To illustrate the bubble behavior in the duct, selected photos of the boiling flow of R-134a from the side and top view covering the whole narrow duct to illustrate the effects of the gap size of the duct are shown in Figures 4.13 to 4.16. The results in Figure 4.13 for the large duct withδ= 2.0 mm clearly indicate that in the relatively upstream region near the duct inlet a great number of bubbles already exist in the flow. Bubbles of varying size can be seen. Specifically, in the upper part of the duct larger bubbles dominate obviously due to the buoyancy effects. These photos suggest that forδ= 2.0 mm the onset of nucleation boiling occurs immediately after the refrigerant enters the heated section of the duct. Because of the presence of the flanges, this bubble cannot be seen from the photos (Figure 2.2). Note that for a high G of 600 kg/m2s and a smallerδof 1.0 mm some delay in the bubble nucleation is seen in Figure 4.14. Besides, the bubble nucleation is earlier in the lower part of the duct. This is due to the difference in the buoyancy effect in different parts of the duct. More specifically, in the lower portion of the duct the flow is heated from above and hence is thermally stable. This in turn results in a lower convection heat transfer coefficient and obviously the heated surface temperature is higher for a fixed wall heat flux.
This higher Tw causes the earlier inception of the bubbles from the surface in the lower portion of the duct. Because the ONB somewhat unstable, the ONB location is not symmetry with respect to the vertical central plane through the duct axis. In fact, the ONB locations move back and forth with time in an irregular manner. But for a further reduction of the duct gap to 0.5mm an opposite trend appears with bubbles nucleated on the heated surface earlier in the upper part of the duct (Figure 4.15). This results from the fact that at the smaller Re for the smaller δof 0.5 and 0.2 mm laminar forced convective liquid flow dominates in the duct since the buoyancy-to-inertia ration is also small (Gr/Re2 = 4.14×10-2
& 6.63×10-3). In this laminar flow (Re = 2670 & 1070) subject to the inner duct heating Ciampi et al. [52] found that a helicoidal flow motion appears in the duct which maintains a supply of warm liquid at the top of the cylinder and cold liquid to the bottom[52], which causes higher Tw in the upper portion of the duct. Thus the bubble nucleation takes place earlier in the upper part of the heating surface. Comparing the results in Figures 4.13-4.15 clearly shows that in the smaller duct the coalescence of bubbles is more pronounced. For δ= 1.0 mm in the exit half of the duct many big bubbles exist. When the duct gap is reduced to 0.5 mm Figure 4.15 shows that big bubbles even dominate in the upper portion of the exit half of the duct. Thus in this region we have a slug flow. For a further reduction
ofδto 0.2 mm the bubble coalescence is so intense that the slug flow dominates the upper portion of the entire duct (Figure 4.16). The above results manifest that in the smaller ducts of 0.5& 0.2 mm the bubbly flow and slug flow coexist.
The photos of the boiling flow taken for the cases at different duct sizes and imposed heat fluxes in the small region around the middle axial location marked on Figures 4.13(a)-4.16(a) are shown in Figure 4.17. First of all, it is noted from the photo taken from the duct forδ=1.0 mm shown in Figure 4.17(a) for the case at Tsat=15℃ and G=600 kg/m2s at the imposed heat flux q=15 kW/m2 that a number of discrete bubbles nucleate from the cavities and slide along the heating surface. As the imposed heat flux is increased to q=25 kW/m2, the active bubble nucleation density increases and a lot more bubbles appear and they move faster (Figure 4.17(b)). Many coalescence bubbles are seen as the heat flux is raised to q=35 kW/m2 (Figure 4.17(c)). Then, the photos taken from the smaller duct with δ=0.5 mm shown in Figure 4.17(d) for the same G, q, and Tsat indicate that a large number of bubbles generated from the cavities in the heating surface tend to merge together to form big bubbles. As the bubbles become larger, they become distorted and elongated as they slide on the heating surface. As the imposed heat flux is increased slightly to q=25 kW/m2 (Figure 4.17(e)), the active bubble nucleation density increases and bubbles collide and coalesce more frequently. The coalescence bubbles rise faster than the tiny bubbles due to the larger buoyancy force associated with them. As the heat flux is raised to q=35 kW/m2 (Figure 4.17(f)), coalescence of the bubbles occurs irregularly at a very high rate. The coalescence bubbles can be very large. In fact, the liquid slugs and discrete bubbles coexist in the duct. At even higher imposed heat flux for q>30 kW/m2, the bubble departure frequency is very high so that it is difficult to clearly distinguish the individual bubbles. In general, the bubble departure frequency increases substantially with the imposed heat flux due to the fact that an increase in the imposed heat flux directly provides more energy to the cavities and more cavities on the heating surface can be activated. Besides, the bubble departure diameter increases slightly at increasing imposed heat flux due to the rise in the wall superheat. Then, the corresponding photo taken from the even smaller duct withδ=0.2 mm are shown in Figures 4.17(g)~(i) for the same G, q and Tsat. Note that for the smallerδmore coalescence bubbles are seen and they are even bigger especially at a higher imposed heat flux. This causes less bubbles nucleated at the heated surface.
The bubble characteristics in the narrow duct around the middle axial location forδ=
0.5 mm affected by the refrigerant mass flux and saturated temperature are illustrated by the photos in Figure 4.18. The results in Figures 4.18(a)-(f) indicate that at a higher mass flux the liquid refrigerant flow moves at a higher speed, which in turn tends to sweep the bubbles more quickly away from the heating surface. Collision and coalescence of bubbles are still significant. Besides, the bubble departure frequency is higher and the bubbles are smaller and in violent agitating motion. However, the active nucleation site density is lower.
Note that at the low mass flux the bubble coalescence is more important and a number of bigger bubbles form in the duct. Then, the effects of the refrigerant saturation temperature on the bubble characteristics are illustrated by comparing the photos in Figures 4.18(a)~(c) with Figures 4.18(g)~(i). The results indicate that at a lower saturation temperature the bubbles grow bigger and move slower due to the higher surface tension. Small bubbles are easier to merge into big bubbles.
To be more quantitative on the bubble characteristics, we move further to estimate the average bubble departure diameter and frequency and the average active bubble nucleation site density on the heating surface for the cases with the bubbly flow dominated in the duct from the images of the boiling flow stored in the video tapes. The results from this estimation are examined in the following. The effects of the three parameters, namely, the refrigerant mass flux, duct size and refrigerant saturated temperature, on the mean bubble departure diameter for the saturated flow boiling of R-134a at the middle axial location (z
=80 mm) in the annular duct are shown in Figures 4.19-4.23 by presenting the average bubble departure diameter against the imposed heat flux for various G,δ and Tsat. First, the effects of the refrigerant mass flux on the average bubble departure diameter shown in Figures 4.19 and 4.20 indicate that the average departing bubble is only slightly larger for a lower refrigerant mass flux. For example, at q =30 kw/m2, Tsat =15℃ and δ= 0.5 mm, the average bubble departure diameter for G =500 kg/m2s is only about 11% larger than that for G =600 kg/m2s (Figure 4.20(a)). Then, the data given in Figure 4.21 also suggest that the average bubble departing from the heated surface is slightly larger in the smaller duct only at the lower G of 500 kg/m2s for the imposed heat flux q > 20 kw/m2. For example, at q =30 kw/m2, Tsat =15℃ and G =500 kg/m2s, the average bubble departure diameter forδ= 0.5 mm is only about 18% higher than that forδ= 2.0 mm (Figure 4.21(a)). Otherwise the effects of the duct gap on the bubble departure diameter are
relatively small. Finally, the results in Figures 4.22 and 4.23 indicate that the average bubble departure diameter is smaller for a higher refrigerant saturated temperature. For instance, at q =30 kw/m2, G =500 kg/m2s and δ= 2.0 mm, and the average departing bubble for Tsat =10℃ is about 18% larger than that for Tsat =15℃ (Figure 4.22(a)).
How the bubble departure frequency is affected by the three parameters for the saturated flow boiling of R-134a at the middle axial location (z =80 mm) in the annular duct are shown in Figures 4.24-4.28 by presenting the average bubble departure frequency against the heat flux for various G, δ and Tsat. Note that the increase of the bubble departure frequency with the imposed heat flux is rather significant for all cases presented here. First, the effects of the refrigerant mass flux on the saturated flow boiling average bubble departure frequency are shown in Figures 4.24 and 4.25. The results indicate that the average bubble departure frequency is somewhat higher for a higher refrigerant mass flux especially at high imposed heat flux. For example, at q =26 kw/m2, Tsat =15℃ and δ
= 0.2 mm, the average bubble departure frequency for G =700 kg/m2s is about 14% higher than that for G =600 kg/m2s (Figure 4.25(b)). Then, the effects of the duct size on the saturated flow boiling average bubble departure frequency shown in Figure 4.26 manifests that the average bubble departure frequency is noticeably higher in the smaller duct. For instance, at q =26 kw/m2, Tsat =15℃ and G =500 kg/m2s the average bubble departure frequency forδ= 0.5 mm is about 23% higher than that forδ= 2.0 mm (Figure 4.26(a)).
Besides, at G =600 kg/m2s and at the same q, and Tsat the average bubble departure frequency forδ= 0.2 mm is about 29% higher than that forδ= 1.0 mm (Figure 4.26 (b)).
Finally, the data given Figures 4.27 and 4.28 indicate that the average bubble departure frequency is significantly higher for a higher saturated temperature. As an example, at q
=26 kw/m2, G =600 kg/m2s and δ= 0.2 mm the average bubble departure frequency for Tsat =15℃ is about 21% higher than that for Tsat =10℃ (Figure 4.28(b)).
The number density of the active nucleation sites for ONB affected by the three parameters for the saturated flow boiling of R-134a at the middle axial location (z =80 mm) in the annular duct are shown in Figures 4.29-4.33 by presenting the average active nucleation site density against the imposed heat flux for various G, δ and Tsat. The data clearly show the substantial increase of the active nucleation site density with the imposed heat flux for all cases examined here. First, the effects of the refrigerant mass flux on the
saturated flow boiling average active nucleation site density are shown in Figures 4.29 and 4.30. The results indicate that the average active nucleation site density is significantly higher for a smaller refrigerant mass flux especially at high imposed heat flux for the smaller ducts withδ= 0.5 & 0.2mm. For example, at q =30 kw/m2, Tsat =15℃ and δ=
0.5 mm, the average active nucleation site density for G =500 kg/m2s is about 22 % higher than that for G =600 kg/m2s (Figure 4.30(a)). Then, the effects of the duct size on the saturated flow boiling average active nucleation site density shown in Figure 4.31 manifest that the average active nucleation site density is significantly higher for a larger duct when the imposed heat flux exceeds 20 kw/m2. For instance, at q =30 kw/m2, Tsat =15℃ and G
=500 kg/m2s the average active nucleation site density forδ= 2.0 mm is about 45% higher than that forδ= 0.5 mm (Figure 4.31(a)). In addition, at G =600 kg/m2s and at the same q and Tsat the average active nucleation site density forδ= 1.0 mm is about 50% higher than that forδ= 0.2 mm (Figure 4.31 (b)). These results seem to introduce adverse effects for heat transfer for a reduction in the duct size. But a large portion of bubbles are merged to become large confined bubbles in the smaller duct forδ= 0.5 & 0.2 mm at higher imposed heat flux. Hence in the small duct the boiling heat transfer is better. Finally, the data shown in Figures 4.32 and 4.33 suggest that the average active nucleation site density increases significantly with Tsat especially at high heat flux. As an example, at q =30 kw/m2, G =500 kg/m2s and δ= 0.5 mm the average active nucleation site density for Tsat =15℃ is about 22% higher than that for Tsat =10℃ (Figure 4.33(a)).
In the small ducts withδ= 0.5 & 0.2 mm subject to a high imposed heat flux, the slug flow prevails and the flow is dominated by the big bubbles. The mean axial speeds of these bubbles are measured for the saturated flow boiling of R-134a at the middle axial location (z =80 mm) in the annular duct. The data are shown in Figures 4.34-4.36. First, the effects of the refrigerant mass flux on the average speed of the bubbles shown in Figure 4.34 indicate that the average bubble speed is substantially larger for higher refrigerant mass flux and higher heat flux. For example, at q =40 kw/m2, Tsat =15℃ and δ= 0.5 mm, the average bubble speed for G =600 kg/m2s is about 30% larger than that for G =500 kg/m2s (Figure 4.34(a)). For instance, at Tsat =15℃, δ=0.2 mm and G =600 kg/m2s, the average bubble speed for q = 28 kW/m2 is about 54% higher than that for q = 46 kW/m2 (Figure 4.34(b)).Then, the data given in Figure 4.35 suggests that the average bubble speed is
substantially larger whenδ is reduced from 0.5 mm to 0.2 mm. This is more pronounced at high heat flux. For example, at q =40 kw/m2, Tsat =15℃ and G =600 kg/m2s, the average bubble speed forδ= 0.2 mm is about 29% higher than that forδ= 0.2 mm (Figure 4.35). Finally, the results in Figure 4.36 indicate that the average bubble speed is significantly higher for a higher saturated temperature especially at a high heat flux. For instance, at q =40 kw/m2, G =500 kg/m2s and δ= 0.5 mm the average bubble speed for Tsat =10℃ is about 25% lower than that for Tsat =15℃ (Figure 4.36(a)).