Numerical investigation on comparative performance assessment of solar air heater with different art

Abstract

Keywords: Artificial Roughness; Renewable Energy; Solar Air Heater; Triangular Duct and Reynolds Number

Introduction

A solar air heater (SAH) utilizes an absorber plate to transform a portion of solar energy into usable thermal output. However, due to the lesser capability of air to conduct the heat transfer lesser heat transfer enhancement is

observed. To enhance the heat transfer phenomenon significantly by maintaining the constant heat flux and constant temperature difference it is possible only with the help of artificial roughness inserted on the absorber plate [1-3]. As the roughness elements break up the viscous sub-layer on the absorber plate, more reattachment sites

*Corresponding author. *E-mail address: santoshgk1988@gmail.com This paper was recommended for publication in revised form by Editor-in-Chief Ahmet Selim Dalkılıç Published by Yıldız Technical University Press, İstanbul, Turkey Yıldız Technical University. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

are created, resulting in the improvement of Nu from the roughness elements to the air flow [3]. However, a negative consequence in the form of a more significant pressure drop is also seen, which boosts the SAH pumping power in rib-roughened cases. On the absorber plate, artificial roughness may be applied in a variety of ways [4-15]. Webb et al. [16] investigated the influence of circular rib with different range of parameters. They reported that the maximum Nusselt number (3.39) and minimum friction factor (2.38) is observed at relative roughness height (e/D) of 0.01[17]. In the open literature similar results are observed [18] with the maximum thermal performance parameter (THPP) of 1.71. With grit type artificial roughness, the Nuavg enhanced by 152% and TPP boosted by 12-36% [20]. Han et al. [21] investigated the impact of inclined ribs in a rectangular duct SAH. They reported that compared with other angle of attacks, 60o inclined rib developed the maximum normalised Nu. However, when the similar analysis is carried out in square duct the maximum heat transfer enhancement is observed in case of 45o inclined rib [22]. This is due to the slowing down of the secondary fluid flow produced by roughness elements moving in the direction of the side walls. In comparison with continuous ribs, discrete ribs allow the secondary flow to merge with main flow appropriately resulting in increase of flow velocity leading to better heat transfer performance [23]. Heat transfer phenomenon may be enhanced even with wire ribs by winding across the absorber plate in different ways, like V ribs [24-27], W-shape [28], and other patterns. The highest Nusselt number is found with V-shaped ribs compared to other roughness elements [29]. Liou et al. [30] studied the effects of artificial ribs and reported that square ribs achieved the maximum Nusselt number with

minimal frictional losses in comparison to other rib shapes. However, when the analysis is carried out experimentally, it is observed that the square ribbed elements have lesser penalty of friction (fp) and higher Nu than the triangle rib [31]. Rokni et al. [32] numerically investigated the influence of ribs on two dimensional SAH model. They reported that when compared with 3D model, 2D models developed better heat transfer characteristics in case of chamfered roughness ribs. Soi et al. [33] investigated the influence of K-shaped ribs by theoretically developing the correlations to calculate the Nu and fp. Maximum Nu is observed at relative pitch of 6.23. To validate the CFD code and the computational domain, quality of the grid and choosing the appropriate turbulence model are the two crucial components required [34]. Additionally, it is observed that the k-ε RNG model is the most effective in simulating turbulent flow in the SAH [35-37]. Only the V-shaped ribbed SAH has more than a ± 3% discrepancy between the values anticipated using CFD models and the actual observations [38]. Korpale et al. [39] investigated the influence of rectangular ribs on the SAH. They reported that the optimal TPP value was 2.77 at 17.22 P/e and 0.044 e/D. The open literature indicates that altering the flow passage (by adjusting the computational domain) and incorporating various artificial roughness elements can substantially enhance the performance of solar air heaters (SAH). This explanation served as the driving force for the authors’ present inquiry, which examines the impact of different rib-roughness shapes to assess the SAH with triangle flow passage and attempts to pinpoint the optimal rib element geometry relative to other investigated geometries. Fewer research on alternative flow passages, including square and triangle. Very less research is carried out on the

triangular duct SAH with different artificial roughness ribs. So, in this research, an effort is made to analyse the impact of different artificial ribs (rectangular rib with backward chamfered, rectangular rib with frontward chamfered, rectangular rib with various e/w, square rib, rectangular and semi-circular rib) on triangular duct to study the fluid flow physics and heat transfer in a systematic manner. The commercial software, ANSYS Fluent 18.1 is employed to simulate the flow domain. The impact of artificial roughness on triangular ducts with intrusions is studied to assess changes in velocity and temperature profiles.

Problem Description

In a standard triangular duct, sharp corners cause flow deceleration. To improve airflow performance and enhance heat transfer, a new approach has been introduced. The solution has been analyzed by considering different artificial roughness ribs by modifying the Re number [55]. In the present case, seven different models are studied, which include different roughness elements on the triangular duct with the same geometry, as shown in Figure 1 (a&b). The optimal roughness geometry is found by creating roughness elements of

various forms and examining their effects in different geometries are considered for the analysis which is represented in Figure 1(b) with geometrical parameters. To analyze the fluid flow pattern, a triangular duct SAH of 0.9m upstream length followed by an absorber plate made of aluminum of 0.5 m length is considered [56]. To avoid the backflow from the exit of the absorber plate, a downstream length of 0.4 m is considered. A constant heat flux of 1000 W/m2 is considered the energy input for the analysis. The simulations were run with the following presumptions to make good use of the computing resources [40-43]:

3. For both roughened and smooth ducts, a thin wall

model is considered for the absorber plate. 4. Throughout the analysis, the thermo-physical properties of the absorber plate and flowing air is assumed to be constant.

5. The losses due to radiation are neglected.

Mesh Generation Mesh impacts the precision, convergence, and speed of a simulation and produces effective solutions. The independent patch approach is used to the triangular cross-sectional duct to produce tetrahedral grids. The accuracy of the findings improves with finer meshing. The interaction between the solid and fluid creates a finer mesh as a result. Figure 2 shows a fine-grained mesh. For all ducts with heat-conducting surfaces, face sizing of a 5mm element size is used for a finer mesh. Each edge’s size is specified independently, with element sizes ranging from 1.5mm to 4.33mm. Tetrahedral grids are made throughout the duct using a patch independent approach. The grid independence test is conducted to see if the numerical analysis described in the next section is accurate. Computational Methodology In order to study the fluid flow and thermal characteristics, the current computational domains for rectangular

roughened, and smooth ducts are separated into structured quadrilateral elements (non-uniform). Numerical equations employed The principles of conservation, encompassing the continuity, momentum, and energy equations (1-3) [57], are solved utilizing ANSYS Fluent™ software, version 18.1. (1) (2) (3)

(4) (5) Spatial discretization gradients are resolved using the least-squares cell method. To accurately capture the viscous sublayer, fine mesh elements are placed near the absorber plate, with quadrilateral cells discretized in non-uniform grids through CFD analysis. Tetrahedral mesh is selected for CR and SCR, while hexahedral elements are utilized for SR, BCRR, RR, and FCRR ribs [59]. Due to the pronounced gradients in the ribbed area. Boundary Conditions A wall boundary condition is assigned to both the insulated and absorber surfaces, while inlet and outlet boundary conditions are applied to the corresponding areas of the computational domain. Based on the Reynolds number range (4000 to 20000), the inlet pressure and temperature of the duct are set to 1 bar and 295 K, respectively. The duct walls are considered insulated, and a uniform heat flux of 1000 W/m² is applied to the roughened absorber plate [56].

Results And Discussion

A solar air heater with different artificial roughness on absorber plate using intrusions by varying the Reynolds number is analysed. The obtained results are compared with fixed and smooth ribs working under identical conditions.

Validation of the Numerical Results A GIT (grid independence test) is performed and it is mandatory to ensure for all cases that the thickness of the 1st layer corresponding to y+ should be equal to 1. The grid cells ranging from 410792 onwards are selected to conduct the CFD simulation, in order to ensure that the obtained results are grid independent [56]. It can be observed from Figure 3 that after 1505191 cells for the case of Reynolds number= 12,000, the percentage variations in average

Figure 3. Grid independence test. friction factor and Nusselt number were less than 0.5%. The deviation from 1825528 to 2123428 cells is 0.003%. It is essential to choose appropriate turbulence models to validate the obtained results. To assess different turbulence models, simulations were performed on a conventional solar air heater [61]. Due to the small rib height relative to the duct height, fluid recirculation is limited to the region close to the absorber plate’s surface. Low-Reynolds number models were chosen for better accuracy near walls, as they are more effective in this region compared to high-Reynolds models [9,10]. This consideration influenced the selection of suitable turbulence models, with several RANS-based

models tested. Five turbulence models were applied across a Reynolds number range from 4000 to 20000 to estimate the Nusselt number. The results were validated against the Dittus-Boelter equation, as depicted in Figure 4. Findings revealed that the RNG model produced results closest to the theoretical Nusselt number. At low Reynolds numbers, the normalized Nusselt number (Nunorm) exhibited a maximum deviation of -6.45% compared to existing literature, while for higher Reynolds numbers, the deviation was less than 5%. These outcomes indicate that the selected mathematical models are both accurate and appropriate for the present study [47-50].

Figure 5. Influence of Nunorm on different roughness elements. Influence of Different Artificial Roughness Ribs on SAH Figure 5 shows the effect of normalized Nusselt number across various roughness elements as the Reynolds number ranges from 4000 to 12000. It is observed that a significant improvement in heat transfer enhancement (Nunorm) with an increase in the Re. At maximum Re of 20000, maximum Nunorm is observed in all the roughness elements. When compared with SCR (6.42%) other elements like CR, SR, RR with different e/W, FCRR and BCRR developed lesser Nunorm like 2.98%, 2.65%, 3.02%, 2.99%, 2.72%, 2.12% and 1.42% respectively. In case of FCRR with e/W 2 experiences poor augmentation due to least Nunorm. Generally, better heat transfer takes places due to breaking up of viscous sub-layers caused by the rib-presence elements inserted above the surface of the plate. Additionally, the various re-attachment locations that the ribs produce across the absorber plate cause the flow to separate upstream of the rib element and reattach downstream of the rib element, which causes a significant change in random mixing of the fluid particles between two successive ribs. Local turbulence is shown to be more intense just after rib components [5154]. The turbulence intensity significantly diminishes with increasing distance from the downstream rib, especially close to the rib surface. It is observed that maximum turbulence intensity of 16.23% is noted in SCR however least intensity is observed in BCRR with e/W of 2 (7.3%). Another crucial characteristic that shows the flow configuration within the SAH duct is the rib-top surface. Due to ribs on the absorber plate greatly influences the fluid flow phenomenon resulting in recirculation of the fluid particles and increase in the reattachment lengths [34-35]. Figure 6 illustrates the variation of velocity streamlines for different roughness elements at a constant Re of 16000 [62]. It is observed that due to the presence of roughness

elements recirculation of fluid particles or secondary flow is developed near to the front and rear end of the rib element. However, stronger eddy formation is observed near to the rear end of the roughness elements rather than the front surface. As the roughness geometry is modified velocity streamline near to the surface of the absorber plate also changes. However, at FCRR with e/w of 2, the rib-top surface correctly directs the flow streamline avoiding the flow separation across the rib. Simultaneously, the rib-leading element’s edge has a chamfered design that directs flow toward the centre, improving the random mixing of viscous layers into the mainstream resulting in the enhancement of convective heat transfer coefficient as shown in the Figure

7. The production of eddies when the flow is detached from

the tip of the roughness elements in case of different shaped rib components leading to a decrease in the hlocal between two ribs. As SCR and CR have no sharp edges, the streamlines run over the rib element without slightly perturbing the flow field, resulting in poor local turbulence intensity and a reduced local heat coefficient. As a result, the SCR and CR acquired a comparatively lesser Nu. With an increase in the heat transfer phenomenon in case of artificial roughness elements SAH friction penalty also increases. Figure 8 illustrates the variation of fp for different roughness elements at different Re. It is observed that with an increase in the Re higher fp values are associated. With an increase in the Re drastic decrease in the fp is achieved. The rib-element with corners (SR, RR with e/w and BCRR and FCRR) often has a higher fp, compared with SCR and CR. For various shaped roughness-geometries, fp in case of CR, SCRR, SR, RR, BCRR and FCRR are 1.96, 2.01, 2.98, 3.12, 3.23, and 3.74 respectively. Due to the development of drastic secondary flow in case of FCRR maximum fluctuation in fp is observed. This can be

Figure 6. Velocity (m/s) profiles on different roughness elements. decreased by reducing the pressure gradient in the opposite direction of the main stream. The main objective of the modified SAH is to maximise the heat transfer phenomenon with minimum fp. However, it is observed that with an increase in the Nu friction penalty also increased substantially. In order to assess the performance of the SAH the effective method is by obtaining the Nunorm or fp. The equation for the thermos hydraulic performance parameter (THPP), which must be calculated as a result, is given in equation. TPP defines that well-constructed SAH predominates fp with an increase in the heat transfer phenomena. Figure 9 illustrates the variation of THPP with an increase in the Re for different roughness elements. It is observed that with an increase in the Re, THPP increased for all the roughness elements. At Re of

20000, maximum TPP is observed in case of SCR (2.99%) in comparison with the other roughness elements like CR, SR, RR with different e/W, BCRR and FCRR are 2.42%, 2.31%, 2.16%, 1.98%, 1.62% and 1.32% respectively. It is interesting to note that the greatest TPP and SCR correlate to the roughness geometry where both Nunorm and fp are maximal. Impact of Velocity and Temperature Distributions To prevent hot spot formation, the triangular duct is modified to three rounded-corner ducts. Figure 10(a) depicts how various roughness elements impact the local temperature gradient at maximum Reynolds number within the triangular duct. It is evident from Figure 10(a) that the highest temperatures are found on the surface,

Figure 7. Influence of hlocal on different roughness elements.

Figure 8. Influence of fp on different roughness elements. decreasing towards the duct’s center, where the lowest temperatures occur across all roughness types. The presence of roughness elements increases surface area, enhancing interaction with fluid particles. Most fluid particles make contact with the heated surface [62]. Modifications to the conventional duct result in a minor temperature gradient at the center but show significant temperature variation at the duct corners. Intrusions thin the laminar sublayer, increasing turbulence and vortex formation, which enhances secondary heat transfer mechanisms [Prasad and Mullick, 1983]. Consequently, the temperature gradient remains nearly uniform from the corners toward the center.

Figure 10(b) illustrates the effect of different geometries on the local velocity gradient within the triangular duct at maximum Reynolds number. The highest velocity occurs at the duct’s center across all configurations [62]. However, in the conventional duct, sharp edges inhibit fluid mixing, leading to zero velocity at the solid-fluid interface. While all roughness elements achieve peak velocity at the center, in the CR configuration, maximum velocity is restricted to the central region, with a gradual reduction moving outward. When compared to the other elements SCR, the velocity profile is greater around the corners. In light of this, it can

Figure 9. Influence of THPP on different roughness elements. be said that the flow close may form in the corner’s rounded shape with roughened surface parts. Validation of the Proposed Model with the Open Literature The typical flow channel of a standard SAH is rectangular. It is observed that area of the SAH plays predominantly important role to enhance the heat transfer and the overall performance. Triangular cross-sectional SAH, also known as unusual SAH, is considered in the present research for this reason. In the present analysis comparison is done with the modified duct with the conventional duct in the open literature. Thus, this comparison can only be used to collect quantitative data. The comparison shows a significant fluctuation in Nunorm and fp when the parameters are slightly changed. It is observed that BCRR geometry

yielded the most significant value of Nuenh for both SAH designs. However, in comparison with triangular duct the obtained fp value [32] is three times larger than the rectangular duct. In case of rectangular duct better Nunorm and fp is obtained in case of CR and SR [18, 32, 42-46] which indicates that the heat transmission outweighs the friction penalty. The TPP levels reported by the various researchers do not agree, and a variance of around 10% is noted. From the proposed roughness elements, it is observed that the roughness elements which possess the sharp corners produced maximum TPP. This is due to an improvement in the secondary flow effectively mixing with primary flow improving heat transfer with minimal frictional cost. As a result, compared to traditional SAH, the suggested SAH yields the best TPP value.

a) b) Figure 10. Changes in Temperature and Velocity Profiles at Z/Itest= 0.7 with Re of 20000 for different models.

1. With the rib height being smaller than the duct height,

fluid particle recirculation is restricted to the region close to the absorber plate’s surface.

2. At lower Reynolds numbers, the predicted normalized

Nusselt number (Nu_norm) shows a maximum discrepancy of -6.45% compared to available literature. For higher Reynolds numbers, however, the deviation in Nu_norm predictions remains below 5%. 3 In case of FCRR with e/W 2 experiences poor augmentation due to least Nunorm. Generally, better heat transfer takes places due to breaking up of viscous sub-layers caused by the rib-presence elements inserted above the surface of the plate.

4. As the distance from the downstream rib increases, the

strength of the turbulence decreases drastically as it is near to the rib surface. It is observed that maximum turbulence intensity of 16.23% is noted in SCR however least intensity is observed in BCRR with e/W of 2 (7.3%).

5. The production of eddies when the flow is detached

from the tip of the roughness elements in case of different shaped rib components leading to a decrease in the hlocal between two ribs.

6. Modifications to the conventional duct result in a slight

temperature gradient at the center, while noticeable temperature differences appear in all duct corners. The intrusions thin the laminar sublayer, increasing turbulence and encouraging vortex formation.

7. The circular rib (CR) has the poorest normalised Nunorm

which is 1.95, compared to other shaped roughness elements. In contrast, standard SAH is a superior choice for other roughness geometries. Future Scope

1. In the SAH-duct domain, transverse vortex generators

for improving heat transfer in duct flows have never been explored.

2. The rectangular SAH-aspect duct›s ratio impact may be

examined to see how it affects the SAH-functionality. duct›s

3. The current analysis was conducted in a controlled,

simulated environment using an electric heating plate; however, further research may allow for real-time operation under direct sunlight.

4. In the SAH-duct domain, higher ranges of Re may be

examined to understand how they affect the thermal hydraulic performance.

Nomenclature

Solar air heater Thermal performance parameter Computational Fluid dynamics Circular rib Semicircular rib Rectangular rib

A Cp Dh e/Dh fp h havg I k ltest linlet loutlet Nu Pr Prt Re T

Duct area Heat capacity Dimeter Relative height of roughness Frictional penalty Height of the triangular duct Average heat transfer coefficient Turbulent Intensity Turbulent Kinetic Energy Length of test section Length of entrance section Length of exit section Nusselt Number Prandtl Number Turbulent Prandtl Number Reynolds Number Temperature

Data Availability Statement

The authors confirm that the data that supports the findings of this study are available within the article. Raw data that support the finding of this study are available from the corresponding author, upon reasonable request.

Conflict Of Interest

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethics

There are no ethical issues with the publication of this manuscript.

Share and Cite