difference in symmetric boundary conditions results

Vikas_22008

Hallo

I have conducted a static analysis on a beam of thickness 5mm. Below are the results. The beam is symmetric about YZ plane. I have used Ux = 0, on the left model at the symmetric plane.  i want to know why there is a difference in the results, when it is symmetric in nature.  I have attached the model file for your reference. Please let me know where i must do changes.

 

Thanks

Vikas M B

loistf

Hi Vikas. If you want to compare results, you should start with displacements before stress. Then, if you go to stress, definitely compare raw stress (no averaging) first. And at al times with the same color scale.

How do displacements look?

Vikas_22008

Hi loistf,

Thank you for the response.
The half beam has 0.072mm displacement and full beam has 0.062mm displacement. 
If i am interested in one of the averaging methods, then why should i compare raw stress?
What do you mean by 'same color scale' ?

 

Thanks 
Vikas M B

loistf

In general, when you are debugging, investigating why a CAE solver is giving you results that are different to what you were expecting (as this case), it is typically good practice to go back to "simpler" (or more basal) results / models, to "hunt" for causes of difference. 

In particular, in your case, the stress at the nodes that are in the middle of the beam will be different if you average. (in one case, this center node has elements at only one side and in the other the same node has elements at both sides). 

For same color scale I mean that the same color should represent the same exact range of values. If you use the automatic values and the max / min are different, any color will represent different values. For example, yellow represents 47 to 55 in one of your images and 48 to 56 in the other). 

By the way, now that I look at it, the results in the full slab don't look symmetric.

Hope it helps. 

 

Rajashri_Saha

Hi Vikas,

I just viewed your query and observed that in full model also the result is not symmetric if you observe the color contour carefully. This might be due to Load and BC issue. I checked the full model and found that at one end where you have put spc, it is connected to the block too, whereas in other end it is connected to independent node of RBE2. Similarly, the force load has to be only on independent node of RBE2 rather than connecting to the block. While updating the load for half model, do not change the loading node. Can you modify these and rerun both the models. 1st try to have symmetrical contour in full model. While comparing both, have the legend values same for both and then compare. 

Thanks

Rajashri

Vikas_22008

loistf said:

In general, when you are debugging, investigating why a CAE solver is giving you results that are different to what you were expecting (as this case), it is typically good practice to go back to "simpler" (or more basal) results / models, to "hunt" for causes of difference. 

In particular, in your case, the stress at the nodes that are in the middle of the beam will be different if you average. (in one case, this center node has elements at only one side and in the other the same node has elements at both sides). 

For same color scale I mean that the same color should represent the same exact range of values. If you use the automatic values and the max / min are different, any color will represent different values. For example, yellow represents 47 to 55 in one of your images and 48 to 56 in the other). 

By the way, now that I look at it, the results in the full slab don't look symmetric.

Hope it helps. 

 

Hi loistf,

thank you for explaining the significance of not using the averaging methods.

 

Vikas_22008

Rajashri Saha_20525 said:

Hi Vikas,

I just viewed your query and observed that in full model also the result is not symmetric if you observe the color contour carefully. This might be due to Load and BC issue. I checked the full model and found that at one end where you have put spc, it is connected to the block too, whereas in other end it is connected to independent node of RBE2. Similarly, the force load has to be only on independent node of RBE2 rather than connecting to the block. While updating the load for half model, do not change the loading node. Can you modify these and rerun both the models. 1st try to have symmetrical contour in full model. While comparing both, have the legend values same for both and then compare. 

Thanks

Rajashri

Hi Rajashri,

I have made some changes in the model as per your suggestions but still the results are not quite symmetric. Please find the attached images of the results. I have adjusted the legend values to check the contour. I still see that there are some minor differences. 

Displacement contour

Stress contour

Could you please tell me what is significance of setting up the same legend values?
I have also attached the model files for your reference. Please feel free to update the model.

Regards

Vikas M B

 

Rajashri_Saha

Vikas_22008 said:

Hi Rajashri,

I have made some changes in the model as per your suggestions but still the results are not quite symmetric. Please find the attached images of the results. I have adjusted the legend values to check the contour. I still see that there are some minor differences. 

Displacement contour

Stress contour

Could you please tell me what is significance of setting up the same legend values?
I have also attached the model files for your reference. Please feel free to update the model.

Regards

Vikas M B

 

By keeping the same legend, you can compare the result by visualization method. now you can see the contours are almost matching.

Thanks

Rajashri 

Vikas_22008

Rajashri Saha_20525 said:

By keeping the same legend, you can compare the result by visualization method. now you can see the contours are almost matching.

Thanks

Rajashri 

Thanks Rajashri