Natural Convection Heat Transfer Enhancement with Single V-Type Partition Plate

Abc

Description

5th European Thermal-Sciences Conference, The Netherlands, 2008COMPUTATIONAL ANALYSIS OF NATURAL CONVECTION WITH SINGLE V-TYPE PARTITION PLATE 1 2 3 R. L. Edlabadkar , N. K. Sane , G.V.Parishwad1PVG’s College of Engg. and Technology, Pune, India-411009 2JSPM’s College of Engineering, Pune, India-411028 3Govt. College of Engineering, Pune, India-411005 Abstract Misumi and Kitamura (1990) have reported an experimental work on enhancement of natural convection heat transfer from vertical plate with a horizontal partition plate and V-plates in water as ambience. This heat transfer enhancing technique was further investigated experimentally in air as ambience by Parishwad et al. (2006). The numerical analysis of this technique is done using Computational Fluid Dynamics (CFD) software, FLUENT, for natural convection adjacent to a vertical heated plate in ambient air surrounding. Because of boundary layer development the vertical fins are inapplicable in the heat transfer enhancement. As compared to conventional vertical fins, this V-type Partition Plate works not only as extended surface but also as Flow separators. This V-type Partition Plate is compact and hence highly economical. It was observed 0that among the three V-type partition plates, subjected to computational analysis, 90 V-partition plate gives the maximum 12% and 15.27% heat transfer enhancement as compared to vertical partition plate and horizontal partition plate respectively. It is ...

Informations

• Publié par: Phion
• Nombre de lectures: 97
• Langue: English

Extrait

COMPUTATIONAL ANALYSIS OF NATURAL CONVECTION WITH SINGLE VTYPE PARTITION PLATE 1 2 3 R. L. Edlabadkar , N. K. Sane , G.V.Parishwad 1 PVG’s College of Engg. and Technology, Pune, India-411009 2 JSPM’s College of Engineering, Pune, India-411028 3 Govt. College of Engineering, Pune, India-411005 Abstract Misumi and Kitamura (1990) have reported an experimental work on enhancement of natural convection heat transfer from vertical plate with a horizontal partition plate and V-plates in water as ambience. This heat transfer enhancing technique was further investigated experimentally in air as ambience by Parishwad et al. (2006). The numerical analysis of this technique is done using Computational Fluid Dynamics (CFD) software, FLUENT, for natural convection adjacent to a vertical heated plate in ambient air surrounding. Because of boundary layer development the vertical fins are inapplicable in the heat transfer enhancement. As compared to conventional vertical fins, this V-type Partition Plate works not only as extended surface but also as Flow separators. This V-type Partition Plate is compact and hence highly economical. It was observed 0 that among the three V-type partition plates, subjected to computational analysis, 90 V-partition plate gives the maximum 12% and 15.27% heat transfer enhancement as compared to vertical partition plate and horizontal partition plate respectively. It is further observed that the base heat 0 transfer coefficient (h b V-partition plates is better than all other configurations. The base) of 90 heat transfer coefficient, h b for plain plate is least among all. 1 Introduction The active heat transfer enhancement techniques have not found commercial interest because of the capital and operating cost of the enhancement devices. The majority of passive techniques employ special surface geometry or fluid additives for enhancement i.e. no direct application of external power. The fins are commonly used whenever it is difficult to increase the rate of heat transfer either by increasing heat transfer coefficient or by increasing the temperature difference between the surfaces and surrounding fluid. The vertical fins are inapplicable to the heat transfer enhancement of a tall vertical plate. This is because the boundary layer developed over the tall plate becomes very thick. To obtain an appreciable improvement of the heat transfer, the fin height must be greater than the boundary layer thickness. Obviously such high fins are not practical. The work by Prasolov (1961), Heya et al. (1982), Bhavnani and Bergles (1990) suggest that the roughness elements whose height is less than the boundary layer thickness will have no appreciable influence on the heat transfer of natural convection and these elements will work as flow retarder rather than the heat transfer promoter. In order to dispose off the boundary layer restrictions and develop a compact high-performance heat transfer plate, some investigators have developed horizontal partition plate and V - shaped plates. Firstly, Misumi and Kitamura (1990) have reported an experimental work on enhancement of natural convection heat transfer from vertical plate having a horizontal partition plate and V- plates in the water ambience.

 base plate with vertical fin base plate with horizontal fin base plate with V fin Figure 1 : Different schematic Test Plates for Computational Analysis Misumi and Kitamura (1990) found that the heat transfer in the downstream region of the partition plate is markedly enhanced when the plate height exceeds certain critical values because of the inflow of the low temperature fluid into the separation region. For vertical plate with V-shaped fins, the heat transfer coefficient obtained was 40% higher than the conventional fins. Further, it was observed that the ratio of the heat transfer enhancement exceeds the ratio of the surface enlargement; moreover the enhancement obtained for horizontal partition plate and vertical fin was less than V-plate. The experimental investigation on V-type partition plate with different included angles, in air as ambience is explained in Parishwad et al. (2006). The schematic test plates for computational analysis are shown in the Figure 1. Following formulae are used to calculate Grashof number Gr, Nusselt number Nu . The air properties at mean film temperature (mean of isothermal temperature of base plate and Ambient temperature) are used. 1. Average heat transfer coefficient h a ,  q convection  h a = θ  Where, q convection - heat flux per unit area(Q convection / A e )  A e - effective area of the base plate and fin, θdifference ( T– Temperature s - T amb ) 2. Base heat transfer coefficient h b ,  q convection  h b = θ  Where, q convection - heat flux per unit area(Q convection / A b )  A b - Area of the base plate θdifference ( T– Temperature s - T amb ) 3. Average Nusselt number Nu a ,  h a H x  Nu a  =  K  Where, H – height of the base plate

4. Grashof number Gr x , 3  gβθx  Gr x = 2 ν  x – height from leading edge of the base plate 2 Numerical Approach and Procedure The laminar air flow over a vertical base plate with length 0.3m, width 0.3m, and V shape fin (the fin limb length is 0.15m and width 0.05m) attached to it was numerically captured using Computational Fluid Dynamics (CFD) software of FLUENT(version 6.2.16) with laminar viscous model. Solid modeling, computational grid generation and meshing were done using software of Gambit Preprocessor 2.2.30. 2.1Computational Grid For comparison, the effective area of the fins and base plate is kept same for all numerical models. Table 1 shows the effective area for 30 mm fin height numerical models. Figure 2 shows the domain for the computational analysis. Taking symmetry into account, only quarter of the actual model (base plate and fin) was developed and solved. Figure 2 also shows thedifferent half fins used for computational analysis. Thus computational domain of size 2.1 m × 0.5 m × 0.45 m was developed using CFD software of Gambit. The domain boundaries were optimized such that any further increase in their limits will have insignificant effect on to the physics of the problem and subsequent results. Thus an optimal approaching domain length of 0.03 m (height of the base plate) was used to achieve fully developed flow conditions upstream of the Base plate with V fin. The top boundary was optimized to five times the height of the base plate with proper successive grading ratio. The side boundary was optimized at a distance equal to the width of the base plate. The front boundary was optimized at a distance of around fifteen times the height of the fin. The grid is made up of around 0.7 millions, hexahedral elements aligned with the flow direction to reduce numerical dissipation errors and thus improve the quality of numerical solutions. Near-wall (base plate and fin), apex of the V fin and region near downstream edges of the V-fin are meshed with fine grid to resolve the high gradients encountered in these regions. The grid quality was checked by critical parameters viz. Skew ness, Area/Volume, Aspect ratio, adjacent cell ratio etc. Proper successive grading ratios were used to control the growth of the cells. 2.2Computational Domain Boundary and Continuum In the Gambit session the base plate and fin were defined as the walls. The right boundary and the boundary in the plane of base plate excluding base plate were defined as the mid symmetry and side symmetry respectively. All other remaining boundaries were defined as either Pressure Inlet or Pressure outlet. All volumes excluding base plate and fin were defined as the fluid continuum. In the early stage of the analysis, the fin and base plate were given no thickness contemplating steady state conduction, thus the problem was analyzed for natural convection only.

Table 1:Effective surface area of set ups 2 Surface Area of Plate(m ) Vertical Plate Vertical Plate Vertical Plate  Plain Vertical with horizontal with vertical with V type Plate fin fin fin Height of -------- 0.03 0.03 0.03 Fin (m) Base Plate 0.18 0.18 0.18 0.18 Area (A b ) Fin Area --------- 0.036 0.036 0.036 Total Area (A e 0.2160.216 0.216 ) 0.18 2.3Computational Parameters and Closure Model The unsteady state segregated solver with implicit formulation was employed as the laminar model because of the inherent laminar nature of the physics of the fluid flow of given problem. The model has reasonably good ability to resolve the complexities of flow separation in the region located near downstream edges of the V-fin and base plate. Because of the fine mesh, flow in the near- wall regions is well represented and no wall functions are utilized. The low Rayleigh number Navier- Stokes equations in conjunction with transport equation for dissipation rate are solved numerically. The absolute velocity Formulation and Superficial Velocity for Porous Formulation was used and Cell Based Gradient option was opted for. To reduce numerical errors, second order spatial discretization schemes are used in the calculations. The air is defined as the fluent fluid material with density following incompressible ideal gas laws and steel was used as the material for base plate and fin. The base plate and fin were kept at constant elevated temperature assuming isothermal condition. Thus the analysis was done for different isothermal conditions of base plate and fin. This isothermal hypothesis will do reasonably for short fin heights. Each computational iteration is solved implicitly. The convergence of the computational solution is determined based on scaled residuals for the continuity and energy equations. The area weighted average of total surface heat fluxes on the fin; fin shadow and base plate were also monitored to check the convergence of the solution. 2.4Numerical Accuracy Uncertainties The process of building fine grids is repeated until further mesh refinements have insignificant effects on the results. This approach is used to reduce the uncertainties associated with the numerical flow field. The experimental results were used as the bench marks to calculate the numerical accuracy

Figure 2:Different half fins for computational analysis 3 Result Analysis For validation of the numerical analysis, the experimental results are compared with computational analysis results for plain plate. The graph in Figure 3 shows the close matching of these results with variation within 10% (maximum variation 8.5% and minimum variation 3%) validating our computational setups and analysis.

Figure 3:Validation of computational analysis graph

Computations were performed for the geometrical configurations shown in Figure 2, for equal base and fin areas dissipating heat under natural convection condition for temperature differenceθ, 0 0 varying from 30 C to 150 C in the steps of 30. Results are obtained in terms of the surface heat flux for various configurations to enable computation of average and base heat transfer coefficients. The performance of different fin configurations is studied by comparing these values. Figure 4, 5 and 6 show the low-pressure regions generated between the limbs of V fin on downstream side. This helps induction of low temperature ambient fluid in the region. The flow separates, destructing the boundary layer and leading to establishment of high heat transfer zone in the downstream side. The fin height at any location must be greater than the local boundary thickness to disturb the same. The low pressure region created at the apex zone of V fin enabling suction effect and the third component (z direction, normal to base plate) of velocity for V configuration helps the V configuration perform better than horizontal and vertical fin configurations. These Figures also show the velocity contours in the region strengthening the high heat transfer zone hypothesis. 0 0 0 Figures 7 and 8 show the base surface heat transfer coefficients for 90 , 120 and 60 V fins. It is 0 evident that the 90 V fin gives least resistance to flow separation in the upstream region and most effective high heat transfer region in the downstream region of the base plate. It was observed that among the three V-type partition plates, the maximum increase in heat transfer enhancement is 12% 0 for 90 V-partition plates as compared to vertical partition Plate and 15.27% as compared to 0 horizontal partition plate. It is further observed that the base heat transfer coefficient (h b V) of 90 Partition Plate is better than all other configurations. The h b for plain plate is least among all

0 Figure 4:Velocity Contours and Pressure Contours for60Vfin

0 Figure 5:VfinVelocity Contours and Pressure Contours for 90

0 Figure 6:Velocity Contours and Pressure Contours for 120V fin

0 Figure 7:V finContours of surface heat transfer coefficient for 90

Table 2:Comparison of base heat transfer coefficient of different configurations Variation with Variation with Base heat Temperature respect to respect to  transfer Difference Horizontal fin Vertical fin Configuration 0 coefficient (θ) C 2 Configuration Configuration (h kb) W/m (%) (%) Horizontal fin 150 5.43 ----- -2.9 Vertical fin 150 5.59 2.95 ----- 0 60 V fin 150 5.78 6.45 3.4 0 90 V fin 150 6.26 15.27 12 0 120 V fin 150 5.72 5.34 2.33

0 0 Figure 8:Contours of surface heat transfer coefficient for 120 and 60 V fins 4 Conclusions The computational analysis done so far confirms the experimental analysis done earlier well within 0 the 10% variation and concurs that 90 V-fin performs better than vertical fin and horizontal fin of the equal dimensions. The study for optimum ratio of base plate height and fin height is already undertaken and is in progress. 5 References 1.Bhavnani S.H., and Bergles A.E., 1990, Effect of Surface Geometry and Orientation on Laminar Natural Convection Heat transfer from a Vertical Flat Plate with Transverse Roughness Elements,Int. J. Heat Mass Transfer, 33, 965-969. 2.Heya N., Takeuchi M., and Fujii T., 1982, Influence of Various Surface Roughness on Free convection Heat transfer from a Horizontal Cylinder, Chem. Engg. J. Vol.23, 185-190. 3.Misumi Toshiyuki and Kitamura Kenzo, 1990, Natural Convection Heat Transfer from a vertical heated plate with a horizontal partition plates, J.S.M.E Int. J. Heat Mass Transfer, 38, 463-470. 4.Parishwad G.V., Edlabadkar R.L., Tasgaonkar G.S., and Sane N.K., 2006, Natural Convection th th Heat Transfer Enhancement with Single V type Partition Plate, 18 National and 7 ISHMT- ASME Heat Mass Transfer Conference, 342-346. 5.Prasolov R.S, 1961, The effects of Surface Roughness of horizontal Cylinders on Heat transfer to air, Inzh.-fiz. (In Russian), 4, 3-8.