Topology Derivative Calculations for Multidisciplinary Structural-Thermal Component Design

Optimal shape design (OSD) is a key element of well-designed multifunctional - multidisciplinary structures. The structural shape should satisfy the design intent and perform in an environment that would include fluid flow and heat transfer. PDM (Product Data Management) and PLM (Product Lifecycle Management) processes would need efficient methodology to be successful for product design and ensure its functionality over its life. The key is to design for functionality as well as ensure product survivability at affordable maintenance cost. The design of shapes to efficiently perform is inclined to be mathematically elegant and computationally challenging as complex multidisciplinary problems are encountered. There is a critical need to have a systematic developed system that will link the physics and mathematics of shape design with the functional intent of the component. In industry optimum design is not a once and for all solution tool because engineering design is made of compromises owning to multidisciplinary aspects of the component and necessity of doing multipoint constrained design. OSD is a branch of differentiable optimization of distributed systems where gradient and Newton methods are natural numerical tools. Existence of solutions and differentiability criteria are the key elements of domain decomposition and shape evolution strategies. As shape evolves oscillations of shapes could lead to non-physical or non-intuitive solutions of the topological optimization problem. OSD is still a numerically challenging because it is computer intensive and in practice one has to make compromises between shapes that are good with respect to more than one criterion. PDM-PLM processes when combined with OSD offers efficient work flow and design information to ensure a systematic design environment within which new products can be designed. This paper focuses on presenting a systematic methodology for design within which product topology can evolve under multidisciplinary loads.