimum diameter of the gasket for flanges. It is when the total Operating moment of the flange under pressure is equal to the gasket Seating moment. For low pressure flanges the diameter should be as close To the bolt circle as possible. For high pressure flanges the diameter should Be as close to the flange I.D. as possible. In this case low pressure is considered to be below 300 psi. High pressure is considered to be approximately 750 psi and higher.


Kettle Pressure Drop

Usually you will see the allowable pressure drop on the specification sheet for the shell side of a kettle reboiler to be stated as "nil". This is close to being true only for the bundle. The inlet and outlet kettle nozzles will have a definite pressure drop. It is best to locate the inlet nozzle on the side of the kettle and above the bundle. This keeps the pressure drop down because there are no tubes in the vicinity to provide a restriction.


Large Boiling Temperature Differences

Large temperature differences in heat exchangers where liquid is vaporized are a warning flag. When the temperature differences reach a certain value, the cooler liquid can no longer reach the heating surface because of a vapor film. This is called film boiling. In this condition, the heat transfer deteriorates because of the lower thermal conductivity of the vapor. If a design analysis shows that the temperature difference is close to causing film boiling, the vaporizer should be started with the boiling side full of relatively cooler liquid. This way, you don't start off flashing the liquid. The liquid is slowly heated up to a more stable condition. If the vaporizer is steam heated, the steam pressure should be reduced which will reduce the temperature difference. With steam heating, take a close look at the design if the LMTD is over 90F This is close to the critical temperature difference where film boiling will start.


Rotated Square Tube Pitch

Some heat exchanger specifications for shell and tube heat exchangers mention square pitch but do not specifically mention rotated square pitch. Engineers with little thermal design experience who are trying to strictly adhere to the specifications may reject this type of tube pitch. The benefits for this type of tube pitch sometimes get lost because of this. Rotated square pitch gives better mixing of the shell fluid and better heat transfer for the heavier fluids. Frequently the shell size can be reduced when there will be heavier liquids on the shell side and the designer uses rotated square pitch. When the shell side flow is below a Re of approximately 5000, use this layout instead of square pitch.


Tube Wall Temperature for Cooling Water

When designing heat exchangers where hot process streams are cooled with cooling water, check the tube wall temperature. Hewitt says that where calcium carbonate may deposit, heat transfer surface temperatures above 140F should be avoided. Corrosion effects should also be considered at hot tube wall temperatures. As a rough rule of thumb, make this check if the inlet process temperature is above 200F for light hydrocarbon liquids and 300-400F for heavy hydrocarbons. Consider using Aircoolers to bring the process fluid temperature down before it enters the water cooled exchanger.


Is an Expansion Joint Required in the Shell?

A fixed tube sheet exchanger does not have provision for expansion of the tubing when there is a difference in metal temperature between the shell and tubing. When this temperature difference reaches a certain point, an expansion joint in the shell is required to relieve the stress. It takes a much lower metal temperature difference when the tube metal temperature is hotter than the shell metal temperature to require an expansion joint. Typically an all steel exchanger can take a maximum of approximately 40F metal temperature difference when the tube side is the hottest. When the shell side is the hottest, the maximum is typically 150F. Usually if an expansion joint is required, it is because the maximum allowable tube Compressive stress has been exceeded. According to the TEMA procedure for Evaluating this stress, the compressive stress is a strong function of the unsupported tube span. This is normally twice the baffle spacing.


Small Temperature Pinch Points in Condensers

Be extra careful when condensers are designed with a small pinch point. A pinch point is the smallest temperature difference on a temperature vs heat content plot that shows both streams. If the actual pressure is less than the process design operating pressure, there can be a significant loss of heat transfer. This is especially true of fluids that have a relative flat vapor pressure plot like ammonia or propane. For example: If an ammonia condenser is designed for 247 PSIA operating pressure and the actual pressure is 5 PSI less and the pinch point is 8F, there can be a 16% drop in heat transfer.


Sometimes Larger Tubes are Better

There is an exception to the rule that a shell and tube heat exchanger service using 3/4 inch tubes will be cheaper than one using 1 inch tubes.This is when the tubeside has a much lower heat transfer coefficient than the outside of the tubes and the following conditions are present:

  1. The flow will be in laminar flow if two(2) tube passes are used.
  2. If four(4) tube passes are used, the tubes in the 3/4 inch selection will have to be significantly shorter than allowed in order to meet pressure drop. On the other hand the 1 inch tube design uses the full allowable tube length.


Special S & T Exchanger Type (NTIW)

A shell & tube heat exchanger with normal segmental baffles has tubes that miss every other baffle. This can lead to long unsupported tube lengths for some applications. A long tube span has a low natural frequency and is prone to vibration. One solution is to design a "no tubes in window"(NTIW). exchanger. This design has no tubes in the baffle cut out. By using intermediate supports between baffles, the natural frequency of the tubes can be raised considerably to resist vibration.


Using Rods For Tube Inserts to Increase Heat Transfer

Use reinforcing rods inserted inside the tubes to increase the heat transfer and tube velocity . It is a quick and economical solution. This is usually done only in clean services. A typical case is using 3/8" rods inside a 3/4" x 14 bwg tube. The tube side heat transfer coefficient is increased bya factor of 1.7. But you have to be able to stand the increase in pressure drop. It goes up by a factor of 9.5. Another example is a 1.0" x 16 BWG avg. wall tube where the heat transfer goes up by a factor of 1.17 and the pressure drop by a factor of 3.5.


Locating Vents on the Shell Side of Vertical Exchangers

Proper venting of equipment is not always given the consideration it deserves. One place where venting is especially a problem is underneath the tubesheet of a vertical exchanger. The problem is that there will always be a space above the vent connection to trap gases or vapors. Besides the poor heat transfer in this region, this can cause corrosion problems. It is important to get the vent connection as close to the tubesheet as possible. Using multiple connections that are smaller is one solution. Another solution is to fabricate the upper tubesheet with a small vent tunnel inside.


Sizing the Vapor Space in Kettle Reboilers

The size of the kettle is determined by several factors. One factor is to provide enough space to slow the vapor velocity down enough for nearly all the liquid droplets to fall back down by gravity to the boiling surface. The amount of entrainment seperation to design for depends on the nature of the vapor destination. A distillation tower with a large disengaging space, low tower efficiency and high reflux rate does not require as much kettle vapor space as normal. Normally the vapor outlet is centered over the bundle. Then the vapor comes from two different directions as it approaches the outlet nozzle. Only in rare cases are these two vapor streams equal in quantity. A simplification that has been extensively used is to assume the highest vapor flow is 60% of the total. One case where this would cause an undersized vapor space is when there is a much larger temperature difference at one end of the kettle then the other. The minimum height of the vapor space is typically 8 inches. It is higher for high heat flux kettles.


Allocation of Streams in Shell & Tube

For those exchangers that need countercurrent flow, the stream with the highest pressure drop is usually best put in the tube side. This is true unless the design pressure is so high for the shell side that there would be material problems. High pressure drop instead of high design pressure is opposite of conventional thinking. If there are gas streams on both sides with mol. weights about the same and a small temperature difference, put the stream in the tubes with the highest value of:

(#//hr)(#/hr)/op. pressure

Othervise calculate the little more difficult term Vel x Vel x Density term for each side and put the stream with the highest value in the tubes.


Connecting Bundles of Existing Coolers for a new Service

When re-using air cooled exchangers in a new service, don't overlook connecting the bundles in a series-parallel arrangement. New air coolers nearly always have the bundle connected in parallel. Arrange the bundles for more series type flow to increase the tube side velocity and get higher heat transfer rates. For example, an air cooler with six bundles could be arranged with four bundles in parallel, connected to two bundles in series. The two series bundles would handle the coldest part of the heat load where higher velocity is needed the most.


Box header design - limit of process temperature change

In the design of an Air cooled heat exchanger, avoid imposing too large a temperature change in the box headers. Too much temperature drop between the inlet and outlet tube passes can cause leakage where the tubes meet the tubesheet. If the temperature change of the tube side stream is over approximately 400F, then use a split header design. This allows a hot top section to slide past a cooler bottom section.


Estimate Gas Heat Transfer Rate for Hydrocarbons

If you need to estimate a gas heat transfer rate or see if a program is getting a reasonable gas rate, use the following:

h = 75 x Sq. Root(Op. pressure/100.)     OR
h = 1.4W0.8

The operating pressure is expressed in Psia. W = #/tube/hr. This is for inside the tubes. The rate will be lower for the shell side or if there is more than one exchanger.


Check Heat Release Curves for Skipping Over Dewpoints & Bubblepoints

Frequently process engineers specify tabular heat release data that skips over dew points and bubble points. If equal increments of heat load or temperatures are used, chances are that the dew points and bubble points will be missed. It is important that the heat content at dew points or bubble points be shown.


Vertical Thermosyphon Recirculation Rate

In the design of vertical thermosyphons, the recirculation rate should be set by the process engineer if there will be anything unusual about the connecting piping. The recirculation rate is especially sensitive to the size and configeration of the outlet piping. If the recirculation rate is left for the thermal designer to set, they will have to make piping assumptions that may be violated later in the actual installation.


Triple Segmental Baffles

There is more than one kind of triple segmental baffles in the shell side of heat exchangers. Be sure you know which kind if you are checking a design that uses them. There is the kind you see in TEMA where there are three different groups in a set. The total number of baffle pieces is six. There is the kind that is like producing two double segmental streams in parallel. There are two groups in a set and a total of five baffle pieces. Another kind has only three pieces in a group and each piece has a different shape.


Increase Capacity of Existing Air Cooler with Fan Drive Changes

If you need to increase the capacity of an air cooler, don't junk it for a new one until you have exhausted the possibilities on changing the fan and the fan motor. The least expensive change is to increase the fan blade angle if it will not overload the motor. But check to make sures the blade angle is not already at the maximum. The next best change in terms of cost is to increase the fan speed by changing the drive ratio between the fan and the motor. If these changes are not enough you could increase the motor size or change the fan for one with more blades.


Cures for Vibration in Existing Bundle

Most flow induced vibration occurs with the tubes that pass through the baffle window of the inlet zone. The unsupported lengths in the end zones are normally longer than those in the rest of the bundle. For 3/4 inch tubes, the unsupported length can be 4 to 5 feet. The cure for removable bundles, where the vibration isn't severe, is to stiffen the bundle. This can be done by inserting metal slats or rods between the tubes under the nozzles. Normally this only needs to be done with the first few tube rows. Another solution is to add a shell nozzle opposite the inlet so as to cut the inlet fluid velocity in half. For non-removable bundles, this is the best solution. Adding a distributor belt on the shell would be a very good solution but it is expensive.

If a u-tube bundle has a vibration problem in the bend area, metal slates or rods can be inserted between the tubes. If a slight decrease in heat transfer is not a problem, encircle the u-bends with a band or heavy wire and squeeze the tube together.


Thermal Evaluation of Long Baffles

The two thermal design problems with using two shell passes with a longitudinal baffle in Shell and Tube heat exchangers are:

1. Heat conduction through the baffle. There is a calculation method by Whistler. It is a correction applied to LMTD.

2. Fluid by-pass around the long baffle. If possible, use an exchanger type where the long baffle is welded to the shell in order to avoid bypass. This should be done with a full penetration weld. The exchanger types, where the long baffles can be welded in, are Fixed Tube Sheets or U-Tubes. If U-tubes, the number of tube passes must be a multiple of four. Then the bundles can be removed. Other designs use multi-leaf long baffles for two shell passes. Since these cannot make a perfect seal, the amount of shell fluid by-bypassing the bundle must be calculated.


kettle Reboiler - Effect of Undersized Kettle Diameter

What effect will an undersized kettle diameter have? The effect will be a decrease in the boiling coefficient. A boiling coefficient depends on a nucleate boiling component and a two phase component that depends on the recirculation rate. An undersized kettle will not have enough space at the sides of the bundle for good recirculation. Another effect is high entrainment or even a two phase mixture going back to the tower.


Shell & Tube or Multi-Tube?

When is it Best to use Multi-tube (Hairpin) Exchangers instead of Shell & Tube?

1. Small surface (less than 400 square feet)
2. When there is a temperature cross
3. Liquid flows less than 150,000 LBS./HR.
4. For natural gas flows less than 1200 X Sq. root(oper. pressure)


Conditions Likely to Cause Shell & Tube Bundle Vibration

Bundle vibration can cause leaks due to tubes being cut at the baffle holes or tubes being loosened at the tubesheet joint. There are services that are more likely to cause bundle vibration than others. The most likely service to cause vibration is a single phase gas operating at a pressure of 100 to 300 PSI. This is especially true if the baffle spacing is greater than 18 inches and single segmental type. Another service that sometimes causes bundle vibration is water in the shell side. Water has a relatively higher momrntum than other most fluids. Then if extra precautions on bundle design are not taken, then there can be a vibration problem later when the exchanger goes into operation.



Increasing Capacity of Existing Shell & Tube Exchangers

To increase heat transfer check out using lowfins or other special tubing. When an increase in capacity will cause excessive pressure drop, you may not have to junk the heat exchangers. Investigate the relatively inexpensive modification of reducing the number of tube passes. Other possibilities are arranging multiple exchangers in parallel.


Lower Limit of Boiling Film Temperature Difference

A reboiler or chiller is best designed so that it doesn't have the lower heat transfer mode of natural convection. The dividing line between natural convection and boiling depends on the type of tubing used. If steel bare tubes are used, the lower limit of temperature difference between the tube wall and the boiling fluid is approximately 5 F. We have designed hydrocarbon chillers down to the temperature difference of 2 F using low-finned tubes. Special enhanced tube surfaces can be used for even lower temperature differences than 2F.


Impingement Rods

When shell pressure drop is critical and impengement protection is required, use rods or tube protectors in top rows instead of a plate. These create less pressure drop and better distribution than an impingement plate. An impengement plate causes an abrupt 90 degree turn of the shell stream which causes extra pressure drop.


What is too Large of Temperature Change in 2 Tube Passes ?

Warning! Large tube side temperature change. A big difference between the inlet and outlet temperature of the tube side causes leakage and bypass problems. The worst case is a shell and tube exchanger with two (2) tube passes where a gasket is used to seal between the passes. A careful analysis should be made if the temperature difference across the pass plate is more than 300 F. For a channel type that has a welded in pass plate, make an analysis if the temperature difference is more than 450 F. If this temperature difference causes an over stressed condition, possible cures are:

Add a unit in series so each unit has a smaller temperature difference.
Use one tube pass if the penalty isn't too great.
For aircoolers, use a split headers design.


Divided Flow LMTD Correction

Something to watch out for is the LMTD correction for Divided Flow Shell & Tube Exchangers. Divided flow (shell type J) does not have the same correction as the usual flow pattern (shell type E). We have seen several instances lately where a thermal design program made this correction factor mistake. True, there is very little difference at correction factors above 0.90. but, there is a difference at lower values. For example:

Equal outlet temperatures Shell type "E" correction Fn = 0.805
Shell type "J" correction Fn = 0.775
Cold outlet 5F higher than hot outlet Shell type "E" correction Fn = 0.765
Shell type "J" correction Fn = 0.65

Contact us if you do not have LMTD correction factor charts for divided flow. TEMA has one chart for a single shell but it gives high values for the above examples and it is hard to read in this range. We have charts for up to 4 shells in series at no charge.


Trouble-Shooting Article

To find out more about heat exchangers, see our article in the 1996 September issue of Hydrocarbon Processing. The title is "Troubleshooting Shell-and-Tube Heat Exchangers." It gives helpful information on diagnosing problems.


Copyright © 1996-2011 Gulley & Associates