Geometric non-linear analysis

[Ronald286]

Hello, support!

I try to make geometric non-linear analysis(just a test) with considering global imperfections(consider buckling mode as initial deformation). And I get the warnings about missing eigenvalues. I tried to increase the number of modes and number of iterations, but it doesn't help. Please, can you look the model, and check if I defined analysis types correctly and help me to solve the problem with warnings. Thank you!

[Artur.Kosakowski · Accepted solution]

Try to mark the Sturm verification on

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[Ronald286]

Thanks, it works!

One more question. As i understood, P-delta effects in RSA named as non-linear analysis. As we know, there are two types of P-delta effects: 1) P-∆; 2) P-δ. Most of codes takes into account the P-δ effect in the formulas of checking stability of the separate members. So, in global analysis, for example, it would be good to have passability to take into account P-∆ effect only(where it neseccary according to the code), and other effects will be considered by formulas. Is it possible to do it in Robot - to take into account P-∆ effect only?(without P-δ)

[Artur.Kosakowski]

In the current version of Robot only P-∆ is implemented. The P-δ can be accounted for by defining some nodes along the bars.

[Tuctas]

  Just a clarification:

  The Non-linear analysis type, as it is explained in Robot’s help topic, takes in to account the so-called “stress-stiffening” effect that actually has to do with the modification of the stiffness matrix according to those effects.

  In the same help topic it also says: “…At the same time, this analysis considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally”.Does it mean that when the structure is already deformed (considering buckling mode as initial deformation) then additional moments due to vertical loads are obtained or that additional moments are caused while the structure is naturally deformed from any kind of loading (not initially -“manually”- deformed through the option in Analysis type – Buckling deformation window)? If it’s the last one then it’s about P-δ effects isn’t it? Or maybe (in the 2nd case) members should be devided through the definition of internal nodes in order those additional moments to be taken in to account? 

[Artur.Kosakowski]

---

  The Non-linear analysis type, as it is explained in Robot’s help topic, takes in to account the so-called “stress-stiffening” effect that actually has to do with the modification of the stiffness matrix according to those effects.

  In the same help topic it also says: “…At the same time, this analysis considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally”.Does it mean that when the structure is already deformed (considering buckling mode as initial deformation) then additional moments due to vertical loads are obtained or that additional moments are caused while the structure is naturally deformed from any kind of loading (not initially -“manually”- deformed through the option in Analysis type – Buckling deformation window)?

Correct. That is P-Δ or P-"big-delta"

If it’s the last one then it’s about P-δ effects isn’t it?

P-δ is something different.

Or maybe (in the 2nd case) members should be devided through the definition of internal nodes in order those additional moments to be taken in to account? 

• P-δ effect, or P-"small-delta", is associated with local deformation relative to the element chord between end nodes. Typically, P-δ only becomes significant at unreasonably large displacement values, or in especially slender columns. An easy way to capture this behavior is to subdivide critical elements into multiple segments, transferring behavior into P-Δ effect.

• P-Δ effect, or P-"big-delta", is associated with displacements relative to member ends. Unlike P-δ, this type of P-Delta effect is critical to nonlinear modeling and analysis.

[Tuctas]

  Thanks for the reply.

  It is now confirmed how much confusion the way that Robot has titled the two types of geometric non-linearity causes... (and i realy don't understand why it isn't changed for so long, i am talking just about the titles: where it says "P-delta" it should write third order effects or large displacements effects and where it says "non linear" it should write "second order" or "stress stiffening").

  I was aware of the differences between P-Δ amd P-δ (the letters Δ and δ are Greek so they are familiar to me..). 

  P-δ effects are taking place in the deformed shape of each member (instead of the whole structure that is in case of P-Δ) so i thought that what help topic says: “…At the same time, this analysis considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally” means that. What does exactly this phrase of the help topic (that corresponds to Robot's explanation about the "non linear analysis") means?

  Finally, when selecting "non linear analysis" type then ONLY the modification of the stiffness matrix is taking place (i.e the so-called "stress stiffening")?  

[Artur.Kosakowski]

Please check http://forums.augi.com/showthread.php?141487-Assorted-queries-non-linear-analysis-for-part-braced-fr... (message #4)

[Artur.Kosakowski]

At the same time, this analysis considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally” means that. What does exactly this phrase of the help topic (that corresponds to Robot's explanation about the "non linear analysis") means?

  Finally, when selecting "non linear analysis" type then ONLY the modification of the stiffness matrix is taking place (i.e the so-called "stress stiffening")?  

See the attached picture.

P-δ is not implemented in the current version of Robot - if needed additional nodes along bar length should be defined.

[t.sautierr]

Does this mean that there is no interest to perform Robot P-Delta (Robot ticking case -> P small delta), if the bars in the model haven't been divided at least in the middle with the dividng tool ?

I mean Robot will not implement automaticaly division with calculations nodes between the ends of the bars and then use it to perform 3d order effect?

By the way, is it useful to divide the bar physically or can we just use the division tool "divide in X part" ticking "without dividing the bars".

Thx Arthur.

[Artur.Kosakowski]

Does this mean that there is no interest to perform Robot P-Delta (Robot ticking case -> P small delta), if the bars in the model haven't been divided at least in the middle with the dividng tool ?

I'm not sure if I understand you correctly but there is no need to have intermediate nodes for P- big Delta.

I mean Robot will not implement automaticaly division with calculations nodes between the ends of the bars and then use it to perform 3d order effect?

There is no automatic division of bars into smaller pieces if you mark either nonlinear or P-Delta check box.

By the way, is it useful to divide the bar physically or can we just use the division tool "divide in X part" ticking "without dividing the bars".

I'd rather use the later option so that the review of results is easier.

[t.sautierr]

Ok, so to perform P-small delta analysis -> first the bars as to be divided into X Part , Otherwise Robot P-delta analysis will not perform something relevant as it needs iinternal nodes in the bars

Thx

[Tuctas]

  O.k now it is clear what exactly Robot does when the user select the one or the the other option of geometrical non linear analysis...

  I hope these 2 "titles" of analysis types will be corrected soon.

  I also think that there should be automatically generated internal nodes along elements (maybe by adding an option how many internal -calculation- nodes to be used) when selecting 3rd order effects, as t.sautier said.

  Regards.  

[mariajodasilva07]

Hello,

Sorry to bother you but I have questions in this topic if you can help me. 

First how do you make geometric non-linear analysis with considering global imperfections (consider buckling mode as initial deformation)? I mean, you creat a vertical load and do the buckling analysis, and you get the buckling modes. Then what? How do you take this results and do an non-linear analysis?

Second question, if non-linear analysis “considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally”,you don't have to intruduce a horizontal load in the node to do the non-linear analysis ( like a initial imprefection), don't you? 

Thank you

Maria Silva

[Artur.Kosakowski]

First how do you make geometric non-linear analysis with considering global imperfections (consider buckling mode as initial deformation)? I mean, you creat a vertical load and do the buckling analysis, and you get the buckling modes. Then what? How do you take this results and do an non-linear analysis?

You set which mode should be used for a structure deformation and set nonlinear analysis for cases / combinations you want to look at.

Second question, if non-linear analysis “considers generation of moments resulting from the action of vertical forces at the nodes displaced horizontally”,you don't have to intruduce a horizontal load in the node to do the non-linear analysis ( like a initial imprefection), don't you? 

You need to have something that causes the structure to move sideways to see the effect so you will need it for a vertical column in case you run such test (a non-symmetrical structure is fine  ). 

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[mariajodasilva07]

Thank you for your answer. It really helps me 🙂

But I have another question. I created in Robot a simple column pinned on the base and fixed in x direction on the top. I introduced a vertical load and do the buckling analysis. Then I did what you said before: I set which mode should be used for a structure deformation and set nonlinear analysis for case with the vertical load. (see attachment)

When I saw the results, it appeared a bending moment, probably that occurred because the robot introduce "something that causes the structure to move sideways", am I right? So my question is, what or where you can know the value of that "something", that force?

Thank you,

Maria Silva

[Artur.Kosakowski]

The final "undeformed" geometry of the column is as shown below:

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.