Learning Forum

[javat33489]

Hi all. Recently I did a calculation with beams and I had a few questions about their work in ansys and the interpretation of the result. In the results, I analyzed Max Combined Stress, which is the combination of forward stress and maximum bending stress. I think this is what you need when analyzing a steel truss.

If it shows me the maximum voltage on one of the corners, can I pay attention to this and say that this is the weakest point (where the maximum voltage is)?

If you exclude the corner, then the max stress points to the junction of two beam nodes, where the mesh intersects and the pipe is in the pipe, can I not take this into account? Here:

Do I need to take into account the negative voltage? After all, in theory, this voltage is simply directed against the axis.

Please do not send me standard ANSYS courses. I know them very well. I would like to hear from ANSYS experts and their answers and recommendations.

Thank you.

[Dave Looman]

In the US there is the AISC code that tells you how to evaluate beam stresses.  Are you working to such a code?  If you don't have a code and just want to stay below the yield stress of the material then the Maximum Combined Stress can be used for that.  Negative stresses would be significant if you were concerned about buckling.

[javat33489]

>> Are you working to such a code? 

No I don't need it

>>If you don't have a code and just want to stay below the yield stress of the material then the Maximum Combined Stress can be used for that.

Thank you.

___

But what about this question?

[Dave Looman]

Are you using Unaveraged stresses?  I would recommend that.

[javat33489]

Thanks for the answer. I replaced as you said. Can I ignore the stresses that occur at the end of a beam element that is inside another?

[peteroznewman]

Beam models are a very idealized model that ignores a lot of detail so that basic sizing calculations can be performed and the design iterations can be rapidly done.  Once the design has converged on a potentially feasible solution, a detailed model is often built to look at the stress around the actual welded connections between the truss ends.  The photo below shows welded ends. A detailed model could be built to analyse the stress, which will probably have higher stresses than the simple beam model due to stress concentrations.

[javat33489]

Yes, but the model can be very large. And the grid can consist of several millions of cells. Then the detailed model cannot be calculated. There are few such computers in the country.

[peteroznewman]

 

Ansys has a capability to do submodeling.  See this course for details.

Submodeling is the process of making a boundary around one joint in the global beam model and transfering the loads going through that boundary into a submodel that can use solid elements and have all the detail of the intersection of the pipes including blends. A submodel is created one joint at a time and each joint can be calculated on a computer typically configured for engineering analysis.

Only a few joints have high stress so only a few submodels are required.

[javat33489]

I understood you. Thank you. I will try. But anyway. When analyzing beam elements, if the stress is on a beam that is inside another beam (at the end), can I turn a blind eye to this?

[peteroznewman]

No.

[javat33489]

Okay, so you need to ensure that there are no such tensions.