See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232862173

Estimation of Nusselt Number in Microchannels of Arbitrary Cross-Section With

Constant Axial Heat Flux

Article in Heat Transfer Engineering · July 2010

DOI: 10.1080/01457630903466647

CITATIONS READS

22 2,454

3 authors:

Ehsan Sadeghi Majid Bahrami

University of Victoria Simon Fraser University
12 PUBLICATIONS 507 CITATIONS 195 PUBLICATIONS 5,465 CITATIONS

SEE PROFILE SEE PROFILE

Ned Djilali
University of Victoria
317 PUBLICATIONS 13,948 CITATIONS

SEE PROFILE

All content following this page was uploaded by Majid Bahrami on 04 December 2014.

The user has requested enhancement of the downloaded file.

 Proceedings of the Sixth International ASME Conference on Nanochannels, Microchannels and Minichannels
ICNMM2008
June 23-25, 2008, Darmstadt, Germany

ICNMM2008-62198

ESTIMATION OF NUSSELT NUMBER IN MICROCHANNELS OF ARBITRARY CROSS-SECTION

WITH CONSTANT AXIAL HEAT FLUX

Ehsan Sadeghi Majid Bahrami

Department of Mechanical Engineering Mechatronic Systems Engineering
University of Victoria Simon Fraser University
ehsans@uvic.ca mbahrami@sfu.ca

Ned Djilali
Department of Mechanical Engineering
University of Victoria
ndjilai@uvic.ca

ABSTRACT
In many practical instances such as basic design, , ,, = Coefficients used in Eq. (10)
parametric study, and optimization analysis of thermal systems, = Hydraulic diameter, m
it is often very convenient to have closed form relations to · = Complete elliptic integral of the second kind
obtain the trends and a reasonable estimate of the Nusselt = Thermal convection coefficient, W/m2K
number. However, finding exact solutions for many practical = Polar moment of inertia, m4
singly-connected cross-sections, such as trapezoidal = Dimensionless polar moment of inertia
microchannels, is complex. = Side of a rhombic cross-section, m
In the present study, the square root of cross-sectional area is = Characteristic length scale, m
proposed as the characteristic length scale for Nusselt number. = Nusselt number
Using analytical solutions of rectangular, elliptical, and = Number of sides of a polygon
triangular ducts, a compact model for estimation of Nusselt
= Perimeter, m
number of fully-developed, laminar flow in microchannels of
" = Heat flux, W/m2
arbitrary cross-sections with “H1” boundary condition (constant
= Temperature, K
axial wall heat flux with constant peripheral wall temperature)
= Mean temperature, K
is developed. The proposed model is only a function of
= Wall temperature, K
geometrical parameters of the cross-section, i.e., area,
= Mean wall temperature, K
perimeter, and polar moment of inertia. The present model is
verified against analytical and numerical solutions for a wide = Velocity in z- direction, m/s
variety of cross-sections with a maximum difference on the = Mean axial velocity of the fluid , m/s
order of 9%.
Greek symbols
= Thermal diffusion coefficient, m2/s
NOMENCLATURE = Aspect ratio
= Dimensionless temperature
= Mean dimensionless temperature
= Cross-sectional area, m2
= Area of limiting rectangle of a cross-section, m2 Δ = Deviation from limiting rectangle
= Area of limiting triangle of a cross-section Δ = Deviation from limiting triangle

1 Copyright © 2008 by ASME

1 INTRODUCTION 1960’s. Several hundred papers on Nusselt number have been
Micro fabrication technologies make it possible to build micro- published which illustrates the importance of this topic, and
fluidic, silicon-based microchannels of different cross-sections also indicates that the development of a general predictive
in micro flow devices such as micro-heat sinks, micro-biochips, model is difficult.
micro-reactors and micro-nozzles [1, 2]. It is evident that the In the present study, a compact predictive model is proposed to
understanding of the micro scale transport phenomena is estimate the Nusselt number for a wide variety of cross-section
important for the designer of microfluidic devices. For this geometries of channels and microchannels. The proposed
reason, many studies have been conducted in order to analyze model is only a function of geometrical parameters of the cross-
the behavior of the convective flow through microchannels. section, i.e., perimeter, area, and polar moment of inertia.
Reviews of these works have been presented in [3- 6].
Using the electrochemical technique to obtain mass transfer 2 PROBLEM STATEMENT
coefficients for rectangular microchannels, Acosta et al. [7] Consider a microchannel with an arbitrary constant cross-
showed that the smooth channel correlations for large-sized section as shown in Fig. 1. To calculate the Nusselt number,
channels hold for smooth microchannels in laminar and following assumptions are made:
turbulent regimes. Rahman and Gui [8] investigated the laminar • Fully developed, steady-state, laminar, and continuum
forced convection of water in etched silicon microchannels; flow
they found that the Nusselt numbers were higher than those • Incompressible flow
predicted by analytical solutions for developing laminar flows • Constant viscosity and thermal conductivity
through rectangular channels. • Negligible viscous dissipation
Shah and London [9] studied heat transfer in ducts for a wide • Negligible rarefaction, slip/jump and surface effects
variety of cross-sectional geometries. They reviewed existing • Constant cross-sectional area A and constant
studies and calculated Nusselt number for isothermal and perimeter P
isoflux wall conditions numerically.
• Constant axial wall heat flux with constant peripheral
Muzychka et al. [10] used the square root of the cross-sectional
wall temperature (“H1” boundary condition)
area of a channel to define the Nusselt number and showed that
this length scale is more appropriate than the hydraulic
diameter. They proposed two bounds for the Nusselt number of
various geometries and developed correlations for these two
bounds based on the Fanning friction factor of a rectangular
channel.
Duncan and Peterson [4] reviewed micro scale convective,
conduction and radiation heat transfer. Peng and Wang [11]
" " ,
reported a review of their extensive research on the one-phase
and two-phase micro scale convective heat transfer. Peng et al.
FIGURE 1. LONG MICROCHANNEL OF ARBITRARY
[12] analyzed the role of the dimensions of the rectangular
CONSTANT CROSS-SECTION, √ .
microchannels on the Nusselt numbers in laminar and turbulent
regimes. They found a strong dependence of the Nusselt
number on the aspect ratio of the microchannel and proposed
Using the above assumptions, the energy equation reduces to
some correlations in which the empirical constants are
[9]:
functions of the microchannel dimensions [12]. Bailey et al. [6]
concentrated their attention on the single phase forced with on the channel wall (1)
convection through microchannels and concluded that the
literature is inconclusive with respect to the effect of
miniaturization on heat transfer and pressure drop. Many of where, is the axial velocity and is the thermal diffusion
these studies [4, 6, 11, 12], highlighted that, in many cases, the coefficient of the fluid. The energy balance can be expressed
experimental results for laminar flow through microchannels as:
deviate significantly from the predictions of the conventional
". (2)
theory.
Our literature review indicates that there is no general
predictive model for the Nusselt number of an arbitrary cross- where, ", , and are the wall heat flux, constant pressure
section microchannel. Also existing experimental data are specific heat, and the mean temperature of z planes,
inconsistent and sometimes contradictory. For instance, Jiang et respectively. is the mass flow rate of the fluid, .
al. [13] and Wu et al. [14] reported that in the laminar regime Applying Eq. (2) and introducing the dimensionless
the Nusselt number increases with the Reynolds number with parameter / " / , the energy equation, Eq.
an exponent ranging from 0.3 to 1.96 whereas Gao et al. [15] (1), after some algebraic manipulations can be simplified to:
and Qu et al. [16] showed that in the laminar regime the Nusselt
number decreases when the Reynolds number increases.
Analytical, experimental, and numerical models have been w, with 0 on the channel wall (3)
developed to predict the Nusselt number in ducts since the

2 Copyright © 2008 by ASME

where P, , are the perimeter, the area, and an appropriate number [20], similar trends are expected for the Nusselt
characteristic length scale, respectively. Also, is the mean numbers of these shapes. However, in spite of similar trends,
axial velocity of the fluid. The Nusselt number based on Eq. (3) the relative difference between Nusselt numbers of rectangular
can be expressed as: and elliptical channels is high and reaches up to 35% for small
values of aspect ratio ( 0.1 .Thus, developing a general
"/ 1 geometric model for the Nusselt number is complicated or
(4) impossible. The goal of the present study is to develop a
geometric model applicable to several cross-sections.
where, is the area-weighted average dimensionless 10
temperature of the channel cross-section. To find the Nusselt
number, must be calculated by solving Eq. (3). This equation 9 ellipse
rectangle
is coupled with the momentum equation. As a result, finding
8
analytical solutions of Eq. (3) for complex geometries of
microchannel cross-sections is highly unlikely. 7

Dh
3 CHARACTERISTIC LENGTH SCALE

Nu
6

To define the Nusselt number, it is conventional to use the ratio

of area over perimeter 4 / ; i.e. the hydraulic diameter, 5

as the characteristic length scale for non-circular channels. 4

Figure 2 compares the analytical solutions of the “H1” Nusselt
number for elliptical and rectangular ducts using hydraulic 3

diameter and square root of cross-sectional area as

2
characteristic length. Elliptical and rectangular cross-sections 0 0.25 0.5 0.75 1

cover a wide range of singly connected microchannels and also ε = b/a

these are two bounds of hyperellipse cross-sections, / (a)
/ 1, within 2 ∞. Marco and Han [17] expressed
“H1” Nusselt number of rectangular ducts in series form which 20

was correlated by Shah and London [9] within 0.03% . 18

ellipse
rectangle
8.235 1 2.0421 3.0853 2.4765 16

(5) 14

1.0578 0.1861 I 12
Nu√A

10
where, the Nusselt number is based on the hydraulic diameter
of the cross-section. It can be converted to the Nusselt number 8
based on the square root of area through:
6

P 4
√ (6)
4√ 2
0 0.25 0.5 0.75 1
ε= b/a
Tyagi [18] found an analytical solution for “H1” Nusselt
(b)
number of elliptical ducts.
FIGURE 2. COMPARSION BETWEEN ANALYTICAL
9π 1 ε 1 ε 6ε SOLUTIONS OF NU FOR ELLIPTICAL AND RECTANGULAR
√ (7)
1 ε √ 17 1 ε 98ε DUCTS USING: (a) HYDRAULIC DIAMETER AND (b)
SQUARE ROOT OF AREA AS CHARACHTERSITIC LENGTH.
where, · is the complete elliptic integral of the second kind.
As can be seen in Fig. 2, the square root of area √ , as the
characteristic length scale, leads to more consistent trends in 4 PROPOSED MODEL
the solutions of elliptical and rectangular ducts. These shapes Bahrami et al. [2, 19] developed a compact model for Poiseuille
cover a wide variety of cross-sections; thus, it can be concluded number, √ , which is only a function of geometrical
parameters of the cross-section, i.e., perimeter, area, and polar
that the square root of area √ is a more appropriate length
scale for Nusselt number as Muzychka et al. [10] proposed. moment of inertia. They started from the analytical solution of
elliptical channel and established their model. It was shown that
Bahrami et al. [2, 19] showed that the difference between the
their model is applicable to a wide range of cross-sections.
Poiseuille number, √ , of elliptical and rectangular ducts is Considering the analogy between the Poiseuille number and the
less than 8% over a wide range of aspect ratio. Considering the Nusselt number [20], the approach proposed by Bahrami et al.
analogy existing between the Poiseuille number and the Nusselt [2, 19] is followed to develop a general model for “H1” Nusselt

3 Copyright © 2008 by ASME

number as a function of geometrical parameters of the cross- Figure 3 shows that Nusselt numbers for different geometries
section. are not in agreement. However, two bounds can be recognized
The solution of elliptical channel is selected to develop a for Nusselt numbers, where, the upper bound is provided by
general model, not because it is likely to occur more frequently elliptical and rectangular cross-sections and the lower bound is
in practice, but rather because of its unique geometrical presented by triangular cross-sections. Thus, different cross-
property of its Nusselt number solution, Eq. (7). Applying the sections can be categorized into three groups: 1) cross-sections
definitions of perimeter, area, and polar moment of inertia for close to rectangle/ellipse, 2) cross-sections close to triangle,
elliptical cross-section into Eq. (7), the Nusselt number for and 3) cross-sections close to both of the previous groups.
elliptical ducts can be rewritten as: Equation (9) shows the Nusselt number as a function of ,
√ / for elliptical ducts. It implies that if there is a general
1
144 √ model for various geometries, it can be a function of , √ / .
4
√
17 P 4 (8) Thus, a new model is proposed as a product function of
17 geometric parameters of the cross-section, i.e. dimensionless
polar moment of inertia , and square root of area over
perimeter √ / .
where, is dimensionless polar moment of inertia about the
center of the cross section, / . The second bracket in the
√
right hand side of the above equation is approximately equal to √
(10)
unity; therefore, this equation can be simplified to: P

The proposed model, Eq. (10), consists of three submodels:

√
√ 8.47 (9) upper, lower, and middle. Coefficients , , and are
P determined by minimizing differences between Eq. (10) and
analytical solutions of elliptical and rectangular ducts for the
The above equation shows that the Nusselt number can be
upper model, and that of right triangular and isosceles
expressed as only a function of geometrical parameters for
triangular ducts for the lower model. The middle model is
elliptical ducts. To validate this geometrical function as a
defined as the average of the upper and the lower models.
general model, the exact values of Nu for some cross-sections
are compared with the geometrical function, Eq. (9). A Coefficients , , and are listed in Table 1 for the proposed
comparison between the analytical solution of rectangular model.
channel and the solution obtained from Eq. (9) shows that:
i. Equation (9) agrees with analytical solution when
TABLE 1. COEFFICIENTS OF THE PROPOSED MODEL.
0.4.
ii. For lower aspect ratios, the difference becomes
Coefficient
relatively high, 17.1% when 0.05.
Model
iii. Moreover, when Eq. (9) is applied to other geometries
such as rhombus and trapezoid, results are far from
analytical or numerical solutions for some aspect upper 108.84 1.04 1.09
ratios and convince us that Eq. (9) is not appropriate
to use as a general model. lower 7.91 0.38 0.15
20

middle
18 Ellipse
Rectangle
2
Rhombus
16
Trapezoid
Circular Segment To apply the model, first we should find out which submodel is
14 Circular Sector
Right Triangle appropriate to use. Relative deviation of the channel cross-
Isosceles Triangle
I 12 sectional area from rectangular and triangular shapes is
Nu√A

Sinusoid
Rectangle with Semicircular Ends proposed as a measure to select the proper submodel. The
10
reference rectangular and triangular cross-section is defined
8 based on the limiting cases of geometry. These limiting cases
generally occur in bounds of aspect ratios or shape angles. For
6 instance, as shown in Table 2, the trapezoid yields rectangle
4
when the aspect ratio goes to zero and/or goes to 90 and
approaches the triangle when the aspect ratio becomes higher
2
0 0.25 0.5 0.75 1 and the upper base goes to zero. Figure 4 illustrates the
ε procedure to apply the model.

FIGURE 3. NUSSELT NUMBER FOR VAROIUS CROSS-

SECTION GEOMETRIES VERSUS ASPECT RATIO.

4 Copyright © 2008 by ASME

TABLE 2. LIMITING CASES FOR TRAPEZOIDAL CROSS-
SECTION. 5 VALIDATION OF THE MODEL
cross-section limiting cases appropriate In this section, the present model is compared with numerical
submodel results of Shah and London [9] for various cross-sections.
Equation (5) is used to convert the reported Nusselt number in
the literature from base to √ base. Geometrical parameters
required for evaluating the Nusselt number are reported for
upper various geometries in Table 3.

5.1 Elliptical and Rectangular Channels

Elliptical and rectangular cross-sections are bases for
developing the upper model. Rectangular cross-section can be
manufactured by photolithographic-based processes such as Si
Atri chemical etching and has various engineering applications [5].
Figure 5 shows the comparison between numerical results [9]
middle and the model for elliptical and rectangular cross-sections. The
difference between the present model and numerical results is
less than 8% for these shapes.
Arec 20

18 data [3]
model
16 model - 10 %
model +10 %

14
lower
I 12
Nu√A
10

6
Start
4

2
0 0.25 0.5 0.75 1
ε= b/a

∆ 1 (a)

∆ 1 20

18 data [3]
model
model - 10%
16
model + 10%

∆ 0.2 ∆ 0.2
∆ ∆ ∆ 14

I 12
Nu√A

10
0.2 ∆ 0.2
8

6
upper model middle model lower model
4

2
0 0.25 0.5 0.75 1
ε= b/a
(b)
√
√
P FIGURE 5. COMPARSION OF PRESENT MODEL AND
NUMERICAL VALUES [9] FOR: (a) ELLIPTICAL CHANNEL,
FIGURE 4. FLOWCHART FOR APPLYING THE PROPOSED (b) RECTANGULAR CHANNEL.
MODEL.

5 Copyright © 2008 by ASME

 TABLE 3. GEOMETRIC PARAMETERS OF STUDIED CROSS-SECTIONS.
cross-section aspect ratio √ /

√ 1
4 1 4

√ 1
2 1 12

√2 1
1 9 12
2 1

√2 1 7
2 1 √1 9 9

1 2 3 3
1 1
φ 2 √ 36

1
φ 4 6

1 1 3
tan /
1 2 n tan / 6 tan /

4 1
2 2 2 4 1

√2 1 12 9
2 5
2 12 32
2

(1)
1 (2)
2 1

6 Copyright © 2008 by ASME

5.2 Triangular Channels 5.4 Other Cross-Sections
The lower model is developed based on the solutions of The present model is also compared with numerical results for
different types of triangular cross-sections. Figure 6 shows the a number of geometries listed in Table 3. The results are
comparison between numerical results [9] and the model. The shown in the following tables and figures.
maximum difference between the present model and numerical
results is 6.6%. TABLE 4. COMPARSION OF PRESENT MODEL AND
NUMERICAL VALUES [9] FOR RHOMBIC CHANNEL.
10
data [9] model % diff.
data [21], isosceles triangle 10 0.087 5.318 5.547 4.13
data [9], isosceles triangle
model, isosceles triangle
20 0.176 4.201 4.511 6.87
8 data [9], right triangle 30 0.268 3.849 4.018 4.19
model, right triangle
40 0.364 3.703 3.721 0.49
45 0.414 3.663 3.615 -1.33
I 50 0.466 3.642 3.528 -3.26
Nu√A

6
60 0.577 3.618 3.398 -6.48
70 0.700 3.610 3.314 -8.93
80 0.839 3.609 3.511 -2.78
4 90 1 3.608 3.489 -3.41

TABLE 5. COMPARSION OF PRESENT MODEL AND

2 NUMERICAL VALUES [9] FOR REGULAR POLYGONAL
0 0.25 0.5 0.75 1
CHANNEL.
ε = b/a
data
model % diff.
FIGURE 6. COMPARISON OF PRESENT MODEL AND [9]
NUMERICAL VALUES FOR ISOSCELES AND RIGHT ANGLE 3 3.546 3.368 -5.27
TRIANGULAR CHANNELS. 4 3.608 3.726 3.17
5 3.678 3.809 3.46
5.3 Trapezoidal Channel 6 3.724 3.872 3.81
Trapezoidal cross-section is an important geometry since some 7 3.766 3.915 3.81
microchannels are manufactured with trapezoidal cross-sections 8 3.780 3.945 4.18
as a result of the etching process in silicon wafers [5]. 9 3.797 3.966 4.26
Furthermore, in the limit when the top side length goes to zero,
10 3.809 3.982 4.32
it yields an isosceles triangle. At the other limit when top and
20 3.852 4.033 4.49
bottom sides are equal, it becomes rectangle/square [2]. Figure
7 shows the comparison between the approximate model and ∞ 3.867 4.051 4.54
the numerical data reported by [9] for different values of . As
can be seen, except for a few points, the agreement between the
model and the numerical values is reasonable (less than 8%). 20

16
18
data, φ = 85 data
data, φ = 60 model
14 data, φ = 45 16 model - 10%
data, φ = 30 model + 10%
model
12 model ± 10% 14

12
Nu√A (model)

I
Nu√A

10

φ 10
8
I
8
6

6
4
4
2
2
0 0.25 0.5 0.75 1
0
0 2 4 6 8 10 12 ε= b/a
Nu√A (data)
I

FIGURE 8. COMPARISON OF PRESENT MODEL AND

FIGURE 7. COMPARISON OF PRESENT MODEL AND NUMERICAL VALUES [9] FOR RECTANGULAR CHANNEL
NUMERICAL VALUES [9] FOR TRAPEZOIDAL CHANNEL. WITH SEMI CIRCULAR ENDS.

7 Copyright © 2008 by ASME

 Literature Review”, Journal of Nanoscale and Microscale
10
Thermophysical Engineering, Vol. 6, pp. 155-173.
9
[4] Duncan, A.B. and Peterson, G.P., 1994, “Review of
data [9]
model
Microscale Heat Transfer”, ASME Appl. Mech. Rev., Vol.
model ± 10% 47, pp. 397–428.
8
[5] Morini, G.L., 2004, “Single-Phase Convective Heat
7 Transfer in Microchannels: A Review of Experimental
I
Results”, Int. J. Therm. Sci., Vol. 43, pp. 631-651.
Nu√A

6 [6] Bailey, D.K., Ameel, T.A., Warrington, R.O., and Savoie,

T.I., 1995, “Single Phase Forced Convection Heat Transfer
5 in Microgeometries-A review”, IECEC Conference ASME-
FL, Orlando, USA, ES-396.
4 [7] Acosta, R.E., Muller, R.H., and Tobias, W.C., 1985,
“Transport Processes in Narrow (Capillary) Channels”,
3 AIChE J., Vol. 31, pp. 473–482.
[8] Rahman, M.M. and Gui, F.J., 1993, “Experimental
2
0 0.25 0.5 0.75 1 Measurements of Fluid Flow and Heat Transfer in
ε = b/a Microchannel Cooling Passages in a Chip Substrate”,
Advances in Electronic Packaging, ASME EEP, Vol. 199,
FIGURE 9. COMPARSION OF PRESENT MODEL AND
pp. 685–692.
NUMERICAL VALUES [9] FOR SINE CHANNEL.
[9] Shah, R. K. and London, A. L., 1978, Laminar Flow
Forced Convection In Ducts, Academic, New York.
6 SUMMARY AND CONCLUSIONS
[10] Muzychka, Y.S. and Yovanovich, M.M., 2002, “Laminar
In this paper, closed form relations have been developed for the
Flow Friction and Heat Transfer in Non-Circular Ducts and
Nusselt number of fully-developed laminar flow in smooth
Channels: Part II - Thermal Problem”, Proceedings of
arbitrary cross-sections channels with “H1” boundary condition
Compact Heat Exchangers, A Festschrift on the
(constant axial wall heat flux with constant peripheral wall
60thBirthday of Ramesh K. Shah, Grenoble, France, pp.
temperature). The square root of area √ , is adopted as the 131-139.
characteristic length scale, as it is superior to the conventional [11] Peng, X.F. and Wang, B.X., 1998, “Forced Convection
hydraulic diameter, . Using existing analytical solutions for and Boiling Characteristics in Microchannels”,
temperature distribution in elliptical, rectangular, and triangular Proceedings of 11th Int. Heat Transfer Conference,
ducts, a compact model including three sub models (upper, Kyongyu, Korea, Vol. 1, pp. 371–390.
lower, and middle models) has been developed. The present [12] Peng, X.F., Peterson, G.P., and Wang, B.X., 1995,
model is only a function of geometrical parameters of the cross- “Frictional Flow Characteristics of Water Flowing
section, i.e., area, perimeter, and polar moment of inertia. The Through Rectangular Microchannels”, J. Exp. Heat
proposed model was compared with numerical results for Transfer, Vol. 7, pp. 249–264.
channels with cross-sections including: ellipse, rectangle, [13] Jiang, P.X., Fan, M.H., Si, G.S., and Ren, Z.P., 2001,
triangle, trapezoid, rhombus, regular polygon, rectangle with “Thermal-Hydraulic Performance of Small Scale Micro-
semi-circular ends, sine, and found to successfully predicts the Channel and Porous-Media Heat Exchangers, Int. J. Heat
Nusselt number for a wide variety of shapes with a maximum Mass Transfer, Vol. 44, pp. 1039–1051.
difference on the order of 10%. [14] Wu, H.Y. and Cheng, P., 2003, “An Experimental Study of
Convective Heat Transfer in Silicon Microchannels with
7 ACKNOWLEDGMENTS Different Surface Conditions”, Int. J. Heat Mass Transfer,
The authors are grateful for the financial support of the Vol. 46, pp. 2547–2556.
Natural Sciences and Engineering Research Council (NSERC) [15] Gao, P., Le Person, S., and Favre-Marinet, M., 2002, “Scale
of Canada, and the Canada Research Chairs Program. Effects on Hydrodynamics and Heat Transfer in Two-
Dimensional Mini and Microchannels, Int. J. Thermal Sci.,
Vol. 41, pp.1017–1027.
8 REFERENCES [16] Qu, W., Mala, G.M., and Li, D., 2000, “Heat Transfer for
[1] Yang, C., Wu, J., Chien, H., and Lu, S., 2003, “Friction Water Flow in Trapezoidal Silicon Microchannels”, Int. J.
Characteristics of Water, R-134a, and Air in Small Tubes”, Heat Mass Transfer, Vol. 43, pp. 3925–3936.
Microscale Thermophys. Eng., Vol. 7, pp. 335–348. [17] Marco, S.M. and Han, L.S., 1955, “A Note on Limiting
[2] Bahrami, M., Yovanovich, M. M., and Culham, J. R., 2006, Laminar Nusselt Number in Ducts with Constant
“Pressure Drop of Laminar, Fully Developed Flow in Temperature Gradient by Analogy to Thin-Plate Theory”,
Microchannels of Arbitrary Cross-Section”, ASME J. of Trans. ASME, Vol. 77, pp. 625-630.
Fluid Engineering, Vol. 126, Sept., pp. 1036-1044. [18] Tyagi, V.P., 1966, “Laminar Forced Convection of a
[3] Obot, N.T., 2002, “Toward a Better Understanding of Dissipative Fluid in a Channel”, J. Heat Transfer, Vol. 88,
Friction and Heat/Mass Transfer in Microchannels- A pp. 161-169.

8 Copyright © 2008 by ASME

[19] Bahrami, M., Yovanovich, M. M., and Culham, J. R., 2007,
“A Novel Solution for Pressure Drop in Singly Connected
Microchannels of Arbitrary Cross-Section”, Int. J. Heat
Mass Transfer, Vol. 50, pp. 2492-2502.
[20] Incropera, F. P. and DeWitt, D. P., 1996, Fundamentals of
Heat and Mass Transfer, John Willey & Sons, New York.
[21] Chen, S., Chan, T.L., Leung, C.W., Yu, B., 2000,
“Numerical Prediction of Laminar Forced Convection in
Triangular Ducts with Unstructured Triangular Grid
Method”, Journal of Numerical Heat Transfer, Vol. 38,
pp. 209-224.

9 Copyright © 2008 by ASME

View publication stats