Modal Frequency response stress probing

Balaji T

Hi, I'm running modal frequency response on a structure, I have to probe stress and displacement results on that. For stress results I would like to plot Frequency Vs Stress.

1. Structure modelled and modal analysis was done.
2. Using Freqi and Rload card created FRF response subcase. Load = 10g acceleration base excitation (SPCD used).
3. Output requested for displacements, stress and GP Stress also.
4. Need clarification in post processing,
   1. For stress probing, I created a derived load case for FRF subcase (selecting all frequencies) and checking grid points stress. Is this correct?
   2. These grid point stress is different for Linear superposition type and Envelope Max type derived load case. Which one to select?
   3. After probing high stress node in the above step, plotting Frequency Vs Stress of that node, these stress are very less compared to above. Why this will happen?

I need support on this process or workflow. One of the tutorial (OST : 1305) requests displacement at a node of interest.

loistf

Hi! Any reason why you are using grid point stress instead of simply stress?

Balaji T

loistf said:

Hi! Any reason why you are using grid point stress instead of simply stress?

Hi,

1. I had a practice of node probing in another tool and want to follow the same procedures.

2. I can probe only the element stress without gpstress card.(Results type >Element Stress>von mises>component>Simple Avg method>apply).

What is the simple stress? is that element stress?

But I request you to give suggestion on best practice for the stress probing.

loistf

Balaji T said:

Hi,

1. I had a practice of node probing in another tool and want to follow the same procedures.

2. I can probe only the element stress without gpstress card.(Results type >Element Stress>von mises>component>Simple Avg method>apply).

What is the simple stress? is that element stress?

But I request you to give suggestion on best practice for the stress probing.

If there is something you were doing in a certain tool, don't hesitate to share, maybe somebody has experience in both tools.

As for my opinion, i could summdsize in: if you want stress, request stress, not gpstress.

Ideally, plot stress on elements. If you need results at nodes, that is another discussion. (you can extrapolate or average or both)

Balaji T

loistf said:

If there is something you were doing in a certain tool, don't hesitate to share, maybe somebody has experience in both tools.

As for my opinion, i could summdsize in: if you want stress, request stress, not gpstress.

Ideally, plot stress on elements. If you need results at nodes, that is another discussion. (you can extrapolate or average or both)

Hi, thanks for your suggestion. I'll check element stress too.

Now can you share your opinion on my above questions, instead of gpstress now I'll go with element stress. So my process is correct?

loistf

Balaji T said:

Hi, thanks for your suggestion. I'll check element stress too.

Now can you share your opinion on my above questions, instead of gpstress now I'll go with element stress. So my process is correct?

I am not in a position to say with certainty if your process is correct or not, but i can point to certain things:

-use stress on elements, as we said, instead of gpstress unless you have a good reason to do otherwise

-i don't see how linear envelope applies to the frequency-based results of an FR analysis. Envelope, (the max across frequencies) sounds like a more typical idea/method

-if you have doubts on what you see after certain postprocessing, i recommend loading the result file in Hg and plot there the raw data.