THEME SDM, Data Analysis SUMMARY

More product variants, the increase in government regulations and the use of stochastic and multidisciplinary optimization causes a drastic increase of the number of numerical simulations performed in the car design process. But the availability of the arising numerical data creates the opportunity to compare thousands of simulation variants with the aim to find optimal design variants (even from previous projects) or to evaluate the impact of specific design changes on the functionality of the car. The huge volume of data (large number of simulations and large data-sizes per simulation) and the diverse content of simulation data are major obstacles for the efficient post processing and analysis of such simulation bundles. Here, we propose a methodology that enables suitable processing. In particular, we extend the existing efficient post processing from one simulation to many hundred simulations. We study two types of realistic scenarios for the approach. The first one assumes that the simulation bundle is managed inside a specific SDM and the second one starts without this assumption, that is the information is supposed to be just saved in a repository in a specific simulation data format. The approach is demonstrated on a simplified frontal car simulation, where a set of geometrical variants is analyzed. A second example shows that the system is able to create a data management capability from simulation data in a specific car design phase