Résumé | Cooling channels are crucial elements in many casting and molding processes as the material cooling stage, i.e., the phase during which the molten or thermally softened material solidifies, is critical for part quality and is often a bottleneck in cycle time reduction. With the advent of additive manufacturing (AM) techniques, the conventional, straight-drilled, channels are increasingly replaced by so-called conformal cooling channels, as the latter can significantly decrease the cooling phase time and temperature gradients, thereby reducing part defects. Nevertheless, the design of conformal cooling channels remains challenging, especially as the geometrical complexity and the number of manufacturing constraints increase. Thankfully, this task can be conveniently handled using a dedicated topology optimization algorithm. In this talk, we present a density-based topology optimization approach for the design of conformal cooling channels for molds and dies, intended to be manufactured using the laser power bed fusion (LPBF) AM technique. In order to keep the surface temperature of the mold cavity uniform during the cooling, the objective function is penalized using the temperature standard deviation of the cavity. This penalization further contributes to the generation of an evenlydistributed network of channels and tends to impede the formation of narrow (large width-to-height ratio) topologies that envelop the cavity. Such topologies are often encountered when seeking to minimize the temperature of a heated surface but, although being efficient for cooling, they are hardly manufacturable, even by AM techniques, and they also make the final design not structurally sound. Furthermore, the manufacturability aspect is enhanced through the use of a projected-perimeter (density-gradient based) constraint, which restricts the emergence of overhanging structures and leads to the generation of channels with teardrop-shaped cross sections. We show that the constraint and penalization control parameters can be tuned in a way that the design cooling efficiency is not significantly reduced while ensuring manufacturability. The computational work is complemented by preliminary results from experimentation in an optimized design fabricated by LPBF. In particular, a comparison of the cooling efficiency is performed between conventional and optimized cooling channel designs. |
---|