FEMZIP compresses data from crash simulations in the car manufacturing industry as precisely as the user specifies. This reduces the total amount of data, yet important information is not lost and is still available after the decompression.
In an effort to improve engineering design larger and more detailed models as well as an increasing number of simulations are performed, yielding a steady growing amount of data. This simulation data has to be analyzed, exchanged among engineers and archived for future reference and re-analysis. Network connections and storage space can then become bottlenecks in workflows used by engineers. The challenge is to handle these large amounts of data in a time- and storage-saving manner in order to eliminate the bottlenecks.
With our FEMZIP compression tools we provide a solution to meet this challenge. FEMZIP tools are specifically designed for the compression of simulation results of various simulation codes.
The high compression factors that can be obtained by employing FEMZIP lead to several benefits:
- Reduced Archive Size
If all simulation data is compressed only a fraction of the storage is required. Storage and backup capacities can hold more simulation results. Hence, investments in the growth of storage and archive infrastructures can be avoided. - Shorter Data Transfer Times
Transferring result data across computer networks can be very time consuming. Transmission speeds are limited by the available network bandwidth. Since compressed simulation results require significantly less storage they can be transferred in a fraction of the time required for transferring uncompressed results. - Quicker Data Loading
Reading data into post-processors can be a time consuming task as well. If data storage is provided on file servers, compressed data can be read directly into post processors significantly quicker due to the faster data transfer. In some cases load times can even be improved when reading from a local hard drive.
Functionality
Simulation results are usually stored in 32- or 64-bit floating point representation. Due to numerical approximation errors and the inaccuracy in the model description itself, the 32- or 64-bit floating point representation is often unnecessarily precise compared with the effective precision of the simulation results. Therefore FEMZIP is mainly based on lossy compression. The loss is controlled by the user by means of precision settings specified for the geometry and for each variable. In addition FEMZIP supports the option to compress specific variables losslessly as well as the option to eliminate variables completely. In general, lower precision settings result in higher com- pression factors.
Process Integration
FEMZIP is made to fit into the workflow used by engineers. In order to maximize the benefit of data compression, FEMZIP is designed to be integrated into batch processes. In this way simulation results can be compressed automatically after they have been generated, reducing file sizes from the start. Although the original data formats can be retrieved by a stand- alone decompression tool, this is often not necessary. The decompression functionality of FEMZIP is integrated in common post-processing tools and enables direct access to the content of compressed files. FEMZIP does not alter existing workflows while providing the benefits of data compression.
