Using Multi-Disciplinary Optimization to Accelerate the Design Process
Design optimization techniques have been utilized at Daimler for many years, helping the organization to maximize performance while minimizing material use and mass; finding the ideal balance between multiple attributes. Although design optimization methods have delivered excellent results throughout Daimler’s programs, the traditional processes of optimizing for different disciplines, such as crash and noise, vibration and harshness (NVH), independently can be slow to deliver a design solution that meets varied attribute targets simultaneously.
During the development of a new vehicle variant, Daimler wanted to explore the potential of utilizing a multi-disciplinary approach to optimization (MDO), whereby several attribute performance targets are considered in a single optimization study, a technique that has the promise to deliver rapid design direction to design teams.
A collaborative project was created to investigate this approach, focusing on a Mercedes-Benz die cast rear cross member that was not yet meeting its crash and NVH targets. The objective of the projects was to increase the stiffness of the casting while minimizing its mass.
Altair's consultants removed the rib pattern from the existing rear cross member design and filled the void with tetra elements to define a ‘design space’ where OptiStruct could add or remove material based on loads and constraints. Using NVH loads from Daimler and the draw direction of the ribs as a manufacturing constraint, OptiStruct created a representation of an optimal rib layout which would meet the NVH performance targets.
A further free size optimization study was conducted to identify unnecessary ribs and areas of ribs which could deliver improvements to the components dynamic stiffness performance. A final size optimization was performed to explore whether the ribs could be made thinner to further reduce mass. Now that the topology, free size and size optimization studies were complete, Altair and Daimler had a base design from which a MDO study taking both NVH and explicit crash models into account could be built. A design of experiments (DOE) was created using HyperStudy. The DOE used the varying thickness of the ribs and sections of the outer shell to explore the design space for further design enhancements while understanding correlations and filtering positive design changes and cross checking crash plots.
Three different rear crash barrier heights were considered in the combined crash and NVH optimization study with more than 2400 optimization jobs run during the DOE. A final optimization study was performed that took all the DOE runs into account in order to find the ideal balanced design that met the NVH and crash targets set while minimizing mass.
The MDO approach proved to be a highly successful approach to solve this design challenge. Including crash objectives in the NVH constraints and the NVH objectives in the crash constraints was a big departure from the more traditional development process where attribute teams operate in relative silos. The NVH performance of the new rear cross member design had been improved significantly. For crash performance the intrusion level and the material/connection rupture was drastically reduced.