In order to support the car design process, a software tool DIFFCRASH has been conceived and recently been implemented by the GMD in the AUTOBENCH project, which is funded by the German Ministry of Research and Technology. This GMD contribution has been done in intense co-operation with project partners BMW and ESI. For a given parameter setting in a carmaker's Input Deck and on the basis of a resulting sample of DSY-files (loosely speaking, a sample of cars), generated by ESI.PAM-CRASH software, so-called stability measures are calculated as statistics (over the car sample, e.g., average, standard deviation, etc,) for different underlying `functionals', The statistics are provided node- and state-wise, the functionals use only geometric information from global node co-ordinates, In a DIFFCRASH run, for each chosen stability measure an ASCll-file of values, ordered according to states and secondly to nodes, is produced and then merged by means of the ESI.DSYMERGE-tool to a common DSY-file (on the basis of a selected DSY-file from the original sample). This so-called DSYMERGE-file is the main output of DIFFCRASH. The extended DSY-file may then be used for a colour-encoded visualisation with PAM-VIEW, in order to detect zones (hot spots) of high statistical values as a suggestive indication of instabilities. In this context, DIFFCRASH-results, from simulations with BMW Model 3, have confirmed the above proposed approach. Concrete results will be presented. (A)
Abstract