Embedded Files
My graduate thesis (download full document) was based on a feasibility study of a means to passively effect liquid motion parallel to a heated surface through surface geometric modifications. Such a passive system is beneficial for electronics cooling applications as it reduces the pumping equipment normally required in flow loops and is desired for space applications, where launch costs greatly restrict the weight of onboard systems. Heat pipes offer similar benefits, but because they rely on the vapor phase, do not have the heat transport capability of liquid cooling systems
The surface geometry considered was a repeated array of asymmetric silicon ratchets with re-entrant cavities located on the shallow face (shown in Fig 1). Due to the sharp angles, bubbles preferentially nucleate in the cavities, then expand and depart the surface in a direction normal to the face with the cavity. Bubbles drag with them a portion of the surrounding fluid, the horizontal component of which can be utilized for fluid circulation purposes in microgravity as it is not driven by gravity-induced buoyancy forces. A serpentine thin film heater provided heat to the surface.
Fig. 1 The concept as proposed to induce fluid advection in microgravity.
I designed and built a complete experimental facility in which to perform experiments as part of this work to comply with requirements set by microgravity flight services in anticipation of future experiments in microgravity. Figures 2-5 show how the ratcheted surface is mounted in the experimental facility. The data acquired is fluid temperature (using thermocouples), applied heat to the test section (measuring current and resistance), and high speed video frames (using a Phantom V310 High Speed Camera), backlit by a custom array of 625 white LEDs.
Figs 2, 3, 4, 5. From left to right: test section showing ratcheted surface and thin film heaters; test chamber showing mounted test section, thermocouple probes, LED array, condensing coils and cartridge heaters; experimental facility showing chamber mounted in aluminum frame, high speed camera, and data acquisition and control equipment.
Experiments were performed using deionized and degassed water at atmospheric pressure at subcoolings of 5°C and 20°C (degrees below boiling temperature). Applied area averaged heat flux was varied between 2 W/cm² and 19 W/cm² at these conditions. Magnified high-speed videos were used to resolve bubble behavior near the surface. A preferential bubble growth and departure direction was confirmed for both subcoolings and lateral fluid motion was confirmed in the high subcooling condition. Repeatability was confirmed with separate experiment performed 58 days later. Tracking of bubbles was accomplished using a custom bubble-tracking algorithm, designed to resolve only bubbles within a two-dimensional plane normal to the viewing direction. Below is a sample of the data showing both the high-speed video frames and the resolved bubble velocities. Instantaneous velocities of individual bubbles parallel to the surface were shown to be in excess of 600 mm/s immediately following departure (see Section 7.2.2 of thesis), and liquid flows with mean velocities between 25 mm/s and 35 mm/s parallel to surface were observed in the plume farther from the surface (see raw and resolved data below).
Fig. 5 Sample of raw and process data showing plume velocities
A simplified semi-empirical model of bubble growth phase is proposed to explain the observed mean liquid velocities (see Section 7.4 of thesis). In summary, the expansion of the bubble normal to the face of the ratchet pushes the fluid ahead of itself. This momentum is imparted on the fluid over an area equal to the average area occupied by each cavity on the face of the ratchet each time a bubble nucleates and grows on the surface. Along with the frequency of bubble departure, the total momentum per unit of time transferred to the fluid can be estimated.
Figs. 6 & 7 Semi-empirical model hypothesized to describe the horizontal component of the plume velocity
The semi-empirical model closely predicts the horizontal component of the plume flow velocities, and suggest this may in fact be the driving factor. Since the concept was proven to work in practice, research on this subject has continued at Oregon State University.
Google Sites
Report abuse