The effect of 10µm microchannel on thermo-hydraulic performance for singlephase flow in semi-circular cross-section serpentine

Transcription

1 Journal o Mechanical Engineering and Sciences ISSN (Print): ; e-issn: Volume 12, Issue 2, pp , June 2018 Universiti Malaysia Pahang, Malaysia DOI: The eect o 10µm microchannel on thermo-hydraulic perormance or singlephase low in semi-circular cross-section S.M. Chan 1, K.H. Chong 1* and Basil T. Wong 1 1 Department o Mechanical Engineering, Faculty o Engineering, Computer and Science, Swinburne University o Technology Sarawak Campus, Jalan Simpang Tiga, Kuching, Sarawak, Malaysia Phone: ; Fax: * kchong@swinburne.edu.my ABSTRACT A microchannel heat exchanger, which oers various engineering applications, such as heating, ventilation and air-conditioning, is increasingly important due to its advantages in cost reduction or material, abrication, and physical size. The current nanotechnology impedes the abrication o microchannel hydraulic diameter at 10 µm and below; however, with rigorous research on nanotechnology, a smaller hydraulic diameter relative to the current microchannel is anticipated. This study simulated the eect o 10 µm transitional microchannel on thermo-hydraulic perormance or singlephase low in semi-circular cross-section, in which the boundary condition or wall temperature is constant 350 K. Its results show that the Dean vortices increase with Reynolds number, leading to a heat transer enhancement in the region o the bend. For Reynolds number o 175, the achieved heat transer coeicient is kj/m 2 K, which is superior to what has been reported in other literature; thereore, the study suggests that a hydraulic diameter channel o 10 µm could greatly improve the heat transer perormance. In addition, it iners the suitability o hydraulic diameter channel o 10 µm or single-phase low in semi-circular cross-section transitional microchannel. Keywords: Thermo-hydraulic; Single-phase low; Semi-circular; Serpentine; Microchannel. INTRODUCTION Microchannel oers a variety o engineering, biomedical application in industries, and one o the critical actors to enhance thermo-hydraulic perormance in microchannel is its hydraulic diameter. There have been many studies on various microchannel geometry with water as working luid where hydraulic diameter is above 10 µm, or example [1-25]. Given that the microchannel or 1-phase low in semi-circular crosssection transitional, with hydraulic diameter below 10 µm, is limited, due to current abrication techniques, it is, however, a challenge that can be overcome in the uture, as a result o rigorous research and development in microluidic area. The development o microluids appeared in 1980s with Tuckerman and Pease [7]. Deined as a channel s dimension in the range o greater or equal to 200 µm and less than 10 µm, where transitional microchannel is greater or equal to 10 µm [26], a microchannel has direct impact on micro-electro mechanical systems, microelectronic cooling and 3724

2 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection biomedical devices. Human blood vessels with a diameter range between 10 and 15 µm mainly involves heat and mass transer [27]. Microchannel could have 1-phase low and 2-phase low, where 1-phase low consists only o single luid low; 2-phase low consists o two-phase luid (e.g. gas and liquid). According to Kandlikar and Grande, or 1-phase (incompressible) channel, a diameter up to 200 µm will be subjected to the same physical change as that o conventional channel (hydraulic diameter more than 3 µm) [26]. The study on 1-phase low is a commonly encountered phenomenon in many applications. Most o the experimental research work on a hydraulic diameter above 10 µm, due to the limited technology to manuacture microchannel below 10 µm, which leads to limited simulation work on microchannel with hydraulic diameter o 10 µm or below. Thereore, this study sets to explore the thermo-hydraulic perormance or 1-phase low in semi-circular cross-section transitional microchannel with hydraulic diameter o 10 µm. Serpentine microchannel provides bend that promotes greater heat transer enhancement (HTE) as a result o intensiied mixing at the bend region. Some researchers studied the impact o microchannel bend on HTE with various radii o curvature [28-29]. Besides the radius o curvature, Reynolds number has a signiicant impact on HTE also. An increase in Reynolds number would increase the Dean vortices at bends. Figure 1 outlines the semi-circular cross section o transitional in this work; it is a repetitive module representing the whole compact microchannel heat exchanger. This module is represented by 2L, radius o curvature ( R c ) and channel diameter ( d ). In this research, the eect o Reynolds number on HTE and pressure drop are conined to ully developed low and constant wall ( T w ). The aorementioned conditions replicate the typical operating condition in microchannel heat exchanger. Other variables such as (L) and radius o curvature remain constant throughout the study. L Bend 2 Bend 3 Figure 1. Semi-circular cross section o transitional. The ollowing section reviews the important actors on thermo-hydraulic perormance in microchannel heat exchanger. Straight Channel Nusselt Number d Bend 1 Bend 4 Ratio o convective to conductive heat transer across a boundary is the deinition o Nusselt number (Nu). It is characterised as the ollowing, 3725

3 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) Nu hd k h (1) where h, D h and k are convective heat transer, hydraulic diameter and luid thermal conductivity, respectively. For semi-circular cross-section, D h is deined by Eq. (2) 1 2 / D h d / (2) Nusselt number o a semi-circular straight channel in the latest inding is displayed in Table 1. H 2 and Tw are constant wall heat lux and constant wall temperature, respectively. Table 1. Latest indings o Nusselt numbers Nusselt number (Nu) Reerence 2 T H w [30] [28] [28] [31] [32] Heat Transer Enhancement and Pressure Drop To investigate the enhancement o dierent path layout compared to straight channel, Venter proposed the Nusselt number o a channel compared to a straight channel [30]. The enhancement actor ( e Nu ) is deined as the ollowing. e Nu Nu (3) Nu path shape straight The same approach is used to measure the pressure drop penalty ( e ) o dierent path layouts. The pressure drop penalty ( e ) o the channel relative to a straight channel is given below. e (4) straight Flow Behaviour o Serpentine Channel The rate o convection heat transer within the channel is dependent on the characteristics o luid low. Channel bend could enhance the rate o convection heat transer. Moreover, in this study, there are our bends or a path layout with each having dierent HTE. At bends 1 and 4, the luid rotates in an anti-clockwise direction, whereas, at bends 2 and 3, it rotates in a clockwise direction (Figure 1). The rotational direction alternates ater every second bend, causing the bends to have dierent perormance o heat transer. Ater the luid low through bend 1, the vortices rotate in a reversed direction in the vicinity o bend 2, causing it to be weak in heat transer. Similarly, the reversal o direction in bend 4 ater bend 3 causes a reduction in 3726

4 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection heat transer. As bends 2 and 3 have the same rotation direction, the strength o the vortices is enhanced within the passage, which causes bend 3 to have good heat transer perormance. Likewise, the rotational direction o bends 4 and 1 is in the same sense, which causes HTE to be better in bend 1. Thereore, bends 1 and 3 are known as curvature reinorcing bends and bends 2 and 4 are known as cancelling bends. The rotation low o luid is called Dean Vortices. The characterisation o Dean vortices ( D ) is expressed as ollows: n n e d R 1 2 D R (5) The above equation is used to indicate the importance o inertial and centriugal orces relative to viscous orces. At low Dean Number or Reynolds number, the low tends towards the straight channel solution, where there will be no secondary low because the viscous orces o luid are larger than the inertial orces. Thereore, the viscous orce suppresses the ormation o secondary low, which is consistent with low heat transer enhancement and pressure drop at low Reynolds number. However, when the Reynolds number increases, the inertial orces become larger than the viscous orces, and a single vortex is developed. The low complexity, as well as the number and strength o vortices, increase with the Reynolds number. Eect o Geometrical Parameters Rosaguti et al. veriied the parameters that aect heat transer enhancement and pressure drop penalty along the path layout: L d and R c d [28]. To determine the eect, one o the parameters will be ixed when increasing another parameter. By increasing the value o L d at ixed R c d, both the heat transer enhancement and pressure drop penalty decrease; the straight passage o the channel becomes longer when L d is increased, which stabilises the low and, subsequently, reduces the mixing o luids. Rosaguti et al. stated that the low development occurs in this longer section, causing the relaxation o HTE towards unity, in other words, towards the value or the straight channel low [28]. Thereore, as the straight section between bends increases, the value o HTE decreases and the dierences o enhancement actor between reinorcing bends and cancelling bends are reduced also by increasing L d. Moreover, the radius o curvature has signiicant eects on the enhancement actor and pressure drop penalty. A decrease in the radius o curvature ( R c d ) leads to greater heat transer enhancement and pressure drop penalty; at lower values o radius o curvature, the bends are sharper, which promotes the ormation o Dean vortices and low separation in the bends vicinity. The ormation o low separation and Dean vortices can cause the pressure drop to increase around the bends. The Equation (5) can be employed to determine the strength o vortices by substituting the value o R c d into the equation. Furthermore, based on the path layout o the channel, the eect o radius o curvature on heat transer enhancement can be explained. When the value o R c d decreases, the bends become sharper and the vortices become more complex, which, in turn, promotes luid mixing and increases HTE at miter bends. However, when the c 3727

5 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) value o R c d increases, the bends become rounded and ewer vortexes emerge in the bends vicinity, which contributes to a lower enhancement actor. Eect o Reynolds Number The eects o Reynolds number on heat transer enhancement and relative pressure drop penalty have been proven by various studies. Rosaguti et al. reported that heat transer and pressure drop increase with the value o Reynolds number [28], whose consequence is attributed to the velocity o luid low being increased when the Reynolds number is increased. The eect o Reynolds number ( R e ) on luid low velocity can be determined with Equation (6): R VD (6) e h On one hand, at low Reynolds numbers, the thermal-hydraulic perormance o the channel is low and the viscous orces suppress the secondary low; on the other hand, at high Reynolds numbers, it causes the strength o Dean vortices to increase, which is consistent with the increase o heat transer enhancement and pressure drop penalty. At low Reynolds numbers, the vortex is weak, due to low development occurring within the straight section o the channel. As the Reynolds number increases, the strength o the vortex increases also, thus, allowing it to survive into the straight section. The strength o the vortex, being dependent on the location o the bends, is the highest at bends 2 and 4, the cancelling bends. Thereore, at high Reynolds number, the vortex strength is high, which in turn increases the distance it can progress into the straight section. METHODOLOGY In this study, the numerical analysis was accomplished with ANSYS FLUENT 16.1 [33]. All the calculations are solved using the second-order bounded dierencing scheme or the convective terms. The simulation was accomplished by setting the entrance channel as velocity inlet, and the exit channel as atmospheric pressure outlet, which was applied on both straight and path layouts. A rotational periodic type boundary condition was applied or the layout. The wall temperature was set constant at 350 K, and the luid inlet channel temperature was set at 300 K. There were three units o path layout, o which the central unit is where ully developed low occurs. Twenty planes were created within the central unit to determine the Nusselt number o straight and microchannels (Figure 6). The average bulk luid temperature and the local wall heat lux were determined rom the created planes. The average bulk luid temperature ( ) is deined as ollows: T b / 2 T max T min The average local wall heat lux ( " q w q '' w ) values was calculated by q q / 2 The local Nusselt numbers ( Nu ) are then determined by x max min 3728

6 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection Nu x " q w T T D k w Constant wall temperature ( T w ), luid thermal conductivity ( k ) and hydraulic diameter ( D h ) are assumed. Once the Nusselt number o and straight path layouts are obtained, the enhancement actors can be determined rom Equation (3). The pressure drop values are deined as thus: P P P P (10) Model Discretization b o, m i, m o m i m, straight e, The ANSYS FLUENT solver utilised an element-based Finite Volume Method (FVM) that involves discretising the spatial domain with mesh, which serves to create inite volumes to conserve mass, momentum and energy. To improve the computational eiciency and accuracy, a swept hexahedral mesh was used or the discretising model. The mesh was extruded rom the inlet along the axial path o the layout to produce the swept mesh, as shown in Figure 2. All the cells were set to quadrilateral shape. Edge division was used to speciy the desired number o divisions within the edge. h (a) (b) (b) Figure 2. Mesh density (a) along axial direction and (b) on a cross-section. RESULTS AND DISCUSSION 3729

7 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) This section reports on the result or Nusselt number o straight transitional microchannel, enhancement actor o channel, and eect o Reynolds number in transitional microchannel. Nusselt Number o Straight Transitional Microchannel In this study, the obtained Nusselt number or a semi-circular cross-section straight transitional microchannel is about under transitional low. Compared to the value 3.323, there is an increase o or 7.34% rom the theoretical value [28]. Moreover, compared to the smallest value 3.221, there is a 10.74% increase [29]. One possible actor that attributes to these dierent values is the applied meshing approach. Rosaguti et al. utilised ICEM computational luid dynamics to create a hexa-mesh or unstructured mesh, which is apparently dierent rom the ANSYS Meshing approach in this study [28]. Figure 3 shows the Nusselt number at dierent locations along the channel. The Nusselt number is the highest at the inlet, then, it decreases gradually until the constant value. It reaches a constant value just beore halway o the 3 mm channel, which is approximately 0.6 mm rom channel inlet. Figure 3. The obtained Nusselt number versus channel length. In Figure 4(a), the temperature proile starts to develop rom the channel entrance until it reaches the channel outlet. The change o temperature is obvious at the channel upstream, whereas, downstream shows almost a single concentration colour, which is the result o the bulk luid temperature being close to the wall surace temperature. As heat transer exists along the channel, the temperature proile keeps changing, even as the low is ully developed. Figure 5 shows that the velocity proile starts to orm at the channel entrance. The velocity keeps growing towards the centre o channel until it is ully developed. Unlike the temperature proile that keeps increasing throughout the channel, there are no changes in the velocity proile when it is ully developed. The velocity magnitude or each layer remains constant until exiting the channel. 3730

8 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection (b) (c) (a) Figure 4. a) Sectional view o temperature boundary layer; b) Inlet channel temperature contour; c) Outlet channel temperature contour. (b) (c) (a) Figure 5. a) Sectional view o velocity boundary layer; b) Inlet channel velocity contour; c) Outlet channel velocity contour. Enhancement actor o channel Figure 6 shows that the eect o Reynolds number on luid low behaviour is very signiicant. At lower Reynolds numbers, the enhancement o heat transer in the path, compared to straight path, are slightly increased with no apparent peak. The average enhancement actor with Reynolds number o 25 was ound to be 0.99 because o the viscous orces being larger than the inertial orces. For a Reynolds number o 100, planes 3, 9, 14 and 18 show the peak value o bend 1, 2, 3 and 4, respectively. By inspecting the curve, the enhancement actor decreases dramatically immediately ater leaving the bends. This curve behaviour obeys the theoretical low o luid o channel as heat transer decreases in the straight section. As bend 1 and 3 are reinorcing bends, they show higher enhancement compared to bends 2 and 4. However, this condition is only applicable or Reynolds numbers up to 150 (Figure 6). Any urther increase in Reynolds number causes the low to become unsteady, in which case, bends 2 and 4 illustrate higher enhancement, compared to bends 1 and 3, which is theoretically not accurate. 3731

9 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) Rosaguti et al. and Fourie reported that Reynolds numbers exceeding 450 tend to produce unsteady low [28-29]. However, in this study, the low becomes unsteady with Reynolds numbers greater than 150. Thereore, it shows that by decreasing the hydraulic diameter, the allowable Reynolds number o luid steady low is decreased. The velocity contours shown in Figure 7 are in the ully developed section o the channel. It was observed that the rotational direction o Dean vortices alternates ater every second bend, causing the bends to have dierent perormance o heat transer. Ater the luid lows through bend 1, the vortices rotate in a reversed direction in the bend 2 vicinity, which causes weak heat transer. Similarly, the reversal direction in bend 4 ater bend 3 causes a reduction in heat transer. As bends 2 and 3 have the same rotational direction, the strength o vortices is enhanced within the passage, causing bend 3 to have good heat transer perormance. Likewise, the rotational directions o bends 4 and 1 are in the same sense, which causes the heat transer enhancement to be better in bend 1. Figure 6. Heat transer enhancement actor as a unction o proportional distance within the transitional microchannel with R c d =1 and L d =4.5 or T boundary condition. The dashed lines indicate the positions o each bend. Figure 7. Velocity contour at the bends downstream location or Reynolds number =

10 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection In Figure 8, the velocity is ound to be higher at bend 2 and bend 4. As these bends are cancelling bends, the rotational direction o the luid that ormed at the previous bend is disturbed, inducing the ormation o vortices in the bends vicinity. In Figure 8, the streamline suggests the vortices at cancelling bends are prominent over reinorcing bends; even with high vortices, bends 2 and 4 show lower heat transer enhancement than bends 1 and 3. Thereore, the heat transer enhancement is not dominated by the concentration o vortices; however, it is inluenced by the mixing reaction at each bend, which iners that the vortices at the cancelling bends support mixing at the reinorcing bends, causing greater enhancement actors. Eect o Reynolds number Figure 8. Fluid velocity streamlines in the studied channel. The eect o Reynolds number on average heat transer enhancement is shown in Figure 9. When the value o Reynolds number increases, the heat transer enhancement increases also. In addition, the eect o hydraulic diameter on heat transer enhancement is displayed in Figure 9, showing that, with lower hydraulic diameter, the enhancement actor o the channel tends to increase higher than or larger hydraulic diameter. For instance, at the Reynolds number o 200, the simulation result is 75% larger than the value provided by Fourie [29]. Furthermore, the maximum Reynolds number or steady low decreases with decreasing hydraulic diameter. Convection heat transer coeicient ( h ) was also determined using Equation (1). The results are as shown in Figure 10. At Reynolds number = 150, the calculated heat transer coeicient is kj/m 2 K. However, the heat transer coeicient as in Fourie s work with Reynolds number 400 is only kj/m 2 K [29], which indicates that the convective heat transer in this study is better than that o the conventional microchannel. 3733

11 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) Figure 9. Comparison between simulation and theoretical results with dierent hydraulic diameters. Figure. 10. Heat transer coeicient or various Reynolds numbers. In Figure 11, the pressure drop does not show any signiicant increase with increasing Reynolds number. At Reynolds number o 25, the pressure drop is about , whereas, at Reynolds number o 175, the pressure drop is approximately Thereore, the pressure drop increases only by The pressure drop rom the simulation has a lower value than that reported by Rosaguti et al. [28]. At lower Reynolds numbers (i.e ), the pressure drop o the straight channel is larger than or the path layout, which resulted in e <1. From Reynolds numbers o 125 onwards, the pressure drop o the layout is prominent over that o the straight channel ( e >1). Thereore, the pressure drop is also dominated by the Reynolds number. 3734

12 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection Figure 11. Eect o Reynolds number on pressure drop o channel. CONCLUSION A proposed methodology was used to investigate ully developed low in both straight and channels o semi-circular cross-section with hydraulic diameter 10 µm. The eect o Reynolds number was more prominent on the heat transer enhancement actor than on the pressure drop penalty. Increasing the Reynolds number led to a greater enhancement actor without incurring a large pressure drop penalty. For Reynolds number o 175, the obtained heat transer coeicient in the study was greater than the theoretical value, which is kj/m 2 K, inerring that hydraulic diameter channel o 10 µm could improve greatly the heat transer perormance in microchannel heat exchanger application. However, physical experiment is needed to enhance the knowledge in this speciic ield. ACKNOWLEDGEMENTS The author would like to acknowledge the assistance and guidance provided by Swinburne University o Technology Sarawak. REFERENCES [1] Zhang P, Yao C, Ma H, Jin N, Zhang X, Lü H, Zhao Y. Dynamic changes in gas-liquid mass transer during Taylor low in long square microchannels. Chemical Engineering Science. 2018;182: [2] Al-Neama AF, Khatir Z, Kapur N, Summers J, Thompson HM. An experimental and numerical investigation o chevron in structures in minichannel heat sinks. International Journal o Heat and Mass Transer. 2018;120: [3] Al-Neama AF, Kapur N, Summers J, Thompson HM. An experimental and numerical investigation o the use o liquid low in microchannels or microelectronics cooling. Applied Thermal Engineering. 2017;116:

13 Chan et al. / Journal o Mechanical Engineering and Sciences 12(2) [4] Filimonov R, Sorvari J. Numerical study on the eect o cross-section orientation on luid low and heat transer in a periodic triangular microchannel. Applied Thermal Engineering. 2017;125: [5] Imran AA, Mahmoud NS, Jaal HM. Numerical and experimental investigation o heat transer in liquid cooling mini-channel heat sink with dierent new coniguration models. Thermal Science and Engineering Progress. 2018;6: [6] Smakulski P, Pietrowicz S. A review o the capabilities o high heat lux removal by porous materials, microchannels and spray cooling techniques. Applied Thermal Engineering. 2016;104: [7] Tuckerman DB, Pease RFW. High-perormance heat sinking or VLSI. IEEE Electron Device Letters. 1981;2: [8] Missaggia L, Walpole JN, Liau ZL, Phillips RJ. Microchannel heat sinks or twodimensional high-power-density diode laser arrays. IEEE Journal o Quantum Electron. 1989;25: [9] Rahman MM, Gui F. Experimental measurements o luid low and heat transer in microchannel cooling passage in a chip substract. Proceedings o the ASME International Electronics Packaging Conerence. 1993; [10] Wang BX, Peng XF. Experimental investigation on liquid orced convection heat transer through microchannels. International Journal o Heat and Mass Transer. 1994; 37: [11] Peng XF, Peterson CP. Eect o thermoluid and geometrical parameters on convection o liquids through rectangular microchannels. International Journal o Heat and Mass Transer. 1995;38: [12] Peng XF, Peterson GP. Convective heat transer and low riction or water low in microchannel structures. International Journal o Heat and Mass Transer. 1996;39: [13] Harms TM, Kazmierczak MJ, Gerner FM. Developing convective heat transer in deep rectangular microchannels. International Journal o Heat and Fluid Flow. 1999;20: [14] Mala CM, Li DQ. Flow characteristics o water in microtubes. International Journal o Heat and Fluid Flow. 1999;20: [15] Lin QW, Gh MM, Li DQ. Pressure-driven water low in trapezoidal silicon microchannels. International Journal o Heat and Mass Transer. 2000;43: [16] Ren L, Qu D, Li D. Interacial electrokinetic eects on liquid low in microchannels. International Journal o Heat and Mass Transer. 2001;44: [17] Celata GP, Cumo M, Gudlielmi M, Zummo G. Experimental investigation o hydraulic and single phase heat transer in mm capillary tube. Journal o Microsale Thermophysical Engineering. 2002;6: [18] Wu HY, Cheng P. Experimental study o convective heat transer in silicon microchannels with dierent surace conditions. International Journal o Heat and Mass Transer. 2003;46: [19] Baviere R, Ayela F, Person SL, Favre-Marinet M. An experimental study o water low in smooth and rough rectangular micro-channels. ASME 2nd International Conerence on Microchannels and Minichannels [20] Hao PF, He FH, Zhu KQ. Flow characteristics in a trapezoidal silicon microchannel. Journal o Micromechanics and Microengineering. 2005;15:

14 The eect o 10µm microchannel on thermo-hydraulic perormance or single-phase low in semi-circular crosssection [21] Shen S, Xu JL, Zhou JJ, Chen Y. Flow and heat transer in microchannels with rough wall surace. Energy Conversion and Management. 2006;47: [22] Jung JY, Kwak HY. Fluid low and heat transer in microchannels with rectangular cross section. International Journal o Heat and Mass Transer. 2008;44: [23] Wibel W, Ehrhard P. Experiments on the laminar/turbulent transition o liquid lows in rectangular microchannels. Journal o Heat Transer Engineering. 2009;30: [24] Mirmanto DBR, Kenning JS, Karayiannis TG. Pressure drop and heat transer characteristics or single-phase developing low o water in rectangular microchannels. Journal o Physics:Conerence Series. 2012;395:1 14. [25] Kottasamy A, Kadirgama K, Annamalai K, Mohanesan K, Ramasamy D, Noor MM. et al. Titanium oxide with nanocoolant or heat exchanger application. Journal o Mechanical Engineering and Sciences. 2017;11: [26] Kandlikar SG, Grande WJ. Evolution o Microchannel Flow Passages- Thermohydraulic Perormance and Fabrication Technology. Journal o Heat Transer Engineering. 2003;24:3 17. [27] Maina JN, West JB. Thin and strong! The bioengineering dilemma in the structural and unctional design o the blood gas barrier. Physiological Reviews. 2005;85; [28] Rosaguti NR, Fletcher DF, Haynes BS. Laminar low and heat transer in a periodic channel with semi-circular cross-section. International Journal o Heat and Mass Transer. 2006;47: [29] Fourie JH. Thermal perormance o periodic channels with semicircular and triangular cross-sections. Potchesstroom: North-West University; [30] Venter J. The optimal hydraulic diameter o semicircular and triangular shaped channels or compact heat exchangers. Potchesstroom: North-West University;2010. [31] Rosaguti NR, Fletcher DF, Haynes BS. Low-Reynolds number heat transer enhancement in sinusoidal channels. Chemical Engineering Science. 2007;62: [32] Geyer PE, Fletcher DF, Haynes BS. Laminar low and heat transer in a periodic trapezoidal channel with semi-circular cross-section. International Journal o Heat and Mass Transer. 2007;50: [33] Liu JT, Peng XF, Yan WM. Numerical study o luid low and heat transer in microchannel cooling passages. International Journal o Heat and Mass Transer. 2007;50: