Analysis software inconsistencies

[Mohanlal0488]

Hi Guys

I know that there is a group for engineering programs but I felt that this would be a better place to post.

I work for a firm where I design structures for dynamic loading, pontoons and complex plate work structures, for my analysis I use Strand7 which I think is extremely powerful as it has allowed me to develop extremely complex models and analysis. I trust the software, as I have performed a number site tests to calibrate my analysis, in addition the support for the software is of a high standard(to name a few reasons). All the other engineers in the office are mechanical engineers who use NX.

Personally due to the type of structures I design I wouldn't feel comfortable to use any other software, maybe Lusas if I had to change...

I have briefly messed around on another package aimed at structural engineers, which I found really poor (keeping in mind that this package is one of the most popular in the country). Developing the same model in ST7, NX and Calculix yielded consistent results while the results on the structural package differed by a large amount. This held true for the linear static and natural frequency analysis. I was unable to attain any sort of real help from the consultants as their knowledge in FEA was really poor, to the extant that they were not sure if mass was considered as lumped or consistent(every package I know off allows both). With regards to natural frequency analysis I was told that the software does not show any modes that are not significant, which made no sense to me. The non-linear analysis was basically at a similar level of my own FEM toolbox that I developed on Matlab/Octave, so not very advance.

So my questions are:
[ul]
[li]what software do you use and what types of designs are you doing with it?[/li]
[li]has anyone had a similar experience with regards to software?[/li]
[li]how confident are you in the software you use and why?[/li]
[li]If you are using software aimed at structural engineers (Bentley products, Master series, etc...), does it adequately cater to your analysis requirements? If not how often are you left with a problem that cannot be analyzed/solved? [/li]
[/ul]

 

[JoshPlumSE]

"what software do you use and what types of designs are you doing with it?"
You're going to get a wide variety of answers here about that. Since this is a structural forum and the membership tends towards US engineers, you'll see a lot of US software packages. In particular, I believe you'll see lumped mass assumptions in most structural engineering packages.

"has anyone had a similar experience with regards to software?"
Similar issue where the assumptions of two different packages are different. Yes, that is pretty common. It can take a lot of investigation into figuring out why. And, a lot of understanding of the two analysis packages. Sometimes some back and forth with tech support and such.

"how confident are you in the software you use and why?"
I'll get on my soap box here a bit and suggest that we're all a little too confident in the packages we use. Not that the packages are bad or what not. But, they all have limits and many of us have too much confidence in OURSELVES and our ability to model properly and / or account for the program limitations.

"If you are using software aimed at structural engineers (Bentley products, Master series, etc...), does it adequately cater to your analysis requirements? If not how often are you left with a problem that cannot be analyzed/solved?"
My guess is that if a company has settled on a software then it adequately caters to their needs. If not, they'd get a different software. There are problems that cannot be analyzed 100% accurately, of course. But, simplifying the problem into something that CAN be analyzed sufficiently to provide for a safe design is something of an art. It requires knowledge and experience. If we don't have the ability to do that then we shouldn't be accepting those types of projects.


Caveat:
I have spend most of my career working for various structural engineering software companies. I have tried to be "software neutral" in my discussion here as this doesn't really focus on any particular package that I have experience with. However, I will say that if the discussion starts to involve specific software then please realize that my voice / opinions are my own and they come with some definite bias.

 

[IDS]

As another Strand7 user I can't comment on how it compares with other commercial packages, but comparing with my own software and with in-house software I have access to I can say:
- Linear static analysis of frames should be in near exact agreement.
- Plate and brick results will have bigger differences, but should still be in good agreement for linear static problems.
- Non-linear properties will introduce bigger differences, especially for soil-structure interaction problems, but the end results should still "make sense".

Have you tried any of the Strand7 Verification examples in the other packages?

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/