Output Forces at Constrained Nodes

[christian.moyer]

Hello,

I am wondering if it's possible to output the reaction forces that enforce displacement constraints at nodes.

My thinking: if simple nodal constraints are enforced using some type of Lagrangian multiplier, where the multiplier is the virtual nodal force enforcing the constraint, then FEBio should be able to output the L.M. at the end of a run or at every time step within a run.

I'm considering an FEBio model where a subset of elements has a prescribed displacement. I'd like to be able to recreate the same amount of displacement using nodal forces (ie, two versions of the same model with the same deformed state, one driven by nodal displacement and one driven by nodal force).

Thanks,
-Christian

[christian.moyer]

Anyone have any thoughts on this? FEBio should be solving for some type of Lagrangian multiplier, or other means to enforce the constraint. It would just be a matter of outputting or storing that information.

[ateshian]

Hi Christian,

In order to plot reaction forces, use <var type="reaction forces"/> in the output request. (See section 3.11.2 in the User's Manual). For example:

	<Output>
		<plotfile type="febio">
			<var type="displacement"/>
			<var type="stress"/>
			<var type="reaction forces"/>
		</plotfile>
	</Output>

This will include reaction forces in the output, for all degrees of freedom which are *not* fixed. So, if you prescribe displacements at a node, you will be able to extract the reaction force. Please note that this reaction force is the negative of the force acting at that location.

Let me know if that answers your question.

Best,

Gerard

[christian.moyer]

Gerard,

Thanks so much. This worked perfectly.

-Christian

[StudentPaul]

Hello,
I have a problem using the method mentioned above, namely an unrecognized tag "var". I use FEBio 1.5. Has this tag been changed lately?
<Output>
<plotfile type="febio" file="testtest27.xplt"/>
<var type="reaction forces"/>
</plotfile>
</Output>

Found the solution : the last slash in the line plotfile type = is not necessary, although it is in the manual 1.5 in section 3.12.2.

Kind regards, Paul

[StudentPaul]

Hello,
I want to use the nodal reaction force to validate a model. But i don't manage to get these of of my file. Should i use the lsdyna option, or the .xplt? I don't know what programs are able to open the .xplt files? And i think this is a binary file and i need to process the output data, for instance to draw a graph, so i would rather have some numerical values. Does anyone have any thought about how to do this?

Thanks, Paul

[maas]

Paul,

You can have the reaction forces written to the logfile instead of the plot file. Since the logfile is a text file, it should be much easier to parse. Please see section 3.12.1.1 of the (online) user's manual and let us know if you have any further questions.

Cheers,

Steve.

[StudentPaul]

Hi Steve,

just to be sure: I can write the reaction force of specific node sets (caused by displacement) to the logfile? Not only the reaction force on a rigid body?
In the case of specific node set reaction force, what is the exact code that i need to use to write the logfile? Lets say for example the reaction force on nodes 3 and 4?

I'm looking at the format of .xplt files at the moment but i hope i can use te logfile because that is indeed much easier.

kind regards, Paul

[maas]

Hi Paul,

Yes, you can request reaction forces for nodes. Although be aware that you will only get a non-zero reaction force for nodes that have a prescribed displacement. You need to add the following line to your logfile section in the input file.

<node_data data="Rx;Ry;Rz">3,4</node_data>

For more details, see section 3.12.1.1 of the user's manual. Hope this helps.

Cheers,

Steve.

[StudentPaul]

Thank you, this helps indeed. And for my understanding: the reaction forces on nodes that are moved by unprescribed displacement (so by connection to other nodes that are prescribed) con NOT be written and found in the logfile or in the plot file?

gr, Paul

[maas]

Yes, that is correct. This is currently a limitation of the implementation. Let me know if this ever becomes a problem.

Steve.