Automotive crash simulations of full vehicle models still constitute a large computational effort which can be a major problem for applications requiring a large number of evaluations with varying parameter configurations. In some applications, highly similar simulations frequently need to be carried out multiple times with only minor local parameter modifications. At the same time, large amounts of numerical simulation data increasingly become available in industrial simulation databases as part of the progressive level of digitalization of automotive development processes. Data-driven modeling methods are an area of active research, aiming to exploit this new “treasure” in order to find interesting patterns and accelerate predictions.
Recent years, CAE plays more important roles in product development than ever. Good CAE models require validation works with various test data. However, the way of comparison needs to be improved in order to meet the market expectations. Usually a set of test data is given to CAE engineers after all specimen are tested, and test machines are cleaned up. In case CAE engineers find out remarkable differences between simulation results and test results are too great due to mistakes in test laboratories, validation process becomes quite difficult in many cases. In this study, quasi static compression of a crash box is used as an example in order to illustrate the proposed process. A preliminary stochastic simulation and a set of tests is conducted, and the deformation of the crash boxes are transformed into the common modal space. This process makes it possible to assess similarity of deformation modes from multiple simulation results and test results at a glance. This analysis can be conducted at test lab, and fundamental difference between test and simulation can be detected every time specimen are tested. In case an issue is detected, CAE engineers and test engineers can start discussion how to improve the test set up and simulation model in the laboratory.
Dimensionality Reduction of Crash and Impact Simulations using LS-DYNA