HEAT EXCHANGER DESIGN MANUAL PROCESS DEPARTMENT
REVIS ION
:1
DATE
: 06.03.2003
Foreword
This Heat Exchanger design Manual discusses about the Heat exchanger design re lated to refin refiner ery y serv servic ices es.. The The disc discus ussi sion on star starts ts with with gener general al infor informa mati tion on on the the Heat Heat exchangers and proceeds into the detailed design aspects focused on Shell and Tube Heat Exchangers. Discussions on Thermal design and some mechanical aspects are also included in the manual. A study report on optimization of exchanger design is included in the final parts of the manual for readers better understanding on the cost implications for exchanger design.
Foreword
This Heat Exchanger design Manual discusses about the Heat exchanger design re lated to refin refiner ery y serv servic ices es.. The The disc discus ussi sion on star starts ts with with gener general al infor informa mati tion on on the the Heat Heat exchangers and proceeds into the detailed design aspects focused on Shell and Tube Heat Exchangers. Discussions on Thermal design and some mechanical aspects are also included in the manual. A study report on optimization of exchanger design is included in the final parts of the manual for readers better understanding on the cost implications for exchanger design.
Table of Contents
1 INTRODUCTION-----------------------------------------------------------------------------------------------------------------------4 2 CLASSIFICATION---------------------------------------------------------------------------------------------------------------------5 3 BASIC DESIGN THEORY----------------------------------------------------------------------------------------------------------10 4 SELECTION OF HEAT EXCHANGERS----------------------------------------------------------------------------------------13 5 TUBE SIDE DESIGN -----------------------------------------------------------------------------------------------------------------16 6 SHELL SIDE DESIGN ---------------------------------------------------------------------------------------------------------------19 7 NOZZLES-------------------------------------------------------------------------------------------------------------------------------27 8 MISCELLANEOUS-------------------------------------------------------------------------------------------------------------------29 9 VIBRATIONS---------------------------------------------------------------------------------------------------------------------------31 10 HEAT EXCHANGERS OPTIMIZATION-------------------------------------------------------------------------------------32 11 GENERIC GUIDELINES FOR HEAT EXCHANGER DESIGNER-----------------------------------------------------49 14 TYPICAL FOULING RESISTANCES USED FOR DESIGNING HEAT EXCHANGER-------------------------57 15 MECHANICAL ASPECTS--------------------------------------------------------------------------------------------------------58
1
Introduction
Heat exchanger is a device, which is used for transferring the thermal energy between two or more fluids that have different temperatures with or without phase change. The wide range of heat transfer services that are required in the energy, process, and environmental control and transport industries has called forth an equally wide variety of heat exchanger configurations to meet the requirements. Each kind of exchanger is suitable for a typical application. Thus, it also includes condensers, reboilers and vapourizers in addition to heaters and coolers where there is no phase change. This manual will however discuss only the shell & tube heat exchanger design in detail.
Based on the transfer process occurring in heat exchangers, they may be classified as Direct transfer type. Storage type. Direct contact type.
A direct transfer type of heat exchanger is one in which the cold & hot fluid flows simultaneously through the device and heat is transferred through a wall separating the fluids. Most heat exchangers employ this mode of heat transfer. Shell and tube heat exchanger is of this type.
A storage type of heat exchanger is the one in which the heat transfer from hot fluid to cold fluid occurs through a coupling medium in the form of porous solid matrix. The hot & cold fluid flows alternately through the matrix, the hot fluid storing heat in it and the cold fluid extracting heat from it. This is not generally used in refineries.
Direct contact heat exchanger is the one in which the fluids are not separated. If heat is to be transferred between a gas & liquid, the gas is either bubbled through the liquid or the liquid is sprayed in the form of droplets into the gas. Cooling towers and scrubbers are two examples of direct contact heat exchangers.
Among the above three types direct type heat exchanger is commonly used in the refinery, chemical and petrochemical applications. Other two types of exchangers have there own limitations. Moreover, the design of storage type and direct type exchanger are not common and thus designed by special vendors.
2 Classification Now based on the geometry of the equipment we can classify the exchangers into following types:
2.1 Tubular Heat Exchangers Tubular heat exchangers are generally built of circular tubes, although elliptical and rectangular tubes have also been used. There is considerable design flexibility because the core geometry can be varied easily by changing the tube diameter, length, and arrangement. Tubular exchangers can be designed for any operating temperatures, pressures, and any temperature and pressure differences between the fluids, limited only by the materials of construction. They can be designed for special operating conditions: heavy fouling, highly viscous flow, corrosiveness, toxicity, radioactivity, etc. These exchangers may be further classified as shell-and-tube, double-pipe, spiral and tube – coil exchangers.
1.1.1
Shell-and-Tube Exchangers
Shell-and-tube exchangers are composed of round tubes mounted in a cylindrical shell with the tube axis parallel to that of the shell. One fluid flows inside the tubes, the other flows across and along the tubes. The major components of this exchanger are tubes (or tube bundle), shell, stationary or front-end head, rear-end head, baffles, and tubesheets. A variety of internal constructions are used in shell-and-tube exchangers, depending upon the desired heat-transfer and pressure-drop performance and the methods employed to reduce thermal stresses, to prevent leakage, to provide for ease of cleaning, to contain operating pressures and temperatures, and to control corrosion.
These exchangers are classified and constructed according to Tubular Exchanger Manufacturers Association (TEMA) standards in the United States or modified TEMA standards in other countries. TEMA has developed a notation system to designate major types of shell-and-tube exchangers. In this system, each exchanger is designated by a three letter combination, the first letter indicating ‘the front-end head type’, the second ‘the shell type’, and the third ‘the rear-end head type’. For instance AEL is fixed tube one shell pass exchanger. The details of this are discussed later.
TEMA has also set up mechanical standards for design, fabrication, and material selection for three classes of heat exchangers: R, C, and B. Class R designates the unfired shell-and-tube heat exchangers generally used for severe conditions in petroleum and related processing. Class C designates the unfired shell-and-tube heat exchangers generally used for moderate conditions of commercial and general processing.
Class B designates the unfired shell-and-tube heat exchangers used for chemical process service. Generally, Class B exchangers employ non-ferrous materials, while the other heat exchangers employ ferrous materials.
1.1.2
Double-Pipe Heat Exchangers
Double-pipe heat exchangers usually consist of concentric pipes. One fluid flows in the inner pipe and the other flows counter-currently in the annulus between the pipes. T his is perhaps the simplest heat exchanger. Flow distribution is not a problem, and cleaning is done easily by disassembly. This configuration is appropriate when one or both of the fluids are at very high pressure because containment in small diameter pipe or tubing is less costly than containment in large diameter shell. Double-pipe exchangers are generally used for the small capacity applications where the total heattransfer surface area is 20 m2 (215 ft2) or less. However, they are at times used for application up to 50 m2 (500 ft2). Stacks of double-pipe or multitube-type heat exchangers are also used in some process applications with radial or longitudinal fins.
Multitube hairpin exchanger consists of bundle of hairpin tubes, with separate shell on the each leg of the hairpin & special cover over the U bend of the hairpin. The tubes may be longitudinally finned or can be plain tubes with baffle to give cross flow. The number of tubes in the bundle is usually much less than conventional shell & tube exchanger.
2.2 Plate- Type Heat Exchangers Plate-type heat exchangers are usually built of thin plates, which are either smooth or corrugated and are either flat or wound in an exchanger. Generally, these exchangers cannot accommodate high pressure and temperature differences. These exchangers are principally used for clean fluids when the temperature difference between process streams is small. The required thermal duty is achieved with the very large surface area per unit volume. The plate-type exchanger can be used when there are several hot or cold process streams present. These types of exchangers are not common in the refinery industry for its limitation on clean fluids and high capital investment. However, where the temperature approaches are very small (about 20°C), this exchanger can find the suitable place. Packinox TM is an example of this type of exchanger manufacturer.
2.3
Extended surface heat exchangers
When additional metal pieces are attached to ordinary heat-transfer surfaces such as pipes & tubes, they extend the surface and enhance the heat exchanger area. Pieces, which are employed to extend the heat transfer surfaces, are known as fins. Extended surface exchangers are used for very high heat loads.
In refinery application, where air is used as cooling media, there extended surface exchanger is commonly used. The fundamentals of the design of Air Fin Cooler are same as of shell and tube heat exchanger, which will be discussed later. Based on the process function also we can classify the heat exchanger to the following: 2.4
Condensers
Condensers are widely used in refinery industries, power generation, chemical process, air conditioning, automotive, and many other industries. Broadly speaking, they are either direct contact type or indirect contact type (stream separated).
In a direct contact condenser, the condensing process stream and coolant come into direct contact with each other and heat is transferred. These condensers are inexpensive, have high heat-transfer rates per unit volume, and fouling is not a problem. However, applications are limited due to mixing of the process stream with the coolant. In a pool condenser, the condensing vapour is injected into the pool of liquid coolant. The spray condenser is the most common direct contact condenser in which sub-cooled liquid is sprayed into the vapour.
The most common types of condensers used in the industry are shell & tube type exchangers. In the process industry, condensation occurs either inside or outside the tubes, the tubes being either horizontal or vertical, and the condensate may be a single component or multi-component (miscible and/or im-miscible) pure vapour with or without non-condensable gases. Condensation may be total (as with a pure vapour) or partial (single or multi-component vapours with or without non-condensable). In addition to condensation, de-superheating and/or sub-cooling may also take place in a condenser.
One of the most important design features of a condenser is to provide a vent for the removal of non-condensable gases (regardless of how small they may be in the vapour). The vent should be located near the coldest part of the condenser to avoid escape of vapour, but should be high enough so that it does not flood with condensate.
An impingement plate should be placed under the vapour inlet nozzle to prevent tube erosion and possible flow-induced vibrations. A variety of shell types have been used depending upon the applications such as E (single- or multi-segmental baffles), G, H, J, and X, the X shell having the lowest pressure drop on the shell side.
2.5
Liquid to vapour phase change exchangers
These are mainly evaporators, reboilers. Evaporators are mainly used for concentrating a chemical solution by evaporation. The final products of evaporators are typically concentrates, crystals, or powder. Vaporisers which supply heat to generate vapours in a distillation column & are located at the bottom of a distillation column are termed as reboilers.
1.1.3
Natural Circulation (Thermosiphon) Reboiler
In a kettle reboiler, the vapour at the exit of the boiling liquid side is dry or almost dry; in other thermosiphon reboilers, it is a two-phase (vapour – liquid) mixture. Proper design of the inlet and outlet piping to the reboiler is essential to maintain minimum pressure drop in the piping. The kettle reboiler (TEMA K shell), usually consists of horizontal bundle of heated U tubes (two tube passes), circular in cross section, placed in an oversized shell; a multiple tube-pass floating head bundle may also be used. The tube bundle diameter ranges from 50 to 70 % of the shell diameter. The liquid to be vaporised enters from bottom and covers the tube bundle; the vapour occupies the upper space in the shell, which don’t have tubes. The large empty space in the shell acts as a vapour disengaging space, and almost dry vapour exits from the top nozzle. Depending upon the length of the kettle and a need to reduce liquid entrainment, one or more vapour nozzles are used. The liquid is fed at a rate greater than the vaporisation rate so that low volatile components do not build up in the shell. The small hold-up space beyond the weir is used to control the removal of excess liquid (which includes non-volatiles); the nozzle in this space is used to drain the excess liquid. Typical weir heights exceed bundle heights by 50 – 150 mm (2 – 6 in). To maintain a high liquid circulation rate, the tube pitch-to-diameter ratio is kept between 1.5 and 2. Generally, the kettle reboiler is considered as a pool-boiling device, however flow boiling prevails in the tube bundle.
The vertical thermosiphon reboiler , has a TEMA E shell with single-pass tubes in which boiling occurs. The tube diameter ranges from 25 to 50 mm (1 – 2 in) for low-pressure operation with wide-boiling mixtures, and tube lengths are usually 2.4 – 3.7 m (8 – 12 ft). The outlet pipe is kept short with its flow area equal to the flow area of all tubes because of the high exit velocity. The inlet pipe area is kept small (up to 50 % of the tube area) to enable flow to be restricted in case of flow instability. The liquid level for non-vacuum operations is maintained to the upper (top) tubesheet. This means the liquid level in t he base of the distillation column must be higher, and the column must be elevated above the ground. For vacuum applications, the liquid level may be dropped to 0.3 – 0.6 times the tube length. Vapour – liquid mixture leaves the reboiler and separation takes place in the column.
The horizontal thermosiphon reboiler , has a TEMA X, G, H, or E shell, single or two-tube passes, and boiling takes place on the shell side. Flow is restricted in the inlet line for stable operation and good control of the reboiler.
1.1.4
Forced Circulation Reboilers
These reboilers are used for heavy fouling, viscous, or solid-bearing liquids, or when the vaporisation rate is low. The pump in the liquid feed line to the reboiler is used to circulate the fluid through the system. The boiling fluid is on the tube side in most cases, but may be on the shell side in special applications. The reboiler may be horizontal or vertical. Generally, very little boiling (< 1 %) takes place within the reboiler due to the high circulation rate.
2.6 Chillers Its function is to cool a fluid to a temperature, smaller than obtainable if water were only used as a coolant. It uses refrigerant such as Ammonia or Freon. 2.7 Heaters Its function is to impart sensible heat to a liquid or a gas by means of condensing steam or Thermicfluid. 2.8 Coolers This term is usually used for exchanger that cools a liquid or gas with water. This term can be applied to exchanger with a purpose of cooling a liquid within range possible by water. In general when a liquid is sub-cooled after condensation or cooling in an air cooler it is then called as trim cooler.
3
Basic Design Theory
Basic heat transfer equation is the relationship of the heat transfer rate, heat transfer surface, and mean temperature difference (MTD) is as given below. Q = U *A * MTD Overall heat transfer coefficient should be calculated to determine how much heat transfer area is needed for the required heat transfer rate at a given MTD. To calculate overall U, physical properties of the material involved, process conditions, and geometry specifications are needed. The prime objective in the design of an exchanger is to determine the surface area required for the specified duty (rate of heat transfer) using the temperature differences available. The overall coefficient is the reciprocal of the overall resistance to heat transfer, which is the sum of several individual resistances. For heat exchange across a typical heat-exchanger tube the relationship between the overall coefficient and the individual coefficients, which are the reciprocal of the individual resistances, is given by
1/Uo= 1/ho + 1/hod + {do.ln(do /di)}/2.kw + (do /di).(1/hid) + (do /di).(1/hi)
Where Uo = The overall heat transfer coefficient based on the outside area of the tube, kcal/hr m 2 °C ho = Outside fluid film coefficient, kcal /hr m
2
°C
hi = Inside fluid film coefficient, kcal / hr m2 °C hod = Outside dirt coefficient, kcal / hr m2 °C hid = Inside dirt coefficient, kcal/ hr m 2 °C kw = Thermal conductivity of the tube wall material, kcal / hr m °C di = Tube inside diameter, m do = Tube outside diameter, m
3.1
Shell side flow correlation
The first heat transfer correlation suggested is due to Colburn in 1933 and is of t he form NuD = 0.33 . (Re)0.6 . (Pr)0.33
Where the Nusselt number and Reynolds number is based in the tube outside diameter and the flow velocity is based in the minimum cross sectional area at the shell inside diameter. The
validity was restricted to Re > 2000 and staggered layout. For baffled shells, where bypass and leakage streams diminish the flow effectiveness, a safety factor of 0.6 was introduced. Eventually the correlation form was slightly modified to include the Sieder-Tate form for nonisothermal effects, which is much easier to handle and within the general data scatter produced equal results.
NuD = 0.2. (Re)0.6 . (Pr)0.33 .(visco. / visco wall )
0.14
Where all properties are based on the average bulk temperature except viscosity wall, which refers to tube wall. The validity was restricted to the t urbulent flow of 2000< Re<40000
3.2
Tube Side Correlation
As compared to shell side, the tube heat transfer & pressure drop can be predicted much more accurately and there are large number of correlation for doing this. Many of these correlations have been obtained experimentally for laminar & turbulent flow. The most commonly used relation for calculating the heat transfer coefficient in turbulent fully developed flows in smooth circular pipes is the Dittus Boelter equation.
NuD = 0.023 . (Re) 0.6 . (pr)n In the above equation, the thermo-physical properties are evaluated at the local bulk mean temperature. The exponent n of the Prandtl number has the value 0.4 if the fluid is being heated and 0.3 if the fluid is being cooled. This equation is valid for 0.7 < Pr<100. The Dittus Boelter equation is not applicable to fluids having Prandtl number greater than 100 because of the fact that the viscosity of such fluids changes rapidly with temperature. Sieder Tate have proposed an equation which is similar to the above and is valid under the same conditions. The essential difference is that it incorporates a viscosity correction factor and is therefore valid up to Pr = 16000.
NuD = 0.027. (Re)0.8 . (Pr)0.33 .(visco. / visco wall)
0.14
In this equation all properties are evaluated at the local bulk mean temperature except viscosity wall, which is evaluated at the wall temperature.
3.3
Average Bulk Temperature
Since physical properties, which are required for calculation of heat transfer coefficient, change with the fluid temperature, first aim is to determine average fluid temperature on which the fluid properties will be based. This bulk temperature is defined as the hypothetical temperature at a given cross section in which all fluids flowing past the cross section are completely mixed and in thermodynamic equilibrium. The bulk temperature changes as the fluid heats or cools, thus, the overall heat transfer coefficient requires either Calculating local coefficients at a series of temperatures between the exit and inlet, followed by graphical or numerical integration. Determining a representative average bulk temperature for average physical property evaluation.
Two methods used for defining average bulk t emperature are:
Arithmetic average bulk temperature: it is a simple method, which gives good results in each increment for incremental calculations if the range of physical property variation is not large. Caloric bulk temperature: For non-incremental calculations in which properties vary greatly along the exchanger’s length, this method is used. This is a more representative method used for overall calculations across the entire exchanger.
3.4
Mean temperature difference
The heat transfer rate is directly proportional to the temperature difference, which provides the driving force for the heat transfer. However, if the fluid temperature variation is plotted against the length of the exchanger it is not a straight line. Therefore determining actual mean temperature difference between the two streams engaged in heat transfer becomes very essential. For convenience, the mathematical expression for LMTD is based on the terminal differences, ∆Tc & ∆Th
∆ Tlm = ( ∆Th - ∆Tc ) / ln ( ∆Th / ∆Tc)
Pure counter flow permits extreme temperature overlap without incurring any special penalty for t he above equation. In case of exchangers with more than one pass, ' F' factor is applied which is always less than one.
Usually actual exchangers are combination of True Counter-current, Co-current and Cross-flow. As such it is required to estimate the “true temperature difference” from the LMTD, by applying a correction factor to account the deviation from true counter current flow.
∆Tm = Ft ∆Tlm
Ft = Temperature correction Factor The temperature correction factor is function of shell and tube fluid temperatures, number of shell & tube passes, shell side leakage and bypass streams etc. The maximum value of F t is 1.0 for true counter current flow. The value for F t is above 0.85 for most of the economical designs. The use of two or more shells in series, or multiple shell side passes, will give closer approach to true counter-current flow, and should be considered when temperature cross is likely to occur .
3.5
Boiling range (BR)
Boiling range is nothing but the difference between dew point & bubble point. Boiling range is very important in case of vaporization services. In case of, vaporization of multicomponent mixtures, based on boiling range, multicomponent mixture correction factor is applied to nucleate boiling heat transfer coefficient. This correction factor is applied due to the effect of the composition gradient on the effective ∆T and can be quite severe if the boiling range is large. Hence it is very essential to enter boiling range of vaporizing side in HTRI so that the correction facto applied will be realistic.
4
Selection of Heat Exchangers
E Shell: The E shell is the most common as it is inexpensive and simple. The tubes may have single or multiple passes and are supported by transverse plate baffles.
F Shell: The F shell has a longitudinal baffle, resulting in two shell passes and nearly a pure counter flow. Although ideally this is a desirable flow arrangement, F shell is rarely used in practice because there are many problems associated with the design. It is difficult to remove or replace the tube bundle. With the F shell, there are also additional problems of fabrication and maintenance. Also problems like internal leakage, unbalanced thermal expansion in case of large temperature difference between inlet & outlet are encountered.
G and H shell: The G and H shells are related to the F shell having variants of the longitudinal baffle. The double-split flow H shell is similar to the G shell, but with two inlet and two outlet nozzles (and two horizontal baffles) to accommodate high inlet velocities. The G and H shells are seldom used for shell side single-phase applications, since there is no advantage over E or X shells. They are used as horizontal thermosiphon reboilers, condensers, and other phase-change applications. The H shell approaches the cross flow nozzle arrangement of the X shell, and it usually has low shell side ∆P compared to that for the E, F, and G shells. Drawback of this type of shell is that no support plate can be given, so maximum tube length gets limited.
J shell: The divided flow J shell has two inlets and one outlet, or one inlet and two outlet nozzles. It has approximately one eighth the pressure drop of a comparable E shell, and is therefore used for low pressure drop applications such as a condenser in vacuum.
K shell: The K shell is used for partially vaporizing the shell fluid. It is used as a kettle reboiler in the process industry and as a flooded chiller in the refrigeration industry. They are used when essentially 100% vaporization is required.
X shell: For a given flow rate and surface area, the cross flow X shell has the lowest shell side pressure drop compared to all other shell configurations except the K shell. Hence, it is used for gas heating and cooling with and without finned tubes and for vacuum condensation. It is also used when shell flows are large
4.1
TEMA Front Head Selection
A – Easy to open for tube side access. For low pressure applications B – Single tube side joint. For Higher-pressure applications, preferred with clean tube side fluid. It is less expensive than Type A, but requires piping dismantling of connected piping, removal of gaskets and of integral cover- or bonnet- which is a heavy item. C –Channel to tubesheet joint is eliminated. Tube side is corrosive, toxic or hazardous and when removable tube bundle is required. It is less expensive than A & normally used for low-pressure operations. N – Fixed tubesheet with removable cover plate. For application where tube side is corrosive, toxic or hazardous and shell side fluid is clean and any leakage possibility is to be eliminated. D – Very High-pressure applications
4.2
TEMA Rear Head Selection
Fixed Head (L, M, N) – Should be used when thermal differential expansion of the shell and tubes is low and shell side fluid is clean. (Low Fouling) L – Same as A type front end. For low pressure applications M – Same as B type front end. For Higher-pressure applications, not requiring frequent maintenance. N – For application where tube side is Corrosive, toxic or hazardous and where leakage of shell to tube side fluid and vice versa, is to be eliminated. U tube – It is a removable bundle without floating head. For thermal differential expansion of the shell and tubes is higher and tube side fluid is clean. For high-pressure applications or, with hazardous/ toxic fluid on shell side. P, S, T, W – Should be used when shell side fluid or both shell and tube side fluid are Dirty. (High fouling) P – Pressure is low and shell side fluid is not toxic or hazardous. Where risk of internal flange leakage is to be avoided. Not commonly used. S – Floating head with backing ring. Normal Pressure requirements, relatively lesser maintenance requirements T – Pull through floating head. High-pressure requirements, frequent need to takeout the tube bundle. W – For low-pressure application
4.3
EXCHANGER Type Selection Criteria & Special GUIDELINES
Whenever ∆T on any side is more than 50°C, normally floating head exchangers are used. It is also preferred whenever cleaning on shell side is required. Exchangers such as process gas trim cooler, where ∆T is less than 50°C & also does not require cleaning shell side, fixed head exchangers can be used. Fixed tube sheet with expansion joint can be used f or high ∆T, in cases where shell side cleaning is not required. This is applicable only for shell pressures up to approximately 5kg/cm2g. Exchangers where very little pressure drop is available on shell side, ‘ J ’ type of shell can be used. Exchangers with good ∆T & no cleaning required on tube side are provided with ‘ U ‘ type of bundle. In case of floating head exchangers with single pass on tube side, expansion bellows are necessary. For e.g. BFW exchangers where normally temperature cross is observed, In these kind
of exchangers’ liquid outlet is given from bottom in order to have free draining. Hence normally expansion bellows are given in the inlet side. Vertical floating head exchangers are provided with floating head on bottom side for free expansion of the shell. Whenever partial pressure of hydrogen is more than 5 bar on any side, tubes should be welded to tube sheet to prevent leakage of hydrogen to the medium at the other side. It is also specified as ‘ strength weld’ on datasheet. Exchanger who requires frequent cleaning on shell side are given with ‘S’ type or ‘T’ of rear head. For example coker naphtha on shell side.
5
Tube Side Design
5.1 Tubes Shell & tube exchangers have either of the two types of tubes, namely (a) plain or bare and (b) finned – external or internal. Plain tubes are most common ones, available in Carbon steel, carbon steel alloys, stainless steel, Aluminum, Aluminum alloys, copper, brass and alloys, cupro-nickel, nickel, monel, titanium, tantalum, carbon, glass, and other special materials.
5.2 Diameter Tube diameters in the range 0.625 in (16 mm) to 2 in (50 mm) are used. The smaller diameters 0.625 in. to 1 in. (16 to 25 mm) are preferred for most duties, as they will give more compact, and therefore cheaper, exchangers. Larger tubes are easier to clean by mechanical methods, and are more rugged, they should be selected for heavily fouling fluids. For mechanical cleaning, the smallest practical size is 0.75 in. (19.05mm). For chemical cleaning, smaller sizes can be used provided that the tubes never plug completely. The frequently used Tube OD is 0.625 in. (16mm), 0.75 in. (19 mm), 20mm, 1.0 in. (25mm) or 1.25 in. (32mm).
5.3 Tube Thickness The tube thickness is selected to withstand the internal pressure and give adequate corrosion allowance. The commonly used thickness corresponds to 20, 18, 16, 14, 12 and 10 BWG, i.e. 0.89mm, 1.25mm, 1.65mm, 2.11mm, 2.77mm, and 3.40mm. Typical tube thickness used are 2 mm or 2.5 mm for 20 mm or 25 mm tube OD for CS. In case of alloyed steel, tube thickness used can be 1.6 mm. For titanium it can be as low as 1 mm.
5.4 Tube Length Tube length used should be maximum as possible so that minimum number of tubes are required which ultimately lead to a smaller size of exchanger. It is also useful in utilising maximum pressure drop available. The tube lengths for both straight and U- tube heat exchangers commonly are 6 ft. (1.829 m), 8 ft. (2.438 m), 10 ft. (3.048 m), 12 ft. (3.658 m), 14 ft. (4.267), 16 ft. (4.877m), 20 ft (6.096 mm). Not normally used, but longer exchangers are made with tube up to 40 to 80 ft. (12.2 to 24.4 m). Tube length also depends on clients' preference such as 6m or 9 m.
5.5
Tube Arrangement
1.1.5
Tube-Side passes
Number of passes can be defined as number of times material flows along the exchanger. The fluid in the tube is usually directed to flow back and forth in a number of “passes” through groups of tubes arranged in parallel to increase the length of the flow path. The number of passes is selected to give the required tube-side design velocity and taking maximum advantage of available pressure drop. Higher velocities in the tube result in higher heat transfer coefficients, at the expense of increased pressure drop compared to that for low velocity. Exchangers are built with from one to up to about sixteen tube passes.
The standard design has one, two or four tube passes. In multi-pass designs, an even number of passes are generally used; odd passes are uncommon, and may result in mechanical and thermal problems in fabrication and operation.
1.1.6 Tube pass geometry There are three main types of tube passes. Ribbon, Quadrant, H – band The number of passes in an exchanger is the number of times one fluid passes though the other fluid compartment.
Ribbon
Quadrant
H - Band
1.1.7
Tube layout patterns
The tubes in an exchanger are usually arranged in a triangular (30°), rotated triangular (60°), square (90°), or rotated square (45°) pattern. For identical tube pitch and flow rates, the tube layouts in decreasing order of shell side heat-transfer coefficient and pressure drop are 30°, 45°, 60°, and 90°. Thus the 90° layout will have the lowest heat-transfer coefficient and the lowest pressure drop.
The triangular pitch (or rotated triangular) pattern will accommodate more tubes than a square (or rotated square) pattern and provide a more compact arrangement, usually resulting in a smaller shell. Furthermore, a triangular pattern produces high turbulence and therefore a high heat transfer coefficients. However, at the typical tube pitch of 1.25 times the tube OD, it does not permit mechanical cleaning of tubes, since access lanes are not available. Consequently, a triangular layout is limited to clean shell side service. For services that require mechanical cleaning on the shell side, square pattern must be used. Chemical cleaning does not require access lanes, so a triangular layout may be used for dirty shell side services provided chemical cleaning is suitable and effective.
A rotated triangular pattern seldom offers any advantage over a triangular pattern, and its use is not very popular.
For dirty shell side services, a square layout is typically employed. Here, a minimum-cleaning lane of 6.35mm (1/4 in) is provided. The triangular pitch is generally used in the fixed tubesheet design because no cleaning is needed.
Where mechanical cleaning is required, the 45° layout is preferred for laminar or turbulent flow of a single-phase fluid, and for condensing fluid on the shell side. For dirty fluids with low shell side Reynolds number (<2000) its usually advantageous to use this type as it produces higher turbulence and higher efficiency.
If the pressure drop is constrained on the shell side, the 90° layout is used for turbulent flow. Where boiling is necessary, the 90° layout, which provides vapour escape lanes, is preferred. As described earlier, fixed-tubesheet construction is usually employed for clean services on the shell side, U-tube construction for clean services on the tube side, and floating-head construction for dirty services on both the shell side and tube side. For clean services on both
shell side and tube side, either fixed tubesheet or U-tube construction may be used, although Utube is preferable since it permits differential expansion between the shell and tube. Hence a triangular tube pattern may be used for fixed-tubesheet exchanger and square (or rotated square) pattern for floating head exchangers. For U-tubes exchangers, a triangular pattern may be used provided the shell side stream is clean and a square (or rotated square) pattern if it is dirty.
1.1.8
Tube Pitch
Tube pitch is defined as the shortest distance between two adjacent tubes. For a triangular pattern, TEMA specifies a minimum tube pitch of 1.25 times the tube OD. Thus, a 25 mm tube pitch is usually employed for 20mm OD tubes and 31.25 is for 25 mm tube. For square pattern, TEMA additionally recommends a minimum-cleaning lane of 0.25 in. (6mm) between adjacent tubes. Thus the minimum tube pitch for square patterns is 1.25 times the t ube OD, or the tube OD plus 6 mm, whichever is larger. For example, 20mm tube tubes should be laid on a 26mm (20mm + 6 mm) square pitch, but 25mm tubes should be laid on a 31.25 mm (25 X 1.25) square pitch. As far as thermal hydraulics is concerned, the optimum tube-pitch to tube OD ratio for conversion of pressure drop to heat transfer is typically 1.25 to 1.35 for turbulent flow and around 1.4 for laminar flow. One can decrease shell side pressure drop by increasing the tube pitch, however, the same is not recommended as it increases the shell diameter. Additionally adopting other methods to reduce pressure drop will result in a cheaper design. Triangular pitch typically holds 15% more tubes than square pitch for same OTL, tube diameter & pitch size. 6
Shell Side Design
6.1 Baffles Baffles are used to support tubes, enable a desirable velocity to be maintained for the shell-side fluid, and prevent failure of tubes due to flow-induced vibration. These are provided in the shell to direct the fluid stream across the tubes, to increase the fluid velocity and so improve the rate of heat transfer by increase in turbulence. They provide support to tubes and minimize the tube vibrations. Support baffles are also provided for extreme vibration condition. They are used to control overall flow direction of Shell fluid (Longitudinal baffle, F shell). They protect the tubes in the top row from erosion, cavitation, and vibration due to the impact of high velocity fluid jet from the nozzle to the tubes (Impingement Baffle).
Here we will discuss baffles primarily with respect to its role in Heat Transfer. The limiting factor for baffle selection is usually the pressure drop and the flow induced vibrations. 1.1.9
Types
Baffles may be classified into primarily two types transverse and longitudinal. The transverse baffle is the one, which is responsible to direct the shell side fluid across the tubes and thus mainly responsible for bringing effective heat transfer. The purpose of longitudinal baffles is to control the overall flow direction of the shell fluid. For example, F, G, and H shells have longitudinal baffles. 6.1.1.1 Longitudinal Baffle These are plates inserted in the shell longitudinally for the purpose of controlling the overall flow direction of the shell fluid as in ‘F, G, and H’ shells. Their major disadvantage is because of increased leakage that may be caused if it is not sealed properly. It is also difficult to remove or replace the tube bundle with such baffles. Although it helps bringing nearly counter current flow but its use is restricted to low pressure drop and low temperature difference applications. 6.1.1.2 Transverse Baffle The transverse baffles may be classified as plate baffles and rod (or strip) baffles. The plate baffles are used to support the tubes, to direct the fluid in the tube bundle at approximately 90° to the tubes, and to increase the turbulence of the shell fluid. T he rod baffles are used to support the tubes and to increase the turbulence of the shell fluid. The flow in a rod-baffled heat exchanger is parallel to the tubes, and the problem of flow-induced tube vibration is virtually eliminated by the baffle support of the tubes. The choice of baffle type, spacing, and cut are largely determined by flow rate, allowable pressure drop, tube support, and flow-induced vibration. 6.1.1.3 Plate Baffles The two types of plate baffles are either segmental or disk-and-doughnut. The single- and double-segmental baffles are most frequently used. The single-segmental baffle is generally simply referred to as a segmental baffle. The segmental baffle is a circular disk (with tube holes) with a segment removed. The baffle cuts, expressed as the percentage of the shell inside diameter, vary from 15 % to 45 %, the most common being 20 – 35 %.
Even though one of the major functions of the plate baffle is to induce cross flow (flow normal to the tubes) to improve heat-transfer performance, this objective is only approximated. Various clearances are required for manufacturing. Fluid can leak through these clearance passages and reduce the heat-transfer effectiveness. Three clearances associated with a plate baffled
shell-and-tube exchanger are tube-to-baffle clearance, baffle-to-shell clearance, and bundle-toshell clearance.
If the primary function of baffles is to support the tubes, they are referred to as support plates. They may be thicker than the baffle, have less tube-to-baffle clearance, and provide greater stiffness to the bundle.
The double-segmental baffle, also referred to as a strip baffle, provides lower shell side pressure drop than that for the single-segmental baffle for the same unsupported tube span. Hence an exchanger with this type of baffle can handle larger fluid flows on the shell side. Multisegmental baffles have a strong parallel flow component, provide a lower pressure drop, and permit closer tube support to prevent tube vibrations.
6.1.1.4 Segmental Baffles with No tubes in Windows. This kind of arrangement mainly comes in to picture because of vibration problem. The lower pressure drop asks for large baffle spacing. Since the tubes in the window zone are supported at a distance of two or more times the baffle spacing, these are most susceptible to vibration. To eliminate the susceptibility of tube vibrations and to reduce the shell side pressure drop, the tubes in the window zone are removed and support plates are used to reduce the unsupported span of the remaining tubes. The resultant design is referred to as the segmental baffle with notubes-in-window (NTIW). The low velocity regions in the baffle corners do not exist for NTIW, resulting in good flow characteristics and less fouling. Thus the loss of heat-transfer surface in the window region is partially compensated. However, the shell size must be increased to compensate for the loss in surface area in the window zone. If the shell side operating pressure is high, this no-tubes-in-window design is very expensive. This can only be envisaged for twophase or gas as shell side fluid.
6.1.1.5 The disk-and-doughnut baffle The disk and doughnut baffle is made up of alternate disk- and doughnut-shaped baffles. Generally, the disk diameter is larger than the half shell diameter, and the diameter of the hole of the "doughnut" is smaller than the half shell diameter. This baffle design provides a lower pressure drop compared to a single-segmental baffle for the same unsupported tube span, and eliminates the tube bundle-to-shell bypass stream C.
The disadvantages of this design are: (1) all tie rods to hold the baffles are within the tube bundle, and (2) the central tubes are supported by the disk baffles, which in turn are supported only by tubes in the overlap region of the larger diameter disk over the doughnut hole.
6.1.1.6 Rod Baffles. A rod baffle is made up of parallel rods (a rod matrix) mounted on a baffle ring. The rod diameter is equal to the tube spacing (zero clearance), so that the baffle touches and constrains all tubes on four sides. Thus, in a square layout, two vertical baffles constrain all tubes from two sides . The following two horizontal baffles constrain all tubes from the remaining two sides (90° to the first side). A set of four-rod baffles (two vertical and two horizontal) is repeated in t he exchanger. The baffle spacing is generally kept at 0.152 m (6 in) for single-phase applications and 0.305 m (12 in) for condensers and reboilers. Generally the tube diameters and tube pitches are selected so that the rod diameter is 4.8 mm (3/16 in) or 6.4 mm (1/4 in). The rod baffle exchanger has several advantages over the plate-baffled exchanger, as follows: (1) it eliminates flow-induced tube vibrations since the tubes are rigidly supported at four successive points. (2) The pressure drop on the shell side is about one-half of that with a double-segmental baffle at the same flow rate and heat transfer rate. Alternatively, the rod baffle exchanger will result in a smaller unit for the same heat transfer and pressure drop. (3) There are no stagnant flow areas with the rod baffles, resulting in less fouling and corrosion, and improved heat transfer compared to that for a plate baffle exchanger. However, the rod baffle exchanger will require longer tubes of smaller diameter. If the tube side fluid is controlling and has a pressure drop limitation, the rod baffle exchanger may not be applicable. This is not common in the refinery application since the fluid is highly fouling and minimum diameter is required ¾”.
6.1.1.7 Impingement Baffle (Impingement Plate) Impingement baffles or plates are generally used in the shell side just below the inlet nozzle. They protect the tubes in the top row from erosion, cavitation, and vibration due to the impact of high velocity fluid jet from the nozzle to the tubes. The most common cause of tube f ailure is the improper location and size of the impingement plate. One of the most common forms of this baffle is a solid circular plate located under the inlet nozzle just in front of the first tube row. The location of this baffle within the shell is critical in order to minimize the associated pressure drop and high escape velocity. For this purpose, adequate areas should be provided both between the nozzle and plate and between the plate
and tube bundle. TEMA Standards specify an impingement protection is required for following cases: For Non abrasive, Single phase Fluids with entrance line values of ρV2 exceeding 1500 (2232). For all other Liquids, including those at its boiling point ρV2 exceeding 500 (744). For all other gases and vapors, including nominally saturated vapors, and for liquid –vapor mixtures. ρV2 the shell or bundle entrance or exit area should be such that they do not produce a value of
in excess of 4000 (5953). Where, ρ = Density in Pounds per Cubic Feet (Kilograms per Cubic Meter)
V = Linear Velocity Feet per second (Meters per Second) 6.1.2
Baffle Spacing (Number of Baffles)
Baffle Spacing is Centerline to centerline distance between adjacent baffles. This is the most vital parameter in STHE design, as it effects cross flow velocity (Velocity across the tubes), therefore heat transfer and pressure drop. This also effects the unsupported tube length and thereby vibration.
The TEMA standards specify for Segmental baffles minimum baffle spacing as one-fifth of the shell inside diameter or 50.8 mm (2 in.) whichever is greater. Closer spacing results in poor bundle penetration by the shell side fluid and difficulty in mechanical cleaning of outside of tubes. A spacing of 50.8 mm (2 in) is required for cleaning the bundle. Spacing closer than one fifth of the shell diameter provide added leakage, which nullifies the heat-transfer advantage of closer spacing. If the foregoing limits on the baffle spacing do not satisfy other design constraints such as D pmax or tube vibration, no-tubes-in-window or pure cross-flow design should be tried.
The maximum baffle spacing is the shell inside diameter. Higher baffle spacing will lead to predominantly longitudinal flow, which is less efficient than cross flow, and large unsupported tube spans which will make the exchanger prone to tube failure due to flow induced vibration. TEMA gives the maximum permissible unsupported tube spans for tube support plates (or Baffles acting as support plates) specified in TEMA standards.
Baffle normally shall be spaced uniformly, spanning the effective tube length. When this is not possible, baffles nearest the ends of the shell, and/ or tubesheets, shall be located as close as practically possible to the shell nozzle, and the remaining baffle are spaced normally. When the baffle spacing is reduced the pressure drop increases at a much faster rate than does the heat transfer coefficient. Thus the optimum ratio lies when there is highest efficiency of conversion of pressure drop to heat transfer. This optimum ratio is normally between 0.3 to 0.6 of shell ID. The number of baffles required gets fixed based on selected tube length and baffle spacing.
6.1.3
Baffle Cut
Baffle Cut is the height of the segment that is cut in each baffle to permit the shell side fluid to flow across the baffle. This is expressed as a percent of the Shell ID. It can vary between 15 to 45% of shell ID in case of single segmental baffle. Both very small and very large baffle cuts result in inefficient heat transfer on the shell side. The baffle cut and spacing should be designed such that the flow velocity is approximately the same for the Cross-Flow (i.e. Flow across the Tubes) and Window Flow (i.e. through the Baffle cut area). It is preferable to have the same within 20%of each other. At large baffle spacing, the percent cut may approach 45 – 49 % in order to avoid excessive pressure drop across the windows compared to the bundle. It is strongly recommended that baffle cuts between 20% to 35% be employed for single segmental baffle. In case parameters such as Pressure drop or heat transfer do not meet other methods should be adopted rather that changing baffle cut beyond these values, such as using double segmental baffles or divided flow or even cross flow. Large or small spacing coupled with large baffle cuts is undesirable because of the increased potential of fouling associated with stagnant flow areas. If fouling is of prime concern, the baffle cut should be kept < 25 % to maintain high velocity through the window zone. Generally, a baffle cut of 20 to 25 percent will be the optimum, giving good heat-transfer rates, without excessive drop. For Double and Triple segmental baffle one should try to select the cut such that equal flow area exists at each baffle position. One needs to take extra care for specifying baffle cut for cases with no tubes in Windows. 6.1.4
Baffle cut orientation
Baffle cut orientation can be Vertical, Horizontal or Inclined (45°C).
The direction of the baffle cut is selected as either horizontal or vertical for a single-phase fluid (liquid or gas) on the shell side. Inclined baffle orientation may as well be selected. For a very viscous liquid, the direction of the baffle cut should be horizontal for better mixing. The direction of the baffle cut is selected as vertical for the following shell side applications: Condensation (for better drainage) Evaporation/boiling (to promote more uniform flow) Entrained solids in liquid (to provide the least interference for solids to fall out) Multishell pass exchangers (includes F shell, because of ease of fabrication).
Depending upon baffle cut direction the horizontal pass partitions will have different effect. For vertical baffle cuts Horizontal partitions are parallel to flow direction Vertical partitions are perpendicular to flow direction For Horizontal baffle cuts Horizontal partitions are perpendicular to flow direction Vertical partitions are parallel to flow direction Thus while selecting tube layout one should take care of baffle cut orientation and try to minimize the ‘F’ Stream. Depending on the same one can decide the number of ‘F’ stream seal rods that are required. With inclined (45°C) Baffle cuts one can assume that F stream by pass as negligible. This type should however be used with single shell pass and preferable single-phase fluid on shell side. Effect of baffle cut orientation is also dependent of inlet outlet nozzle position. The inlet outlet nozzle orientation should be selected in order to ensure shell side fluid properly passes through the tube bundle. For example if the inlet nozzle is in vertical position it is better to have horizontal single-segmental baffles, however this is not the governing criterion. One should be cautious while placing the first baffle with respect to inlet-outlet nozzle position. It should be placed such that minimum bypassing occurs.
6.1.5
Baffle Clearances
Baffles have maximum impact on the various shell side flow streams. The selection of baffle cut, orientation, spacing and clearances can decide how effective is the actual heat transfer across the exchanger.
Shell to Transverse Baffle Clearance This is the factor which brings about the Leakage ‘E’ stream (discussed lr). It permits double the value, for cases where its impact on Mean Temperature Difference and shell side coefficient is not significant. But usually these values are assumed for all practical purposes. The baffle-toshell clearance should be kept to a minimum to minimize the E leakage stream.
Tube to Transverse Baffle Hole Clearance This is the factor, which brings about the Leakage ‘A’ stream. The tube-to-baffle hole clearance should be kept at a minimum to reduce tube vibration and resultant damage as well as to minimize the 'A' leakage stream. This stream is fairly effective in heat transfer as it is contact with the tubes. Its flow is not only dependent on the clearance but also on the baffle thickness. All support plates should also be accounted for. TEMA indicates the minimum Baffle and Support plate thickness depending on shell ID and plate spacing. The clearance in terms of the difference between the baffle hole ID and tube OD is 0.40 mm (1/64 in)
for unsupported tube length L > 0.91 m (36 in)
or for the tube OD 31.8 mm (1.25 in) or smaller; 0.80 mm (1/32 in)
for the unsupported tube length L < 0.91 mm (36 in)
or for the tube OD > 31.8 mm (1.25 in). If the baffles are acting as support plates they may be thicker than the baffle, have less tube-tobaffle-hole clearance, and provide greater stiffness to the bundle.
6.2
Sealing Trips
Sealing trips are trips placed along the shell to reduce the bundle to shell leakage and improves the shell side heat transfer coefficient. In some cases like floating head type, because of floating head bonnet flange and bolt circle, many tubes are omitted from the tube bundle near the shell. This results in the largest bundle-to-shell clearance and the shell must be made greater than in the fixed and U-tube designs to accommodate the floating-head flange. This allows the shell side fluid to bypass the tubes. In order not to reduce the exchanger performance, sealing strips (or dummy tubes or tie rods) in the bypass area are essential. However, localised high velocities near the sealing strips could cause flow-induced tube vibration; hence, proper care must be exercised for the design.
6.3
Shell Flow distribution
Fluid flows through the shell in 5 paths: A stream: Through gap between tube & baffle. This fraction is also effective for heat transfer since it is in contact with tube surface.
B stream: Between tubes across the bundle. Most effective stream for heat transfer. C Stream: Through gap between bundle & shell. This stream is partially effective in heat transfer since it contacts the outer surface of the bundle
E stream: Through gap between baffle & shell. Not effective in heat transfer. F stream: Through pass partition lane. Partially effective in heat transfer. Out of these fluid flow streams, cross flow (B stream), mainly leads to good heat transfer. In good design this stream is maintained between 0.3 to 0.7 in order to minimise shell side flow leakage. ‘A’ stream can be minimised by giving tight tolerance to tube OD & baffle hole OD. ‘E’ stream can be minimised by reducing the tolerance between bundle & shell ID. ‘C’ stream, which is also called as bypass, is reduced by introducing sealing strips. Sealing strips are normally given longitudinally between first baffle to last baffle. ‘F’ stream is reduced by introducing dummy tubes or seal rods. They physically block the flow but have no holes in the tubesheet. No tube side fluid passe s through them.
Flows ( percent ) for industrial exchangers ( HTRI data) Stream Cross flow (B) Bypass (C+F) S - B Leakage T - B leakage
Turbulent 40 – 70 % 15 – 20 % 6 – 20 % 9 – 20 %
Laminar 25 – 50 % 20 – 30 % 6 – 40 % 4 – 10 %
7 Nozzles 7.1 Shell Side 1.1.10 Size Standard pipe sizes will be used for the inlet and outlet nozzles. Nozzle sizes are related to the piping size, the exchanger shell design, and the escape flow area into the tube bundle. This
escape area should have ρ v2< 6000kgm –1s –2 (4000 lbmft –1s –2). It is the total free area between a nozzle and the projected area on the tube bundle. Sometimes tubes may be removed to obtain sufficient area. One should also take care that the percentage pressure drop across the exchanger nozzles is not very high as this drop is not utilised in heat transfer, thus giving loss in its efficiency.
Nozzles have an impact on the exchanger thermal design as they may influence its geometry. In case if required nozzle size is very large it may be helpful and economical to use TEMA ‘J’ or ‘H’ type of shell.
Another important feature of nozzle size is its impact on inlet-outlet baffle spacing. Large inlet or outlet baffle spacing may cause vibrations related problems and a well-designed exchanger may turn out to be unsuitable because of this reason. Therefore one should take this factor into account before getting to finer optimization details.
1.1.11 Orientation Another important parameter is inlet nozzle orientation. It is usually dependent on the baffle cut orientation. If the baffle cut is parallel to the centerline of the nozzle, bypassing of the inlet and outlet baffle spacing can take place. If its necessary to select this arrangement one should take appropriate penalty factor as the performance of unit gets reduced. The direction of shell side flow with respect to gravity is also dependent on inlet nozzle location. For some cases, especially for 90°-tube pattern with laminar f low on shell side, performance can be enhanced or reduced due to natural convection and other effects. One should be cautious while placing the first baffle with respect to inlet-outlet nozzle position. It should be placed such that minimum bypassing occurs. In case of Horizontal baffle cut and top/bottom inlet nozzle arrangement depending upon whether selected baffles are odd or even in number the outlet nozzle position has to be determined. With vertical cut baffles and side inlet nozzle arrangement, outlet nozzle position should accordingly be decided. For other cases ease of flow should be given preference. One may select inlet-outlet nozzle position such that the hot fluid travels from top to bottom (such as condensers) and cold fluid from bottom to top (such as reboilers). This however should be selected only if other parameters do not get effected.
1.1.12 Location Nozzle location is very important in case of stacked exchangers, it is desired that the exchangers are designed such that the exchangers can be mounted directly nozzle to nozzle. Depending on shell type selected the inlet-outlet nozzle location are on same sides, opposite sides or in the center. When in need of stacked exchangers, one should take care of this factor while selecting exchanger type. Depending upon whether first tube pass is required to be co-current or counter, the shell side inlet nozzle location is on same side or opposite side of the tube side inlet nozzle. This factor can often influence the Mean temperature difference. For 1-1 pass exchangers shell side tube side inlet nozzle should be placed on opposite sides. When ‘U’ tube or Floating head full support plate is located care should be taken to locate the nozzle (if at all present) before (i.e. on the side of shell) the support plate. The spacing should be specified accordingly.
7.2 Tube side Standard pipe sizes will be used for the inlet and outlet nozzles. One should take care that percentage drop across tube side nozzles is not very high as compared to that on exchanger as this pressure drop is not utilized for heat transfer. Tube side nozzles are generally located on the circumference, orientation depending on tube passes and partition plate locations. They may also be located axially if required. Here again one should try to select inlet-outlet nozzle position such that the hot fluid travels from top to bottom (such as condensers) and cold fluid from bottom to top (such as reboilers). In case of odd number of passes the inlet-outlet nozzles are located on opposite sides. However, most of the times the number of passes are even and they are located on the same sides.
8 Miscellaneous 8.1 Outer tube limit (OTL): The outer tube limit contains all tubes of the tube layout. Distance between OTL & shell diameter is very important & is different for floating head, fixed head or U tube bundle. For values refer TEMA.
8.2 Tie Rods and Spacers The baffles and support plate are held securely in position by tying them with tie rods and spacers. Their number is important in thermal design, as spaces for them need to be kept in the
layout, thus effecting the overall geometry. The number of rods required will depend on the shell diameter. The recommended numbers for a particular diameter are tabulated in accordance to that of TEMA.
9
Vibrations
Fluid flowing through a heat exchanger can cause the heat exchanger tubes to vibrate. The tube vibration problem is complicated because it concerns fluid dynamics, structural dynamics, and the mechanical properties of the metals involved.
Different types of vibration mechanisms are as follows. Fluid elastic instability. Vortex shedding Acoustic resonance Turbulent buffeting Flow pulsation
Fluid elastic instability is important for both gases & liquids. This is not a resonance effect & occurs above a critical flow velocity. In this case increasing the flow velocity makes vibrations much worse. There are different methods to avoid fluid elastic instability such as increase tube natural frequency by decreasing the span lengths or increasing the tube diameter. Sometimes reducing clearance between tube & baffle or increasing tube pitch also helps in minimizing tube vibrations. After trying all these options, if still vibrations exist, tubes in the window region can be removed so t hat all tubes are supported & no tube has double length span. If fluid elastic instability is found at the shell entrance, then height under inlet nozzle can be increased. If still vibration is not going then rod type of impingement baffle can be used. However, HTRI vibration analysis uses longest tube span at the inlet space (inlet baffle spacing + central baffle spacing) for the bundle entrance analysis. This little conservative with perpendicular cut baffle, in which tube span at the inlet is equal to inlet baffle spacing. In such cases, in order to get more realistic bundle entrance velocities & amplitudes, exchanger can be assumed to be NTIW design with same number of tube spans. But here care should be taken that, it is only for checking bundle entrance/exit velocities for perpendicular baffle cut exchangers. Vortex shedding is caused by the periodic shedding of the vortices from the tubes and can lead to damage of tubes if vibrations coincide with the tube natural frequency. Some measures such as changing span lengths can be taken to avoid vibrations.
Acoustic resonance is very important in case of gases. It occurs when the frequency of an acoustic wave in the heat exchanger coincides with tube natural frequency. Even if acoustic wave does not cause any vibrations, it can lead to intolerable noise.
It can be avoided by changing acoustic wave by changing span lengths. Generally deresonating baffles are used for avoiding acoustic vibration problem. A general criterion for placing deresonating baffle is parallel to direction of cross flow & generally it is placed 5 % off center.
Turbulent buffeting mechanism is very important in case of two-phase flow. The turbulence in the flowing fluid contains a broad range of frequencies and can coincide with the tube natural frequency to cause tube vibrations.
Flow pulsation is because of periodic variations in the flow. This can become very important in case of two-phase flow or reciprocating machine discharge exchangers.
Vibration output is generally checked for following specific points in HTRI output:
•
Unsupported lengths span to TEMA maximum span should not exceed 0.8 at the designing stage at any point of time.
•
Fluid elastic instability is very important vibration problem, it should be checked for following ratios.
Baffle tip cross velocity ratio < 0.8 Average cross flow velocity ratio < 0.8
• Acoustic vibrations : Vortex shedding ratio & turbulent buffeting ratio should not be more than
0.8 & Chen number to be less than 1300 •
For tube vibration check: Vortex shedding ratio should be less than 0.5. If not then, cross flow amplitude is checked for its ratio with tube gap to be less than 0.1.
•
In any case, cross flow RHO-V-SQR value in any pf the span should no exceed 3000 kpa.
10 Heat Exchangers Optimization Heat exchanger thermal checking can be of two different kinds, rating and design. In all the two cases basic guidelines are essentially same.
10.1 Rating Rating is nothing but performance estimation for a given exchanger geometry where process conditions & heat balance is known & check is made to see whether heat exchanger will perform to the desired conditions.
Thus we can say that rating is nothing but identification of resistance distribution, investigation of pressure drop utilisation & its distribution, checking shell side flow distribution & carrying out vibration analysis.
10.2 Sizing Sizing is a series of sequential ratings. It basically starts with allocation of fluid, shell type, baffle design, bundles, tube diameter, length, layout etc. It involves optimization of exchanger to get desired results in best possible way. Thus optimization involves minimising the size of the unit, minimise initial cost, operating cost (pressure drop), maintenance cost (velocity, pressure drop
In both the case of thermal design, over design is required to consider to absorb any upset to the process condition. This is the additional performance of the exchanger over desired performance. Overdesign can be given on the surface, or duty or flow as per requirement.
10.3 Optimization (based on a technical paper) The word “Optimum” is originally Latin, meaning “best” in case of a shell & tube heat exchanger, optimum process design would mean that combination of various parameters yielding the most suitable piece of equipment to perform t he given duty.
The parameters under designer’s control may be of different types:
Purely process parameters such as inlet and/or outlet temperatures of one or both fluids, flow rate/s of one or both fluids etc. Performance related parameters such as pumping energy cost of one or both fluids, amount of heat to be exchanged with reference to the overall process. Geometric parameters such as shell diameter, tube length, tube diameter, tube pitch, type of tube layout (square, triangular or their rotated versions), number of shell passes (with longitudinal baffle/s), number of tube passes, type and number of baffles, baffle cut, etc.
The parameters of type (a) and (b) as above form a part of optimization process where “Pinch Technology” is now days employed. This is a new area dealing with the optimization of the entire plant and not just one heat exchanger. The topic of this paper is optimization of single equipment and hence these parameters are not discussed here.
Only the geometric parameters are
considered. However, there are some overlapping areas where the parameters of (a) and (b) type are relevant. They are considered only in those contexts.
Coming back to the definition of optimum once again, the “most suitable” selection becomes a vague reference.
It is therefore necessary to first pinpoint what is most suitable or desirable.
Generally, minimum or lowest cost is taken as the most desirable attribute of the design.
Here again, depending upon which cost is to be minimized, there are two ways a design can be optimized.
Annualized total cost to be minimized. Only fixed cost to be minimized.
A)
Minimization Of Annualized Total Cost
Purchase and installation of a heat exchanger requires some capital expenditure. This is needed only once. Therefore, this is called as the fixed cost of the exchanger.
This fixed cost, in-turn, implies certain annual cost. For the purchase of the heat exchanger, at least part of the capital is to be borrowed. Interest is to be paid on the borrowings. This payment is to be made every year till the loan is repaid. This becomes the annual cost.
On usage of the equipment, its value depreciates. Hence, depreciation also becomes annual cost. Then again, there is repairs and maintenance to be provided. In case of heat exchangers, this cost is usually low. In some cases, especially in corrosive service, repairs and maintenance cost could be significant.
All these annual costs are related with fixed cost. Usually these annual costs are expressed as percentage of fixed cost of the heat exchanger.
The fixed cost of the exchanger is strongly related with its heat transfer surface.
It is general
accepted that tubes form the single largest cost component in a shell & tube heat exchanger. As all the heat transfer surface is provided in the form of tubes, it is no wonder that the fixed cost of a heat exchanger is directly proportional to its area.
The next costliest components in a heat exchanger are tube sheets and main flanges. Cost of these components depends on the diameter of the heat exchanger.
The diameter of a heat
exchanger is in turn a function of its area. Thus, major cost components in a heat exchanger are all functions of its heat transfer area.
The concepts of the foregoing paragraphs indicate that annualized fixed cost of a heat exchanger is a direct function of its area. If this relationship is plotted in the form of a graph, the curve appears as shown in figure-1. At zero area, the cost is naturally nil. As the area of the heat exchanger increases, its cost also increases. This establishes the shape of the curve.
Operating a heat exchanger necessarily involves flow of two fluids through it.
The flow is by
pumping (for liquids) or by blowers or compressors (for gases). Gravity flow for liquids and pressure flow for vapours and gases are not uncommon but in majority of cases, at least one of the fluids is pushed through the heat exchanger. The passage of the fluid through the heat exchanger means a drop in its pressure. This pressure drop plus the pressure drop in the pipe leading to and after the heat exchanger is to be made up by a pump or a blower. Thus, this pushing through the heat exchanger involves spending energy. Cost of this energy becomes the operating cost of the heat exchanger.
The parameters of the type (b) mentioned in the previous topic thus become relevant in considering the variable cost. This is the reason they were called performance-related parameters.
Heat exchanger is static equipment.
Therefore, the only variable cost (i.e. the operating cost)
comes not from operating the equipment itself but from operating other machinery to make use of the static equipment!
From flow of fluids concept, it is known that the pressure drop for a fluid to flow through a pipe is inversely proportional to the fifth power of its (pipe’s) internal diameter. Common sense indicates that the area of a heat exchanger is proportional to the square of its shell diameter. Combining these two, it stands to reason that the operating cost (i.e the pumping energy cost) of a heat exchanger is inversely proportional to approximately the 2.5 th power of its heat transfer area. This, of course, is over simplification. Tube side operating cost is also a function of the number of tube passes. This fact introduces some amount of non-linearity in the operating cost curve.
The curve of operating cost as a function of heat transfer area is shown in figure-2. At zero area, (the area for the flow of fluids is zero and hence) this cost becomes infinite. At the other extreme, at
infinitely large area, this cost becomes zero. This establishes the asymptotic shape of the variable cost curve. The shape resembles what is known as a “rectangular hyperbola”.
Adding the annualised fixed cost to the annual operating cost gives the total annual cost. This total cost is to be minimized for optimum process design of the heat exchanger. Combining figure 1 and 2 cost into figure-3, the total cost function would appear to be the curve 3. For ready reference, curves 1 and 2 are also given in the same figure. There is a minimum on the curve 3. Object of the process designing activity is to locate this minimum. Practically, it is not possible to reach the exact minima. This is so because the parameters do not vary continuously but have discrete values. Attempts are then made to reach as near the minimum as possible. The way to reach this is given in the following paragraphs.
It is assumed that the flow rates and inlet-outlet temperatures are fixed by process requirements. Thus, we are assuming that the type (a) parameters of the previous topic have been frozen. This assumption is reasonable because the optimization of the process takes place before optimization of the individual equipment.
A particular value of pressure drop is then assumed for each fluid. Each assumed value of pressure drop corresponds with a particular annual operating cost.
With the assumed pressure drops, minimum heat transfer area solution is worked out.
Fixed cost for this solution is calculated. With the know percentage of the fixed cost as the annual cost, the annual cost is also worked out. The total annual cost is then worked out and recorded.
A new value of pressure drop is now assumed for one of the sides, (say, shell side) keeping the other side pressure drop (tube side) as constant. Entire set of calculations is repeated and the total annual cost is again recorded. This procedure is repeated a number of times with changed values of the shell side pressure drop, keeping the tube side pressure drop constant. Among the recorded annual cost values, minimum is located.
The pressure drop value (for shell side here)
corresponding with this minimum is the optimum pressure drop for this (shell) side.
This (optimum) value of (shell side) pressure drop found as above is now kept constant and the other (tube) side pressure drop is now varied for optimization. By several trials, optimum value of pressure drop for the second (tube) side is found out.
Earlier, when the first (shell) side pressure drop was optimized, the constant value for the other (tube) side pressure drop may not have been at its optimum. One more cycle of trials is conducted to locate (perhaps) a new optimum value of pressure drop for the first side.
The process is repeated till for both sides, the optimum values found as above do not change. These values then become the real (or absolute) optimum values. Till then, all earlier values can only be called as temporary or local optimum values.
It can be readily perceived that a lot of trials may be needed to reach the optimum. The entire procedure is quite tedious, even with the help of a computer. Only if, for some process reason, one of the sides pressure drops is fixed, the optimization procedure becomes relatively simpler.
Sometimes, instead of type (b) parameters being involved in the optimisation, type (a) parameters are involved. For example, for a given fixed pressure drop value, effect of variation in the flow rate (with or without change in its outlet temperature) can be studied and the calculations for the total annual cost are made just as before.
As the cost of energy is proportional to the product of
pressure drop and flow rate, even flow rate variation introduces variations in the pumping energy cost.
10.4
Minimization Of Only Fixed Cost
The total annual cost minimization of previous section is somewhat rare. Minimization of the fixed cost is very common. Furthermore, fixed cost minimization is required even during the total cost minimization. Therefore, this type of optimization is very common. It can be safely said that as a designing standard, each heat exchanger must be optimized for minimum fixed cost. Fir fixed cost minimization, we are talking about type © i.e. geometrical parameters. The type (a) and (b) parameters being already frozen. That is, flow rates, inlet and outlet temperatures, both side fouling factors and both side pressure drops are finalized. Now, geometrical parameters are to be selected in such a way as to result in a minimum cost heat exchanger for the given duty. The parameters under the designer’s control are: Placement of fluids, i.e. which one on shell side and which one on tube side Tube length Type of layout – triangular, square or rotated versions Tube diameter and thickness
Pitch for the layout Number of baffles, i.e. baffles spacing Baffle cut Number of shell passes Number of tube passes
Truly speaking, not all the parameters are under designer’s control. For process requirements or plant space requirements, some of the parameters are not to be varied. The restrictions re called as constraints. The optimization is then called constrained optimization. It no long remains true optimization. Some of the very common constraints areSpace considerations dictate that the length of the heat exchanger be less than a certain value. Space considerations dictate that the diameter of the heat exchanger be less than a certain value Tube side fluid velocity should not be below a certain value to avoid erosion or abrasion of tube wall. Pressure drop on tube side should be below a certain given value Mass flow velocity on shell side should be below the given value Mass flow velocity through the baffle window should be below the given value
These constraints are general constraints. There are some particular constraints also which have to be considered for each individual parameter. These will be discussed along with discussion on each parameter.
Let us now see the effect of each parameter on the optimization.
10.5 Placement of Fluids Some guidelines are given in the textbooks. Generally smaller flow rate fluid is placed on the shell side.
This facilitates provision of adequate turbulence by increasing number of baffles. But
increased turbulence, one aim to increase the heat transfer co-efficient for unit pressure drop is higher on tube side compared with the same on shell side.
Other considerations may also dictate placement. For instance, cooling water (which is likely to deposit scales) is generally placed on tube side. This facilitates mechanical cleaning of tubes from inside. Shell side has no such access except in U-tubes or full floating head (removable tube bundles). Even then, mechanical cleaning from outside is quite difficult.
In general, highly befouling fluids that need frequent mechanical cleaning of the heat exchanger are usually placed on tube side.
Highly Highly corrosive corrosive fluids fluids are preferabl preferably y placed on tube side.
Whicheve Whicheverr side a corrosive corrosive fluid fluid is
placed, the tubes, where the heat transfer takes place, have to be of corrosion resistant material. Even the tubesheet tubesheet has to be in corrosion corrosion resistant material. material. But, compared with with the cost of entire shell, the cost of channel and heads in corrosion resistant material is substantially less.
Fluids with very high operating pressure (when the other f luid has relatively much less pressure) are preferably placed placed on tube side. side. There are two reasons reasons for this. First, only channel channel and heads are to be in higher thickness thickness as against the entire shell. shell. Second the tube thickness thickness to withstand internal internal pressu pressure. re.
Of course course,, the diamete diameterr of tubes tubes is gener general ally ly small small and the thickne thickness ss requir required ed to
withstand higher pressure is not much more than that for low pressure but when all the tubes are thicker, it may add considerably to the cost of the heat exchanger.
10.6 10.6 Tube Tube Leng Length th For the same heat transfer area, area, less number of longer longer tubes are required. This makes it possible possible to fit the tubes in a smaller smaller shell. As the cost of tube sheet, flanges flanges and the shell are are dependent dependent on the inside diameter of the t he shell, this results in a less expensive heat exchanger.
It is however however not possi possibl ble e to increase increase the tube length length indef indefini initel tely. y. cumberso cumbersome me to handle. handle.
Too long long an exchan exchanger ger is
Restricti Restriction on on maximum maximum tube length may be there from plant plant layout layout
consider consideratio ations ns (item (I) above). above). For a longer longer heat exchanger exchanger (with smaller smaller shell diameter), diameter), the pressure drop, especially especially on the shell side, side, is much higher. If the allowable pressure pressure drop limit is relatively low, long tube lengths do not yield an economical solution.
Except for expensive and made-to-order tubes, there are certain standard lengths only in which the tubes are available available.. Therefore, variation variation of tube length is generally generally in steps. steps. Infinitesimal variation is not possible.
10.7 10.7 Tube Tube Diamete Diameter, r, Pitch Pitch And And Layout Layout Larger the tube diameter, diameter, it is easier to clean the tube mechanically mechanically from inside. inside. On fouling service therefore, it is desirable desirable to go for larger tube tube diameters. On cooling water water service, it is desirable desirable to go for tubes with minimum outside diameter of 25.4 mm
Larger the tube diameter, in general, more thickness is required to withstand the pressure, whether it acts internally or externally. externally. For a moderate pressure pressure and limited range range of tube diameters diameters under consideration, consideration, there may not be any difference in the tube thickness t hickness required. Larger the tube diameter, less is the total heat transfer area accommodated in a given shell. Therefore, a solution satisfying the process duty is reached with larger diameter shell.
Also, wherever minimum velocity constraint is specified (e.g. item (iii) as above), use of large diameter tubes usually fails to yield an optimum solution.
Pitch is usually 1.2, 1.25 or 1.33 times the tube Outside Diameter (O.D) as per the standards of TEMA (Tubular Exchangers Exchangers Manufacturers’ Manufacturers’ Association). Association). Larger the pitch, pitch, less number of of tubes can be fitted in a shell of given diameter. diameter. Hence, smaller smaller pitch yields a more economical economical solution. solution. This, however, is subject subject to availability availability of adequate pressure pressure drop on shell side. side. When unusually unusually small pressure drop limit is specified on shell side fluid, larger pitch may be desirable.
Triangular pitch is more compact. compact. More number of tubes can be accommodated accommodated in the given shell, shell, as compared with with square pitch layouts layouts (for same tube OD and pitch). Correspondingly, Correspondingly, the shell side becomes more crowded and the pressure drop on shell side is more.
Heat exchangers requiring shell side cleaning use removable tube bundles and also use square pitch layouts/ These bundles are relatively easier to clean than with triangular layout, or, use of only particula particularr tube OD and pitch). pitch). Correspo Correspondin ndingly, gly, the shell side becomes more crowded crowded and the pressure drop on shell side is more.
There are certain combinations that are industry standards¾” (19.05mm) OD tubes on 15/16” (23.81mm) triangular pitch 10.05mm OD tubes on 1” (25.4mm) triangular pitch 19.05mm OD tubes on 25.4mm square pitch 25.4mm OD tubes on 11/4” (31.75mm) triangular pitch 25.4mm OD tubes on 31.75mm square pitch
Sometimes, additional constraints are placed on the design, such as, use of only particular type (square or triangular) of layout, or, use of only particular tube OD, or, use of only one specific combination of OD, pitch and type.
10.8 10.8 Baff Baffle le Spac Spacing ing Baffles usually mean mean a small percentage percentage of total cost. This does not mean that their effect on the cost can be ignored. ignored. Difference Difference in the cost of 6 baffles baffles and 9 baffles baffles may not be signific significant ant but differenc difference e between between the cost of 6 and 30 befalls befalls may be very significa significant. nt. Significa Significant nt or not, every every baffle added is an extra cost not only of the baffle itself but of the spacer tubes, tie rods and the labour labour cost for assembly into the shell also. also. Furthermo Furthermore, re, consideri considering ng that the baffles and tube sheets are drilled together, every baffle adds to the thickness of the bunch for drilling. drilling.
More the number of baffles, higher is the heat transfer co-efficient and so(higher) is the pressure drop too. This leads to the philosophical philosophical observation observation that nothing is is free in nature. nature. One has to pay the price for better heat transfer co-efficient.
Increasing the number number of baffles does not always always result in less heat transfer area. For condensers without desuperheating and subcooling (on shell side) there is no change in the heat transfer area. For boiling in a kettle type type reboilers also, there is is no change in the heat transfer transfer area. In both these cases, the baffles act more as tube bundle supports than heat transfer promoters.
In case of condensa condensation tion in presence presence of non-cond non-condensa ensable ble and all applicati applications ons of sensibl sensible e heat transfer (including desuperheating desuperheating and/or subcooling subcooling in total condensers), the additional turbulence turbulence helps to increase the heat transfer co-efficient.
In all the cases when the number of baffles has been increased, the price of added pressure drop has to be paid, whether there has been an increase in the heat transfer co-efficient or not!
Good engineering practice dictates that there are limits to the range over which the number of baffles can be varied. Standards of TEMA suggest suggest that baffles should not be placed closer than D/5 D/5 or 2” (50mm), (50mm), whichever whichever is more. more. Here, Here, D is the internal internal diameter diameter of the shell. shell. Similarl Similarly, y, baffles baffles should not be placed farther than D or 24” (600 mm whichever is lower). This 600mm is not an absolute value. value. It depends upon the tube tube diameter and the unsupported unsupported span the tube tube can sustain without deflecting deflecting too much. much. Also, these are are suggestions suggestions only. There can be exceptions. exceptions.
Then again, the mass velocity constraints [item(vii)above] impose further limits on the range over which the baffle spacing can be varied.
The baffles spacing cannot be varied infinitesimally. The number of baffles must be an integer. To place these baffles in the shell in (as far as possible) an equidistant manner, the spacing becomes a step value type variable. Therefore, it is more prudent to work out the maximum and minimum number of baffles that can be placed, considering all the constraints. Then, vary the number of baffles within this range, starting at minimum number and increasing this number by one each time.
Again, I wish to emphasise that without adequate allowable pressure drop on shell side fluid, it is meaningless to try and optimize the number of baffles.
10.9 Baffle Cut Only bearing this factor has on optimization is in cases where pressure drop on shell side becomes critical. It is desirable to place more number of baffles to get smaller heat transfer area but the pressure drop limit is somewhat unsuitable. In such a case, by increasing the baffle cut, it is possible to get lower pressure drop even with more number of baffles. Of course, the extent to which this can be done is limited because the contribution of baffle window pressure drop in the overall pressure drop on shell side is less than the cross flow pressure drop. The range of baffle cuts over which the variation is 15% to 40% in steps of 5%. Generally, 33.3% is also is also included as a distinct value within this range. In some cases, where constraint item (viii) as above applies, the lowest cut/s may not be admissible. In this case, the range becomes narrower.
For segmental baffles, beyond 40% (some people take this value at 45%) cut value, the shell side flow becomes more axial than cross. For this type of flow, the procedure for calculation of heat transfer co-efficient and pressure drop is different. It is an uncommon procedure and hence not easily available in the standard textbooks.
All this discussion assumes that design algorithms are available in which effect of baffle cut on heat transfer co-efficient and pressure drop can be calculated.
Many manufacturers and designers
prefer to maintain the baffle cut at only 25% which happens to be the most common value and for this, heat transfer co-efficient and pressure drop calculations are readily available.
The percentage baffle cut need not imply only segmental baffles. For other types (e.g. disc and doughnut) also, variations in cut can apply.
As a matter of detailing, it should be ensured that each baffle is supported at minimum three points. This becomes difficult when the cut is large, say, 35% or 40%. For this reason also, some prefer to omit 40% and more cut baffles from optimisation options.
10.10Number Of Shell Passes Usually, single shell pass heat exchangers are used in the industry. This is because providing a pass partition on shell side is difficult, costly and perhaps unnecessary.
The alternative of two shell passes becomes relevant only when there is a large temperature cross for the fluids, which the standard 1-2 pass configuration heat exchanger is unable to handle effectively. Another requirement before considering a second shell pass is the availability of liberal pressure drop on shell side fluid. This aspect becomes critical as the shell side cross flow area becomes half what it would be without the shell baffle and at the same time, the number of times the fluid has to change direction at baffle becomes double. The new pressure drop is nearly eight times the original. If the shell inside diameter (I.D.) is small enough, generally upto 400mm, the shell partition plate can be put using sealing strips. If the shell I.D. is large enough for a person to go inside, generally above 1,000 mm, the partition can be welded to the shell. For these two ranges, it is possible to think of a shell with two passes made with longitudinal baffle. For intermediate diameters 400
If it is not possible to provide two shell passes but the process requirement is very much there, on account of large temperaturte overlap, two or more shells in series are thought of instead of two shell passes in one shell.
10.11Number Of Tube Passes This parameter plays an important role in the optimisation of the design and is at the disposal of the designer. Common values of number of tube passes used in the industry are 1,2,4 & 6 and 8. This is as recommended by TEMA. For these values of number of passes and the five standard tube layouts mentioned earlier, tube count tables are easily available in the textbooks and handbooks. Alternatively, tube count procedures are also available.
This gives a definitive picture of the
configuration before checking suitability for process duty.
For some applications, e.g. juice heaters in sugar industry, the passes in these heat exchangers are very high, often about 16 or 20. In these cases, the exact working out of the tube counts is a
tedious job, worked out on case-to-case basis. No standard table or simple procedure is available for tube count.
Use of odd number of passes, e.g. 3 or 5 is very rare in the chemical industry/ it is certainly not totally absent. For these passes also, it is not easy to get the tube counts nor is it easy to calculate the value. It is to be worked out by plotting and counting process.
More the number of passes, more is the heat transfer co-efficient and less is the heat transfer area required to perform the duty. This is because the turbulence on tube side increases. This benefit is not without its usual price of higher-pressure drop. By doubling the number of passes, the pressure drop increases by about eight fold while the heat transfer co-efficient increases by only about 1.74 times. Even then, this cost benefit ratio is better than the same available on shell side.
For creating tube side passes, pass partition plates have to be provided on the tube side (inside the channels).
This makes it necessary to remove certain number of tubes to accommodate the
partition plates. More the number of passes, more tubes are to be removed. Thus, there is a reduction in the available heat transfer area within the given shell of particular diameter. The area reduction is considerable for shells of smaller diameter; the effect is very nominal for large diameter shells.
As with many parameters in designing of heat exchangers, this parameter also can be varied in certain steps. Infinitesimal variation is not possible. This places some restrictions on the degree of control over the optimization process.
Increasing the number of passes does not always help. For processes in which the shell side heat transfer co-efficient is very poor, substantial increase in the tube side co-efficient also fails to make enough impact. On the other hand, the tube side pressure drop that has been increased to achieve this co-efficient continues to use more energy and consequent more operating cost.
The case
presented as figure-4 will illustrate this point numerically.
Conversely, where the shell side heat transfer co-efficient is very good, e.g., condensers for vapour, it is always beneficial to increase the heat transfer co-efficient on tube side by increasing the number of tube passes. Even for cooling of non-viscous liquids (on shell side) with cooling water, it is always possible to derive some benefit by increasing the tube passes. This feasibility comes from
the fact that heat transfer co-efficient on both sides become comparable and at the same time, large pressure drop is available on cooling water. Second half of figure-4 illustrates this point numerically.
There may be some constraints over the number of passes. For example, in vertical thermosiphon type (tube side) reboilers, only single pass design is acceptable.
In U-tube bundle type heat
exchangers, only 2 or 4 pass design is required. The very construction makes it impossible to have single pass. It may be theoretically possible to have 6 or 8 passes but practically it is very difficult to provide these number of tube passes in a U-tubes heat exchanger. In all these cases, full variation of tube passes during optimization is not possible.
For small diameter shells, it becomes difficult and cumbersome (and sometimes even impossible) to accommodate the partition plates for large number of passes. Therefore, for small shells, it is a standard practice to place an upper limit on the number of tube passes. There are no known standards. Usually the designer’s opinion of what constitutes good engineering practice becomes the guiding principle. Different manufacturers also set different standards.
Another factor that
comes into play is the tube diameter. In a shell of given inside diameter, if the tube diameter is small, it is generally possible to accommodate more number of tube passes than when larger diameter tubes are used,
10.12General Strategy First attempt for finding the optimum solution must start with the minimum number of shells. This strategy is consistent with the knowledge that tube sheets and flanges cost double for two shells if the required heat transfer area is provided in two shells instead of in one shell.
Keeping only one shell is not a difficult task to fulfil for most of the case an engineer encounters normally. For problems involving large temperature overlap for fluids, one has to take recourse to using more number of shells in series, or, using shells with two shell passes, or both. The last is a very rare occurrence. This point has been discussed adequately in the previous section.
When the flow rate of one (or both) of the fluids is very large and when the corresponding allowable pressure drop is less, it is sometimes better to think of two shells in parallel. No doubt, generally it is an expensive alternative but surprisingly, sometimes this solution comes out to be less expensive than trying to fit the design in one shell. The reason is not easily apparent. By taking two shells, the flow in each is halved. The pressure drop is one fourth! This makes it economical when overall heat transfer area requirement and its cost are considered. Alternatively, for low-pressure drop designs, one may think of split flow or divided flow patterns in a single shell.
Having started with minimum number of shells of a certain I.D., the designer now has to select all the other parameters within the shell in such a way that minimum area can be fitted in. After fixing all the parameters (no doubt within the given constraints), the geometry is now checked for satisfying the performance, i.e., adequacy of heat transfer area and pressure drops within allowable limits.
If this combination of parameters is suitable from process performance point of view, well and good. Otherwise, progressively, other combinations with increasing heat transfer area are tried out. The first combination, which satisfies the given duty, is then selected. This way, one can ensure that always an optimum (or nearly an optimum) design is reached.
FIGURE - 1
FIXED COST
ANNUAL COST
HEAT TRANSFER AREA
OPERATING COST FIGURE - 2 ANNUAL COST
HEAT TRANSFER AREA
TOTAL COST
FIGURE - 3 TOTAL Cm ANNUAL COST
FIXED
OPRATING Ao HEAT TRANSFER AREA
10.13Cost benefit analysis of heat transfer improvement and effect of tube material Shell side film Tube side film coefficient ho coefficient kcal/hr m2°c hio kcal/hr m2°c
2000
Overall clean Overall clean Remarks coefficient coefficient c w/o tube Uc tube Ucr kcal/hr m2°c kcal/hr m2°c
95.2
100 5000
2000
98
666.7
1000 5000
2000
833.3
1578.9
100 5000
3000
Tube conductivity in Kcal/ hr (m2/m) °C
Tube thickness 16 BWG (1.65 mm)
94.2
S.S.
94.9
M.S.
95.2
Cu
97 97.6 98
S.S. M.S. Cu
620
S.S.
648.8
M.S.
664.5
Cu
763.4
S.S.
805.6
M.S.
829.9
Cu
1339.9
S.S.
1482.4
M.S.
1566.6
Cu
2255.6
S.S.
2669.6
M.S.
2995.7
Cu
SS 15 MS 40 Cu 330
11 Generic guidelines for heat exchanger designer • As a normal practice, streams being heated flow from bottom to top or streams being cooled
normally flow from top to bottom. •
Selection of tube or shell side:
The fluid with high fouling tendency preferably to be taken at tube side.
Corrosive fluids to be taken at tube side.
The fluid of smaller volumetric flow rate to be taken on shell side.
High-pressure fluids preferably to be kept on tube side to minimize the shell
thickness.
Fluids with high viscosity to be kept on shell side so that turbulence can be induced
by introduction of baffles.
•
Two-phase flow to be kept on shell side.
Property input to HTRI: Whenever there is water present along with liquid hydrocarbon, always conductivity of the dry stream needs to be inputted i.e. without water. Reason behind is normally water forms slugs while flowing with hydrocarbon therefore heat transfer coefficient is mainly governed by hydrocarbons & not by water. Moreover, hydrocarbon conductivity is lower than water conductivity making design more conservative.
•
For pressure above 15bar, property curve is inputted at only one pressure level, but below 15bar it is advisable to enter at least two-pressure level property curve with same inlet temperature.
•
Whenever heating cooling curve property points given are more than 10, plot of enthalpy versus temperature is plotted, & points are decided.
•
Floating head heat exchangers are typically limited to the shell ID of 1.4-1.5m. But we can still go for higher shell ID in case of fixed head heat exchanger.
•
In case of exchangers where very low-pressure drop is allowable, shells in parallel can be used.
•
More than one number of shell can be used as unit in order to avoid local temperature cross. However, F type of shell can also be used to reduce more number of shells. F shell with 2 passes on tube side acts as a pure counter current exchanger.
•
Longer shell is always cheaper than big shell diameter. It can also be used to improve velocity & hence heat transfer coefficients.
•
Typical tube sizes used are 20mm or 25mm. But design can be st arted by using 20mm tube OD.
•
Tube length should be maximized as far as possible to utilize maximum available pressure drop.
•
Maximum number of tube passes can be used keeping watch on the tubes side pressure drop & design velocity.
•
Tubes can be arranged in square, triangular rotated square or rotated triangular patterns. Triangular pitch provides more compact design & good heat transfer coefficient. But whenever shell side cleaning is required square pitch is preferred which enables better cleaning of the shell side. Also some cases such as, low allowable pressure drop on shell side or shell side vaporisation, square pitch can be preferred.
•
•
Normally used tube pitches.
Square pitch with 20mm tube OD: 26 mm tube pitch
Square pitch with 25mm tube OD: 32 mm tube pitch
Triangular pitch with 20mm tube OD: 25 mm tube pitch
Triangular pitch with 25mm tube OD: 31 mm tube pitch
In case H2S, or H2 is present in the process, welded tubes are required. For welded tubes minimum tube gap between adjacent tubes should be at least 6.35 mm. In that case, 3
rd
point
covered above will also require at least 26 mm, as a tube pitch is the exchanger is under H2 service. •
Single segmental baffle can be used for normal exchanger designs. In cases where allowable pressure drop on shell side is low, double segmental baffle can be used.
•
Number of baffles & baffle pitch can be decided based on allowable shell side pressure drop. Minimum baffle spacing should not be less than one-fifth of shell diameter or 50.8 mm (2 inch) whichever is greater.
•
Baffle cut can be adjusted to maximize cross flow, maintain allowable pressure drop and avoid bypass & leakage fractions.
•
Baffle cut can vary between 15 % to 45% of the shell inside diameter. But, both very small & very large baffle cuts are detrimental to efficient heat transfer. Generally, is recommended that only baffle cuts between 20% and 35% be employed.
• Any change in the baffle geometry should be done by watching pressure drop on shell side. •
Baffle cut orientation should be based on process conditions. Normally phase change operation on shell side is provided with vertical baffle cut. It is also called as parallel cut or inline baffle. Similarly horizontal baffle cut is also called as perpendicular cut or transverse baffle. Horizontal baffle cut can be used for vaporizing services. If horizontal cut is given for condensing service, care should be taken that gravity separation zone is avoided on the shell side.
•
Sealing strips or seal rods can be inserted in order to minimize bypass fraction (stream B) & inline bypass fractions (stream F).
•
Pressure drop across nozzles is considered as a part of heat exchanger pressure drop hence, generally nozzle size should be such that pressure drop contributed by nozzle should not be more than 15% of exchanger pressure drop.
•
Exchanger which are subject to steam out or where there is a possibility of being subject to vacuum conditions during operation or shutdown should be designed for full vacuum.
•
For cooling water exchangers with treated or inhibited cooling water service, carbon steel is normally specified. However some clients prefer admiralty brass, as it is their hardware. For other fresh water services, admiralty metal is normally specified. For sea or brackish water, material such as aluminum brass 90:10 or 70:30 cupronickel or titanium is used.
Fluid
Tube material
Preferred velocity (m/sec)
Cooling water
Carbon steel / Admiralty
2 - 2.5
Sea water
90:10 Cupronickel
2 – 2.5
Sea water
Aluminum brass
1.5 – 2
Brackish water
Carbon steel
1.2
•
In general very high velocities lead to erosion. The minimum recommended liquid velocity inside tubes is 1.0m/sec &, while maximum is 2.5-3.0m/sec.
11.1 What exchanger designer should analyze in HTFS / HTRI output
Output produced by HTFS / HTRI reflects all the data fed by designer, hence it should be checked whether the data considered by HTFS is appropriate.
Area ratio of actual to required should be at least 1in case of HTFS. Based on clients requirement area ratio need to be decided (for example 1.10 , 1.15). For HTRI % overdesign should be be checked i.e. 10% or 15%.
Among three figures of overall heat transfer coefficient (clean / dirty / service), dirty heat transfer coefficient gives the real picture of the design. Clean coefficient is without considering any fouling & service coefficient is a back-calculated value from the heat transfer equations.
Calculated pressure drops on both the sides (shell / tube) should be less than allowable pressure drops for their respective side.
Check whether fouling coefficients considered are appropriate & are as per Technip standard or client requirements.
Output should be read for warnings messages / fatal errors. In general, if the fatal error exists, the program will not run successfully. But warning messages should be read carefully & attempt should be made to rectify them. Some warnings can be ignored after due consideration such as 1.
The most fouling fluid is on the shell side. You
may wish to put it on the tube side, which is easier to clean. 2.
The shell side design pressure is greater than
that of the tube side. You may wish to consider putting the high-pressure stream in the tube side.
Requirement of impingement should be identified based on TEMA guidelines & it should be checked for the given design.
In general, nozzle size should be checked for pressure drop not more than 15% of the total.
Minimum baffle spacing should not be less than D/5 or 50mm. But in practice, it is always better to use at least 150mm-baffle spacing. Based on this number of baffle should be decided.
Clearances such as tube/baffle, bundle/shell or baffle/shell used by HTFS / HTRI are as per TEMA. These clearances could be changed only after discussion with mechanical department.
Inlet / outlet end lengths should be sufficient enough to accommodate nozzle, shell flange etc. Thus end lengths can be different for front head & rare head, based on type such as fixed or floating head.
Heating / cooling curves provided should have as many property points as possible while designing 2-phase exchanger. In case single-phase exchanger, two property points also gives proper result.
In case of vaporizers it should checked that bubble point & dew point considered by HTFS / HTRI is same as that in property package. Best way is to give immediate first preceding property point of bubble/dew point.
Normally in vaporizers it is observed that approximately 90% area is required for between bubble point & dew point. Hence it is advisable to give maximum number of points between these two to get best possible result.
Output should be checked for vibration analysis & if present any, attempt should be made to avoid it by changing the span length as discussed earlier.
Shell side flow fractions should be checked for cross flow (stream B). Generally it should be more than 0.45 in good designs.
All other fractions are leakage, which should be reduced by changing geometry such as baffle cut, clearances etc.
Tube layout should be checked on OPTU (HTFS) for checking whether tubes are fitting in the given shell ID. Also tube layout should be checked for incorporation of tie-rods, impingement plate, sliding strip, sealing strips etc.
11.2 Important Points specific to software HTRI In case of vaporizers in HTRI, boiling range should be necessarily specified since the same affects heat transfer coefficient. If this is omitted HTRI calculates the same but BR calculated can be over or under predicted which will in turn affect heat transfer coefficient. In some of the cases where there is a presence of inert gases, mole fraction of inert gas should be specified. It is possible to put very minute data in HTRI, hence it is useful for running existing exchanger such as revamp. Unlike TASC / OPTU, tube layout produced by XIST / XTLO is identical. RPM (resistance proration method) & CPM (Composition profile method) are two important methods available for condensation. RPM is a default method can be used widely & is applicable for most of the cases, whereas CPM should be used in case of presence of inert gases in the condensing side, but it asks for accurate VLE data. Physical property based method which uses theoretical boiling range mixture correction factor, is used in case of vaporisation & gives satisfactory results. In general it can be said that HTRI gives better result in case of phase change operations. HTFS
Design mode of HTFS is more user friendly & gives better options. Different constraints can be forced to get desired results. While designing exchanger minimum certain end length is necessary for accommodating nozzle, shell flange etc. If in case given end length is not sufficient enough, HTFS does not give any warning & hence it can lead to a fabrication problem. In certain cases, seal rods needs to be incorporated to minimise inline pass partition stream. It is not possible to incorporate seal rods in HTFS other than sealing strips. Can not force inlet / outlet dome height for the exchanger.
In case of NTIW design, if it is giving vibration simple remedy such as incorporation of additional support plate per baffle spacing solves the problem. But this kind of option is not available in HTFS. Variable baffle spacing can not be used in HTFS. Maximum numbers of tubes, which can be fitted in given shell ID, differ in TASC & OPTU. Hence output given by TASC should be checked in OPTU.
. 12. Reboiler 12.1 Reboiler Selection: Generally kettle type reboilers, vertical & horizontal thermosiphon reboilers are main types of reboilers.
Kettle type reboiler: These can vaporise fully column bottoms & acts as a additional stage for distillation. Kettle type reboilers can achieve high heat flux. Since kettle type reboilers principle is based on pool boiling it is less sensitive to hydrodynamics. Kettle type is preferred near critical pressure region over others. But disadvantages of the kettle type are, they are very costly & they have very high fouling tendency since there is little turbulence.
Horizontal thermosiphon reboiler: Generally ‘G or H’ shell is used for minimizing the pressure drop, sometimes ‘E or J‘ can also be used. Generally for clean heating mediums this can be preferred with U tubes. Horizontal reboilers can achieve high circulation rates, high convective heat transfer component, & high boiling heat transfer coefficient. Since these reboilers require lower static head, lower boiling point elevation is there which saves some heat transfer area. It has a lower fouling potential as compared with kettle and are lesser sensitive than vertical thermosiphon reboilers. Disadvantages of horizontal thermosiphon reboiler are more fouling on shell side compared with vertical type.
Vertical thermosiphon reboiler: Vertical thermosiphon reboilers are generally single tube pass since boiling is inside the tube. Since fixed tube sheet is generally used, heating medium needs to be clean. Maximum allowable heat flux is generally 30000 kcal/hr m2 °C for Pr = 0.3 and 2000 kcal/hr m2 °C for Pr = 0.9. Vertical thermosiphn reboilers are cheapest of all since they are generally fixed tube sheet exchangers. Fouling is less pronounced since high circulation & shear stress. It has highest MTD due to pure counter current operation. Disadvantages of this type of reboiler are it give lower heat flux than shellside boiling especially near critical pressure. They are very sensitive to changes in operating conditions.
12.2 Design of reboilers To start with, reboiler is designed as a normal exchanger to carry out specified heat duty. Once exchanger design is ready it is checked for its thermosiphon reboiler duty. But this method is limited to HTFS software only. In case of HTRI software, it can be directly runned for thermosiphon reboiler operation. Designing thermosiphon reboiler includes fixing of inlet / outlet geometry. Once typical inlet / outlet circuit geometry is fixed, recirculaton quantity gets fixed. Driving force for this circulation is level in column & resistance are pressure loss in exchanger and inlet / outlet pipeline of exchanger. This circulation flow is checked for its adequacy. If required, inlet / outlet piping can be changed to achieve stability. Thermal adequacy of this exchanger is checked by seeing whether reboiler is capable of carrying out given heat duty i.e. if it can generate required amount of vapor. If required, heat transfer area can be altered. In case of reboilers it is very essential to provide heating curves at second pressure level also. In these cases, both the pressure levels should have same inlet temperature point. General criteria for checking thermosiphon reboiler output: Inlet piping: Pressure drop < 0.2 to 0.4 bar/km Velocity < 1 m / sec Outlet piping: RHO-V-SQ < 6000 kpa RHO-V-SQ (only for vpour) > 100 kpa …required for separation. Frictional plus acceleration pressure drop should be less than 0.3 to 0.35 times the available static head. Recommended weight fraction of vapor at outlet is 0.3 for vertical & 0.4 for horizontal thermosiphon reboiler. In some of the cases where heat transfer area required is very large, it is not possible to fit it in one shell. In such cases, if two numbers of shells are required then reboiler should be run for the half of the duty & circulation. This is because HTRI does not allow common inlet & outlet piping for parallel exchangers. Thus thermosiphon reboiler needs to be checked for both recirculation quantity & thermal adequacy.
13.0 Typical values of heat transfer coefficient
Shell Side Boiler feed water Boiler feed water Blow-down Demin water Fuel oil Gasoline Hydrogen Ethylene HT-shift Kerosene / gas oil Naphtha Natural gas Nitrogen Process condensate Process steam (superheating) Process steam (superheating) Steam Water Process water
Tube side Blow-down HT-shift product Cooling water Water Water Water Cooling water Cooling water Boiler feed water Water Water Brackish water Cooling water Water Ht-shift product Process gas Naphtha Water MP condensate
U (kcal/ hr m2 °C) 600 400 500 2000 100 400 500 325 300 250 250 225 175 500 175 250 200 1000 800
14 Typical fouling resistances used for designing heat exchanger Following are the typical values of fouling resistances, which should be used unless otherwise dictated by client.
Process side Sour Naphtha
Fouling factor (Kcal/m2 hr °C) 2000
Sweet naphtha Cooling water Sea water Boiler feed water Blow down water Steam Demin water Fuel oil Gasoline Kerosene LPG MEA solution Methanol Process gas (shift feed / product) Hydrogen / Nitrogen Refinery gas Regenerating gas
2500 2500 1650 5000 2500 5000 5000 1000 5000 5000 5000 2500 5000 5000 5000 1650 1650
15 Mechanical aspects The mechanical aspects to be considered during thermal design of the exchangers and after the Mechanical design of the exchangers are discussed below, 1. Inlet/outlet baffle spacing 2. Dome height top/bottom 3. Impingement plate placement 4. Tubesheet thickness
Inlet/outlet baffle spacing Once the central baffle spacing (baffle pitch) has been frozen during the thermal design of the exchangers the inlet and outlet baffle spacing can be calculated approximately. Because to calculate the spacing distance the information on tubesheet thickness and flange thickness are necessary which can be calculated during mechanical design.
During thermal design of shell and tube exchangers inlet baffle spacing can be calculated roughly P t u
b
a d
w
i d
t h
e ts h h k e e t N
o
zd zi a l e
I m p i n g e B a f f l e s p
f l a n t h g ke t u b e p r o j e c t i o n o u t st u i d b e e s h e e t 3 m m
and used in the design of the exchanger. To calculate roughly the inlet spacing sum the following, Flange thickness (take from previous projects for similar exchanger) Clearance between shell flange and nozzle reinforcement pad (50-100mm) Twice the Nozzle dia for Reinforcement pad width on both sides. Nozzle diameter. Clearance between reinforcement pad and first baffle (30-50mm). This distance normally depends on the impingement plate presence and its placement. Because the distribution of the shell side fluid towards the impingement plate and towards the tube bundle, will have an effect based on the clearance. The check for the ρv² value in this region to be carried out in later stage during mechanical design to ensure that they are with in TEMA range. This can be done by the spreadsheet available in the process console. In case of floating head exchangers if the shell fluid enters at the floating head side and support plate is present the tube area between the support plate and the tube sheet is considered as lost area and the inlet spacing will be considered only upto support plate instead of tubesheet.
Outlet end spacing This will be calculated as same as that of the inlet end spacing, once the inlet baffle spacing and central baffle pitch have given to HTRI/HTFS as input the program calculates the outlet end spacing.
End lengths
The end length describes the tube length between the baffle and the tube end. For example the inlet end length for the diagram shown above will be the summation of tube projection (3mm), tubesheet thickness, flange thickness, clearance between the shell flange weld to the reinforcement pad, reinforcement pad width (twice the nozzle dia), clearance between the pad and the first baffle. In case of a floating head exchanger the additional backing ring thickness will be added to the above dimensions. These dimensions are available exactly only after the mechanical design.
B
a c
S
h
k
i n
e l l
f
Dome height Top/Bottom: Dome height describes the flow area for the shell side fluid between the nozzle and the impingement plate. The dome height as calculated by the HTRI/HTFS program must be analyzed and checked for the ρv² values that they are within TEMA range (5953 kg/ms²). The shell entrance and exit ρv² values are also available in the HTRI output. Insufficient area for the flow of shell side fluid at shell entrance, bundle entrance, shell exit and bundle exit will lead to vibration of tube bundle and unwanted pressure drop at the shell entrance and exit.
Impingement plate The need for the provision of impingement plate has been discussed before, now the dimensions and placement of the same will be discussed. The impingement plate can be either a circular plate or a rectangular plate and have the dimension of 1” (25 mm) more than the nozzle diameter on all the sides. The impingement plate is generally placed on the tube bundle supported by the tie rods (welded). If tie rod is not available on that position then local tie rods will be provided
