Abaqus Element Strain Energy from SET in Large Model

[mbgm8ndb]

Morning all,
I have a fairly large model I am looking at (~200,000 elements), which I have used to predict natural frequencies and mode shapes using a simple frequency analysis. This works great and takes 1-2 hours to run.

As well as predicting displacement (U) to give the mode shapes, I also predict things like strain energy in the elements. I then split the model up into different sets, say floor, sides and roof, and extract the Strain Energy for every element in a set and sum them in abaqus, using operate on xy data, to give the total strain energy in each set for each mode. Then I save these to a .rpt file. This then shows for each mode in which area the most deformation is occurring and I use the data to rank the importance of each mode.

The problem I have is to sum all the elements in a set takes my computer about 6-12 hours for each set, and during this time Abaqus just hangs so is unusable. Even to delete the unwanted element data, just leaving the SUMs takes 2-3 hours. I think this is due to the large number of elements.

The computer I am using is a Dual Core 2.4GHz 64 Bit machine with 6 Gb of ram. It runs the analysis fine, but it just seems to take ages to process the results into a format I can use them.

Does any one have any ways for speeding up this post processing of the Abaqus results?

Thanks,

Nigel

 

[IceBreakerSours]

Why don't you simply plot the strain energy of the whole model that is available as a default history variable? There is no need to carry out an addition in Viewer.

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[mbgm8ndb]

For each mode I need to compare the Strain Energy in different regions and calculate what % of the energy is in each component.

 

[IceBreakerSours]

Okay, then why not simply request strain energy output for just those regions and the whole model?

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[mbgm8ndb]

I am not sure what you are saying here. I am already requesting SE measurements from these regions and from the whole model. But I believe you then have to sum the SE in the elements for each region. If there is another way without summing each element, then this is what I am looking for?

 

[IceBreakerSours]

Hmmm. I may be wrong but I would be surprised if ABAQUS doesn't return the sum of the strain energy from a region. Why would it output strain energy for each element if you requested the energy for a region? I would find it odd, if that is the case.

I would try a simple problem with 3..4.. elements in 2D under a simple load and perform a static analysis, and see what's really going on. Let us know what you find out.

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[mbgm8ndb]

This is kind of what this post is asking. I am currently obtaining the sum of an output in a region using post processing by generating ODB Field Output and then Operate on XY Data. Can any one out there advise a quicker and easier way?

When I set up the Field Output I can ask for output from a set rather than the whole model and I can set up an output for each Set that is no problem, but I am not sure then how best to sum the findings to give the total SE in each set?
The only way I know currently is by generating ODB Field Output and then Operate on XY Data.

 

[mbgm8ndb]

Is there something I need to do in the Step, or the Field Output requests to tell it to SUM the strain energies?

 

[IceBreakerSours]

Although I am still not sold on whether the energy is summed or not, you could try Python scripting for post-processing. Whatever you do in CAE gets recorded in the .rpy file. You could use that code to have ABAQUS do the dirty work for you.

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083