| dc.description.abstract |
The increased demand for high computing power coupled with advances in microprocessor and electronic technology has led to increased amount of power delivery demand and heat dissipation requirements in ever-shrinking volumes. In many cases, the resulting temperature rise excludes operating at maximum performance level to prevent failure due to overheating whereas power delivery to these components encounters significant losses. Redox flow battery (RFB) is a technology that addresses both the power delivery and the heat removal issues simultaneously, by using a redox couple dissolved coolant electrolyte. These batteries are essentially electrochemical cells, each of which use two electrolytes (anolyte and catholyte) to flow through different compartment (anode and cathode) separated by a semi-permeable membrane. RFBs were originally developed, and are being extensively used, for the large-scale energy storage application. Traditionally, therefore, RFBs were studied within the context of large scale grid energy storage with a view to optimizing designs and designing control strategies. However, due to their nature, the RFBs optimized for large scale grid energy storage are not automatically optimized for miniaturized versions that can be used for electronic power delivery and heat dissipation. Specifically, the trade-off between power density and pumping power required to move the electrolytes through the RFBs need to be better understood. For these applications, the batteries are mounted on a side of the computer chip, with components such as flow channels, manifolds, supply tubes, electrodes, membranes and current collectors. Since experimentation with such a micro-scale component is expensive and time-consuming, there is need to develop a computational model to study key cell parameters of miniaturized RFB, and to limit the experimental variables. In this study, anolyte and catholyte flows are modeled using Stokes-Brinkman equations of mass and momentum conservation for free channel and porous media. Electrochemical reactions are modeled using Butler-Volmer reaction kinetics, whereas species mass transfer is modeled using species dependent diffusion coefficients. The decoupled energy equation is solved to estimate the temperatures in the domain. From this study, it is observed that electrolyte flow rate has a significant effect on cell voltage and heat transfer in the field. The trade-off between flow rate and pressure drop also has been analyzed to get high power output from the cell. Effect of different operating conditions like state-of-charge of electrolyte, operating current density are also investigated and we found that these parameters affect the battery performance significantly. The microchannel size and shape also influences the pressure drop and thereby the parasitic pumping power and heat transfer. Effect of electrode thickness on the ohmic overpotential and thus the performance of the battery is also studied. Based on the above findings, we propose optimum dimensions of the channel, flow rate and electrode thickness to get maximum output power from the battery. Flow through, flow by and interdigitated designs for fluid flow are studied in order to compare their performance under similar operating conditions. We expect these simulations will guide experimental studies and inform designers of RFBs for microprocessor and printer circuit board cooling solutions. |
en_US |