Pushover Analysis

[sribeiro]

Dear STKO Team,

I've been trying to do a pushover analysis on the annexed document. Even though I get a curve, its not as smooth as it should be and also the displacement correspondent to 0.8Fmax is too big.

Am I doing the pushover correctly?
Is there a way that I can apply an acceleration proportional to the mass of the structure?

Thanks,
S

[kesavapraba]

Hi, Penalty factors in the analysis settings seem to be high. You might consider using 14 in your case.

[sribeiro]

Hi!

I tried reduce the penalty and not big changes happend.
Am I doing the pushover correctly?

Also, is there a way that I can apply an acceleration proportional to the mass of the structure?

Best,
S

[STKO Team]

Here are the issues:

• Use a smaller tolerance (you are modeling in meters, so 1.0e-5 - 1.0e-6 should be fine)
• Increase the reference load. The value you input is something like 100k smaller than the expected reaction. Use a reference load which is close (in terms of order of magnitude) to what you expect. It will make the lambda (unknown in the Displacement control) more close to 1.0 and more stable. Try to multiple your hload by 1.0e4
• Finally pay attention when recording reaction near multi-point constraints! In your selection set, you included all the fixed nodes, but the 2 nodes at the 2 ends of the bridge are directly attached to a multi-point constraint. Reaction might be also on the other node of the multi-point constraint. To be sure select both of them.

[sribeiro]

Here are the issues:

• Use a smaller tolerance (you are modeling in meters, so 1.0e-5 - 1.0e-6 should be fine)
• Increase the reference load. The value you input is something like 100k smaller than the expected reaction. Use a reference load which is close (in terms of order of magnitude) to what you expect. It will make the lambda (unknown in the Displacement control) more close to 1.0 and more stable. Try to multiple your hload by 1.0e4
• Finally pay attention when recording reaction near multi-point constraints! In your selection set, you included all the fixed nodes, but the 2 nodes at the 2 ends of the bridge are directly attached to a multi-point constraint. Reaction might be also on the other node of the multi-point constraint. To be sure select both of them.

Hi again!
I did the proposed changes (and fixed other mistakes) but I still get some "noise" on my analysis.
As it can be seen on the picture, the curve has two weird "breaks" that I don't understand where they came from.
Also attached the model. I also tried with load control and the curve has the same shape.

Best,
S

[sribeiro]

Also did a TH analysis to "compare" the analysis and I'm quite concerned about the results.
As it can be seen in the image bellow, the displacements and the loads are lower on the TH analysis.
Shouldn't they be similar?

Am I doing any of these analyses wrong?

[STKO Team]

In the selection set "base nodes" you only included the nodes fixed at the ground, and the zeroLength elements at the 2 ends. But not the slave nodes of those zeroLength!
This is how your selection set should look like:

BaseNodes.png (92.98 KiB) Viewed 3181 times

Regarding the time history analysis.
It doesn't match the pushover just because in the monitor in the time history you are monitoring the SUM of the Uy displacments of all the "top nodes" [selection set 6]. You should use instead the "node control" [selection set 5]

[sribeiro]

Regarding the time history analysis.
It doesn't match the pushover just because in the monitor in the time history you are monitoring the SUM of the Uy displacments of all the "top nodes" [selection set 6]. You should use instead the "node control" [selection set 5]

Dear STKO team,

I did that selection change and the values of the results didn't change, I still get around 0.15m displacement on the TH analysis and on the pushover is ~1.3m max. displacement.

[STKO Team]

Sorry but your blue curve is wrong.
It comes from your pushover analysis with the old (wrong) selection set.
Please re-run the pushover (or re-extract the results from the post processor) and you will see that the pushover curve is actually much higher (same stiffness of your dynamic) because previously you were missing some reaction from those nodes

[sribeiro]

Sorry but your blue curve is wrong.
It comes from your pushover analysis with the old (wrong) selection set.
Please re-run the pushover (or re-extract the results from the post processor) and you will see that the pushover curve is actually much higher (same stiffness of your dynamic) because previously you were missing some reaction from those nodes

I actually changed the selection set (only for the dots now) and the results are not as you predict.. I get extremely high reaction values and only a linear response
I attached the stko file