How do I model rotational spring or give rotational stiffness at a node in Abaqus/standard?

[njain094]

I am trying to model a normal spring and a rotational spring to find Pcr of the system. I have attached a figure for a better understanding of my problem. (The actual problem is different than this one. This problem is just to know how to model a rotational spring in Abaqus). I am able to model a normal spring because the measurement of delta(deformation) is pretty simple in this case which is the change in the distance between two nodes. I am having trouble modeling a rotational spring at a node because I am not able to understand the proper measure of rotation. Please let me know if somebody knows anything in this regard.

 

[FEA way]

Spring elements SPRING1 and SPRING2 (available only in Standard) can be used to model torsional springs. The first one acts between node and ground while second one acts between two nodes (in fixed direction). To associate the spring with rotational degree of freedom you have to specify the number of this DOF as the first entry in the first data line of *SPRING keyword. Rotational DOFs in Abaqus are measured with radians. If you have some difficulties with modeling this using springs keep in mind that you can use connector elements instead.

 

[njain094]

Hello,

Thanks for replying.

I used both of those options, one being torsional spring between two nodes (DOF 6 at both nodes) (SPRING1) and the second one being the spring between a node (DOF 6) and the ground (SPRING2). But in both these cases, when I increase the stiffness, it does not change the end result. If I increase the stiffness of the torsional spring there must be some change in the value of the Pcr but there is not. I also tried using the connector element. The story is the same. If I increase the stiffness, it does not change Pcr of the system. That is why I am curious to know as to how Abaqus is finding the amount of rotation. I understand that the measure is radians. But what is its value? How to see the amount of rotation in torsional spring.

And just to make sure, with SPRING1 and SPRING2 spring elements I should be able to model torsional spring correct?

 

[FEA way]

Can you attach your model (.inp or .cae file) ? I would take a look to see how you modeled this system in Abaqus and if BCs, connections and spring definitions are correct. Is that based on some example from literature ? Do you have analytical solution for that ?

 

[njain094]

I have attached the CAE file. In the model, I have introduced a normal spring, and a rotational spring in between two nodes. Here I am doing buckling analysis to find Pcr of the system. (Please refer to the jpg image.)

Yes, I do have an analytical solution for this system.

I am working with this system to find out how I can use rotational spring in Abaqus. My original problem is different than this one.

Boundary conditions: pinned at the bottom and a vertical load of value -1 is applied to the top.

In addition to this, I have attached one more CAE file where I have used normal spring and transverse spring. This works fine.

 

[njain094]

Spring Model

 

[njain094]

Transverse spring CAE file

 

[njain094]

The question is, does Abaqus know that the measure of theta is w.r.t to the vertical? I think that is the issue here.

 

[Rakz.mech]

Dear njain094,

Before starting further studies in buckle analysis, you should know that buckle analysis is very different from general static analysis. End restraints, imperfect members a and member with combined loading can be solved in buckle analysis, but the eigenvalues you get finally will be unaltered, meaning that Pcr is not correct. If you are seeking to find Pcr using Buckling analysis you should consider ideal case such as fixed-fixed, fixed-free and so on. If you wish to know the effect of rotational restraint I suggest you go by static general or Riks method (Post buckling analysis). I am looking forward to your reply, and I might as well wish to know more about the solution you got..

 

[njain094]

Hello Rakz.mech,

I was planning to do static general but in Abaqus if I directly use static general the spring will just compress because there is no imperfection in it. Imperfection is introduced in the model by updating the nodes by incorporating modeshapes. Am I right? Please let me know what you think.

 

[njain094]

Hello Rakz.mech,

I added an artificial imperfection in the location of the reference point (top). But the rotational spring still does not take any load. I have attached a CAE file here. Please let me know if you find anything. (The coordinate of top reference point is 0.001, 10.0)

 

[FEA way]

Imperfection can be also introduced in form of small lateral load applied in the step prior to buckling analysis. But you will need additional transverse spring to handle such load.

 

[njain094]

Yes FEA way, I understand that. Thank you for the suggestion.

Were you able to find anything regarding rotational stiffness? Please let me know.

 

[Rakz.mech]

The model you have sent is in CAE format and it is of a higher version of abaqus. It would help if you copy the input file to notepad and send it using txt format. And again providing rotational spring and imperfection will not alter buckling load if you use Linear Perturbation (Buckle analysis), however, you can use static general or static risk. Let me know further.....

 

[njain094]

hello, please find the attached input file.