As part of the Norwegian research project AluPart, an automated process has been developed to optimize an automotive wheel suspension part. The extruded and formed aluminium part is optimized with respect to cost, performance and manufacturability.
The cost value is based on use and reuse of aluminium, performance is measured as durability with respect to cyclic loading and manufacturability puts a constraint on the maximum allowed plastic strain experienced during the forming. 22 parameters were used to describe the crosssection of the extruded prole, the shape of four large cutouts and the shape of the forming tools.
A hybrid optimization strategy, consisting of an initial multiobjective global search (cf. gure 1) followed by a singleobjective local renement, was used. The main optimization phase used a model giving a 10x speed-up compared to the original, while the nal phase used the original model. In the context of product development, the applied multi-level modeling approach was well justied and is recommended for future design work.
modeFRONTIER linked the ABAQUS simulations of the multiple cutting and forming operations together into an automatic process and guided it towards the specied objectives. While the analysis was soundly built, the success rate of simulations was less than 10 percent due to multiple sources such as geometry build failures, elements exceeding the distortion limits and non-convergence in spring-back calculations. Despite the harsh environment, the optimization algorithm found the global optimum in a reliable way.
The optimized design reduced the material cost by 25 percent while still fullling the constraint on the maximum allowed plastic strain. At the same time, its durability increased from 13 000 to 1 700 000 load cycles. |
