FORCE FROM REACTION VS COUNTER REACTIONS. Control of support removal time.

[Sylwester]

If yo do it like this, and in the analysis at point 4 you reset the time to 0, you can simply use the same timeSeries for both load patterns

Assuming I do it this way (with 2 transient analyses)
At point 4 the transient analysis duration is about 5s, how can I control that the second tranisent at point 7 starts when the first transient analysis for example is only halfway through about 2s.

What I mean by this. By removing the first column, the structure goes into vibration and the long analysis time allows to track after what time the structure stabilises.

‘Switching on’ the second analysis (removal of the second column) after a certain time (the time during which the structure is not fully stabilised (the 2s) will allow to follow the asymmetric response, which will depend on the dynamic properties of the structure.

[STKO Team]

In OpenSees, any analysis keeps track of what happened to the previous ones.
For example, performing one analysis that lasts 4 seconds is the same as performing two subsequent analyses with a duration of 2 seconds each.

Just explain in detail the time at which each column starts failing, and the duration of the reduction of the counter-reaction and we will explain how to set it up

[Sylwester]

Just explain in detail the time at which each column starts failing, and the duration of the reduction of the counter-reaction and we will explain how to set it up

I modified your model according to the algorithm with 2 analyses.

Assuming that the second column is removed after 2s the first one, the first transient analysis duration is 2s.

In the first tranisient analysis i keep the load costatns but idont use pseudo time (In recorder I want deflections within 5 s interval).

Then I remove the second column and in the second analysis I give a time of 3s so that the total time is 5s.
Can you please check if this way is correct?

stko modified.rar

(125.7 KiB) Downloaded 95 times

[STKO Team]

Try this:

stko modified.zip

(184.68 KiB) Downloaded 94 times

There are 2 path time series:

1. first column, removal starts at t=0 and finishes at t=0.35
2. second column, removal starts at t=2 and finishes at t=2.35

The first analysis doesn't need the loadConst because at t=2, the first time series is anyway at lambda=0
Don't reset the time to zero because the second time series already accounts for the total time

[Sylwester]

Try this:

Thanks !!!
Best wishes!!!

[STKO Team]

You're welcome

[Sylwester]

Actually, it is enough to just remove the support. However in this case you cannot control the duration of the removal (it will coincide with the time step).

The alternative is to use the force-from-reaction to automatically compute the reaction of the removed constraint. Then you can give it a Path time series that goes from 1 (100% of the reaction) to 0 (full removal) over a given time step.

If you want to do it parametrically, you can use the Python API in STKO (if you don't know how to use it, please watch the webinar #20 Intro to STKO v2.0's Python API.

Here's a file with the second approach:
MODEL 1 FORCE FROM REACTION.zip

Here's a python script that runs your analysis parametrically:
ParametricRemovalTime.zip

Here's what you get:
param.png

DEAR STKO,

The file you shared with me parametricremovaltime works well for beams. The problem I have is that even for a single span slab, setting tr times from 0.001s to 3s in 0.005s increments causes opensees to crash and shut down. At the same time, 100 analyses are run.
I have to divide my analyses into tr max intervals every 0.4s.

I tried swapping the script so that each successive analysis starts only when one analysis comes to a complete end. Unfortunately, each successive anlize only starts when I manually close STKOMONITOR. I tried to set automatic closing of the application but failed.

On the other hand I wanted to use openseesmp but the script you gave me only works for one processor.

Is it possible for me to get this script so that I can use openssesmp ? Preferably in such a way that each subsequent analysis is only carried out when the previous one finishes.

One more request, can you modify the script so that it closes the opensees solver windows automatically ? If at the same time performs 100 analyses I have a black screen of running solvers in addition, some analyses are not completed because the file shuts down sooner.

Best regards

[STKO Team]

Try this approach:

run_analysis_pool.zip

(1.25 KiB) Downloaded 84 times

In this example, I use the queue to run at most 4 jobs at a time of 10 total jobs.