Heat transfer and flow characteristics have been determined for a single-phase rectangular loop thermosiphon. The plane of the loop was vertical, and tests were performed with in-plane tilt angles ranging from 3.6° CW to 4.2° CCW. Velocity profiles were measured in one vertical leg of the loop using both a single-component Laser Doppler Velocimeter (LDV), and a commercial Particle Image Velocimeter (PIV) system. The LDV data and PIV data were found to be in good agreement. The measured average velocities were approximately 2–2.5 cm/s at an average heating rate of 70 W, and were independent of tilt angle. Significant RMS fluctuations of 10–20% of the mean velocity were observed in the test section, in spite of the laminar or transitional Reynolds numbers (order of 700, based on the hydraulic diameter). These fluctuations have been attributed to vortex shedding from the upstream temperature probes and mitre bends, rather than to fully developed turbulence. Animations of the PIV data clearly show these large scale unsteady flow patterns. Multiple steady state flow patterns were not observed.

A model single phase loop thermosiphon was designed and built such that velocity profiles and heat transfer rates could be studied by obtaining temperature data, along with a detailed set of non-intrusive Laser Doppler Velocimetry measurements. The quantities measured were the instantaneous velocity profiles for the water flow. These velocities were then reduced into sets of mean and root mean square velocities, for different angles of tilt of the experimental setup. Nominal heating and cooling rates were held constant. The single phase thermosiphon consisted of an aluminum flow loop, fitted with a test section made of Plexiglas, that was rectangular in cross section. The entire pipe loop was filled with distilled water for this set of experiments. Flow was generated by thermoelectric coolers fixed on both sides of the top leg and thermoelectric heaters fixed on both sides of the bottom leg of the thermosiphon. The flow direction was reversed by changing the angle of tilt of the experimental setup. The data were taken across the 1.5 inch and 0.75 inch dimensions of the test section at both the center and at the right end of the wall as tilt angle was varied from 3.6 degrees clockwise to 4.2 degrees counter clockwise. The predicted velocities from a one dimensional flow model were 3-4 times bigger than the observed mean velocities. Average heating rate was 70 Watts, while measured average cooling rate was 27 Watts. RMS velocities were typically 10-20% of the mean velocities. It was found that the mean velocity profiles across the 1.5 inch dimension of the flow area were consistently higher near the back wall and lower near the front wall of the test section.