Generally, unindirectional solidification experiments of transparent alloys are conducted using thin film samples sandwiched between two glass slides with a small channel height[1]. In such systems, natural convection and melt flow can be assumed negligible and the solification process is diffusive in nature [2,3]; these physical realities allow fundamental simplifications in theoretical/numerical modeling, without losing physical significance. However, natural convection and melt flow do exist in all actual solidification processes and have a significant effect on interface morphology and microstructure formation and development. Recognizing the latter, a great deal of effort went in recent years towards the investigation of the effect of melt flow on interface dynamics and morphology [3–5]. The objective of this paper is to study the natural convection and melt flow near the solid-liquid interface during horizontal unidirectional solidification. In particular, the authors are interested in the melt flow and solid-liquid interface under various channel heights (H) and temperature differences across the hot and cold ends (ΔT) of the samples. A horizontal unindirectional solidification experimental system was constructed. The samples used here are rectangular ampoules made of borosilicate glass that is 3.2 mm thick (on bottom and top sides) and 2.3 mm thick (on the vertical sides). The channel formed in the sample is 75 mm long and 50 mm wide. Three ampoules with channel heights of 1, 3.2 and 5 mm, respectively, are used in these experiments. The ampoules are filled with succinonitrile (99% pure) seeded with polyamide tracer particles (5 μm in diameter, density ρ=1030 kg/m3); the latter are used to resolve and visualize the fluid velocity in the melt. Surface temperatures of the sample on the hot end and cold end are measured with J-type thermocouples. The unidirectional solifification setup is mounted on a microscope stage so that the interface can be observed from the top with the regular microscope. A long distance microscope (LDM) affixed either to a photo- or video- camera is used to observe the vertical shape of the interface, as well as to qualitatively and quantitatively assess the flow. During experiments, the sample is allowed to reach both thermally and flow-wise a steady state situation. The heating and cooling systems are adjusted to make the solid-liquid interface stay at the center of the gap between the heating and cooling chambers for case of observation. The density of polymide particle being close to that of succinonitrile melt allows an almost neutrally buoyant behavior of the tracing particles and thus minimizes the error in flow velocity calculations as well as enhances confidence in the observed qualitative flow patterns. With the help of proprietary computer software, the flow velocity is obtained by evaluating the difference in successive two images of the same particle at time intervals consistent with the sampling speed of video-camera (0.033 sec).

Due to the experimental difficulties brought about by the high pressure and temperature growth conditions flow and heat transport in a hydrothermal autoclave for the growth of single quartz crystals has been studied mostly numerically. To date, most of the numerical models and associated results are not validated, or only qualitatively validated through results derived from crystal growth production. In this study, the authors used a simulated model reactor represented by an enclosure with the two lower half sidewalls uniformly heated while the upper halves are uniformly cooled. Flow in the reactor is qualitatively visualized using a full field flow lighting and seeding technique and quantitatively evaluated using particle image velocimetry. Finally, based on the physical setup and experimentally determined boundary conditions, flow is numerically simulated and compared to the experimental results. The agreement between the experimental and the numerical data was used to validate the numerical model. The ensuing parametric study shows the changing of flow pattern and velocity magnitudes for two differential temperature cases: (ΔT = 10°C and ΔT = 1°C) and a variety of enclosure aspect ratios.

Bayonet tubes, simple refluent heat exchangers, are widely used to heat or cool a media when the heating/cooling agent is readily accessible from one side only. Many studies have been conducted to evaluate the heat transfer performance of bayonet tubes. The majority of these studies focus on the heat transfer in the annular section and little on the end surface. This paper presents a numerical simulation of the laminar flow and heat transfer in a bayonet tube. The simulation is first validated by the experimental data in the literature. The flow and heat transfer in bayonet tubes are then investigated with both flat and curved end surfaces. Both local and average Nusselt number on the end surfaces are calculated under various Re and geometry conditions. Effect of the end surface curvature is studied by comparing the performances of the flat and curved ended bayonet tubes.

The nature and patterns of solution flow in hydrothermal autoclaves are critical to the quality, growth uniformity, and growth rates of synthetic single crystals. Small horizontal temperature deviations, which exist in industrial practice, were found to be critical in establishing flow patterns. However, the mechanism that determines how temperature deviations affect flow pattern is not well understood. In this study, an experimental system is set-up to study the flow in a model reactor (an enclosure). Temperature in the enclosure is visualized using liquid crystals. With the experimental results, a numerical model is validated and then used to simulate flows in enclosures that are subjected to similar thermal condition as industrial autoclaves. Flow patterns are obtained with various temperature deviations, for various aspect ratios and various Rayleigh (Ra) number between 4.05E8 to 3.24E9. Flows studied are unsteady in nature. Without temperature deviations, the overall flow pattern is anti-symmetric. With a temperature deviation, the wall layers are un-balanced. The impingement of streams on the wall layers does not affect the wall layer flow at low Ra numbers. At high Ra number, wall layers are broken by the impinging streams. The dominant heat transfer mechanism in the enclosure changes significantly as the aspect ratio of the enclosure changes. In enclosures of high aspect ratios that heat transfer resistance is mainly at the fluid exchange between the two halves, temperature deviations significantly affect heat transfer by stabilizing the direction of the streams.

No abstract