TitleSingle Phase and Flow Boiling Heat Transfer and Flow Characterization in Microscale Pin Fin Heat Sinks
Publication TypeThesis
Year of Publication2015
AuthorsRasouli, E
UniversityOregon State University
CityCorvallis, Or.
Thesis TypeDissertation

The ever increasing requirements for heat dissipation in various thermal management applications such as computer chip cooling and high power electronics have necessitated the need for novel thermal management techniques. Thermal management using heat sinks with microscale features is amongst the prominent techniques developed over the past two decades. In this dissertation, single and phase change heat transfer and pressure drop through one such heat sink, namely microscale pin fin heat sinks (µPFHS), is examined experimentally. In particular, effects of pitch-to-diameter and aspect ratio variations are studied on the thermofluidic performance of studied µPFHSs. Single phase heat transfer and pressure drop of two distinct fluids, liquid nitrogen and Performance Fluid (PF5060) are characterized experimentally through the µPFHSs with staggered diamond shape pin fins. The LN₂ and PF5060 experiments' Reynolds number (Re_Dh, based on pin fin hydraulic diameter) is in range of 108-570 and 8-462, respectively. Results are presented in a non-dimensional form in terms of the friction factor (f), Nusselt (Nu), and Reynolds numbers and are compared with the predictions of existing correlations in the literature for micro pin fin heat sinks. Heat sinks with the higher pitch ratio (coarser array) not only show lower pressure drops at a fixed Re_Dh, but also enhance significantly heat transfer rate when compared against the heat sink of the same pin fin size but denser arrangement. Flow visualization experiments using an infrared camera on PF5060 single phase tests are performed to understand the counter-intuitive trends seen in the global results. Flow through heat sinks with the same aspect ratio but larger pitch ratio exhibit unsteady vortex shedding in the wake region of pin fins, which markedly enhances convective heat transfer rate. Existing correlations developed for µPFHSs (such as that by Prasher et al. [1] and Koşar and Peles [2]) are capable of predicting the f and Nu data with good agreement only in the absence of vortex shedding, while the unsteady flow past the transition Re_Dh results in poor comparison of correlations with experimental data. A comparison of the experimental Nu data of PF5060 (Pr≅12.2) with the data of LN₂ (Pr≅1.9) shows significant change between the slopes of the curves of two fluids only in the heat sinks without vortex shedding. In the heat sinks with unsteady vortex shedding, the Nu_Dh curves show significantly decreased dependency on Pr number. Consequently, separate correlations are developed for predicting Nu in the case with and without unsteady vortex shedding using data from two distinct fluids and four PFHS geometries over a range of Re_Dh from 8 to 643. Given the clear heat transfer enhancement that occurs for certain pitch ratio designs of PFHSs in single phase flows, flow boiling experiments with PF5060 are performed to clarify whether additional changes to the pressure drop and two-phase heat transfer coefficient occur upon the introduction of the unsteady vortex shedding. Subcooled (∆T_sub=12.5℃) and saturated flow boiling of PF5060 through the micro pin fins are investigated. The heat sinks are tested at three constant mass fluxes of 30, 60, and 100 kg/m².s with heat fluxes ranging from 1.1 to 17.8 W⁄(cm² ) based on the planform area of the heat sinks. Flow regimes are studied with high speed imaging. Nucleate boiling heat transfer is the dominant mechanism for exit vapor qualities less than 0.5; at higher qualities annular film evaporation becomes dominant. The salient effect of unsteady vortex shedding is in elimination of wall temperature overshoot. In nucleate boiling regime, the heat sinks with unsteady flow flapping show higher two-phase heat transfer coefficients. The predictions of existing correlations for h_tp in literature are not in good agreement with the experimental data (MAE>30%) and show a systematic deviation depending on the µPFHSs dimensions.