Applied reanalysis techniques for large scaled structural mechanical optimization problems covering automotive needs

摘要

Simulation of automotive vehicles has recently become an important step in car development. Rapid prototyping, which offers a close to market product, with less experimental testing, requires a fast and accurate computer simulation to guarantee a quality product. Several optimization techniques were developed over the years and are today available in many commercial software packages and play an important role in the design process. With the increasing capabilities for simulation, enabled by evolving algorithms and improved computer hardware, the demand for optimization processes covering all the simulated functions has increased. Most simulations for real world problems, such as static and dynamic behaviour of an automotive vehicle or its acoustic behaviour due to road and engine excitation loads, require detailed and large multi million degrees of freedom (DOF) finite element models to ensure accurate results. Calculation using these models is still a significant challenge. Multidisciplinary optimization is not achievable within a reasonable time using the current analytical methods and computer hardware. The following paper presents an approach for solving one aspect of the problem mentioned above. It is based on the reduction technique proposed by Kirsch and is referred to as the combined approximations (CA).