$+HDW6SUHDGLQJ0RGHOIRU'RXEOH6LGHG&URVV)ORZ0DQLIROG0LFURFKDQQHO+HDW([FKDQJHUV

Raphael Mandel, Martinus Arie, Amir Shooshtari, Michael Ohadi

Advanced Heat Exchangers/Process Intensification Laboratory

Department of Mechanical Engineering
University of Maryland
College Park, MD 20742

k Thermal conductivity, W/m-K

$%675$&7 L Flow length, m
Manifold-Microchannel Heat Exchangers (M2HXs) take ݉ሶ Mass flow rate, kg/s
advantage of the high heat transfer rates of microchannels NTU Number of transfer units, -
without the associated high pressure drops by reducing the flow Nx Number of nodes in the x-direction
length through the microchannel. However, a consequence of Ny Number of nodes in the y-direction
the short flow lengths is that the assumption of negligible Q Heat transfer rate, W
spreading in the fluid-separating wall (base plate) may no t Thickness of the fin, m
longer be valid, and conventional heat exchanger correlations T Temperature, K
may not be applicable to compute the heat exchanger U Conductance, W/m2-K
effectiveness. In the limiting case of extremely short flow w Width of the channel, m
lengths, spreading dominates, and a constant base plate x X-coordinate direction
temperature can be assumed. This is the approach taken by y Y-coordinate direction
previous authors to determine heat exchanger effectiveness for z Z-coordinate direction
a cross-flow M2HX. However, for practical designs, the actual *UHHN/HWWHUV
base plate temperature distribution is somewhere between ε Effectiveness, -
cross-flow with zero spreading and constant base temperature 6XEVFULSWV
(infinite spreading), the value of which can be determined only B Base (solid)
by a computationally expensive, full three-dimensional C Cold-side
simulation. Accordingly, a custom-coded two-dimensional H Hot-side
model was developed to predict the effects of spreading on S Single-sided
overall heat exchanger effectiveness. The governing partial
differential equations are presented, and a custom-solution ,1752'8&7,21
algorithm is defined. The model was validated by testing While microchannel heat exchangers boast low mass and
conventional cross-flow correlations under the conditions of volume due to their high surface area to volume ratios, the
zero spreading. The model was then used to predict the associated high pressure drops and pumping powers could
effectiveness of a designed heat exchanger and compared to a reduce system efficiency. One way to reduce microchannel
full three-dimensional simulation, with agreement to within pressure drop is to divide the microchannel into an array of
1.4%. Parametric studies on the effect of heat exchanger parallel microchannels using manifolds to guide and distribute
metrics—such as NTU and heat capacity ratio—were the flow over the microchannels. Such systems are known as
performed with various heat spreading assumptions, and the manifold-microchannels, and have been extensively studied in
trends described. the literature [1-22]. The main advantage of manifold-
microchannel heat exchangers is their ability to simultaneously
.(<:25'6 Thermal Management, Additive reduce pressure drop and increase heat transfer. Due to the
Manufacturing, Manifold-Microchannels, Heat Spreading, division of the flow over many parallel channels, the flow rate
Longitudinal Conduction and flow travelling length through each channel are reduced,
which reduces the pressure drop proportional to the square of
120(1&/$785( the number of divisions [2]. In addition, due to the short flow
A Base area, m2 lengths, the flow is in the developing regime, which enhances
Cp Specific heat capacity, J/kg-K heat transfer coefficient. Due to the poor thermophysical
Cr Heat capacity ratio, - properties of air, it is particularly advantageous to apply the
dx Mesh spacing in the x-direction, m manifold-microchannel concept to air-to-air heat exchangers.
dy Mesh spacing in the y-direction, m A double-sided manifold-microchannel is shown in Fig. 1(a).
H Height, m

978-1-5386-1272-9/$31.00 ©2018 IEEE 147 17th IEEE ITHERM Conference

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
 large number of microchannels in the heat exchanger and the
large number of cells per microchannel required to achieve a
mesh-independent result, doing such a simulation is either
computationally expensive or computationally prohibitive,
depending on the size of the heat exchanger. Accordingly, a
simplified model is desired.
To that end, the main objective of the present work is to
develop a computationally-efficient model for spreading in a
double-sided, cross-flow manifold-microchannel heat
exchanger, such that heat exchanger effectiveness can be
accurately predicted.

02'(/
6SUHDGLQJ0RGHO
The following assumptions and simplifications were made in
the model:
(1) Constant fluid properties and isotropic solid properties
(2) Negligible heat transfer in the manifold
(a) (3) Constant temperature Nusselt number
(4) Uniform microchannel flow distribution

Due to the symmetries in the system and the assumption of

uniform flow distribution, the pattern of multiple inlets and
outlets on the hot and cold sides can be used to reduce the
domain to a smaller “unit cell,” shown in Fig. 1(b). An
isometric view of the unit cell is shown in Fig. 2(a). Due to the
large number of microchannels in each unit cell (geometry in
Fig. 2(a)) not to scale), it was assumed that the effect of discrete
fins and channels on flow and heat transport was small, and the
fins and channels could be removed. Instead, an effective heat
transfer coefficient boundary condition could be used to model
the fins and channels on each side of the fluid-separating wall
(base), and the domain simplified to a flat plate of thickness
equal to ‫ܪ‬௕ , as shown in Fig. 2(b).
(b)
Using the control volumes shown in Fig. 2(b) and
Fig. 1: (a) Double-sided cross-flow manifold-microchannel
concept, and (b) resulting unit-cell pattern. neglecting thermal diffusion due to the large Peclet numbers,
the governing equations for the cold-side and hot-side
Fabrication of an M2HX using conventional conservation of energy, respectively, are given by
manufacturing methods is a challenge because the manifold ߲ ݉ሶ஼
and microchannel sections must be bonded. However, by using ൬ ‫ ܶ ݌ܥ‬൰ ൌ ܷ஼ ሺܶ஻ െ ܶ஼ ሻ (1)
߲‫ܮ ݔ‬ு ஼ ஼
additive manufacturing, both sections can be fabricated as a
single component, simplifying the fabrication process. Several ߲ ݉ሶு
works have shown successful fabrication of an M2HX using ൬ ‫݌ܥ‬ு ܶு ൰ ൌ ܷு ሺܶ஻ െ ܶு ሻ (2)
߲‫ܮ ݕ‬஼
additive manufacturing [11, 13].
where ݉ሶு and ݉ሶ஼ are the hot-side and cold-side mass flow
However, due to the short flow lengths, heat spreading
occurs in the fluid-separating wall (base), causing the rates, ‫݌ܥ‬ு and ‫݌ܥ‬஼ are the hot-side and cold-side fluid specific
conventional correlation for cross-flow heat exchanger heats at constant pressure, and ܷு and ܷ஼ are the effective hot-
effectiveness to be inaccurate [23, 24]. For extreme cases, such side and cold-side conductances between the hot-side and cold-
as those studied by [10, 12], heat spreading in the base causes side fluids, ܶு and ܶ஼ , are the hot-side and cold-side fluid
the base to approach a uniform temperature. For less extreme temperatures, and ܶ஻ , is the base temperature. It is worth noting
cases, the base temperature is somewhere between constant that due to the complicated flow path present in manifold- d-
temperature (infinite heat spreading) and that predicted by microchannels, the cold-side and hot-side base conductances,
conventional cross-flow correlations (zero heat spreading). ܷ஼ and ܷு , cannot be obtained from a developing flow
While a full three-dimensional simulation of the heat exchanger correlation. Instead, they must be obtained from a single-sided
could be performed to determine base temperature, due to the manifold-microchannel CFD simulation, which will be

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
described in the next section. It is also worth noting that while iterative procedure was repeated until an energy balance was
local values of base conductance could be used, for simplicity, obtained to a tolerance of 1e-4. The sensible heat gain was
an effective value was used instead, as will be described in the computed on the cold- and hot-sides, respectively, from
ே೤
next section.
ͳ
Similarly, the governing equation for conservation of ܳ஼ ൌ ݉ሶ஼ ‫݌ܥ‬஼ ෍ ܶ஼ ሺ݅ ൌ ܰ௫ ሻ െ ܶ஼ ሺ݅ ൌ ͳሻ (8)
ܰ௬
energy in the base is given by ௝ୀଵ
ேೣ
ͳ
߲ ଶ ܶ஻ ߲ ଶ ܶ஻ ܳு ൌ ݉ሶு ‫݌ܥ‬ு ෍ ܶு ሺ݆ ൌ ͳሻ െ ܶு ሺ݆ ൌ ܰ௬ ሻ (9)
݇௫ ‫ܪ‬஻ ൅ ݇௬ ‫ܪ‬஻ ܰ௫
ଶ ௜ୀଵ
߲‫ݔ‬ ߲‫ ݕ‬ଶ (3)
ൌ
ܷ஼ ሺܶ஻ െ ܶ஼ ሻ ൅ ܷு ሺܶ஻ െ ܶு ሻ

where ‫ܪ‬஻ is the height of the base, and ݇௫ and ݇௬ are the
effective thermal conductivities in the x- and y-directions,
respectively. Due to the different geometries and orientations
of the fins, the effective thermal conductivities in the x- and y-
directions are not equal. The effective thermal conductivities
can be computed as

‫ݐ‬஼ ‫ܪ‬஼
݇௫ ൌ ݇஻ ൬ͳ ൅ ൰ (4)
‫ݓ‬஼ ൅ ‫ݐ‬஼ ‫ܪ‬஻

‫ݐ‬ு ‫ܪ‬ு
݇௬ ൌ ݇஻ ൬ͳ ൅ ൰ (5)
‫ݓ‬ு ൅ ‫ݐ‬ு ‫ܪ‬஻
where ݇஻ is the isotropic thermal conductivity of the solid
material, ‫ݓ‬ு and ‫ݓ‬஼ are the hot-side and cold-side
microchannel widths, ‫ݐ‬ு and ‫ݐ‬஼ are the hot-side and cold-side
fin thicknesses, and ‫ܪ‬ு and ‫ܪ‬஼ are the hot-side and cold-side
heights of the microchannel, respectively. These equations for
effective thermal conductivity are derived from the control
volumes given in Fig. 2(b), which assumes that all the heat is (a)
transported through the base height, ‫ܪ‬஻ . It is worth noting that
the effective conductances appearing in Eq. (3), ܷு and ܷ஼ ,
include the effect of base conductance in the z-direction, as will
be described in the next section.
Since fluid properties were assumed to be constant, the
temperature boundary condition does not affect the
effectiveness. For simplicity, the boundary conditions were
chosen to be

ܶ஼ ሺ݅ ൌ ͳሻ ൌ Ͳ (6)

ܶு ሺ݆ ൌ ͳሻ ൌ ͳ (7)
Eqs. (1)-(3) form a linear system of PDEs, which was
formulated and solved using a custom algorithm in Matlab.
This custom algorithm is shown in Fig. 3.
First, the base temperature was initialized as ܶ஻ ሺ‫ݔ‬ǡ ‫ݕ‬ሻ ൌ
ͲǤͷ. The cold-side and hot-side temperatures were then updated
by solving Eqs. (1)-(2), respectively, by using the latest values (b)
of the base temperature. The base temperature was then Fig. 2: (a) Unit-cell pattern and geometric definitions, and
(b) final model domain and sample mesh. Geometry and
updated by solving Eq. (3), and an energy balance was
mesh enlarged to show details.
computed by measuring the difference in the sensible heat loss
and gain on the hot-side and cold-side outlets, respectively. The

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
 It is worth noting that when the simulation converges, ܳ ൌ for the single-sided microchannel simulations are shown in Fig.
ܳ஼ ൌ ܳு . The heat exchanger effectiveness can then be 4. It is also worth noting that the height of the base selected in
computed from the single-sided manifold-microchannel simulations was only
half the total base height, as shown in Fig. 4, since two single-
ܳ sided manifold-microchannel simulations—one for the hot-
ߝൌ (10) side and one for the cold-side—are required.
‹ሺ݉ሶ஼ ‫݌ܥ‬஼ ǡ ݉ሶு ‫݌ܥ‬ு ሻ
The governing continuity, momentum, and energy
equations were solved to a tolerance of 1e-4 for continuity and
due to the unit temperature difference applied at the base and momentum, and 1e-10 for energy. The mass-averaged fluid
inlet. temperature on the pressure outlet was then computed directly
Mesh independence was conducted by increasing the mesh in the CFD software and used to calculate the single-side
and observing the effect on heat exchanger effectiveness. A effectiveness. For a constant base temperature, the single-side
grid of 64 x 64 was found to be required to predict effectiveness effectiveness, ߝௌ , is related to the inlet, outlet, and base
to within < 1%. With the required mesh, the simulation was temperatures, ܶ௜௡ , ܶ௢௨௧ , and ܶ௕௔௦௘ , respectively, by
timed to take less than 0.050 s on a 4-GHz desktop computer,
allowing one to quickly perform parametric studies and ܶ௢௨௧ െ ܶ௜௡
optimization. ߝௌ ൌ (11)
ܶ௕௔௦௘ െ ܶ௜௡

For simplicity, a unit temperature difference between ܶ௜௡

and ܶ௕௔௦௘ was chosen: that is, ܶ௜௡ ൌ ͳ‫ ܭ‬and ܶ௕௔௦௘ ൌ Ͳ‫ܭ‬, as
shown in Fig. 4. However, it is worth noting that the resulting
effectiveness value will be independent of the selected
temperature difference due to the assumption of constant fluid
properties.

Fig. 3: Solution algorithm.

6LQJOH6LGHG0DQLIROG0LFURFKDQQHO6LPXODWLRQ Fig. 4: Domain of a single-sided manifold-microchannel

The effective hot-side and cold-side conductances, ܷு and ܷ஼ , CFD simulation.
are determined separately for the hot-side and cold-side by
performing a three-dimensional CFD simulation of a single- The single-side effectiveness was then used to compute the
sided manifold-microchannel “unit-cell”, as was done by [2, single-side conductance, ܷௌ , according to the relationship for
10, 12, 17, 18]. Due to the assumption of a uniform constant base temperature [25, 26]:
microchannel flow distribution, the hot-side and cold-side
conductances are only computed once at the nominal hot-side
and cold-side flow rates. The domain and boundary conditions

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
 ܷௌ ‫ܣ‬ௌ
ߝௌ ൌ ͳ െ ‡š’ ൬െ ൰ (12)
݉ሶௌ ‫݌ܥ‬

where ‫ܣ‬ௌ is the single-side base area in the x-y plane, and ݉ሶௌ
is the single-side fluid mass flow rate at the velocity inlet. Eq.
(12) can be easily manipulated to provide an explicit solution
for ܷௌ as a function of ߝௌ . It is worth noting that since
conduction in the base is simulated in the CFD model, the
conductance computed by Eq. (12) will include the effect of
conduction in the z-direction. Further details about the single-
side microchannel simulations are provided in [12].

9$/,'$7,21 Fig. 5: Validation of numerical model with zero spreading.

To validate the numerical model, the model was compared to
the available cross-flow correlations. Since the correlations are
&$6(678'<
only applicable when heat spreading is negligible—that is, In this section, a case study, representing a double-sided
when ݇௫ ൌ ݇௬ ൌ Ͳ—the numerical model was validated with manifold-microchannel heat exchanger designed to meet
݇௫ and ݇௬ artificially reset to zero. specific targets, is presented. The geometry and operational
The effectiveness of a cross-flow heat exchanger with both conditions are given in Table 1.
fluids unmixed is given by [25, 26]: The results from the case study are shown in Fig. 6. The
results show that the effectiveness is bounded by two limiting
ͳ
ߝ ൌ ͳ െ ‡š’ ൬ ܷܰܶ଴Ǥଶଶ ሺ݁‫݌ݔ‬ሺെ‫ܥ‬௥ ܷܰܶ଴Ǥ଻଼ ሻ െ ͳሻ൰ (13) cases: zero heat spreading and infinite heat spreading. For the
‫ܥ‬௥
case of zero heat spreading, the present model was modified by
where ‫ܥ‬௥ is the heat capacity ratio and is defined as artificially reducing ݇௫ and ݇௬ to zero. For the case of infinite
heat spreading, ݇௫ and ݇௬ were artificially increased by orders
‹ሺ݉ሶ஼ ‫݌ܥ‬஼ ǡ ݉ሶு ‫݌ܥ‬ு ሻ of magnitude until the results converged. A value of ~100,000
‫ܥ‬௥ ൌ (14) W/m-K was found to be sufficient. The results indicate that
ƒšሺ݉ሶ஼ ‫݌ܥ‬஼ ǡ ݉ሶு ‫݌ܥ‬ு ሻ
heat spreading reduces effectiveness, which is consistent with
and ܷܰܶ is the number of transfer units, and is defined as the literature [23, 24].

ܷ‫ܣ‬ Table 1: Case study geometry, operational conditions, and

ܷܰܶ ൌ (15) results.
‹ሺ݉ሶ஼ ‫݌ܥ‬஼ ǡ ݉ሶு ‫݌ܥ‬ு ሻ
&ROGVLGH +RWVLGH
where the overall conductance, ܷ, is defined as ݉ሶ [mg/s] 41.45 32.59
ܷ [W/m2-K] 3671 3003
ͳ ͳ ିଵ ‫[ ݌ܥ‬J/kg-K] 1038 1071
ܷൌ൬ ൅ ൰ (16) ‫[ ܮ‬mm] 5.1 6
ܷ஼ ܷு
‫[ ݓ‬μm] 180 190
‫[ ݐ‬μm] 100 100
While Eq. (13) provides the analytical solution for ‫ܥ‬௥ ൌ ͳ, ‫[ ܪ‬μm] 810 520
the predictions are reasonably accurate for Ͳ ൑ ‫ܥ‬௥ ൑ ͳ [26]. ‫ܪ‬஻ [μm] 300
The model was therefore run over a wide range of ܷܰܶ and ‫ܥ‬௥ ݇஻ [W/m-K] 21.5
and compared with Eq. (13), with the results shown in Fig. 5. ܷܰܶ [-] 1.448
Figure 5 demonstrates that the model provides accurate results ‫ܥ‬௥ [-] 0.8112
when there is zero spreading, indicating the correct
In addition, the results indicate that the model prediction
implementation and solution of Eqs. (1)-(2). To validate correct
lies between the two extreme cases, with zero heat spreading
implementation and solution of the remaining equations, a full
providing higher effectiveness values than infinite heat
three-dimensional simulation was performed, which will be
spreading.
discussed later.

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
 effectiveness always lies between the assumption of zero heat
spreading (݇ ՜ Ͳ) and infinite heat spreading (݇ ՜ λ). In
addition, it was observed that for a low NTU values—or
equivalently, low effectiveness—the effect of heat spreading is
negligible, as both the infinite heat spreading and zero
spreading assumptions provide the same prediction. However,
as NTU exceeds ~1, the curves begin to deviate from one
another, and the heat spreading effectiveness is observed to lie
below the zero heat spreading effectiveness and above the
infinite heat spreading effectiveness. The deviation is observed
Fig. 6: Case study results and three-dimensional CFD to increase as NTU increases, and therefore, high effectiveness
validation. heat exchangers will be more severely affected by heat
spreading than low effectiveness heat exchangers.
To validate this, a full three-dimensional simulation was
performed. A diagram of the full three-dimensional model is
given in Fig. 7.

Fig. 8: Effect of NTU on effectiveness for various heat

spreading assumptions.
Fig. 7: Full three-dimensional model for case study.
In addition, while the effectiveness for the simulated heat
To obtain mesh independent results, more than 1,000,000 spreading (݇ ՜ ݇஻ ) and zero heat spreading (݇ ՜ Ͳ) cases
cells were required, and the required computational time was increase as NTU increases, for the case of infinite heat
approximately 1 hour on a 12-core, 24-thread, 2.6-GHz Intel spreading (݇ ՜ λ), the effectiveness is observed to
Xeon processor. The results are included in Fig. 6, indicating asymptotically approach a maximum value. Increasing the
that the present model tends to over-predict heat spreading, and NTU further does not increase effectiveness. This maximum
therefore under-predict effectiveness. Heat spreading is over- r- value is observed to be a function of the heat capacity ratio, ‫ܥ‬௥ ,
predicted by the model due to the assumptions made in Eqs. and the results are shown in Fig. 9.
(4)-(5) that the fins and the base contribute equally to heat
spreading; in reality, this is not the case, especially as the fins
get taller relative to the spreading distance. However, the
predictions from this model are still more accurate than the
assumptions of zero heat spreading and infinite heat spreading.
More importantly, since the model tends to over-predict heat
spreading, it will always provide a conservative estimate of
effectiveness. Since manifold-microchannels rely on short
microchannel pitches to reduce pressure drop and increase heat
transfer coefficient, the effect of spreading needs to be
accounted for in the design stage.

3$5$0(75,&678',(6
The effect of NTU on effectiveness was then studied for
Fig. 9: Effect of ‫ܥ‬௥ on maximum effectiveness for the case
various heat spreading assumptions. The results are displayed
of infinite heat spreading.
in Fig. 8, and indicate that regardless of NTU, the actual

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
 The asymptotical behavior observed can be easily the effectiveness increases. In particular, for the heat capacity
explained by considering a balanced (‫ܥ‬௥ ൌ ͳ), high-NTU, ratio of ~0.8 given in the case study (see Table 1), the
counterflow heat exchanger. With zero heat spreading, the maximum obtainable effectiveness from Fig. 9 is 55%, which
temperature distribution in such a heat exchanger will take the is in agreement with the asymptote observed in the infinite heat
form of three parallel lines, as shown in Fig. 10(a). Thus, at the spreading result presented in Fig. 8.
cold-side inlet, the base temperature is significantly lower than Lastly, for heat exchangers with extreme heat spreading, a
at the hot-side inlet, resulting in a high effectiveness. However, “multi-pass” M2HX can been utilized. Since multi-pass heat
for infinite heat spreading, the base temperature approaches a exchangers obtain their desired effectiveness by combining
constant value, as shown in Fig. 10(b). The base temperature multiple low-effectiveness heat exchangers, heat spreading will
will always operate at the temperature at which both fluids have have a lesser effect. A sample temperature distribution for a
experienced the same difference in sensible heat. If the heat multi-pass M2HX is given in Fig. 10(c), and shows that even
capacity ratio is one, this point will be halfway between the though within each pass the base temperature is uniform, the
inlet temperatures, as shown in Fig. 10(b). Since the cold and overall effectiveness of the heat exchanger is higher than would
hot fluids cannot get hotter or colder than the base, respectively, have otherwise been possible with a single heat exchanger.
the maximum obtainable effectiveness will be 50%, which is
consistent with Fig. 9.
&21&/86,21
In conclusion, a heat spreading model for cross-flow manifold-
microchannel heat exchangers was presented in this study. The
model equations were defined, the solution algorithm
described, and the model validated. A case study was then
provided, and parametric studies performed, with the following
observations:
(1) the model presented allows heat exchanger designers to
predict the amount of heat spreading in a computationally
(a) efficient manner;
(2) for the case study, a deviation of 15% between the limiting
cases of infinite spreading and zero spreading was
observed;
(3) for the case study, the model matches the three-
dimensional CFD validation to within 1.4%;
(4) for the case of low NTU/effectiveness heat exchangers,
this model is not necessary to predict their performance, as
(b) heat spreading is not important for such heat exchangers;
(5) as NTU increases, the deviation between the heat
exchanger performance, and the limiting cases of infinite
heat spreading and zero heat spreading increases;
(6) for the limiting case of infinite heat spreading, a maximum
effectiveness exists, and is dependent of the heat capacity
ratio;
(7) to obtain a high effectiveness heat exchanger when heat
spreading is dominant, a multi-pass flow configuration can
(c) be utilized;
Fig. 10: Sample temperature distribution in a balanced,
high-NTU counterflow heat exchanger with (a) zero heat Future work includes abandoning the assumption that the
spreading (b) infinite heat spreading, and (c) infinite heat
microchannel flow distribution is uniform. If the flowrate
spreading in a multi-pass configuration.
through each channel was allowed to vary, it could be predicted
For heat capacity ratios less than one, the maximum by a separate flow distribution model, and the effect of
effectiveness is larger than 50%, as shown in Fig. 9. Since the manifold design on flow distribution and resulting heat
base must exist at the temperature at which the fluids exchanger performance could be determined, allowing the heat
experience the same change in sensible heat, the base exchanger to become even more compact.
temperature will shift towards the fluid with the higher heat
capacity ratio. This, in turn, will cause the fluid with the lower
heat capacity ratio to experience a larger temperature difference
as it is heated or cooled to the base temperature, and therefore,

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
5()(5(1&(6 [12] M. A. Arie, A. H. Shooshtari, V. V. Rao, S. V.
[1] M. M. Ohadi, K. Choo, S. Dessiatoun, and E. Cetegen, Dessiatoun, and M. M. Ohadi, "Air-Side Heat
1H[W *HQHUDWLRQ 0LFURFKDQQHO +HDW ([FKDQJHUV. Transfer Enhancement Utilizing Design Optimization
Springer Publishing Co., 2012. and an Additive Manufacturing Technique," -RXUQDO
[2] E. Cetegen, "Force Fed Microchannel High Heat Flux RI+HDW7UDQVIHUvol. 139, pp. 031901-1 - 031901-12,
Cooling Utilizing Microgrooved Surfaces," Ph.D., 2017.
Department of Mechanical Engineering, University of [13] M. A. Arie, A. H. Shooshtari, and M. M. Ohadi,
Maryland, College Park, 2010. "Experimental Characterization of an Additively
[3] E. Kermani,, ""Manifold Micro-Channel Cooling of Manufactured Heat Exchanger for Dry Cooling of
Photovoltaic Cells for High Efficiency Solar Energy Power Plants," $SSOLHG 7KHUPDO (QJLQHHULQJ vol.
Conversion Systems," Master of Science, Department 129, pp. 187-198, 2018.
of Mechanical Engineering, University of Maryland, [14] R. Mandel, A. Shooshtari, S. Dessiatoun, and M.
College Park, MD,, U USA, 2008. Ohadi, "Streamline Modeling of Manifold
[4] E. Kermani, S. Dessiatoun, A. Shooshtari, and M. Microchannels in Thin Film Evaporation,"" presented
p
Ohadi,, "E
"Experimental investigation of heat transfer at the ASME 2013 Summer Heat Transfer
performance of a manifold microchannel heat sink for Conference, Minneapolis, MN, 2013.
cooling of concentrated solar cells," presented at the [15] R. Mandel, S. Dessiatoun, P. McCluskey, and M.
Proc. IEEE Electronic Components and Technology Ohadi, "Embedded Two-Phase Cooling of High Flux
Conference, San Diego, CA, USA, 2009. Electronics via Micro-Enabled Surfaces and Fluid
[5] L. M. Boteler,, "Microfabrication
"M and Analysis of Delivery Systems (FEEDS),"" ppresented at the Proc.
Manifold Microchannel Coolers for Power ASME 2015 International Technical Conference and
Electronics,"" PPh.D.,, Department
D of Mechanical Exhibition on Packaging and Integration of Electronic
Engineering, University of Maryland, College Park, k, and Photonic Microsystems collocated with the
2011. ASME 2015 13th International Conference on
[6] L. Boteler, N. Jankowski, P. McCluskey, and B. Nanochannels,, M Microchannels, and Minichannels,
Morgan,, ""Numerical investigation and sensitivity San Francisco, CA, USA, 2015.
analysis of manifold microchannel coolers," [16] R. Mandel, S. Dessiatoun, and M. Ohadi, "Embedded
,QWHUQDWLRQDO-RXUQDORI+HDWDQG0DVV7UDQVIHUvol. Two-Phase Cooling of High Flux Electronics Using a
55, no. 25-26, pp. 7698-7708, 2012. Directly Bonded FEEDS Manifold,"" presented
p at the
[7] W. Escher, T. Brunschwiler, B. Michel, and D. Intersociety Conference on Thermal and
Poulikakos, "Experimental Investigation of an Thermomechanical Phenomena in Electronic
Ultrathin Manifold Microchannel Heat Sink for Systems, Las Vegas, NV, 2016.
Liquid-Cooled Chips," (in English), ), -RXUQDORI+HDW
), [17] R. K. Mandel,, "Embedded
" Two-Phase Cooling of
7UDQVIHU7UDQVDFWLRQVRIWKH$VPH H Article
A vol.
l. 132,
1 High Flux Electronics via Micro-Enabled Surfaces
no. 8, AUG 2010 2010, 0, Art.
A no. ARTN 081402. and Fluid Delivery System (FEEDS)," Doctor of
[8] M. Arie, A. Shooshtari, S. Dessiatoun, M. Ohadi, and Philosophy,, Department
D of Mechanical Engineering,
E. Hajri,, "Simulation
"S and Thermal Optimization of a University of Maryland, College Park, College Park,
Manifold Microchannel Flat Plate Heat Exchanger," ," MD, 2016.
presented at the ASME 2012 International Mechanical [18] R. Mandel, A. Shooshtari, and M. Ohadi,, "A "2.5-D"
Engineering Congress and Exposition, Houston, Modeling Approach for Single-Phase Flow and Heat
Texas, USA, 2012. Transfer in Manifold Microchannels," ,QWHUQDWLRQDO
[9] M. Arie, A. Shooshtari, S. Dessiatoun, and M. Ohadi, -RXUQDORI+HDWDQG0DVV7UDQVIHU,Q5HYLHZ2017.
"Thermal optimization of an air-cooling heat [19] D. G. BaeHWDO, "Embedded Two-Phase Cooling of
exchanger utilizing manifold-microchannels," High Heat Flux Electronics on Silicon Carbide (SiC)
presented at the Proc. Thermal and using Thin-Film Evaporation and an
Thermomechanical Phenomena in Electronic Systems Enhanced Delivery System (FEEDS) Manifold- d-
(ITherm), 2014. Microchannel Cooler," presented at the Intersociety
[10] M. Arie, A. Shooshtari, S. Dessiatoun, E. Al-Hajri, Conference on Thermal and Thermomechanical
and M. Ohadi, "Numerical modeling and thermal Phenomena in Electronic Systems, Orlando, FL,
optimization of a single-phase flow manifold- USA, 2017.
microchannel plate heat exchanger," ,QWHUQDWLRQDO [20] D. Bae, R. Mandel, and M. Ohadi, "Effect of bonding
-RXUQDORI+HDWDQG0DVV7UDQVIHUvol. 81, pp. 478- structure and heater design on performance
489, 2015. enhancement of feeds embedded manifold-
[11] M. A. Arie, A. H. Shooshtari, and M. M. Ohadi, "Air- microchannel cooling,"" ppresented at the International
Side Enhancement of Heat Transfer in an Additively Technical Conference and Exhibition on Packaging
Manufactured 1kW Heat Exchanger for Dry Cooling and Integration of Electronic and Photonic
Applications," presented at the Thermal and Microsystems (InterPACK), San Francisco, CA,
Thermomechanical Phenomena in Electronic Systems 2017.
(ITherm), Orlando, 2016.

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.
[21] K. P. Drummond HW DO, "Evaporative Intrachip
Hotspot Cooling with a Hierarchical Manifold
Microchannel Heat Sink Array,"" ppresented at the
Intersociety Conference on Thermal and
Thermomechanical Phenomena in Electronic
Systems, Las Vegas, NV, USA, 2016.
[22] K. P. Drummond HW DO, "A hierarchical manifold
microchannel heat sink array for high-heat flux two-
phase cooling of electronics," ,QWHUQDWLRQDO-RXUQDORI
+HDWDQG0DVV7UDQVIHUvol. 117, pp. 319-330, 2018.
[23] A. F. Mills, E. Jones, Ed. +HDW DQG 0DVV 7UDQVIHU,
First ed. Boston: RICHARD D. IRWIN, INC., 1995.
[24] W. M. Kays and A. L. London, &RPSDFW +HDW
([FKDQJHUV, Third ed. Malabar, Florida:: K KRIEGER
PUBLISHING COMPANY, 1998, p. 335
[25] Y. A. Cengel and A. J. Ghajar, +HDW DQG 0DVV
7UDQVIHU )XQGDPHQWDOV DQG $SSOLFDWLRQV, Fifth ed.
New York, NY, USA: McGraw Hill Education, 2015,
p. 1208.
[26] F. P. Incropera, D. P. DeWitt, T. L. Bergman, and A.
S. Lavine, V. A. Vargas,, EEd. )XQGDPHQWDOVRI+HDW
DQG0DVV7UDQVIHU, Sixth ed.. H Hoboken, New Jersey,
USA: John Wiley and Sons, 2007, p. 1070.

Authorized licensed use limited to: Batangas State University. Downloaded on June 26,2024 at 06:39:43 UTC from IEEE Xplore. Restrictions apply.