We consider the general problem of convective cooling of high-power-density electronic equipment. We seek to answer the questions “what is the theoretical maximum thermal performance for an air-cooled compact heat sink and what is the general configuration of this heat sink?” A follow-on question of equal importance addresses the optimization of this heat exchanger simultaneously considering thermal performance and pressure drop. We approach the solution to this problem using the Constructal theory of Bejan. Constructal theory attempts to explain the origin of structure in the nature by optimizing the solutions from the conservation laws governing the problem at hand. With it, a set of design formulas are developed by considering first lower, and then higher, orders of construction. The formulas may then be applied to design a device with improved performance. The objective function of the Constructal optimization is the heat transfer rate from the heat source and the constraints are those associated with geometry, the theory of heat conduction and convection, and pressure drop or its associated cost. Constructal theory is currently applied to a porous-metal-matrix heat sink. We find that optimal distribution of porosity follows a power-law function of distance measured from the base of the heat sink where the exponent in the power law varies from 1 (linear distribution of porosity) to about 2.5. “T”-cell and building-block models for simulating the thermal and fluid-flow performance of a porous-metal-matrix heat sink are being developed and tested to independently verify these results.

Data centers consume nearly 3% of the all electrical power produced worldwide. Most of this power is used for cooling electronic components so they perform reliably. We are using computational models, and laboratory and full-scale experiments in this NSF-funded work (an I/UCRC) to investigate the performance of an organic Rankine cycle (ORC) and LiBr absorption refrigeration (AR) to produce electric power (ORC) or offset the cooling load (AR) for large data centers. The power production from the ORC and/or reduction in cooling load from the AR will improve data center energy efficiency and reduce global energy consumption.

Gravity-driven water networks, which draw water from a source at a high elevation, such as a natural spring or a stream, and deliver it through a branching pipe network to household taps or public tapstands, can reliably provide potable water to developing communities. For a demand-driven design, the central problem is to select proper diameters for the pipes to satisfy desired flow and pressure conditions. Passive flow controls must be used in these networks. For an existing network, where pipe diameters are known, improvements in performance will also rely on passive flow control. In this work, we consider the development and design of smart, inexpensive, passive controllers applied to networks designed for minimum cost, as well as the data needed and instrumentation required to access performance of these networks. Instrumentation includes in-line flow meters and pressure gages to measure fill levels of water tanks.

In order to evaluate the spatial and temporal removal of heat from the parison during blank mold contact, forming simulations are completed using the forming code, Polyflow. These simulations include large surface deformations, viscoelastic behavior, conjugate heat transfer, and the non-trivial contact phenomena at the glass-mold interface. Due to the complexity and size of these models, they are computationally expensive and very time consuming. In addition to this, there is some ambiguity in the applied heat transfer coefficient (or thermal contact resistance) required to complete the simulations. It is desirable to analytically and experimentally quantify this heat transfer coefficient. A one-dimensional finite-difference model of transient glass to mold heat transfer has been developed and validated for this purpose. This model was modified to calculate heat flux from mold surface temperature measurements taken in both a controlled experimental setting, as well as on the Emhart Glass Research Center (EGRC) production machine. This heat flux was in turn, used to predict the temperature distribution in the glass and a corresponding heat transfer coefficient. The purpose of the experimental data collection was to isolate the effects of operating conditions in a controlled environment. This study reports the effects of initial mold temperature, initial glass temperature, and pressing pressure on the resulting heat fluxes and heat transfer coefficients. The data collected in these experiments, in conjunction with the numerical reduction codes have shed light on the importance of different parameters governing heat transfer during the forming process, and is the focus of this thesis. With this knowledge, a simplified first order predictive model for heat flux and heat transfer coefficients was developed. Upon completion of this thesis, the simplified model will be used in order to develop a more sophisticated mechanistic model, which will attempt to capture the physics of the process, as well as eliminate the need to perform a full forming- model in Polyflow.

In aircraft and aerospace applications reducing weight is important to improve the overall performance of the craft. Carbon-fiber composite materials are known to be lightweight and have high in-plane thermal conductivities. These are used in the manufacture of air-to-liquid and radiant-panel heat exchangers which are the focal points for an ongoing study. Key among the analysis and design of these units is the need to take advantage of the high in-plane thermal properties and the minimization of surface contact resistance between mechanically bonded heat exchanger constituents.

Miniature thermoelectric coolers are employed along the periphery of a bipolar plate in a proton exchange membrane fuel cell to cool the adjacent membrane exchange assemblies where most of the waste heat is generated. We have produced and run numerical models for the thermal performance of fuel cells that use this strategy. We find that this strategy maintains the cell stack operating temperature between 45 and 60 C; an acceptable range that precludes the need for any internal liquid cooling or external humidification of feed gases.