Unit II Mechanical Design

8/9/2019 Unit II Mechanical Design

1/19

UNIT II

Thermal design of heat exchangers

The mechanical design is done by the mechanical engineers on the inputs of chemical engineers

and using the codes. The most widely used code in Tubular Exchanger ManufacturesAssociations (TEMA). This USA code along with ASME selection !!! (unfired pressure "essel)

code is used together for the mechanical design of the heat exchanger. The !ndian code for the

heat exchanger design !S #$%&.

'ere we would discuss about the process design (or thermal design) leading to the siing of the

heat exchanger. efore understanding design steps* it is necessary to understand the following

for the heat exchanger.

8.2.1 Overall heat transfer coefficient

As understood by the pre"ious discussion that generally heat exchangers are tubular in nature(+ote, we are not discussing about plate type heat exchangers). Thus we can easily find out the

o"erall heat transfer coefficient based on our pre"ious -nowledge. igure /.0 shows a simplestform (double pipe heat exchanger) of tubular heat exchanger* where fluid A is being heated by

fluid in a co1current flow pattern. The inside and outside radii of the inner tube is represented riand ro. The length of the exchanger for heat transfer is considered as 2 for section 0 to /.

Fig.2,1: (a !chematic of a do"#le $i$e heat exchanger (# thermal resistance net%or& for overall heat

transfer

Thus the rate of heat transfer from the hot fluid to the cold fluid will be represented by e3.4.0*


2/19

The o"erall heat transfer coefficient5

ased on inside area of the inner pipe (e3.4./)

ased on outer side area of the outer pipe (e3.4.&)

'h m"lti)$ass exchangers*

The simplest type of heat exchangers is double pipe heat exchangers* which is inade3uate forflow rates that cannot readily be handled in a few tubes. !f se"eral double pipes are used in

parallel* the metal weight re3uired for the outer tubes becomes so large that the shell and tube

construction* such as 010 exchanger will be helpful. !n that one shell ser"es for many tubes* is

economical. The heat transfer coefficient of tube side and shell side fluid is "ery important and

the indi"idual heat transfer coefficients must be high enough to attain high o"erall heat transfercoefficient. As the shell would be 3uite large as compared to the tubes* the "elocity and the

turbulence of the shell side fluid is important.

!n contrast* the 010 exchanger has limitations also. 6hen the tube side flow is di"ided e"enly

among all the tubes* the "elocity may be 3uite low* resulting in low heat transfer coefficient.

There it may be re3uired to increase the area to ha"e the desired heat exchange for this low heat

transfer coefficient. The area may be increased by increasing the length of the tube. 'owe"er* thetube length re3uirement may be impractical for a gi"en situation. Thus the number of tubes

should be increased without increased the tube length. The increased number of tubes would also

pro"ide the increased "elocity in the shell side resulting in the higher heat transfer coefficient.

Therefore* multi1pass construction is needed* which would permit to use the practical andstandard tube lengths. 'owe"er* the disad"antages are that*

0. The construction of the exchangers become complex.

/. 7arallel flow cannot be a"oided.

&. Additional friction losses may occur.


3/19

!t should be noted that generally e"en number of tube passes are used in multi pass exchanger.

+T- correction factor

!n the earlier chapter* we ha"e seen for co1current or counter current flow system. The a"erage

dri"ing force for heat transfer was defined by log mean temperature difference (2MT8). Thusthe 2MT8 can be used for 010 exchangers for co1current and counter current. 'owe"er* for multi

pass exchangers (01/* /1#* etc.) the fluids are not always in co1current or counter current flow.

The de"iation for co1current or counter current flow causes a change in the a"erage dri"ing force.Therefore* in order to use true heat transfer dri"ing force* a correction factor is re3uired into the

2MT8. Thus* the heat transfer rate can be written as (e3.4.4)*

q = UdA(FTTm)

where* Ud9 o"erall heat transfer coefficient including fouling:dirt

A 9 heat transfer area

T ;Tm9 true a"erage temperature difference.

T9 2MT8 correction factor

!t is to be noted that the following assumption ha"e been considered for de"eloping 2MT8*

0. The o"erall heat transfer coefficient is constant throughout the exchanger

/. !n case any fluid undergoes for phase change (e.g.* in condenser)* the phase change occurs

throughout the heat exchanger and the constant fluid temperature pre"ails throughout the

exchanger.

&. The specific heat and mass flow rate and hence the heat capacity rate* of each fluid is

constant.

#. +o heat is lost in to the surroundings.

$. There is no conduction in the direction of flow neither in the fluids nor in the tube or shell

walls.


4/19

!t should be noted that in case of condensation or e"aporation the correction factor becomes

unity (T 90). 6hile designing a heat exchanger* the rule of thumb is that the T should not be

less than %.4.

Fig. 2.1: FT$lot for 1)2 exchanger t: cold fl"id in the t"#e T: hot fl"id in the shell 1: inlet 2: o"tlet


5/19

Fig. 2.2: FT $lot for 2)/ exchanger t: cold fl"id in the t"#e T: hot fl"id in the shell 1: inlet 2: o"tlet

Individ"al heat transfer coefficient

!n section* we ha"e seen that the o"erall heat transfer coefficient can be calculated pro"ide the

parameters are -nown including indi"idual heat transfer coefficients. !n this* section we will

discuss how to find out the indi"idual heat transfer coefficient* which is basically based on thewell1established correlations and discussed earlier also.

The heat transfer coefficient (hi) for the tube side fluid in a heat exchanger can be calculated

either by Sieder1Tate e3uation or by =olburn e3uation discussed in earlier chapter.

'owe"er* the shell side heat transfer coefficient (ho) cannot be so easily calculated because of

the parallel* counter as well as cross flow patterns of the fluid. Moreo"er* the fluid mass "elocity

as well as cross sectional area of the fluid streams "ary as the fluid crosses the tube bundle. The

lea-ages between baffles and shell* baffle and tubes* short circuit some of the shell fluid thus

reduces the effecti"eness of the exchanger.

>enerally* modified 8onohue e3uation (e3.4.?) (suggested by 8.@. ern) is used to predict the

ho*


6/19

where* h%9 shell side heat transfer coefficient

8h9 hydraulic diameter of the shell side

-%9 thermal conducti"ity of the shell side fluid >s9 mass flow rate of the shell side

The 8hand >scan be easily calculated if the geometry of the tube arrangement in the shell is-nown. The tubes may be generally arranged as a s3uare or triangular pitch* as shown in figure

Fig.2.0: T"#e arrangement in the shell (a triang"lar $itch (# s"are $itch

The hydraulic diameter (Dh) for tubes on s3uare pitch9

Dhor


7/19

Shell side flow area can be calculated using baffle information number of tubes in the shell and

tube arrangement. !f /$C cut baffles are used* that means the shell side flow will be from this

/$C area. 'owe"er we ha"e to reduce the area of the pipes which are accumulated in this

opening. So depending upon the information we may determine the shell side fluid flow area. !t

may also be found out by the following way*

where*

= 9 tube clearance 9 baffle spacing 8s9 inside diameter of shell

p 9 pitch of the tube

Fo"ling factor or dirt factor

D"er a time period of heat exchanger operation the surface of the heat exchanger may be coated

by the "arious deposits present in the flow system. Moreo"er* the surfaces may become corroded

or eroded o"er the time. Therefore* the thic-ness of the surface may get changed due to these

deposits. These deposits are -nown as scale. These scales pro"ide another resistance and usually

decrease the performance of the heat exchangers. The o"erall effect is usually represented by dirt

factor or fouling factor* or fouling resistance* f (Table 4.0) which must ha"e included all the

resistances along with the resistances due to scales for the calculation of o"erall heat transfer

coefficient.

The fouling factor must be determined experimentally using e3. 4.#*

Thus to determine the f* it is "ery important to -now Ucleanfor the new heat exchanger. The Uclean

must be -ept securely to obtain the f* at any time of the exchangerFs life.

Table-8.1 Fouling factor of a few of the industrial uids


8/19

Tem$erat"re $rofiles in U)T"#e heat exchangers

ig. 4.? shows the temperature profile along the length of a 01/ exchangers and /1# exchangers.

exchangers.


9/19

Fig.8. Tem$erat"re)length c"rve corres$onding to (a 1)2 exchanger (# 2)/ exchanger

The nomenclature used in the fig.4.? is described below

Tha, !nlet temperature of hot fluid

Thb, Dutlet temperature of hot fluid

Tca, !nlet temperature of cold fluid Tcb, Dutlet temperature of cold fluid

Tci, !ntermediate temperature of cold fluid

!n the abo"e arrangement it is assumed that the hot fluid is flowing in the shell side and cold

fluid is flowing in the tube side of the exchangers. The fig.4.? (a) shows the 01/ exchangers in

which the hot fluid enter into the exchanger from the left side and exits from the right side. The

cold fluid enters concurrently that is from the left side to the tube of the exchangers and goes up


10/19

to right end of the exchangers and returns bac- to ma-e two tube pass* and exits from the left end

of the exchangers. The temperature profile all along the length of the exchanger is shown in the

corresponding temperature length profile. igure4.? (b) shows the flow direction and

corresponding temperature length profile for /1# exchangers. The shell side fluid two passes and

the tube side fluid has #1passes in the exchangers.

!t can be easily understood that whene"er the number of passes is more than one* the flow cannot

be truly co1current or counter current. Thus it will be a mix of co1current and counter current

flows in any multi pass heat exchangers.

Though the temperature profile of the hot and cold streams can be easily predictable for single

pass heat exchangers but for the complex flow modes* the prediction of temperature distribution

will be difficult as shown in fig.4.?. As can be seen when 01/ exchangers was (fig.4.? (a)) used in

co1current mode* the temperature profile was gi"en in the figure. 'owe"er* if the fluid streams

enter in counter current mode a temperature cross may occur sometimes. Temperature cross isdescribed as the positi"e temperature difference between the cold and the hot fluid* when these

fluids lea"e the exchangers. !n that case the cold fluid will attain the maximum temperature

inside the exchanger instead of at the exit (fig.4.0%).

Fig. 8.10: 1-2 ow pattern and temperature prole in echanger showing cross ow

At this temperature cross* the cold fluid temperature reaches the maximum at a point inside the

exchanger and not at its exits. This temperature cross point also coincides with the point of

intersection of the temperature profile of the hot fluid and the co1current one of the cold fluid.

The difference (Tc/ 1 Th/) is called the temperature cross of the exchanger. 'owe"er* if the


11/19

temperature cross does not appear then the (Tc/ 1 Th/) is called the approach. Moreo"er* on

careful e"aluation it can be seen that for the multi shell side pass a significant length of the

exchanger ha"e cross flow pattern in the tube flow when the shell side fluid is migrating from

one shell pass to another shell pass. Thus calculating heat transfer co1efficient for shell side

becomes little challenging and will be explained in section 4./.


12/19

Fig.8.11: Tem$erat"re $rofiles of (a $arallel flo%, and (# co"nter flo%, for different ine"alities


13/19

FOU+IN3 F45TO6

Material deposits on the surfaces of the heat exchanger tube may add further resistances to heat

transfer in addition to those listed abo"e. Such deposits are termed fouling and may significantly

affect heat exchanger performance. The heat exchanger coefficient* Uc* determined abo"e may

be modified to include the fouling factor f.!caling is the most common form of fouling and is associated with in"erse solubility salts.

Examples of such salts are =a=D&* =aSD#* =a&(7D#)/* =aSiD&* =a(D')/* Mg(D')/* MgSiD&*

+a/SD#* 2iSD#* and 2i/=D&. The characteristic which is termed in"erse solubility is that*

unli-e most inorganic materials* the solubility decreases with temperature. The most important of

these compounds is calcium carbonate* =a=D&. =alcium carbonate exists in se"eral forms* but

one of the more important is limestone. The material fre3uently crystallies in a form closely

resembling marble* another form of calcium carbonate. Such materials are extremely difficult to

remo"e mechanically and may re3uire acid cleaning.

5orrosion fo"lingis classified as a chemical reaction which in"ol"es the heat exchanger tubes.

Many metals* copper and aluminum being specific examples* form adherent oxide coatingswhich ser"e to passi"ate the surface and pre"ent further corrosion. Metal oxides are a type of

ceramic and typically exhibit 3uite low thermal conducti"ities. E"en relati"e thin coatings of

oxides may significantly affect heat exchanger performance and should be included in e"aluating

o"erall heat transfer resistance.

5hemical reaction fo"lingin"ol"es chemical reactions in the process stream which results in

deposition of material on the heat exchanger tubes. 6hen food products are in"ol"ed this may be

termed scorching but a wide range of organic materials are subHect to similar problems. This is

commonly encountered when chemically sensiti"e process fluids are heated to temperatures near

that for chemical decomposition. ecause of the no flow condition at the wall surface and the

temperature gradient which exists across this laminar subIlayer* these regions will operate at

somewhat higher temperatures than the bul- and are ideally suited to promote fa"orable

conditions for such reactions.

Free7ing fo"lingis said to occur when a portion of the hot stream is cooled to near the freeing

point for one of its components. This is most notable in refineries where paraffin fre3uently

solidifies from petroleum products at "arious stages in the refining

process* obstructing both flow and heat transfer.

iological fo"lingis common where untreated water is used as a coolant stream. 7roblems range

from algae or other microbes to barnacles. 8uring the season where such microbes are said to

bloom* colonies se"eral millimeters deep may grow across a tube surface "irtually o"ernight*impeding circulation near the tube wall and retarding heat transport. iewed under a microscope*

many of these organisms appear as loosely intertwined fibersImuch li-e the form of fiberglass

insulation Traditionally these organisms ha"e been treated which chlorine* but present day

concerns on possible contamination to open water bodies has se"erely restricted the use of

oxidiers in open discharge systems.


14/19

9artic"late fo"ling results from the presence of rownian sied particles in solution. Under

certain conditions such materials display a phenomenon -nown as thermophoresis in which

motion is induced as a result of a temperature gradient. Thermodynamically this is referred to as

a cross1coupled phenomenon and may be "iewed as being analogous to the Seabec- effect. 6hen

such particles accumulate on a heat exchanger surface they sometimes fuse* resulting in a

buildup ha"ing the texture of sandstone. 2i-e scale these deposits are difficult to remo"e

mechanically.

Most of the actual data on fouling factors is tightly held be a few specialty consulting companies.

The data which is commonly a"ailable is sparse. An example is gi"en below,

9ress"re dro$ in the heat exchanger

7ressure drop calculation is an important tas- in heat exchanger design. The pressure drops in the

tube side as well as shell side are "ery important and 3uite a few co1relations are a"ailable in the

literature. Dne such co1relation is gi"en below in the subse3uent subsection.

4./.J.0 =orrelation for tube side pressure drop (e3. 4.0%)

where*

;7t*f 9 total pressure drop in the bundle of tube

f 9 friction factor (can be found out from MoodyFs chart)

>t9 mass "elocity of the fluid in the tube

2 9 tube length n 9 no of tube passes

g 9 gra"itational acceleration

Kt9 density of the tube fluid

di9 inside diameter of the tube

m 9%.0# for e G /0%%

%./$ for e L /0%%


15/19

The abo"e correlation is for the pressure drop in the tubes owing to the frictional losses.

'owe"er in case of multi pass flow direction of the flow in the tube changes when flow is from

01pass to another pass and the pressure losses due to the change in direction is called return1loss.

The return1loss (;7t*r) is gi"en by e3.4.00*

n 9 no of tube pass

"t9 "elocity of the tube fluid

Kt9 density of the tube fluid

Therefore* the total tube side pressure drop will be*

;pt9 ;7t*f ;7t*r

8.2..2 5orrelation for shell side $ress"re dro$

The following correlation (e3.4.0/) may be used for an unbaffled shell*

The abo"e e3uation can be modified to the following form (e3.4.0&) for a baffled shell*

6here*2 9shell length

ns9 no of shell pass

nb 9 no of baffles

Ks 9 shell side fluid density

>s 9 shell side mass "elocity

8h 9 hydraulic diameter of the shell

Dsi = inside diameter o shell

s = shell side riction actor


16/19

The hydraulic diameter (8h) for the shell can be calculated by the following e3uation (e3. 4.0#)*

where, nt= number o tubes in the shell

do= outer diameter o the tube

The riction actor (s) can be obtained by the Moodys chart or the corresponding Reynoldsnumber

;eat transfer effectiveness and n"m#er of transfer "nits (NTU

The 2MT8 is re3uired to be calculated for the e"aluation of heat exchanger performance.

'owe"er* the 2MT8 cannot be directly calculated unless all the four terminal temperatures (Tc*i*

Tc*o* Th*i* Th*o) of both the fluids are -nown.

Sometimes the estimation of the exchanger performance (3) is re3uired to be calculated on the

gi"en inlet conditions* and the outlet temperature are not -nown until 3 is determined. Thus the

problem depends on the iterati"e calculations. This type of problem may be ta-en care of using

performance e3ui"alent in terms of heating effecti"eness parameter (N)* which is defined as the

ratio of the actual heat transfer to the maximum possible heat transfer. Thus*

7D2EM,

0. ind the D"erall heat transfer coefficient for a shell and tube counter flow heat exchanger wherethe heat exchanged is 00./J >O:hour with the heat transfer area of ?< m/. Assume the 2MT8 as

J


17/19

to cool the oil to a lower exit temperature by increasing the length of the heat exchanger.

8etermine the minimum temperature to which the oil may be cooledP

#. The flow rates of hot and cold water streams running through a parallel flow heat exchanger are%./ -g:. and %.$ -g:s respecti"ely. The inlet temperatures on the hot and cold sides are J$B= and

/%B= respecti"ely. The exit temperature of hot water is #$B=. !f the indi"idual heat transfer

coefficients on both sides are i"e your choice for a parallel flow

or counter1flow heat exchanger* with reasons. =alculate the surface area of the heat exchanger* if

the o"erall heat transfer coefficient is /# 6:m/B=. Ta-e =p of water 9 #.04 -O:-gB=.


18/19

&. =alculate the eynolds number and e"aluate the heat transfer coefficient* hi* using the co1

relations gi"en in the chapter.

#. Similarly* calculate the eynolds number of the fluid flowing through annulus.

=alculation the e3ui"alent diameter of the annulus and find the outside heat transfer

coefficient* ho.

$. Using hi and ho* calculate the o"erall heat transfer coefficients. +ote that it will be a

clean o"erall heat transfer coefficient. !n order to find design outside heat transfer

coefficient using a suitable dirt factor or fouling factor. The tube fouling factor is

suggested by TEMA (table 4.0).

The calculations are based on trial and error. !f the heat transfer coefficient comes out to be "ery

small or the pressure drop comes out to be "ery high* this procedure to be redone for different set

of diameters in the step0.

!hell and t"#e heat exchanger:

The shell and tube heat exchanger also in"ol"es trial and error but it is not as simple as in case of

double pipe heat exchanger.

The design of shell and tube heat exchanger includes*

a, heat transfer re3uired for the gi"en heat duty

b, tube diameter* length* and number*

c, shell diameter*

d, no of shell and tube passes*

e, tube arrangement on the tube sheet and its layout* and

f, baffle sie* number and spacing of the baffles.

The calculation of 2MT8 can be done if the terminal temperatures are -nown. 'owe"er* the

design heat transfer co1efficient (i.e.* heat transfer co1efficient including fouling factor) and the

area are dependent on each other and thus challenges in"ol"e for the estimation. The also

depends upon eynolds number* which depends upon the li3uid flow rate* sies and the number

of tubes. Therefore* is a function of diameter and the no of tubes and the parameter pro"ides the

area.

Moreo"er* can also be calculated is based on shell side co1efficient but then it re3uires tube

number* diameters and pitch. Thus* the abo"e discussion shows that and A are not fully explicit

and re3uires trial and error method of calculation.

The guideline for shell1and1tube calculation is shown in below*

0. energy balance and exchanger heat duty calculation*

/. find all the thermo1physical properties of the fluid*

&. ta-e initial guess for shell1and1tube passes*

#. calculate 2MT8 and T*


19/19

$. Assume (or select) Udirty* based on the outside tube area. =alculate corresponding heat

transfer area* A.

Date post:	01-Jun-2018
Category:	Documents
Upload:	balusapps
View:	218 times
Download:	0 times

Unit II Mechanical Design

Documents