BRANCHING GENERATIVE DESIGN TO CODIFY TOPOLOGY-OPTIMIZED HEAT CONDUCTION STRUCTURES

Topology optimization (TO) allows the generation of high-performance thermal paths in which solid material dissipates heat towards a predefined area, usually under predefined volume constraints. Depending on the boundary conditions, constraints and heat sources, the final geometry can vary in design and performance. The way in which the solid material is distributed across the domain reflects a gradient based optimization process that develops branches in such a way as to increase the heat transfer area by moving away from the thermal noise, e.g., applied heat flux. This distribution can be coded and all the main parameters, i.e., number of branches, position, length, position of intersection point, etc., can be listed to obtain a set of data capable of recreating similar structures that can achieve almost the same results. Thus, the methodology consists of creating an in-house tool that starts with the generated topology-based design and then applies a framework to codify the structures. The goal is to explore possible relationships between the features of these optimized branches and traditional fractal structures. By doing so, this study aims to derive new fractal-based relationships that can replicate the performance of TO designs. Such a connection could bridge the gap between TO-generated geometries and fractal theory, offering a novel approach to designing efficient heat conduction paths that mimic natural fractal patterns.