When you’re running Abaqus, the output files (.odb files in particular) can get really huge and out of control without proper attention. A lot of the default settings err on the side of caution and provide you output variables you’ll never look at, at time intervals you’re not interested in. Once these two factors multiply (especially if you have a lot of elements to begin with), your file size really balloons, which makes for slow file transfers and slow operation in post-processing. The good news is that it’s very easy to control what outputs you get and when, so you can keep your Abaqus .odb’s small.
Here are some tips to dramatically decrease your Abaqus ODB file size:
- Think about the field output variables you will actually need from the results. Don’t request VARIABLE=ALL or VARIABLE=PRESELECT which adds many unnecessary outputs. For example, don’t request PEEQ if your materials are elastic or CPRESS if there is no contact. If you forget to request something you may extract it from the restart files using *Post output. (do request *Restart, write, frequency=999 in Abaqus Standard)
- You probably don’t need field output results at every increment, especially if the convergence starts to falter and increments get very small. Reduce the frequency of output to the output database as described here.
- Create useful history output requests for a small number of important nodes, elements, or integrated surfaces. A reasonable number of history output requests can be written at a much higher frequency without significant impact on file size.
- Define a node to monitor whose selected degree of freedom will be written to the .sta and .msg files. This helps to observe what’s happening in a running analysis job without retrieving the results. Similarly, *Node print and *El print can be used to write useful results to the .dat file for review before retrieving the output database.
- If possible, request field output only for specific node or element sets.
- The *Node output and *Element output keywords have a parameter “EXTERIOR” which will reduce file size by omitting output for interior nodes and elements which are often unseen.
- Zipping your .odb files can save transfer time at the cost of processing time. Depending on your connection speed it may be worthwhile. Large .odb files typically compress by 50%-70%.