Hi, what is the best way to measure P-small delta in a column? I'm modeling a SCBF with brace buckling, and I have several columns on their weak axis with large gravity loads. At one point, my model stops converging. I believe this is due to a column buckling, specifically the 6th one, as shown in the figure below.
http://i.imgur.com/HkIOIg4.png
Right now I have the base of my each column connected to the end of my rigidlinks through a corotTruss element. Ideally to measure the displaced shape of my column I think I would need to discretize my truss chord the same way i decretize my column, but you cannot discretize the truss element without causing stability problems in the model.
Secondly I tried to use many elasticBeamColumns element with low E and high A and I, set to equalDOF 1 and 2 at the base of the column and the end of the rigid link, but these elements deformed and weren't able to remain straight like I want them to (like a truss).
Essentially I am wanting to use the truss element as a LVDT. What are some solutions to this?
Secondly, what are some ways to verify my column has truly failed at this point? I am getting warnings at MANY elements at the point of no convergence, not just the ones in the column in question. No matter how I change my column discretization, solution scheme, fibers, iterations, etc, this is the case. Should I just use moment curvature example from the wiki and vary my axial load to piece together the P-M interaction curve for my column?
Measuring column P-small delta.
Moderators: silvia, selimgunay, Moderators
-
wuhaoshrek
- Posts: 122
- Joined: Tue Oct 28, 2008 4:01 am
Re: Measuring column P-small delta.
Joesh wrote:
> Hi, what is the best way to measure P-small delta in a column? I'm modeling
> a SCBF with brace buckling, and I have several columns on their weak axis
> with large gravity loads. At one point, my model stops converging. I
> believe this is due to a column buckling, specifically the 6th one, as
> shown in the figure below.
>
> http://i.imgur.com/HkIOIg4.png
>
> Right now I have the base of my each column connected to the end of my
> rigidlinks through a corotTruss element. Ideally to measure the displaced
> shape of my column I think I would need to discretize my truss chord the
> same way i decretize my column, but you cannot discretize the truss element
> without causing stability problems in the model.
>
> Secondly I tried to use many elasticBeamColumns element with low E and high
> A and I, set to equalDOF 1 and 2 at the base of the column and the end of
> the rigid link, but these elements deformed and weren't able to remain
> straight like I want them to (like a truss).
>
> Essentially I am wanting to use the truss element as a LVDT. What are some
> solutions to this?
>
> Secondly, what are some ways to verify my column has truly failed at this
> point? I am getting warnings at MANY elements at the point of no
> convergence, not just the ones in the column in question. No matter how I
> change my column discretization, solution scheme, fibers, iterations, etc,
> this is the case. Should I just use moment curvature example from the wiki
> and vary my axial load to piece together the P-M interaction curve for my
> column?
yeah, same question. corotational geometric trasf. considers only equilibrium in the
deformed configuration (coord), but doesn't include P-small delta, does it?
> Hi, what is the best way to measure P-small delta in a column? I'm modeling
> a SCBF with brace buckling, and I have several columns on their weak axis
> with large gravity loads. At one point, my model stops converging. I
> believe this is due to a column buckling, specifically the 6th one, as
> shown in the figure below.
>
> http://i.imgur.com/HkIOIg4.png
>
> Right now I have the base of my each column connected to the end of my
> rigidlinks through a corotTruss element. Ideally to measure the displaced
> shape of my column I think I would need to discretize my truss chord the
> same way i decretize my column, but you cannot discretize the truss element
> without causing stability problems in the model.
>
> Secondly I tried to use many elasticBeamColumns element with low E and high
> A and I, set to equalDOF 1 and 2 at the base of the column and the end of
> the rigid link, but these elements deformed and weren't able to remain
> straight like I want them to (like a truss).
>
> Essentially I am wanting to use the truss element as a LVDT. What are some
> solutions to this?
>
> Secondly, what are some ways to verify my column has truly failed at this
> point? I am getting warnings at MANY elements at the point of no
> convergence, not just the ones in the column in question. No matter how I
> change my column discretization, solution scheme, fibers, iterations, etc,
> this is the case. Should I just use moment curvature example from the wiki
> and vary my axial load to piece together the P-M interaction curve for my
> column?
yeah, same question. corotational geometric trasf. considers only equilibrium in the
deformed configuration (coord), but doesn't include P-small delta, does it?