Resource constrained heat sink optimization.

摘要

There exist several different paradigms to optimize the thermal management of electronic components. The present effort seeks to develop methodologies that enable least-material optimization, and design for sustainability, manufacturability, and thermo-economics, and provide a framework for their integration. The requisite microprocessor heat dissipation has been most often achieved via a heat sink (fin structure), with or without a fan. The design of such heat sinks is the focus of this thesis.; Existing single fin mass minimization techniques pioneered by A. D. Kraus, were extended to design optimal multiple-fin least-material arrays, for forced and natural convection air-cooling. A thermal Total Coefficient-of-Performance (COP_T), relating the cooling capability of the heat sink to the energy invested in both its creation and operation, was used to determine the degree of sustainability of specific heat sink designs. Least-energy optimization was carried out to maximize the COP_T, and this approach was combined with the classical Entropy Generation Minimization method.; Information gathered from industry was used identify manufacturing process limitations, and was combined with theoretical relations to construct work and cost models. The data and the models were subsequently utilized, to maximize the thermal performance capability of each process, and evaluate some of them with respect to sustainability and thermo-economics. The forced convection computations were performed using a well-validated analytical model from literature, while analytical models were constructed to characterize the thermal performance of the natural convection plate and pin fin arrays.; The least-material optimization was successfully applied as a heat sink design heuristic. An expression was derived for the optimal lateral spacing of isothermal staggered cylindrical pin arrays in natural convection. The least-energy optimization revealed the existence of an optimal work allocation ratio, for forced convection, that represented the most beneficial distribution of a fixed amount of input energy between pumping power and heat sink creation, over a given product life cycle. For the design space explored, both the bonded and folded fin processes were found to be best for the COP_T metric. On considering a broader set of metrics, the folded fin designs were seen to the best process.