02-17-2021, 09:39 PM
Thanks a lot for this information! I've used Femap before, and I've been looking closely at how we can improve MYSTRAN on the output front (e.g. formats and automatic comparison/testing).
It's great to know that the neutral files work with Femap. I've had some issues with them before, but maybe it's because I was using a model edited by me to work with NASTRAN-95.
About the vonMises stress differences, I think we'll need to have Bill look into those. Are you at liberty to disclose the model(s)/result file(s)? Those could be useful when looking into those differences and what could be causing them.
And the scaling... I thought Femap inferred the scales from the values in the neutral file, but after looking at the PDF where they define how the format works, it appears there are some fields that the solver can use to "suggest" display modes for vectors and such. Same goes for result vector numbering and the transparent/weird display, although the latter can be manually adjusted in the Display specifics dialog.
But there's a serious hindrance when investigating anything like this: the only "standard" format MYSTRAN outputs is NEU (OP2 support is on its way), and _there seems to be no open-source parser/writer library for those._ Since neutral files are made of blocks that can be in any order and there's lots of deprecated/redundant features (about one fourth of all data block types were obsolete **by the time Femap 6.0 was released**), looking at their contents directly is a nightmare. To make matters worse, the modern standards are really, _really_ hard to come by. Took me a while to find the PDF for v6.0...
Due to all that, it's really hard to look into how NEU output can be improved because the closest we can get to inspecting the files is Femap, and it abstracts away a lot of details. That's why I've been searching all around for the latest NEU standard -- this way, there could finally be a compliant, open parser/writer library and we could inspect and compare output files to fix those issues.
All in all... thanks for all the info. I'll tell Bill about those differing values in case he hasn't read this already. As for the other issues, their fixing will be much easier once we can inspect those files properly, and I'm working on that. If you know anything about where I can find the latest standard for Femap NEU files, that'd be much appreciated, just like any other feedback.
It's great to know that the neutral files work with Femap. I've had some issues with them before, but maybe it's because I was using a model edited by me to work with NASTRAN-95.
About the vonMises stress differences, I think we'll need to have Bill look into those. Are you at liberty to disclose the model(s)/result file(s)? Those could be useful when looking into those differences and what could be causing them.
And the scaling... I thought Femap inferred the scales from the values in the neutral file, but after looking at the PDF where they define how the format works, it appears there are some fields that the solver can use to "suggest" display modes for vectors and such. Same goes for result vector numbering and the transparent/weird display, although the latter can be manually adjusted in the Display specifics dialog.
But there's a serious hindrance when investigating anything like this: the only "standard" format MYSTRAN outputs is NEU (OP2 support is on its way), and _there seems to be no open-source parser/writer library for those._ Since neutral files are made of blocks that can be in any order and there's lots of deprecated/redundant features (about one fourth of all data block types were obsolete **by the time Femap 6.0 was released**), looking at their contents directly is a nightmare. To make matters worse, the modern standards are really, _really_ hard to come by. Took me a while to find the PDF for v6.0...
Due to all that, it's really hard to look into how NEU output can be improved because the closest we can get to inspecting the files is Femap, and it abstracts away a lot of details. That's why I've been searching all around for the latest NEU standard -- this way, there could finally be a compliant, open parser/writer library and we could inspect and compare output files to fix those issues.
All in all... thanks for all the info. I'll tell Bill about those differing values in case he hasn't read this already. As for the other issues, their fixing will be much easier once we can inspect those files properly, and I'm working on that. If you know anything about where I can find the latest standard for Femap NEU files, that'd be much appreciated, just like any other feedback.