Author Topic: existing event file format  (Read 16434 times)

azing

  • Newbie
  • *
  • Posts: 33
    • View Profile
existing event file format
« on: April 06, 2021, 05:10:56 PM »
Hi,

What is an example of a valid existing wind event format?

Thank you,
Azin

fmk

  • Administrator
  • Full Member
  • *****
  • Posts: 233
    • View Profile
Re: existing event file format
« Reply #1 on: April 06, 2021, 09:19:21 PM »
I attach an example. This was generated from WE-UQ using the stochastinc option for a 2 story building. It is in JSON format, basically there is an EVENT object that defines some patterns and time series, a pattern for each floor for each dof (the example only provides for dof 1). If you understand OpenSees the idea of a time series and load pattern should be familiar. If you need more keep posting.

ps. when you run the examples locally if you go to your Documents folder there will be a new directory WE-UQ. thta dir wil contain a file debug.out we use to track down issues and another folder tmp.SimCenter. In there there will be. folder templatedir, some dakota files and if it ran succesfully a whole bunch of workdir.X directories. You can go into any of those directoris and have a look atthe files. The workflowdriver file is the series of commands that is run for each sample. If you ever have problems, you can run that script to see the errors. I like to sometimes look at the .tcl file here to see what the errors are.

azing

  • Newbie
  • *
  • Posts: 33
    • View Profile
Re: existing event file format
« Reply #2 on: April 07, 2021, 05:20:23 AM »
This is great, thank you. Is the "StaticWindLoad" defined within each load pattern a load coefficient that will be applied to the data points in the corresponding time series?

fmk

  • Administrator
  • Full Member
  • *****
  • Posts: 233
    • View Profile
Re: existing event file format
« Reply #3 on: April 08, 2021, 10:04:55 PM »
it should be. it should be an initial load applied under static analysis before the dynamic analysis begins so the building does not get a sudden impact load. it is that way in current code base; not 100% sure it is in what you have. we are releasing a new version of this app (should have been done this week but late) .. you can check the example.tcl file in a workdir.X dir to see how it is currently being handled.

azing

  • Newbie
  • *
  • Posts: 33
    • View Profile
Re: existing event file format
« Reply #4 on: April 10, 2021, 04:17:40 AM »
Well that is great! I didn't know about that, so I have added a ramp at the beginning of my loading histories to prevent sudden dynamic impact.

I couldn't find anything in the example.tcl file about "StaticWindLoad". I am just wondering what is the timeSeries type using which this load pattern is applied to the structure? Is it a linear timeSeries to create a ramp? If that is the case, then it's so nice since I don't have to add the ramp manually to the loading histories.

fmk

  • Administrator
  • Full Member
  • *****
  • Posts: 233
    • View Profile
Re: existing event file format
« Reply #5 on: April 12, 2021, 06:49:10 PM »
if you look at the example.tcl script, you see the first analysis step after the wind loading is created is not actually a transient step, but a static one .. which is why no ramp was needed

azing

  • Newbie
  • *
  • Posts: 33
    • View Profile
Re: existing event file format
« Reply #6 on: April 14, 2021, 07:09:21 AM »
ok thank you, I saw the static analysis parameters in the example.tcl file now. But I don't understand what this means:

integrator LoadControl 0.0
analyze 1

This is an OpenSees question, I guess. But I don't know what a load factor increment of zero does. Does this take the first input in the wind loading time series and apply that statically to the building? If yes, then the value assigned to "staticWindLoad" in the event.jason file is unused.

fmk

  • Administrator
  • Full Member
  • *****
  • Posts: 233
    • View Profile
Re: existing event file format
« Reply #7 on: April 17, 2021, 04:55:35 AM »
in an analysis step it increments the pseudo time by 0 .. in opensees the applied load is given by the load factor obtained from the time series for the current time multiplied by the reference loads in the load pattern .. so for this case it uses the very first value in the time series as the loadFactor.