The purpose of the CFD FLUENT analysis was to determine the enclosure geometry most conducive to natural convection heat and mass transfer, given the limitations of a distillation enclosure. As the available geometric shapes are somewhat infinite, the processing time to examine each and every possibility was unrealistic. Rather than trying various geometric shapes randomly, a more logical approach was taken. The process is summed up as follows —

1. The basic geometric shapes that could conceivably meet the limitations and demands of a distillation enclosure were drawn and meshed. These included a rectangle, oval, circular cylinder, and trapezoid.

2. Each of these basic geometries was assigned reasonable boundary conditions representing the realistic operating conditions of a low temperature solar distillation enclosure.

3. The meshed models of these basic geometries were then solved for their steady state behavior in 2D.

4. The heat transfer between the hot-sink and cold-sink, due to natural convection, was determined in each case and compared with the heat transfer in the other geometries.

5. Various parameters such as aspect ratio, tilt angle, fillets, and other minute geometric considerations were varied in order to find the maximum amount of heat transfer possible for each basic geometry.

6. The basic geometries were then compared with respect to their maximum heat transfer to determine the optimum basic geometry.

7. Once the optimum geometry type was determined it was remodeled in 3D and its final volumetric geometry was determined.

The above iterative optimization process resulted in the geometry shown in Figure 2. It can be best described as an “ovalized rectangle” as it benefits from the "best” of both of these two geometries. A rectangle in general had high values of heat transfer when compared to the other basic geometries, this agrees well with other studies done [4]. In the case of a simple rectangle however, with a horizontal top and bottom, the heat transfer rate was slightly reduced due to the severe deflection of flow away from the vertical walls. The dome shaped top cover allows the convective current to move more smoothly over the curved surface, which Fl9ure 2 — Ovalized rectangle enhances the heat transfer from the surface. These results are in agreement with a similar study done which also examined the natural convection in dome shaped enclosures [5].

Further examining and modifying led two additional improvements. Inserting a partition between the two heat transfer surfaces yields further improvement of ~ 5-10% in heat transfer. This is due to forcing the air to enter the hot-sink from its coldest point and not allowing it to "cut across” the center. The same is true for the hot air on top. This way large temperature gradients for driving heat transfer are kept. An additional benefit of the partition is that it prevents ‘cross mixing’ between the saline solution and the condensing (and dripping) distillate. The tilt angle at which the enclosure is at with respect to gravity also has an influence on the natural convection. Certain tilt angles were shown to improve the overall heat transfer due to natural convection.

The partitioned ovalized rectangle was determined to be the optimal geometry for a distillation enclosure. It was remodeled in 3D (Figure 3) and solved iteratively, so as to determine the exact volumetric parameters that would provide the best heat transfer by natural convection through the enclosure.

In this examination three factors were tested — separation distance between heat transfer surfaces, enclosure tilt angle (0), and aspect ratio (H/L). The following graphs (Figure 4 — 6) illustrate the results.

Figure 5 — Effect of tilt angle Figure 6 — Effect of aspect on heat flux ratio on heat flux

The results described previously show that by optimizing the geometry in 3D we can improve the heat transfer by natural convection. As for the separation distance and aspect ratio there is a significant influence on heat transfer while the influence of the tilt angle seems to be less. When applying the results we can propose an enclosure geometry that is optimal for heat and mass transfer, as shown in Figure 7.

3.40Є+02

3.24e+02

Figure 7 — Contours of static temperature for the proposed distillation enclosure

Experimental Device

Initially, the enclosure with optimized geometry was constructed and insulated. To prevent corrosion the material primarily used in the enclosure was stainless steel with the few features that demanded more flexibility made from plexi-glass. Once the enclosure "shell” was complete the evaporator and condenser sections were assembled into it.

The evaporator is made of 15 individual evaporation sheets, each fed hot brine from thin, holed, feed pipes each connected to the incoming main pipe. The best material for the evaporation sheets, after experimentation, was found to be a tight-weave cotton cloth. Each individual cloth was "tunneled” at the top to improve the uniform spread of the flow as it leaves the feed pipes onto the cloths. The bottoms of the cloths are V-shaped so as to centralize the down flowing brine into one "drip line” at the bottom (refer to Figure 8).

The condenser, on the opposing side, is a tube bank made up of seven, parallel, 9 mm copper tube sections. Each copper tubing section is fed from the seawater main entering at the bottom of the enclosure and feeds into the central pipe supplying the solar collectors at the top of the enclosure. The enclosure volume is divided into half by a partition made of plexi-glass which prevents the humid air flows from prematurely mixing and which also keeps the distillate from being lost or contaminated. The bottom of the enclosure is divided into two by a dividing wall that completely separates the two pools of fluid — brine and distillate — with each pool draining out of its respective pipe. Figure 8 shows the open evaporator / condenser.

Bottom of evaporation cloths

Results

The salt content of the brine feed water (Ms) was made to be approximately 23,000 ppm by adding table salt. The salt content of the distillate (Md) produced by the device was found to be ~ 16.7 ppm. These results were measured on the feed and distillate samples by titration, and were corroborated by an electric conductance measurement.

The device was tested while operating at various feed water flow rates, enclosure tilt angles, and temperatures. Following each experiment the optimal efficiency (GORopt — equation (3)) and the ideal efficiency (GORi — equation (1)) were calculated to serve as indicators to the device’s actual performance. Results for the actual efficiency (GORa) and
the ideal efficiency (GOR), as a function of the optimal (GORopt) (both calculated at the same operating temperatures), are shown in Figure 9.

Stated differently, Figure 9 represents a comparison between three efficiencies for each

Figure 9 — Actual efficiencies compared to the optimal and ideal calculated efficiencies

experiment. The uppermost line, GOR, is higher than the other two as it corresponds to an efficiency representing the ideal situation — an ideal heat engine and heat pump. The lower line, GORopt represents the efficiency that is theoretically possible in an optimal but realistic regeneration device, and the scattered data GORa represents the actual measured efficiencies. A definite correlation can be seen between the actual measured GORa in the laboratory device and the GOR attainable in an optimal device operating under the same temperature limitations. The device’s efficiency was also measured as a function of the tilt angles starting at 0° degrees through 25° degrees (at 5° degree intervals). This test sequence was conducted twice, the results are shown in Figure 10.

Figure 10 — Effect of enclosure tilt angle on overall efficiency

Discussion

The dissolved salt content in the distillate of 16.7 ppm shows the device’s ability to fulfill its basic requirement — desalinate water. When considering the actual thermal efficiency of the laboratory device one can see that it corresponds relatively well with the optimal efficiencies possible at the same temperature values (see Figure 9). In many cases however, the actual efficiency exceeds that of the optimal efficiency — an "impossibility”. By examining this contradiction statistically we see that more than likely the inherent error in thermocouple measurement is responsible for the "overshooting” of GORa. However, even when taking the worst possible error in measurement and subtracting it from the least-squares average, the data representing GORa still corresponds well to the optimum curve (GOR0pt). On one hand the experimental data does seem to have an inherent amount of error due to measurements, but on the other hand even when taking this error into account the device is still operating at very near the best possible efficiency.

From figure 10 it is clear that there does seem to be some influence of the tilt angle on the overall device efficiency. As the errors in measurement are on the order of the experienced improvement, however, it is difficult to state a definite conclusion as to the exact contribution of the tilt angle. The tilt angle that seems to be most conducive to natural convection is somewhere in the vicinity of 15° degrees which is less than predicted by Fluent (15° vs. 35°). In any case, further investigation is necessary to determine the precise effect the enclosure tilt angle has on the overall efficiency.

The lag time the device takes to reach steady state conditions (approx. 8 hours) is a detriment. In order for the device to operate at its optimum the fluctuations in temperature need to be prevented. In addition to this, as the device is completely insulated from the environment, it is not dependant on outside temperatures and may take advantage of 24- hour a day operation. For these two reasons it seems that the only way to effectively operate an efficient regenerative solar distillation device is to have it supplied hot water from an intermediate (storage) reservoir.

Conclusion

This research project, in contrast to other research [6] investigating low temperature distillation devices employing regeneration, incorporated a CFD analysis that formed the base for the optimization process. By using such an analysis the optimum geometry for the distillation enclosure was determined. Along with the basic geometry, other features such as aspect ratio, tilt angle, and partitions that also contribute to the overall thermal efficiency, could be tested for their contribution and incorporated into the design. As a CFD analysis was incorporated, not only could the distillation enclosure be optimized, but the behavior of the natural convection flow field could be predicted with relative ease.

The product of the CFD analysis was an optimized low temperature distillation device. The device was built, tested experimentally to verify its functionality, and finally tested for its thermal efficiency. The following points conclude the performance of the device.

— The proposed optimized design is functional and is capable of producing distilled water of high quality (dissolved salt content of distillate ~ 17 ppm).

— The actual operating efficiency of the proposed design compares well with the efficiency of an optimal device, thus the design may be considered optimized.

— The device built is durable and simple to operate — two additional benefits that add to the overall effectiveness of the device.

This research has shown that by incorporating a CFD analysis the originally complex natural convection flow found in distillation enclosures could be investigated. This way a low temperature distillation device incorporating regeneration was optimized and designed.