| This thesis is a study of design and optimization of thermal systems and components, typically used in the thermal management of electronic systems. Our focus herein is the development and verification of procedures, methodologies, and design approaches that facilitate the computer driven or automated design of critical thermal management components. Typical case studies or problem domains will involve state of the art electronic device heat spreaders, heat sink bases, and finned heat sinks.; Initially, existing design and optimization, methods will be surveyed for their applicability and practicality to our chosen problem domains. In particular the issue of how well classical optimal designs (for relatively simple, known solutions) can be recreated or duplicated with common or existing computer driven methodologies will be determined. Optimal electronic heat sink designs for contemporary “challenge problems”, representing the state of the art in the design of electronic heat sink components will be subsequently solved. Finally, novel techniques for the computationally efficient design and optimization of thermal components will be derived, formulated, and verified with additional case studies also representing the leading edge of contemporary thermal design technologies.; The intellectual challenges (and solutions) of this study include the conceptual formulation and problem statement, mathematical derivation, and detailed software implementation of novel topological optimization techniques, representing fundamentally new approaches to the design and optimization of thermal components. Additionally, sophisticated non-linear function approximation techniques are borrowed from the statistics community and effectively deployed in the development, and implementation of generalization based optimization; leading to the computation of entire families of optimal designs. Finally, advanced (adaptive) parametric modeling techniques are derived and implemented for the optimal solution of a contemporary heat sink design challenge problem. |
