Shape Optimization in Vehicle Design with Experimental Validation 982842

For cast iron automotive parts, the automated shape optimization technique is a powerful design tool that usually enhances performance and reduces overall cost. Possessing the solid finite element model of what could be considered a starting point, first trial or initial design, commercially available structural synthesis software are able to optimally relocate the nodes, thus creating an optimal final shape. This is very suitable for parts obtained by metal melting for two main reasons: first, such components generally tend to be massive (posing the challenge to reduce weight) and second, the manufacturing method itself presents wide freedom regarding the shapes possible to be obtained (augmenting the chances the computationally generated shape is feasible in practice).

This paper presents a case study in which a front suspension support of a heavy truck has its shape optimized resulting in significant weight reduction besides considering strict stress constraints related to the vehicle's safety. The procedure was conducted using VMA's GENESIS version 3.0, a general purpose structural optimization software that relies on the finite element method to perform the needed calculations (that, on their turn, are usually simplified by a module of the software responsible for approximating the finite element model through a proper set of techniques). The algorythm for shape optimization is the DOMAIN method, and its basic characteristics will be briefly discussed and compared with alternative approaches.

Finally, the overall procedure is verified by means of a photoelastic experimental stress evaluation, which certifies the validity of the computational optimization effort. It is shown that the region displaying the highest stress concentration factors remains the same, and the factor values are kept in a range that does not compromise the strength of the part. These considerations meet the predictions of the finite element analysis performed over the model of the final design presented by the optimization software.