Characterization of the Liquid Film in Slug- and Annular-Regime Microchannel Flows

Ravi S Patel, Purdue University

Abstract

Improved methodologies for modeling of microscale phase-change phenomena are necessary to provide insights that can enable the design of technologies that operate at such small length-scales. In particular, multiphase microchannel heat sinks are an attractive option for the thermal management of excess waste heat generated in electronics due to their compact size, effective handling of high maximum heat flux levels, and ability to minimize temperature gradients across the heat rejection surface. However, current mechanistic models rely extensively on simplifying assumptions and empirical correction factors due to a lack of knowledge regarding liquid-gas interface shapes in two-phase microchannel flows. This gap in the literature exists because current experimental techniques to characterize this phase boundary are unusable or poorly resolved when applied to microchannel flows, due to the order of magnitude differences in length scale between the small size of objects and high speeds at which they are moving. In order to characterize these processes and develop improved mechanistic performance models, new metrological capabilities must be developed specifically for the investigation of phase-change phenomena at the microscale. The slug and annular flow regimes play a critical role in determining the performance of a microchannel heat sink as they are the predominant regimes that occur in small channels due to the strong effects of surface tension in governing flow morphology. Mechanistic models for the annular flow regime propose that thin-film evaporation is the dominant heat transfer mechanism; however, the film geometry is typically greatly simplified. In this work, a novel characterization technique is developed for the detection and three-dimensional reconstruction of the liquid-gas interface in annular microchannel flows. The flow field is visualized through the use of an optical microscope, relying on seeding particles and the thin focal plane of the objective lens to isolate and investigate discrete slices at varying depths within the channel. The liquid-gas interface is detected within the focal plane by identifying particle locations along the interface. By scanning the focal plane location across the depth of the channel it is then possible to obtain a three-dimensional time-averaged reconstruction of the complete annulus. The technique is deployed in a parametric characterization of adiabatic annular-regime microchannel flows. Next, an improved mechanistic model is developed that predicts the liquid-gas interface shape with greater fidelity than prior models by taking into account interface curvature and surface tension effects. Interface shape predictions generated using this model are compared to the experimental measurements. The model is also used to predict experimental heat transfer data obtained from the literature and is shown to perform with an accuracy that is comparable to previous predictive models, but without relying on any empirical corrections to improve model accuracy; this drastically expands the general applicability of the present model compared to prior approaches. The slug flow regime is characterized by a series of elongated gas bubbles surrounded by a thin liquid film, spaced apart in the streamwise direction by a liquid plug. Extant mechanistic models for this flow regime similarly rely on inexact approximations of the liquid-gas interface structures in the two-phase regions of the flow. These approximations are based on previous film measurements taken in unrelated environments and empirically corrected to yield a best match between experimental heat transfer coefficient measurements and model predictions. To experimentally investigate these structures, the previously developed interface characterization technique is adapted for implementation in the slug flow environment. Key challenges specific to this regime are addressed, such as the increased difficulty of performing visualizations and resolving the streamwise-varying interface structures. Frequency-domain based image filtering techniques are applied to the images obtained in order to address issues related to increased background noise generated by greater reflection and refraction from solid-liquid and liquid-gas interfaces present in this flow environment. Additionally, advanced active contour segmentation approaches are implemented for the interface detection within visualized frames in order to tractably visualize multiple slug bubbles to arrive at a single time-average reconstruction. The characterization technique is then used in the parametric investigation of the liquid film in slug-regime adiabatic microchannel flows, and an empirical model for the corner liquid film thickness is developed specifically for a microchannel flow environment. New methods for the characterization of dynamic film structures and improved film models are developed in the present work. These resources will be suitable for use by the practitioner in the design and implementation of microscale technologies that rely on phase-change.

Degree

Ph.D.

Advisors

Garimella, Purdue University.

Subject Area

Engineering|Mechanical engineering

COinS