Improved design of microchannel plate geometry for uniform flow distribution.

Balaji, S. ; Lakshminarayanan, S.

The need to achieve high throughput in micro devices leads to high velocities through the microchannels leading to high pressure drops. Due to the nature of the microchannel plate geometry, the highest pressure drop may get localized to cause uneven flows along the microchannels leading to performance reduction. To maintain uniform flow distribution even under high throughput conditions, a two-dimensional model has been constructed to study the flow distribution along the microchannels for various plate geometries of a micro heat exchanger. A novel micro heat exchanger configuration to achieve uniform flow distribution under all operating conditions has been proposed, modelled and tested. From the results obtained, it can be seen that the plate geometry with inlet and outlet valves inline with the microchannels along with two inlets and four outlets (for possible counter-current heat exchange operation) provides uniform flow distribution under a wide range of flow conditions.

Le besoin d'atteindre une capacite elevee dans les micro-dispositifs mene a des vitesses elevees dans les micro-canaux, ce qui produit des pertes de charge elevees. Du fait de la geometrie des plaques de micro-canaux, la perte de charge la plus elevee peut se produire en des points precis et causer des ecoulements inegaux le long des micro-canaux, reduisant ainsi la performance. Afin de maintenir une distribution de l'ecoulement uniforme dans des conditions de capacite elevee, on a construit un modele bidimensionnel pour etudier la distribution de l'ecoulement le long des micro-canaux pour diverses geometries de plaques dans le cas d'un micro-echangeur de chaleur. Une nouvelle configuration de microechangeur de chaleur permettant de realiser une distribution d'ecoulement uniforme dans toutes les conditions operatoires est proposee, modelisee et testee. D'apres les resultats obtenus, on peut voir que la geometrie de plaques munies de vannes d'entree et de sortie alignees avec les micro-canaux dans le cas de micro-dispositifs a deux entrees et quatre sorties (pour un eventuel fonctionnement d'echange de chaleur a contre-courant) offre une distribution d'ecoulement uniforme pour une large gamme de conditions d'ecoulement.

Keywords: microchannel, micro heat exchanger, microreactor, plate geometry, flow uniformity

INTRODUCTION

In recent years, the concept of micro systems has found applications in a large number of scientific fields. With each passing day, the benefits of micro devices are being established by many researches around the world. The main reason behind this enticing improvement is the exceeding potential of such systems over conventional macroscopic systems. The peculiar features (Hessel et al., 2003) that make these systems superior (in performance) to macroscopic systems are 1) greater surface to volume ratio, 2) laminar flow, 3) low holdup, 4) safe operation with highly reactive or hazardous chemicals, 5) good controllability, 6) minimal energy consumption, 7) accurate analysis of expensive chemicals utilizing very small quantities, etc. All the above mentioned characteristics exist because of the smaller dimensions of the system. When a microreactor is used for producing a specific chemical, then the quantity produced per second will be in the order of microlitres. Thus, a number of micro systems should be used in parallel (Hasebe, 2004). Multiple repetitions of small, precisely fabricated micro systems (numbering-up strategy) are required to match the outcome obtainable from conventional processes.

Apart from the concept of numbering-up, when high throughput is considered, the small volume of these systems results in high flow velocities (> 1 m/s) along the microchannels inside the micro systems. These high velocities cause localized pressure drops (> 30 N/[m.sup.2]) in the system causing irregular flow distribution in the microchannels. In general, the flow through the microchannels depends on various factors like channel length, channel diameter, geometric shape, physical properties of the fluid, size and location of the inlet and outlet ports. Any non-uniformity in the flow affects the output or system performance. For example, if a micro heat exchanger is considered, when there is uneven flow distribution, the flow with less velocity will experience greater heat exchange than the flow with high velocity. As a result, an undesired temperature difference will originate in the system. Similarly, in a chemical reactor, the residence time of reactants varies with respect to the flow rate. Thus, the extent of reaction in each microchannel may vary if the flow is non-uniform. However, when catalytic reactions are considered, the extent of reaction is a strong function of channel diameter and the amount of catalyst loaded. For example, Delsman et al. (2005) showed that the influence of flow maldistribution on the overall reactor conversion is relatively small, while the influence of variations in the channel diameter (due to variations/defects introduced during the fabrication process) and the amount of catalyst coating on the channel walls (due to variations in the wash coating) are more pronounced. An extensive study on the effect of temperature non-uniformity in a microreactor by Rebrov et al. (2003) elaborates the consequences of non-uniformity in microchannels. Thus, it is imperative to maintain uniform flow distribution throughout the micro system even under high throughput conditions. So far, optimized plate geometries are proposed by many researchers focusing mainly on the channel diameter, channel length and geometric shape. However, the obtained results tend to fail under high throughput conditions. Thus, a new geometry is proposed in this study by considering the size and the location of inlet and outlet ports where the flow distribution becomes a weak function of the velocity and on the physical properties of the fluids used.

Flow distribution in the microchannel plate has already been studied by many researchers (Commenge et al., 2002; Delsman et al., 2004a-c; Tonomura et al., 2004). Optimization of the microchannel plate for uniform flow distribution has also been demonstrated in the past (Rebrov et al., 2003; Commenge et al., 2002; Delsman et al., 2004a). All the works that have been done so far were focused on a plate with the same geometry (shown in Figure 1). Figure 1 shows the plate geometry for a single flow stream. The plate geometry is coarsely meshed in order that the reader may easily visualize the regions of fluid flow. The overall set-up consists of two inlet and two outlets with separate inlet and outlet valves for hot and cold streams. A geometry with which the flow distribution varies based on the flow velocity has been taken and the optimum shape of each sections like manifold (region between the inlet and the microchannels), channel length, channel width, inlet and outlet width were determined such that uniform flow distribution is obtained.

[FIGURE 1 OMITTED]

Commenge et al. (2002) modelled the flow distribution in a microchannel plate using an approximate pressure drop model. They analyzed the influence of channel length and channel width on the velocity distribution in the microchannel plate. In general, the work highlighted the influence of geometric parameters and characteristic dimensions on the quality of fluid distribution among the channels of the micro structured plate. However, the velocity distribution determined using the proposed approximate model is independent of the flow rate as the pressure drop is calculated only based on the wall friction. The model developed did not reflect the effects caused due to increasing inlet velocity. Hence, the conclusions made using the approximate model will be specific to operating condition. For instance, when the inlet velocity is increased, the flow distribution along the microchannels will no longer be uniform using the optimal geometry obtained from their strategy.

Delsman et al. (2004a) developed a three-dimensional model to study the flow distribution in a microchannel plate. They showed that there exists a transitional velocity below which the flow distribution is independent of flow rate and above which the inertial effects dominate making the flow distribution to be highly dependent on flow rate. They found an optimal geometry by varying the flow distribution area, length of the microchannels and the plate height. They showed that uniform flow distribution can almost be obtained by doubling the cross sectional area of the inlets and outlets and doubling the channel lengths. Also, a small extra space at the outlet side of the channels helps in increasing the flow uniformity further. Delsman et al. (2004b) studied the operation of a micro device for preferential oxidation of carbon monoxide in hydrogen rich reformate gas using both modelling and experimentation. The micro device consisted of two heat exchangers of a reactor that is integrated with a heat exchanger. The 3-D model of the heat exchanger proposed in Delsman et al. (2004a) was used for optimizing the geometry of the flow distribution chambers for equal flow rates among all the microchannels. An integral one-dimensional heat exchange model was used to design the device. Subsequently, a two-dimensional model to represent the heat exchange in the system was tested in Delsman et al. (2004c). They showed that the one-dimensional model fails to represent the system behaviour accurately as it ignores the temperature differences at the inlet and at the channels further from the inlet.

Tonomura et al. (2004) characterized the flow pattern in a plate fin micro device using computational fluid dynamics. They demonstrated that the flow uniformity in the microchannel is highly dependent on manifold shape (section of the plate between the inlet and the microchannels), length and location of fins and inlet flow rate. A CFD based optimization technique was illustrated to obtain an optimal design for the micro device concerned. Altogether, this work neatly described the importance of computational study which has been unrecognized or given less emphasis in micro device design.

Apart from the above, many other works have focused on micro heat exchangers. Stief et al. (1999) investigated the performance of counter-current micro heat exchangers with respect to wall materials and their conductivity. Hardt et al. (2003) carried out studies of enhancing heat transfer in microchannels with fins in a checkerboard arrangement. The idea of different optimization techniques and model predictive control concepts was also applied to control the output temperature in micro heat exchangers (Okabe et al., 2003 and Roudsari et al., 2005). A substantial research has also been done in micro cross-flow heat exchangers (Harris et al., 2002). They fabricated both polymer and nickel based micro cross-flow heat exchangers and showed that a significant increase in heat transfer can be obtained just by decreasing the heat exchanger dimensions by one order of magnitude. In addition, Liu et al. (2004) designed and fabricated a micro flow heated channel with perfect insulation on the wall. The main aim of this work is to acquire a uniform heat flux boundary condition on the channel wall. Other interesting results on Micro Heat Exchangers can also be found in Alm et al. (2005), Hegde et al. (2005), Lerou et al. (2005).

The plate geometry taken for optimization in all the previous works is a single input, single output design for each flow stream (hot stream and cold stream). The flow uniformity is highly dependent on the direction of the flow from the inlet and through the microchannels. When the inlet is inline with the flow in microchannels (Delsman et al., 2004a), uniform flow along the channels can be acquired with ease. When the inlet is perpendicular to the microchannels (Commenge et al., 2002) then maximum flow non-uniformity will be seen under high flow rates. Thus, the plate geometry with inlet lined up with the microchannels is considered in the present work.

The main contributions of this paper are: (i) to determine an alternate geometry for the heat exchanger such that the flow uniformity along the microchannels is easily attained and (ii) to provide a comparison of the flow uniformity along the microchannels for different types of existing and proposed micro heat exchanger plate geometry.

MICROREACTOR GEOMETRY

The plate geometry proposed by Commenge et al. (2002) is shown in Figure 1. As the flow through the micro system is initially at right angles to the flow in microchannels, under high flow rates, most of the fluid will be driven towards the right corner. This results in increasing velocity at the right end inducing a non-uniformity in the flow distribution. In Figure 2, the most optimal geometry as proposed by Delsman et al. (2004a) is shown. In this configuration, the inlet flow is aligned (parallel) to the flow in microchannels. Such a configuration comparatively (compared to the geometry with inlet location as shown in Figure 1) reduces the pressure drop along the system and hence acceptable flow distribution (uniform) can be obtained. The pressure drop during high flow conditions can further be reduced when the width of the inlet and outlet are increased. However, for a counter-current operation, alternate stack of symmetric plates will be assembled such that in each plate two distinct fluid streams can be passed. Hence, the length of the inlet and outlet channels cannot be increased completely.

[FIGURE 2 OMITTED]

The pressure drop along the micro system can be reduced in numerous ways such as increasing the channel width, changing the location of the inlet and outlet ports, etc. All these options were tried and it was found that some non-uniformity still persists. One could also decrease the pressure drop by fixing the length of the inlet and increasing the length of the outlet without diminishing the capability of the system to operate in a countercurrent fashion. The non-uniformity cannot be reduced significantly when the number of channels is increased and the channel length is approximately equal to the plate width (Delsman et al., 2004a). For such cases, the configuration proposed in this work (as a sequel to the work done by Delsman et al., 2004a-c) can be used efficiently for uniform flow distribution. A new configuration of the microchannel plate that can be used for counter-current operation with uniform flow is proposed here. When real micro devices are considered, the entire system is made up of thousands of such microchannel plates. Realizing uniform flow distribution in single microchannel plate geometry results in the flow distribution along other similarly fabricated plates to also be uniform. Thus, the flow in the entire micro device (which contains thousands of plates arranged in parallel) will also be uniformly distributed. This explains the practicality of the proposed approach in designing a micro device based system.

To begin with, the effect of the number and location of the inlet ports on the uniformity of the flow distribution is studied with a microchannel as shown in Figure 3. The microchannel consists of three inlet and three outlet ports as shown. For simplicity, inlet and outlet ports are represented as IP (IP1, IP2, IP3) and OP (OP1, OP2, OP3), respectively. The effect of using different number of ports has been studied and the difference in maximum and minimum velocities along the microchannel has been tabulated (Tables 1 & 2). In Table 1, the one inlet and one outlet port combination is studied. Circular ports are considered (see Figure 4) and as discussed earlier, the inlet and outlet ports located on opposite sides are not allowed to be aligned in order that counter-current operation may be facilitated (see Figure 5). From the results obtained (Table 1), it can be seen that uniform distribution can be obtained when the inlet and outlet ports are not aligned. The difference in distribution is less affected by the distance between the inlet and outlet ports (i.e. when either IP2 or IP3 is used, the velocity difference is almost the same).

[FIGURES 3-5 OMITTED]

The microchannel is also simulated for one inlet and two outlet port combinations. The maximum and minimum values of the velocity attained in the microchannels are determined and the difference in each case is shown in Table 2. Other combinations with inlet port 3 opened are not shown because of its symmetry with inlet port 1. From Table 2, it can be seen that when the inlet is located at the centre of the plate geometry, the flow is well distributed and the uniformity is much more pronounced. The port combination IP2-OP1&OP3 provides the minimal velocity difference.

It must be pointed out that we have considered the inlet and output ports to be of circular shape (as in practical fabrications). This is in contrast to other papers that consider these ports as rectangular slits in their CFD simulations. While this approximation (circular ports considered as rectangular ports) may be acceptable in macro scale systems, the nature of each component should be reflected accurately in the model for microsystems where shapes are known to have a significant effect on the system performance (discussed further under the Results and Discussion section). A small change in the shape might produce completely different results. Figure 4 shows the proposed plate geometry with circular cross sections for the inlets and outlets (which represents the true nature of the system) based on the results obtained from Tables 1 and 2. Again, the plate geometry is coarsely meshed for the convenience of the reader to visualize the flow regions. The geometry with only one flow (cold or hot stream) is shown in Figure 4. The empty circles in the figure denote the inlet and outlets for the stream (hot or cold stream) in the next plate. By locating the inlet at the centre and the outlet at the corners the pressure drop is spread throughout the reactor. Due to this spreading, a significant decrease in the pressure drop can be attained along the microchannels. As the outlet is twice as big as inlet, the overall pressure drop can also be reduced.

MATHEMATICAL MODEL

A micro heat exchanger (8.6 mm x 7 mm) with microchannels of 200 mm wide and 4 mm length has been modelled. The plate dimensions and the number of microchannels in each plate are chosen based on the practical values seen in the literature (Hessel et al., 2003; Arias et al., 2001; Kleiner et al., 1995; Commenge et al., 2002; Delsman et al., 2004a, b, c, 2005). It is assumed that the mean free path of the fluid being tested is low enough such that the Knudsen number is very less (< 0.01). Under low Knudsen number, the effects due to rarefaction can be neglected (Commenge et al., 2002; Harley et al., 1995; Piekos and Breuer, 1996). Based on this assumption, a continuum model can be constructed. Navier Stokes equation is used to represent the flow characteristics inside the microchannel (Olofsson, 2004). In general, the Reynolds number of fluid flow inside the channels will be very small (< 200) as the diameter of the channels is in the order of micrometres. Hence, the flow is completely laminar throughout the system (this has been verified in our simulations).

The momentum balance and continuity equations are given by:

[rho] [partial derivative]u/[partial derivative]t - [nabla].[eta]([nabla]u + [([nabla]u).sup.T]) + [rho] (u.[nabla])u+[nabla]p = F (1)

[nabla].u = 0 (2)

where, [eta] denotes the dynamic viscosity, u the velocity vector, [rho] the density of fluid, p the pressure and F is the body force term (COMSOL[R], 2005). In the simulation study, it is assumed that there is no body force and the values of density and viscosity are assumed as 1 kg/[m.sup.3] and [10.sup.-5] Pa.s, respectively. The inlet conditions specified are in terms of pressure to vary the velocity of the fluid stream over a wide range (4 to 4000 mm/s). Similarly, the outlet boundary condition is also specified in terms of pressure with p = 0 and n. ([nabla].[eta]([nabla]u)+[([nabla]u).sup.T]), where n is the normal to the boundary surface. All other boundaries are specified as "no slip" conditions. Eighteen straight microchannels are used.

A 3-D view of the microchannel modelled has been shown in Figure 5. The boundary condition for the circular pipes which connect the alternate stack of micro plates, has been taken as circular shape. The three pipes in the figure are connected to one micro plate (bottom) and the other three void circles are connected to another micro plate (top). A surface plot of the velocity distribution at a high flow velocity of 4000 mm/s in the proposed geometry is shown in Figure 6. The model is simulated with 35 248 mesh elements and 166 272 degrees of freedom (COMSOL[R], 2005).

[FIGURE 6 OMITTED]

RESULTS AND DISCUSSION

The model developed in COMSOL[R] is tested for varying velocities and the flow velocity distribution along the microchannel at the centre is plotted. Different numerical solvers, each employing several mesh refinements, were employed in order to verify that numerical artifacts do not affect the results. Initially the configuration in Figure 1 is modelled to indicate the drawback inherent in the original geometry (Commenge et al., 2002). The flow distribution for varying flow velocities is shown in Figure 7. The x-axis (arc Length) represents the distance along the width of the microchannel plate (with x = 0 denoting the left edge and x = 7 x 10-3 representing the right edge). The velocity measurements are obtained at the centre line halfway along each microchannel. The distribution is consistent only for low flow rates and becomes significantly non-uniform under high flow rates. Thus, the configuration in Figure 1 will provide optimal results only for low velocity conditions. The importance of circular port geometry is shown in Figure 7.

[FIGURE 7 OMITTED]

Figure 7 and Figure 8 represent the velocity distribution for the plate geometries with circular and rectangular inlet and outlet ports, respectively. Comparing the flow profiles, it is seen that there exists a significant difference when the model is solved with different port geometries (circular vs. rectangular). Practically speaking, the circular port interprets the physical geometry more accurately than the rectangular slit boundary. As reported earlier, the shape of the boundary plays a significant role in obtaining meaningful results from the developed model.

[FIGURE 8 OMITTED]

The configuration shown in Figure 2 is then tested with circular inlet and outlet ports. From the results obtained, it can be seen that the non-uniformity in the flow occurs as the velocity is increased. The velocity distribution in the microchannel is shown in Figure 9. The velocity varies from 2 to 4 m/s. Thus, when real and appropriate geometries (circular cross section ports) are used for simulations, the major non-uniformity that exists in the plate geometry can be seen.

[FIGURE 9 OMITTED]

Moreover, the optimal space at the inlet and outlet manifolds (denoted as [delta] in Figure 2) might vary with respect to the physical properties of the fluids used. However, the value of [delta] is fixed for a geometry and hence the fabricated micro heat exchanger can be used efficiently for certain fluids only. Thus, it is difficult to think of a general microchannel configuration which can be operated with distinct flow conditions for a wide range of fluids and for a wide range of flow conditions.

Under high velocity conditions, in the above mentioned plate geometries, the pressure drop is very high in some regions making the flow distribution inconsistent. This can be seen in Figures 7, 8 and 9. On the other hand when the configuration in Figure 4 is considered, the velocity is distributed evenly throughout the microchannel region. This is because the inlet is at the centre and the outlets are at the two corners of the opposite ends of the plate. The proposed configuration is modelled and tested for uniformity. The results are shown in Figure 10 for the same values of [rho] and [mu] used to obtain the results in Figures 7 and 9. From Figures 7, 9 and 10, it can be clearly seen that the velocity is more evenly distributed (although not completely) throughout the microchannels in the new configuration than in the other two configurations. Dead zone on either sides of the inlet port as seen in Figure 6 is the only limitation of the proposed geometry. This limitation is of less significance in the light of the fact that the improved design proposed here ensures consistent flow along the microchannels.

[FIGURE 10 OMITTED]

The principal aim is to find a general configuration which can work for several fluids even under high flow rates. Thus, the proposed configuration is also tested for high density fluid. The same operating conditions (pressure values) are used and the results are shown in Figure 11. Because of the high density fluid, when same pressure conditions as applied in low density fluid are applied, the velocity obtained is less when compared to the velocity of the low density fluid (Figure 10). Nevertheless, the proposed designs are examined for a reasonable range of velocity. The results in Figures 10 and 11 demonstrate that the proposed microchannel plate geometry works well for both low and high density fluids and even at high flow rate conditions. It can be clearly seen that the flow uniformity in the micro system is a strong function of the size and the location of inlet and outlet ports. Thus, by determining the suitable size and location of the ports using numerical simulations, flow uniformity can be obtained through out the system.

[FIGURE 11 OMITTED]

CONCLUSIONS

A two-dimensional model has been used to study the flow distribution along the microchannels for various plate geometries of a micro heat exchanger. The flow uniformity throughout the system is assessed and a novel micro heat exchanger configuration is proposed to achieve uniform flow distribution under a wide range of operating conditions (robust for velocity and density variations). From the results obtained, it is seen that the plate geometry with inlet and outlet inline with the microchannels along with two inlets and four outlets (for possible counter-current heat exchange operation) to distribute the pressure drop evenly throughout the system proves to be a robust design. It can also be concluded from the results that complete understanding and substantial improvements can be obtained through the application of the systems engineering tools in micro systems design.

REFERENCES

Alm, B., R. Knitter and J. Haubelt, "Development of a Ceramic Micro Heat Exchanger Design, Construction and Testing," Chemical Engineering and Technology, 28(12) 1554-1560 (2005).

Arias, F., S. R. J. Oliver, B. Xu, R. E. Holmlin and G. M. Whitesides, "Fabrication of Metallic Heat Exchangers Using Sacrificial Polymer Mandrils," Journal of Microelectromechanical Systems, 10(1) 107-111 (2001).

Commenge, J. M., L. Falk, J. P. Corriou and M. Matlosz, "Optimal Design for Flow Uniformity in Microchannel Reactors," AIChE Journal, 48(2) 345-358 (2002).

COMSOL Version 3.2, Reference Manual, (2005).

Delsman, E. R., A. Pierik, M. H. J. M. De Croon, G. J. Kramer and J. C. Schouten, "Microchannel Plate Geometry Optimization for Even Flow Distribution at High Flow Rates," Chemical Engineering Research and Design, 82(A2) 267-273 (2004a).

Delsman, E. R., M. H. J. M. De Croon, A. Pierik, G. J. Kramer, P. D. Cobden, C. Hofmann, V. Cominos and J. C. Schouten, "Design and Operation of a Preferential Oxidation Microdevice for a Portable Fuel Processor," Chemical Engineering Science, 59 4795-4802 (2004b).

Delsman, E. R., M. H. J. M. De Croon, G. J. Kramer, P. D. Cobden, C. Hofmann, V. Cominos and J. C. Schouten, "Experiments and Modelling of an Integrated Preferential Oxidation Heat Exchanger Microdevice," Chemical Engineering Journal, 101 123-131 (2004c).

Delsman, E. R., M. H. J. M. De Croon, G. D. Elzinga, P. D. Cobden, G. J. Kramer and J. C. Schouten, "The Influence of Differences Between Microchannels on Microreactor Performance," Chemical Engineering and Technology, 28(3), 367-375 (2005).

Hardt, S., W. Ehrfeld and V. Hessel, "Strategies for Size Reduction of Microreactors by Heat Transfer Enhancement Effects," Chemical Engineering Communication, 190 540-559 (2003).

Harley, J. C., Y. Huang, H. H. Bau and J. N. Zemel, "Gas Flow in Micro-Channels," Journal of Fluid Mechanics, 284 257-274 (1995).

Harris, C., K. Kelly, T. Wang, A. McCandless and S. Motakef, "Fabrication, Modeling, and Testing of Micro-Cross-Flow Heat Exchangers," Journal of Micro Electromechanical Systems, 11(6) 726-735 (2002).

Hasebe, S., "Design and Operation of Micro Chemical Plants - Bridging the Gap Between Nano, Micro and Macro Technologies," Computers and Chemical Engineering, 29 57-64 (2004).

Hegde, P., K. N. Seetharamu, G. A. Quadir, P. A. Aswathanarayana, M. Z. Abdullah and Z. A. Zainal, "Thermal Analysis of Micro-Channel Heat Exchangers with Two-Phase Flow using FEM," International Journal for Numerical Methods in Heat and Fluid Flow" 15(1), 43-60 (2005).

Hessel, V., S. Hardt and H. Lowe, "Chemical Micro Process Engineering (Fundamentals, Modeling and Reactions)," WILEY VCH, (2003).

Kleiner, M. B., S. A. Kuhn and K. Haberger, "High Performance Forced Air Cooling Scheme Employing Microchannel Heat Exchangers," IEEE Transactions on Components, Packaging and Manufacturing Technology--Part A, 18(4) 795-804 (1995).

Lerou, P. P. P. M., T. T. Veenstra, J. F. Burger, H. J. M. ter Brake and H. Rogalla, "Optimization of Counterflow Heat Exchanger Geometry Through Minimization of Entropy Generation," Cryogenics, 45 659-669 (2005).

Liu, C. W., C. Gau and B. T. Dai, "Design and Fabrication Development of a Micro Flow Heated Channel with Measurements of the Inside Micro-Scale Flow and Heat Transfer Process," Biosensors & Bioelectronics, 20 91-101 (2004).

Okabe, T., K. Foli, M. Olhofer, Y. Jin and B. Sendhoff, "Comparative Studies on Micro Heat Exchanger Optimization," Proceedings of Congress on Evolutionary Computation, 1, pp. 647-654 (2003).

Olofsson, J., J. Pihl, J. Sinclair, E. Sahlin, M. Karlsson and O. Orwar, "A Microfluidic Approach to the Problem of Creating Separate Solution Environments Accessible from Macroscopic Volumes," Analytical Chemistry, 76 4968-4976 (2004).

Piekos, E. S. and K. S. Breuer, "Numerical Modeling of Micromechanical Devices Using the Direct Simulation Monte Carlo Method," Journal of Fluids Engineering, 118 464-469 (1996).

Rebrov, E. V., S. A. Duinkerke, M. H. J. M. de Croon and J. C. Schouten, "Optimization of Heat Transfer Characteristics, Flow Distribution, and Reaction Processing for a Microstructured Reactor/Heat-Exchanger for Optimal Performance in Platinum Catalyzed Ammonia Oxidation," Chemical Engineering Journal, 93 201-216 (2003).

Roudsari, F. H., M. J. Kharaajoo and M. Khajepour, "Design of Intelligent Predictive Controller for Electrically Heated Micro Heat Exchanger," Proceedings of Advances in Intelligent Computing: International Conference on Intelligent Computing, Lecture Notes in Computer Science, 3645 286-295 (2005).

Stief, T., O. U. Langer and K. Schubert, "Numerical Investigations of Optimal Heat Conductivity in Micro Heat Exchangers," Chemical Engineering Technology, 21(4) 297-302 (1999).

Tonomura, O., S. Tanaka, M. Noda, M. Kano, S. Hasebe and I. Hashimoto, "CFD Based Optimal Design of Manifold in Plate-Fin Microdevices," Chemical Engineering Journal, 101 397-402 (2004).

S. Balaji and S. Lakshminarayanan *

* Author to whom correspondence may be addressed. E-mail address: chels@nus.edu.sg

Department of Chemical and Biomolecular Engineering, 4 Engineering Drive 4, National University of Singapore, Singapore 117576

Manuscript received July 31, 2006; revised manuscript received August 23, 2006; accepted for publication August 28, 2006.

Table 1. Velocity data for one inlet--one outlet port microchannel
configuration

Input-Output Port Maximum Minimum Velocity
Combination Velocity Velocity difference
 (m/s) (m/s) [DELTA]V (m/s)

IP3-OP1 2.105 1.689 0.417

IP2-OP1 2.042 1.615 0.427

IP1-OP1 2.565 1.639 0.926

Table 2. Velocity data for one inlet--two outlet ports
microchannel configuration

Input-Output Port Maximum Minimum Velocity
Combination Velocity Velocity difference
 (m/s) (m/s) [DELTA]V (m/s)

IP1-OP1&OP2 2.302 1.580 0.722

IP1-OP1&OP3 2.267 1.644 0.623

IP1-OP2&OP3 2.167 1.687 0.480

IP2-OP1&OP2 2.050 1.676 0.374

IP2-OP1&OP3 2.012 1.749 0.263

IP2-OP2&OP3 2.048 1.684 0.364