Abstract
Thermal Dissipation is a critical issue in the performance of semiconductor devices. The current practice is to use forced convection by air over a heat sink which is bonded to the microelectronic device. With increased packing density of the circuits inside a chip, large amounts of heat is generated and air cooling is no longer sufficient. Forced liquid convection using microchannels is considered to be a viable option for cooling of these microprocessor chips. This work deals with the evaluation of the single phase pressure drop in microchannels. There are two types of microchannels under consideration. Plain microchannels which have basically long uninterrupted flow channels while the enhanced channels which have the interrupted flow lengths. Enhanced microchannels, because of their offset strip fin geometry, significantly increase both the heat transfer as well as the pressure drop. This work deals with the evaluation of single phase flow pressure drop in both plain and enhanced microchannels. For plain microchannels there have been a few investigations in the literature which suggest that the microchannel performance can generally be predicted using the classical fluid flow equations. However there are some experiments that still show departure from the classical theory that cannot be explained. It is proposed in this work that the reason for this discrepancy can be traced to the effects due to flow maldistribution in plain microchannels. A systematic experimental investigation is performed to study the effects of slight variations in channel dimensions and their influence on the flow maldistribution in an attempt to validate the applicability of classical theory to microchannel flows. Enhanced microchannels however have not been investigated thoroughly. There is very few data available in the literature. FLUENT is a CFD software which can be used as a tool to design and optimize these enhanced channels. However it has to be first validated with experiments. Thus pressure drop experiments are carried out on an offset strip fin silicon microchannel and the data is predicted using FLUENT, which is CFD software. Also existing predictive models for friction factor in offset strip fin minichannels are tested to check their validity for microchannel flows. For plain microchannels, it seen that with uniform flow assumption, the friction factor is either underpredicted or overpredicted using the theory depending upon the reference channel dimension. However by accounting flow maldistribution in plain microchannels, friction factor can be accurately determined using theoretical equations. For enhanced microchannels it is observed that FLUENT can predict the pressure drop within 10%. In this work only adiabatic flows are considered. It is recommended that this work should be extended to flows with heat transfer.
Library of Congress Subject Headings
Fluid-structure interaction--Computer simulation; Fluid dynamics--Computer simulation; Integrated circuits--Cooling; Heat-transfer media; Heat--Transmission
Publication Date
11-1-2007
Document Type
Thesis
Department, Program, or Center
Mechanical Engineering (KGCOE)
Advisor
Kozak, Jeffrey
Advisor/Committee Member
Venkataraman, P.
Recommended Citation
Bapat, Akhilesh V., "Experimental and numerical evaluation of single phase adiabatic flows in plain and enhanced microchannels" (2007). Thesis. Rochester Institute of Technology. Accessed from
https://repository.rit.edu/theses/5938
Campus
RIT – Main Campus
Comments
Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TA357.5.F58 B37 2007