0% found this document useful (0 votes)
41 views

IISRT zAbhijit-Devaraj

Research paper

Uploaded by

VIPAN KUMAR
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
41 views

IISRT zAbhijit-Devaraj

Research paper

Uploaded by

VIPAN KUMAR
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

Fin Geometry for Minimum

D. Poulikakos Entropy Generation in Forced


A. Bejan Convection
Department of Mechanical Engineering, This paper establishes a theoretical framework for the minimization of entropy
University of Colorado; generation (the waste of exergy, or useful energy) in extended surf aces (fins). The
Boulder, Colo. 80309 entropy generation rate formula for a general fin is derived first. Based on this
general result, analytical methods and graphic results are developed for selecting the
optimum dimensions of pin fins, rectangular plate fins, plate fins with trapezoidal
cross section, and triangular plate fins with rectangular cross section.

Introduction
Extended surfaces (fins) constitute one of the most effective output from the engine cycle (the lost power is the difference
design features for promoting heat transfer between a solid between the theoretical Carnot output and the actual power
surface and a stream of fluid. The importance of this thermal output). Therefore, it is the engineer's job to focus on each
design technique in the general area of heat transfer component of the system and, by design, to try to minimize
augmentation and energy conservation is fully recognized by the irreversibility (Sgen) of that component.
the heat transfer community [1,2]. Heat exchangers serve as components in a wide range of
The traditional approach to the optimization of fins power and refrigeration applications. Therefore, in order to
consists of minimizing the consumption (investment) of fin conserve available work (exergy), it is necessary to approach
material for the execution of a specified heat transfer task. the design of such heat exchangers from the point of view of
More than a half-century ago, Schmidt [3] stated intuitively entropy generation minimization. It is important to go further
that a two-dimensional fin must have a parabolic-law pointed and focus on the components of heat exchangers and con-
cross-sectional profile if it is to require the least material ceptualize the design of each such "building block" for
(volume) for a certain heat transfer rate. Schmidt's design minimum irreversibility.
principle was later proved by Duffin [4] who relied on the This paper focuses on the design of fins for minimum
formalism of variational calculus. This design principle has entropy generation in forced convection heat transfer. This
been steadily brought closer to the realities of fin manufac- design philosophy allows us to properly account for the fact
turing and heat exchanger operation by a number of con- that, in addition to enhancing heat transfer, extended surfaces
tributors who have analyzed the role of radiation, two- increase fluid friction. The trade-off between heat transfer
directional heat transfer (curvature), temperature-dependent and fluid friction is a classical dilemma in heat exchanger
thermal conductivity and variable heat transfer coefficient design [2, 12]: the irreversibility minimization philosophy
(see, for example, references [5-7]). Many of these con- places this trade-off on a solid foundation, as heat transfer
tributions have been summarized by Kern and Kraus [8]. The and fluid drag are both mechanisms for entropy generation.
essence and practical limitations of this design philosophy are In this paper, the competition between enhanced thermal
discussed in a recent paper by Kraus and Snider [9]. contact and fluid friction is settled when the heat transfer
The objective of this paper is to outline an entirely different irreversibility and the fluid friction irreversibility add to yield
approach to the optimization of fins. This approach consists a minimum rate of entropy generation for the fin.
of calculating the entropy generation rate of one fin, and This paper begins with a derivation of the formula for the
minimizing it systematically. rate of entropy generation in an arbitrary fin engaged in
The first and second laws of thermodynamics, taken forced convection heat transfer. Based on this general result,
together, state that the entropy generated by any engineering it is shown how the geometric parameters of common fin
system is proportional to the work lost (destroyed) irreversibly shapes can be selected so that the fin saves the most exergy
by the system. This truth is expressed concisely as the Gouy- (available work) while performing its specified heat transfer
Stodola Theorem [10] function. For simplicity, throughout most of this study the
classical fin heat transfer model [13] is adopted, whereby the
" l o s t — -"O Lj 'Jgen (1) fin is slender enough so that the conduction process can be
all regarded as unidirectional. It is further assumed that the
system
components properties of the fin material and those of the external fluid
are constant. The external flow is assumed uniform and
where WlQSt is the lost available work (lost availability, or lost parallel to the base surface of the fin.
exergy) [11], T0 is the absolute temperature of the en-
vironment, and Sgen is the entropy generated in each com- Entropy Generation Due to Convective Heat Transfer
partment of the system. Equation (1) implies that the ther- From a Single Fin
modynamic irreversibility (entropy generation) of each system
component contributes to the aggregate loss of available work The entropy generated by a single fin in crossflow can be
in the system (Wlosl). For example, in a heat engine the en- evaluated based on the general model presented in Fig. 1.
tropy generated in one component (e.g., the condenser) is Consider, an arbitrary fin suspended in a uniform stream with
responsible for a proportional share of the loss in power velocity, U„, and temperature, T„. The heat transfer, qB, is
driven by the temperature difference between the fin base, TB,
and the free stream, TK. In addition, the crossflow
Contributed by the Heat Transfer Division for publication in the JOURNAL OF
arrangement is responsible for a net drag force, FD, which is
HEAT TRANSFER. Manuscript received by the Heat Transfer Division October' transmitted through the fin to the base wall.
30, 1980. As shown in Fig. 1, we choose a control volume which is

616/Vol. 104, NOVEMBER 1982 Transactions of the ASME


Copyright © 1982 by ASME
Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
fixed relative to the fluid. The environment (the wall) "drags"
the fin with a speed U^ through the stationary fluid, by
applying a net force FD tangentially to the control surface. In
the steady state, the first and second laws of thermodynamics
dictate, respectively,
QB-QC ,+FDUa3=0 (2)
q« QB
S„ n =
To, TB >o
0)
where q„ is the net heat transfer rate from the control volume c=>Q-
to the rest of the fluid, and Sgen is the rate of entropy
generation associated with the heat and fluid flow
arrangement. Combining equations (2) and (3) yields
QBO. B FDUX
S = +• (4)
Tj(l + 8BITK) TL
In this expression 6B is the temperature difference between the CONTOL VOLUME
base of the fin and the free stream, TB — TX.
The fin entropy generation rate, (4), is a remarkably simple
result which demonstrates that inadequate thermal con- Fig. 1 Schematic of a general fin in a corrective heat transfer
ductance and fluid friction contribute hand-in-hand to the arrangement
degrading of fin thermodynamic performance. The first term
on the right-hand side of equation (4) represents the entropy irreversibility contributions, heat transfer and fluid friction
generation due to heat transfer across a nonzero temperature (equation (5)), are both relevant only in the case of forced
difference, while the second term is the entropy generation convection. This is the case considered in the present paper. In
associated with fluid friction, natural convection the designer must be concerned with only
the heat transfer part of expression (4): in this case the fluid
*^gen ~~ WgenJheauransfer "*" Wgen/fluid friction (5) friction irreversibility is due to the dissipation of work
Note that the fluid friction entropy generation rate is equal to produced not by man, but by the gravity-driven circulation
the mechanical power needed for dragging the fin through the seen as a "heat engine cycle" [14].
fluid, FD t/o„, divided by the absolute temperature of the fluid. An examination of expression (4) leads to the conclusion
It is proposed to size an individual fin so that the irrever- that in order to minimize the heat transfer contribution to Sgen
sibility contributed by the fin to the larger system (the heat one must minimize the base stream temperature difference, 6B
exchanger) is a minimum. Mathematically, this design ap- (note that (S gen ) heat transfer varies monotonically with 6B). In
proach consists of minimizing expression (4). The power of practical terms, however, to minimize 6B would imply the use
expression (4) is that it brings together the conflicting issues of of an infinitely large fin: such a fin would be impractical and
excessive thermal resistance versus loss of fluid pumping thermodynamically undesirable, because an infinitely large
power: in equation (4), these two effects are weighed relative fin would have an infinite (S gen ) fluid friction> hence, an infinite
to one another, not as heat transfer rate versus pumping S gen . We learn that although the fin size influences
power (such a comparison would not be appropriate [12]), monotonically the base stream temperature difference, it
but as heat transfer irreversibility versus fluid friction plays a crucial trade-off role in the thermodynamic per-
irreversibility. It is important to keep in mind that the two formance (Sg<;n).

Nomenclature (cont.)

m = fin conduction parameter, t/oo velocity of free stream


slenderness ratio, LID, of equations (8) and (16) V parameter, equation (20)
triangular plate fin (Fig. M = property group, (k/X)i/2/ Wlo* rate of lost available work
2(d)) prl/6 (exergy, availability), [W]
b = breadth of plate fin (Figs. Ns = entropy generation number, X longitudinal coordinate
2(b), 2(c)) defined in equations (10), (Fig. 2)
fluid friction irreversibility (18), (24), and (28) y transversal coordinate (Fig.
coefficient, equations (11), Nu = Nusselt number 2(d))
(19) Pr = Prandtl number a half-angle (Fig. 2(c))
B' fluid friction irreversibility QB = base heat transfer 7 slenderness ratio LID (Fig.
coefficient, equation (25) Qco = heat transfer from control 2(a)) or L / b (Fig. 2(b))
B" = fluid friction irreversibility volume to remaining fluid 8 plate thickness (Figs. 2(b),
coefficient, equation (29) (Fig. 1) 2(d))
Cn = drag coefficient, equation R = parameter, equation (26) &B base plate thickness (Fig.
(12) Re, = Reynolds number, jU^/v 2(c))
c, = skin friction coefficient, *^gen = entropy generation rate 8e tip plate thickness (Fig.
equations (17) and (27) [W/K] 2(c))
D = pin diameter (Fig. 2(a)), TB = absolute temperature of fin eB base-stream temperature
and base width (Fig. 2(d)) base difference, TB — T„
FD = drag force To = absolute temperature of A thermal conductivity of fin
h = heat transfer coefficient environment material
k = thermal conductivity of r. = absolute temperature of Me> A'B parameters, equation (26)
fluid free stream V kinematic viscosity of fluid
L = fin length (Fig. 2) u = parameter, equation (29) P density of fluid

Journal of Heat Transfer NOVEMBER 1982, Vol. 104/617

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Fig. 2 Four common fin geometries: (a) pin fin; (b) rectangular plate
fin; (c) rectangular plate fin of trapezoidal profile; (d) triangular plate fin
of rectangular profile
P i>2U„ReLReDCD
In what follows, we examine the fin size trade-off in detail, (9)
by minimizing the rate of entropy generation of some of the 27V
most common fin geometries encountered in practice. The
additional assumption that the temperature difference, 8B, is where the drag coefficient is CD=FD/(\/2 p U„lDL). Since
small compared with the absolute temperature is made the external flow is assumed known, the Reynolds number is
used as dimensionless notation for the two dimensions of the
6B<<Tm (6) pin fin, Re fl = U^D/v, ReL = U^LIv.
This assumption is fully justified in the case of "enhanced In this study, we are adopting the design viewpoint that the
thermal contact" applications such as the main counterflow " j o b " of the individual fin is to transfer heat at a specified
heat exchanger for a helium liquefier, where 6B/Ta <0.1 [15]. rate, qB, from the wall to the known stream, in the least
Consequently, the entropy generation expression takes the irreversible manner possible (with minimum generation of
simpler form entropy). Therefore, the entropy generation number [16] for
this design problem is constructed as
QB&B +FDU«
e —
•""gen T 1 ' T
* oo -* oo
Ns = Sgm/[
The size trade-offs presented later in this paper are based on \kvTj)
minimizing equation (7) subject to the following constraints:
(fc/\)>
• c o n s t a n t ^ , T„, Um
• constant fluid properties
In applications where assumption (6) is not valid, as in the
case of steam-to-air heat exchangers, the designer must start
Nu 1 W.4^(A)"'^]
with the complete expression, equation (4). BCDReLReD (10)

Pin Fins where B is a fixed dimensionless parameter that accounts for


Consider first the pin fin geometry shown in Fig. 2(a). This the importance of fluid friction irreversibility relative to heat
geometry is one of the simplest, because it depends on only transfer irreversibility,
two dimensions: the length, L, and the diameter of the cir-
B= pv3kTa>/qB2 (11)
cular cross section, D. According to the unidirectional heat
conduction model described in the Introduction, the Note that parameter B is known as soon as the fluid
relationship between base heat flux and base stream tem- properties, temperature, and the base heat transfer rate are
perature difference is [8] specified. If the pin fin is slender, the Nusselt number and the
1/2 drag coefficient can be evaluated from the results developed
QB / Ah \
6B= m (8) for a single cylinder in cross flow [17]
- 2
' =\-kF)
kD m ta.nh(mL) K R e D < 4 , Nu = 0.998 R e £ , 0 " P r 1 / 3

Substituting this expression into equation (7) gives the total CD = 10ReD-°-6
entropy generation rate as 4 < R e f l < 4 0 , Nu = 0.919 Re£,°-385Pr1/3
c
QB C i , = 5.484 ReD-0246
4 0 < R e £ ) < 4 x l 0 3 , N u = 0.683Re jD a466 Pr 1/3
•^ r„ 2 (XAr) 1 / 2 Nu 1 / 2 Re z > -f-tanhr2Nu 1 / 2
CD = 5.484 Re B -°- 2 4 6

618/Vol. 104, NOVEMBER 1982 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


CD
en
CD

Q"
CD
cr

Re D
Fig, 3(a) Minimization of entropy generation in a pin fin of optimum
length, equation (13) (M = 100)

e
10

5
- O.OI
10 100 1000

s ^
Re D
N
N Fig. 4 Thermodynamic optimization of a pin fin subject to fixed
\ slenderness ratio: entropy generation number versus pin diameter (B =
Nj^euopt
IO 10 - 7 , M = 100, range 40 < Re < 4000)

straightforward manner by solving dNs/dReL=0. The op-


id - Recppt\^ \. timum pin length calculated in this manner is

ReL.opt "
Rej
2Nu 1/: (4)i/ariDh-[(T 8
3 yn]
(13)
IO - The engineering significance of result (13) is that the op-
timum pin length can be calculated immediately, provided
Re^, is specified. Substituting equation (13) into equation (10),
we obtain the minimum Ns corresponding to optimum pin
IO i i i i i i length, Ns (Re Lopt , Re^,). This function was minimized
numerically, as shown in Fig. 3(a): the entropy generation
10 icf id 0 io8 10 io4 number Ns [ReLopt (ReD), ReD] has a clear minimum with
respect to pin diameter. Figure 3(b) summarizes our
B numerical results for optimum pin diameter Re Dopt . The same
graph shows also the optimum pin length, calculated by
Fig. 3(b) Optimum pin fin diameter and optimum length for absolute
minimum irreversibility (M = 100)
substituting Re flopt into equation (13). The dashed portions of
the curves on Fig. 3(b) show the domain in which Lopt/Dopt <
5, i.e., where the "slender pin fin" model [equation (12)] loses
4x 103 <Re f l <4x IO4, Nu = 0.195 Refl°-618Pr1/3 its accuracy.
CD = 1.1 An alternative approach to sizing a pin fin for minimum
irreversibility consists of determining the optimum diameter,
4x 10 <Re D <2x IO , Nu = 0.0268 R e ^ ^ P r 1 7 3
4 5
ReD _opt, subject to fixed slenderness ratio,
CD = \A (12)
7=_L (14)
The entropy generation number, Ns, emerges as a function
of five dimensionless groups, two pertaining to fin geometry This constraint stems from practical limitations encountered
(ReL, ReD), and three accounting for the working fluid and in the process of manufacturing a surface covered with a large
for the fin-stream convective arrangement (Pr, k/\, B). number of fins. The entropy generation number (10) can be
Minimization of Ns with respect to ReL is achieved in a expressed as a function of ReD and y,

Journal of Heat Transfer NOVEMBER 1982, Vol. 104/619

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Fig. 5(a) Optimum pin diameter versus friction irreversibility
parameter B(M = 100)

is summarized in Fig. 5(a), in the wide range 1 < ReD < IO6,
IU -14 10 " 1 6 < B < 10 and 5 < 7 < 15. When the slenderness ratio
is fixed, the optimum pin diameter decreases as B increases
|6'°— ——___ (i.e., as the fluid friction irreversibility becomes more im-
4 id" —•—-——__________ portant in the irreversibility total, equation (5)). For design
IO
:=:::::::::: purposes, the same information is reported in Fig. 5(b) as

id
if == Re A o p t versus 7.

Plate Fins
$ — - — — This section focuses on fin geometries modeled as thin
conducting plates parallel to the flow direction. Fins of this

- if S = = =
"5. type vary widely with respect to the shape and cross section of
CD the surface swept by the flow: in many cases the fin is wide at
cr id the base and narrow at the tip, in approximate agreement with
Schmidt's principle of material reduction in a fin with fixed
heat transfer [3]. Below, we consider the minimum entropy
generation design of three frequently used plate fin
geometries, illustrated in Figs. 2(b), 2(c), and 2(d).
IO

Rectangular Plate Fin. The simplest plate fin geometry is


shown schematically in Fig. 2(b). The minimum irreversibility
i • • I . . design of this fin requires the selection of three geometric
0 IO 15 parameters, the length, L, the breadth (length swept by fluid),
b, and the plate thickness, 5, where L > > b > > 8. The
t relationship between heat transfer and base fluid temperature
difference is [8]:
Fig. 5(b) Optimum pin diameter versus slenderness ratio (M = 100)

0.77 M QB
Nx=- + 2.74257R60 1 (15) (16)
1.65 kbbm tanh (mL) • * " & )

Rer 3 tanh -Re f l c


M As in the preceding section, the study is placed in the limit
where M is the combined property group M = where the plate fin is slender (b < < L); in this limit we rely
(A:/X) 1/2 /Pr 1/6 . on laminar heat transfer and skin friction results for two-
Representative results of the numerical work of minimizing dimensional flat plates [18]
expression (15) are shown in Fig. 4. The entropy generation
number, Ns, has a clear minimum with respect to pin h = 0.664 Re i , 1 / 2 Pr 1 / 3 ,C / = = 1.328 Re„- 1 / 2
diameter, Re fl , when 7, B, and M are fixed. The Ns-ReD pUjbL
dependence (Fig. 4) is reported not for design purposes, but to (17)
show the designer the "sharpness" of the irreversibility
Based on equations (16, 17), the entropy generation number
minimum to be achieved through the precise selection of pin
for a rectangular plate fin in laminar flow becomes
diameter.
Figure 4 shows also that the optimum pin diameter, Re A o p t , qB2Ua
Ns = Sgm/
increases if the slenderness ratio decreases. This general trend V kvTj )

620/Vol. 104, NOVEMBER 1982 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


(k/\y however convenient, are valid only in the laminar regime\
1.15 Re 6 1 / 2 Re 6 3 / 4 Pr 1 / 6 tanh(mL) ReA,opt < 5 x 10 5 , and in the slender shape limit 7 opt > > 1.
Similar results can be developed numerically for the turbulent
+1.328 BReLRcb (18) regime, by replacing equations (17) with appropriate
where correlations for turbulent heat transfer and skin friction.
B = pv3 kT^lqs1, Rectangular Plate Fin of Trapezoidal Profile. A relatively
more complex plate fin geometry is represented in Fig. 2(c).
mL=lA5
(I) Pr 1 / 6 Re,Re„

From the outset one notes that the thickness, 5, appears


4
Re, (19) This time the fin longitudinal section is trapezoidal, again, in
the spirit of E. Schmidt's principle of material (volume)
minimization [3]. This geometry has another important
only in the heat transfer term of TVj, equation (18), con- advantage over the rectangular shape of Fig. 2(b): the tapered
sequently, Re 6 does not play a trade-off role in the profile makes trapezoidal fins accessible to metal-cutting
minimization of Ns. Since in most practical applications 6 is operations designed to remove the fin-to-fin material. The
determined by considerations such as price, availability, and minimum entropy generation design of this class of fins can
machinability of sheet metal, it makes engineering sense to be approached along the same lines as the design of rec-
regard Re s as fixed. The minimization of Ns with respect to tangular fins. However, trapezoidal fins have an additional
Reft and Re L is achieved by solving the simultaneous set of geometric parameter in their constitution, namely, the half-
equations dNs/dReb = 0 and 8Ns/dReL = 0. Eliminating angle, a.
Re L between these two equations leads to an implicit result of In the interest of brevity, only the key formulas are
the form reported. The entropy generation number can be expressed as
1/2
ln[y1/2+(y+l)1/2]- / k \
.-[ (l + u) 1/2 J ' 0.614?^-J Re, 1
N* :
Sgen/| 2
u = (1.328BRe 6 3 / 2 Re 6 ) - ' (20) F ArPr'
The numerical solution to equation (20) is v = 0.7717. From
the system dNs/dReb = 0 and 8Ns/dReL = 0, we conclude 2 sin a -1 1/2
that the optimum geometry for minimum entropy generation
is explicitly given by [ tana(2 Re L tan a + Re s ) J

Re b, opt = 0.9845 - 2 / 3 R e 6 " : (21) r /oQtj)*•!(/*,)+/,(p,)*ifa,) 1 2Rei+Re4 ,„,,


+ B' r^ ^ (24)
M / 2 R e 1/3 p r -
6B- (22) L KJne)LUR)-LUe)K,(aR) J 2 cos a
Re L , opt =0.685(/t/X)
Dividing equation (22) by equation (21) yields the optimum where
slenderness ratio of the plate fin Pr 1 '' 6 (fcX) l '' 2 piJ co 2 i'T oc
B' = 2.164- 3/4 (25)
1/2
2
QB Reb
7op, = ( ^ ) = 0.696 MB Re 6 (23)
\ O / opt and
In conclusion, the optimum plate fin dimensions, L and b, ^ r 5,(1- tan a) n
can be calculated directly using equations (21) and (22), as /*« =
soon as the base heat flux (qg)t the flow and the sheet L 2 tan a • 1
thickness, 5, are known. The optimum length and the op-
;(1-tan a)
timum breadth both decrease as B increases, in accordance
with the trend discovered in the optimization of pin fins (Figs. V-B
- , "[- +L + k'
2 tan a
3-5). As the plate thickness 5 increases, the optimum length
increases while the breadth decreases, hence, the optimum
slenderness ratio, equation (23), increases. (26)
It is necessary to keep in mind that the formulas (21-23), \ k sin a /

10
*\ B Re.
0a.«»t

Q5 200 ,
Ql 320 \
-' ^ Q05 405\

\X
Ns
• \ \
Re, (5B,opt

\ \

10' - \

I— 1 1 1 1 1 1

Re,,

Fig. 6 Thermodynamic optimization of plate fins of trapezoidal profile


(« = 10 deg, Res = 100). Left side: entropy generation number versus
geometric aspect ratio. Right side: optimum aspect ratio versus friction
irreversibility parameter B ' .

Journal of Heat Transfer NOVEMBER 1982, Vol. 104/621

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


As in the preceding example, we have assumed laminar function of longitudinal position (x). It is worth pointing out
boundary layer flow [18] that the D << L limit is in full agreement with the
unidirectional heat conduction model adopted in the In-
be
L+- troduction.
1.328 The analytical path leading to the entropy generation rate
FD = CfPU, b,Cr
cos a RiT 172 " formula is similar to the method used in the earlier examples.
Again, in the interest of brevity, only the key result is reported
h = 0.664 Re „ 1 / 2 Pr (27) QB1- "— GO

~b Ns = Sgcn/
In addition, the Harper and Brown approximation [8] was Tjv(k\y
employed, whereby the heat transfer from the tip of the fin is 0.868 /o(M)
taken into account by using the "augmented" fin length L + +0.885 5 " Re f l 1 / 2 R e i (28)
Re 6 l / 2 Re D 3 / 4 Pr l / 6 I^u)
Se/2 in place of L (see Fig. 2(c)).
The entropy generation rate (24) depends on four geometric with
parameters Re L , Reft, Re 6 , and a. Whether or not all four Tx^{k\ynn
parameters may be considered independent depends on B"
QB'
specific circumstances, particularly on the constraints faced
by the designer (fixed cost, volume, base area, etc.). In Fig. 6, Re,
we show a sample of numerical results obtained in the case
when b, 5e, and a are fixed by design: the only geometric
M=1.536
(1) Pr 1
Re„ Re, 1/4
(29)

variable in this case is the fin length, L (the dimension per- The entropy generation rate depends on three geometric
pendicular to the wall). Choosing the optimum Re L for parameters, ReD, ReL, and Re 6 . Note that Re 5 does not play a
minimum entropy generation is geometrically equivalent to trade-off role, because it appears only in the heat transfer part
choosing the optimum base thickness Re 6 since Re s = Re5 of expression (28).
+ 2 Re L tan a. The left side of Fig. 6 shows that Ns has a Figure 7 presents a sample of optimum fin size results for
sharp minimum with respect to Re s . The optimum values of cases where the triangle aspect ratio a = LID is fixed. The fin
base thickness Reynolds number are reported in the right half irreversibility, Ns, reaches a clear minimum at a specific value
of Fig. 6 for the case Res = 100 in the range 5 deg < a < 10 of fin base width, ReD; the optimum,fin size, ReD, depends
deg. We find that the optimum fin size (Re* ) decreases as the on the relative importance of fluid friction irreversibility
fluid friction effect (B) becomes more pronounced, in (B"), on the metal-fluid combination (M), and on the plate
agreement with conclusions reached in previous examples. It thickness (Re 6 ). The right-hand side of Fig. 7 is a summary of
is also apparent that in the a range considered, the angle a has Ns results obtained for a number of common metal-fluid
a relatively minor impact on the optimum fin size for combinations involving copper, aluminum, water, and air.
minimum irreversibility. Regardless of combination, the optimum fin size (Re A o p t )
decreases as the triangular shape of the fin becomes, by
Triangular Plate Fin of Rectangular Profile. To the design, more slender.
thermodynamic designer, this geometry is challenging due to
the absence of convenient correlations for heat transfer and
Concluding Remarks
fluid friction in the three-dimensional boundary layer flow
which, in most certainty, will cover the triangular faces of the In this paper we have applied the minimization of entropy
fin. This analytical difficulty can be partially dealt with in the generation (exergy waste) to the design of extended surfaces.
limit D < < L, where the three-dimensional effects will be Using the first and second laws of thermodynamics, the
minor. In this limit, we can approximately treat the sharp- entropy generation rate associated with a single fin of un-
pointed triangular plate as a flat plate in parallel flow, with specified shape and properties was calculated. This general
the special property that the length swept by the flow (y) is a result was then applied to the minimization of irreversibility
7MO*

R^o^,,,,
5'lC

KT" -J I L_J_ 3-»"


10" w
Re0
Fig. 7 Thermodynamic optimization of triangular plate fins (Re s =
100). Left side: Entropy generation number versus base width (Air-Cu,
8 " = 5 x 1 0 " 9 ) , Right side: optimum base witwidth versus geometric
9
aspect ratio a = L/D, D.(Air-Cu.B" = 5 x 1 0 - 9 , Ail
Air-AIB" 4 x 10"
H 2 0 - C u B " = 3.5 x 10 , H20-AIB' 2.5 x 10 " 9 ) .

622/Vol. 104, NOVEMBER 1982 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


in four of the most frequently used fin shapes. The con- this regard the development of heat transfer technology for
clusions of our irreversibility minimization procedure have helium refrigeration and large scale superconducting systems,
been reported in analytical or graphical form. where "good design" has become practically synonymous
In the analysis of the individual fin we have assumed that with "least irreversible design" [19].
the heat transfer rate and the fluid flow parameters are It is important to recognize the fundamental value of the
prescribed. This decision is necessary in order to isolate exergy-conservation design philosophy relative to the local
(remove) the fin from the aggregate system (the heat ex- and temporal character of the economic optimizaton
changer). It is important to note that heat exchangers, too, are procedure followed by individuals in industry. The con-
isolated from their respective aggregate systems (power servation of exergy is clearly what the heat transfer com-
plants) for the purpose of analytical optimization. The munity must learn first and teach others through its technical
decision to isolate a component from the aggregate system journals, if we are to agree on a reasonable course toward
leads, necessarily, to the adoption of design constraints fuel-sufficiency in this world [20].
around the component. For example, in the design of a heat
exchanger we use the notions of "required heat transfer" and
"fixed pressure drop;" in the thermodynamic optimization of Acknowledgment
the fin we have regarded qB and the flow as fixed. The This work was sponsored by the Office of Naval Research,
alternative to optimizing the heat exchanger component-by- The Power Program.
component, as in this paper, would be to minimize the
irreversibility of the heat exchanger, as a whole. However,
this alternative is neither practical nor of fundamental value References
in view of the diversity of heat exchangers and the large 1 Bergles, A. E., Webb, R. L., Junkhan, G. H., and Jensen, M. K.,
number of geometric features which would have to be op- "Bibliography on Augmentation of Convective Heat and Mass Transfer,"
timized simultaneously. Report HTL-19, ISU-ERI-AMES-79206, Iowa State University, May 1979.
2 Junkhan, G. H., Bergles, A. E., and Webb, R. L., "Research Workshop
The chief conclusion of this study is that the size of in- on Energy Conservation Through Enhanced Heat Transfer," Report HTL-21,
dividual fins can be chosen in a way that combines a specified ISU-ERI-AMES-80063, Iowa State University, Oct. 1979.
heat transfer function with a desired (optimum) ther- 3 Schmidt, E., "Die Warmeubertragung durch Rippen," Zeit. d. ver
Deutch Ing., Vol. 70, 1926, pp. 885-889, and 947-951.
modynamic function, namely, the function of conserving 4 Duffin, R. J., " A Variational Problem Relating to Cooling Fins,"
exergy. It was also shown that the thermodynamic-optimum Journal of Mathematics and Mechanics, Vol. 8,1959, pp. 47-56.
fin dimensions can be subjected to additional design con- 5 Maday, C. J., "The Minimum Weight One-Dimensional Straight Fin,"
straints, such as the fixed slenderness ratio for pin fins, ASME Journal of Engineering for Industry, Vol. 96,1974, pp. 161-165.
6 Guceri, S., and Maday, C. J., " A Least Weight Circular Cooling Fin,"
equation (14). Built into all the analytical and graphic results ASME Journal of Engineering for Industry, Vol. 97,1975, pp. 1190-1193.
is the message that the thermodynamic optimization of fins 7 Razelos, P., and Imre, K., "The Optimum Dimensions of Circular Fins
requires a detailed knowledge of the fin-fluid flow in- with Variable Thermal Parameters," ASME JOURNAL OF HEAT TRANSFER, Vol.
teraction. As a first step, we accounted for this interaction by 102, 1980, pp. 420-425.
8 Kern, D. Q., and Kraus, A. D., Extended Surface Heat Transfer,
using well-known correlations developed for two-dimensional McGraw-Hill, New York, 1972.
external flow arrangements. It is felt, however, that much 9 Kraus, A. D., and Snider, A. D., "New Parametrizations for Heat
remains to be done experimentally with the objective of Transfer in Fins and Spines," ASME JOURNAL OF HEAT TRANSFER, Vol. 102,
documenting the local heat transfer and drag characteristics 1980, pp. 415-419.
for fins of various shapes bathed by flows of various direc- 10 Szargut, J., "International Progress in Second Law Analysis," Energy,
Vol. 5, 1980, pp. 709-718.
tions. 11 Kestin, J., "Availability: The Concept and Associated Terminology,"
As a final comment, we would like to address the issue of Energy, Vol. 5, 1980, pp. 679-692.
12 Bejan, A., Discussion of a previously published paper, ASME JOURNAL
thermodynamic optimization versus economic optimization OF HEAT TRANSFER, Vol. 102,1980, pp. 586, 587.
(cost minimization). The former has been the subject of this 13 Gardner, K. A., "Efficiency of Extended Surfaces," Transactions of the
paper, as applied to the sizing of individual fins engaged in ASME, Vol. 67, 1945, pp. 621-631.
convective heat transfer. The latter is, of course, the 14 Tritton, D. J., Physical Fluid Dynamics, Van Nostrand Reinhold, New
philosophy of those who are competitive in the industrial York, 1980, p. 141.
15 Collins, S. C , "Liquefaction Techniques," Ch. 3 in Technology of
world. The two design philosophies, thermodynamic versus Liquid Helium, edited by R. H. Kropschot, B. W. Birmingham and D. B.
economic, do not necessarily lead to identical fin-sizing Mann, NBS Monograph 111, Oct. 1968.
decisions. This is due to the fact that the total cost formula for 16 Bejan, A., "General Criterion for Rating Heat Exchanger Performance,"
a certain piece of heat exchange equipment contains many International Journal of Heat and Mass Transfer, Vol. 21, 1978, pp. 655-665.
17 Gebhart, B., Heat Transfer, McGraw-Hill, New York, 1971, pp.
items in addition to the cost of lost available work (exergy). 212-214, 270.
However, as the cost of fuel ("bottled exergy") continues to 18 Gebhart, B., op. cit., pp. 198, 243.
rise and dominate the total cost figure of equipment for power 19 Hubbell, R. H., and Toscano, W. M., "Thermodynamic Optimization of
and refrigeration, the thermodynamic optimization Helium Liquefaction Cycles," Advances in Cryogenic Engineering, Vol. 25,
1980, pp.551-562.
philosophy leads to designs which come closer and closer to 20 Cambel, A. B., "Preface," Second Law Analysis of Energy Devices and
those demanded by the cost minimization formula. Witness in Processes, Energy, Vol. 5,1980, pp. Ill, IV.

Journal of Heat Transfer NOVEMBER 1982, Vol. 104/623

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

You might also like