Automated approach for optimization of the aluminum extrusion process

The technical achievement described in this paper is to optimize all the variables in the extrusion of complex aluminum profiles using finite element (FE) simulation, statistical regression analysis, and analytical equations in a software platform to account for these variables and their interactions with each other as a group to achieve an ideal extrusion practice. This approach is done by repetitive computer simulations based on varying die modifications conducted through a CAD model of the die and then applying incremental changes to extrusion parameters (several die variables, extrusion speed, temperature, etc.) via an FE code and regression analysis in an optimization program that takes account of thousands of "tweaks" to the die configuration and extrusion process parameters. Computation time varies depending on extrusion complexity but is reasonable in terms of industrial practice. As has been determined in this paper, this optimization procedure was applied to the extrusion of two sample extrusions, a hollow tube and a complex rectangular hollow with screw bosses at its corners. In each case, die pressure was reduced and weld quality and the level of stress in the die were improved, while maintaining or increasing exit speed.