Documentation for the element class

[rjaeger]

Hi Frank,

I'm interested in getting documentation on the methods and procedures for adding new elements in opensees. Could you explain the following functions: Element::getResistingForce(), Node::getTrialDisp(), Node::getIncrDisp()? I'm confused about whether getTrialDisp returns the displacement from the beginning of the simulation, or is it from the last committed state?

Is there any formal documentation on these functions?

Thanks,
Robbie

[rjaeger]

Hi Frank,

Just bumping this thread in case you missed it. I have your thesis and I have looked through the wiki and manual, but maybe I'm just not seeing these terms.

Also, for elements that have pressure stiffness terms, how can we determine which elements are subjected to a pressure boundary condition?

Thanks,
Robbie

[fmk]

getTrialDisp() returns from the beginning.
getIncrDisp() returns change from last committed step
getIncrDeltaDisp() returns change from last trial step

the elements with pressure end with a UP or UPU .. the authors of the code were not too consistent and duplicated each other sometime!
as for which of these can handle a pressure boundary condition, i don't know.

[rjaeger]

Hi Frank,

Thank you for your reply. In regards to the pressure boundary conditions, I'm considering axisymmetric elements (like for pressure vessels) where the internally and externally applied loads are in terms of pressures rather than forces. In these cases, the nodal forces are displacement dependent since the surface area changes with nodal location. Is there any functionality in OpenSees to handle these types of loads?

Thank you,
Robbie

[fmk]

Robbie, sorry for the delay. no is the answer. Nodal Loads are currently constant. You could of course create a new subclass of NodalLoad.

[rjaeger]

Hi Frank,

Thanks again for your help. When developing an element with multiple DOF per node, does the ordering of the DOF have to follow a specific organization such as [d1x, d1y, d2x, d2y] (d is displacement) when building the mass, damping, and stiffness matrices? Or is it only important to be consistent throughout (e.g. when building M, C, K, and resistingForce vectors)?

Thanks again.

[fmk]

it's up to the elements .. of course, you need to keep them consistant across different elements if they appear in your model and you use the same nodes.