Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Compression of NVH simulation results

Presentation held at NAFEMS World Congress (NWC), Crete, Greece, 16th and 19th of June 2009
: Lidzba, J.; Rettenmeier, M.; Hahn, D.; Thole, C.-A.; Iza-Teran, R.

Postprint urn:nbn:de:0011-n-1180456 (209 KByte PDF)
MD5 Fingerprint: 21856f463c72c9288c5c9245b6bb72dc
Created on: 3.3.2010

International Association for the Engineering Analysis Community -NAFEMS-:
The analysis advantage. Perspectives on engineering simulation for today and beyond : NWC 2009, International Conference on Simulation Technology for the Engineering Analysis Community, NAFEMS World Congress, June 16th-19th, Crete, Summary of Proceedings & CD-ROM
Glasgow: NAFEMS, 2009
ISBN: 978-1-874376-42-2
12 pp.
International Association for the Engineering Analysis Community (World Congress) <2009, Crete>
International Congress on Simulation Technology for the Engineering Analysis Community (NWC) <2009, Crete>
Conference Paper, Electronic Publication
Fraunhofer SCAI ()
data compression; NASTRAN; NVH; OP2 file

Crash specific data compression has become a standard technology, which is widely used in automotive industry. This paper describes lossy data compression of NASTRAN-OP2 files as they emerge from NVH simulations in the automotive sector. The largest part of these files contains real or complex eigenmodes but also element energies and stress or strain tables. During the automotive design process a large number of simulations is performed. As a result huge amounts of data have to be stored. Using data compression the size of the files can be significantly reduced with less storage space required. The reduction in file size resulting from the compression also leads consequently to faster file transfer and I/O-times of post processing tools, which are able to read compressed files via a specific library. Therefore, efficient compression of OP2-files means a large benefit for NASTRAN users. To achieve very high compression ratios a lossy data compression scheme was chosen. During compression floating point data is quantized using a user controlled precision. This allows the user to meet his requirements as a tradeoff between the magnitude of errors and the compression factor. To achieve the best com-pression ratio at a given precision the entropy of the data is reduced further by using specially designed prediction algorithms. The technologies involved exploit the shell and solid element connectivity information derived from the OP2 files. To accomplish seamless workflow integration the compressed data must be fast and easily accessible. Therefore the HDF5 format was chosen as output file format. The advantages of the HDF5 file format will be exploited during decompression. Easy and fast data access facilitates the parallelization of the decompression using OpenMP. This is especially interesting for post processing tools, which directly read OP2 data from the compressed files. The software FEMZIP-N which makes use of all these features showed very satisfying compression results on NVH simulations conducted by different automotive companies. Compression ratios between 5.0 and 40.0 have been achieved.