Error while sizing Uniaxial components

[ULBsha]

Hi,

I'm sizing a structure containing different type of elements (beam and panels - Unstiffened, grid stiffened, uniaxial, ...) and for one load case, I get the following error on the uniaxial component:

An error occurred while executing a user-defined analysis:
"Stiffener buckling, flat, flexural-torsional stability, argyris"




In the *.HDG file, I get:

A local buckling length to width span ratio is less than 1.
Length A has conservatively been set equal to the larger B length for this optimisation attempt.

Could someone please help me ?


PS: if I toggle off the "Buckling, Stiffener" failure analysis mode of all the selected concepts, the component can be analysed without any error...


Thanks.
Regards

[Ryan]

What version of HyperSizer are you running? Previous versions had issues with load cases with low -N_x and high +N_y load. What do the design-to loads look like for the problematic load case?

Could you post the HDG file as well?

Thanks,
Ryan

[ULBsha]

I'm using V6.4.53.

I don't have a clear idea of the order of magnitude for 'low' and 'high' load values. Please find attached a screenshot of the Design-to load tabs for one case that works correctly (left) and the case with the issue (right).

I also attached the HDG file.

Thanks again

[Ryan]

There was an issue in V6.4 for cases with compressive N_x and tensile N_y in ratios where the +N_y load stabilized the skin from buckling (buckling load = infinity). This caused an error with the flexural-torsional buckling routine because it requires a local buckling as an input. This has been addressed in V7.0, so there's a good chance this case will work in 7.0.

In your screenshot, the load case numbers for the left and right are the same. Also, the loads in the right hand side are pink which indicate that the data has not been updated - in order to the view the load case that it causing the issue, first you will have use process of elimination to figure out which load case is the problem (deactivate half the load cases, rerun, if the error occurs, deactivate half of the active load cases, and repeat). Once you have the load case figured out, you can deactivate FTB, rerun, and look at the design-to loads to have a look at the load state that is causing the issue.

-Ryan

[ULBsha]

Dear Ryan,

In your screenshot, the load case numbers for the left and right are the same.

Actually they are not the same. The mechanical load cases (500) are the same, however, the thermal loads are different, even if the ID is the same (501). In the right hand side (the analysis that doesn't work), the structure is applied to an ambient temperature of 295K. But since the initial temperature is also 295, I guess that we can assume there is no thermal load since we will not have any thermal stress and the material properties are still not degraded. On the other hand (the left hand side analysis that works fine), a 800K thermal load is applied to the structure, and as I said, this one works without any problem.

Also, the loads in the right hand side are pink which indicate that the data has not been updated -

But isn't this because the analysis can not be performed ? Even if I refresh the window, it remains pink..

in order to the view the load case that it causing the issue, first you will have use process of elimination to figure out which load case is the problem (deactivate half the load cases, rerun, if the error occurs, deactivate half of the active load cases, and repeat). Once you have the load case figured out, you can deactivate FTB, rerun, and look at the design-to loads to have a look at the load state that is causing the issue.

If, with the same load cases, I turn off FTB of all the selected concepts, there is no more issue as I said previously. Please find attached the Design-to load tab. They do not look too high/low to me...


PS: And what about the HDG file ?


Thanks

[Ryan]

It would be helpful to which load case is causing the FTB error. It's likely that the load case has tensile Ny load. When you disable FTB, there's no way to get this information.

-Ryan

[ULBsha]

I don't understand so... what should I do ?
There is only one mechanical load case since the thermal loads are zero... The mechanical load case is representative of pressure load on the fuselage skin during the cruise...


Thanks.

[Ryan]

OK. If there's only one load case then that's it's the one causing the error. I will contact you later to see if we can run this component in V7.

-Ryan

[ULBsha]

Dear Ryan,

I have a very strange error with uniaxial components. I thought here was the most appropriate topic to ask for.

Just after setting up a project, the analysis can be done without any issue. HyperFEA also works fine. However, when I update the bounds and re-run HyperFEA, I get the following error:

"Application Error in clsHyperSizerImport::DhsImport [ Sub ]( ) "

Actually the project has several groups, with only one component per group. I have several uniaxial groups, and by changing step by step the value of the bounds, I noticed that for some combinations of only one of these uniaxial groups, this error occurs.
For instance, here are 5 different tries:

1) This one works:
- Top face thickness range: 1-200mm
- Web thickness: 1-100mm
- Stiffener spacing; 1-1200mm

2) This one works:
- Top face thickness range: 1-103mm   <-- decrease
- Web thickness: 1-100mm
- Stiffener spacing; 1-1200mm

3) This one DOESN'T work:
- Top face thickness range: 1-102mm   <-- decrease
- Web thickness: 1-100mm
- Stiffener spacing; 1-1200mm

4) This one works again:
- Top face thickness range: 1-102mm
- Web thickness: 1-200mm                     <-- increase
- Stiffener spacing; 1-1200mm

5) This one works as well:
- Top face thickness range: 1-102mm
- Web thickness: 1-100mm                  <-- back to 100
- Stiffener spacing; 1-1000mm            <-- decrease


Not that the buckling span is 4500mm.


Thanks for your help
Regards

[Ryan]

It's probably the same issue as before.

Your bounds are extreme which are causing numerical issues - probably with postbuckling. I would tighten up the bounds to be more in line with the reality and disable postbuckling.