timeseries integrator

Forum for OpenSees users to post questions, comments, etc. on the use of the OpenSees interpreter, OpenSees.exe

Moderators: silvia, selimgunay, Moderators

Post Reply
elisa silvestri
Posts: 5
Joined: Sat Jan 29, 2011 11:39 am
Location: padova

timeseries integrator

Post by elisa silvestri »

Hi everyone!

I have a question,if it's possible: how can I write the timeseries integrator in the load pattern.tcl file?
The fact is that I have only the acceleration's file for the plain ground motion and I would integrate the velocities and the displacements from the accelerations.

Thank you so much.
Bye.
vesna
Posts: 3033
Joined: Tue May 23, 2006 11:23 am
Location: UC Berkeley

Re: timeseries integrator

Post by vesna »

Would you explain your problem again? I'm not sure I understand it right.
elisa silvestri
Posts: 5
Joined: Sat Jan 29, 2011 11:39 am
Location: padova

Re: timeseries integrator

Post by elisa silvestri »

I'm studying for my thesis a bridge, modeled with pile foundations.
To simulate the earthquake motion, I used the load pattern "multiple support" (with "plan ground motion" and "imposed motion SP"). The command "plan ground motion" requires three series: one for displacements, one for velocities and one for accelerations. I have only the accelerations' series and the Opensees' manual says:

"<-int (IntegratorType intArgs)>
If only the acceleration record is specified, the user has the option of
specifying the TimeSeriesIntegrator (page 333) that is to be used to
integrate the acceleration record to determine the velocity and
displacement record (optional, default: Trapezoidal)"

but I don't understand how to introduce the integrator in the files .tcl (for example in the load pattern file).

I hope I'm explaining better.

Thank you.
vesna
Posts: 3033
Joined: Tue May 23, 2006 11:23 am
Location: UC Berkeley

Re: timeseries integrator

Post by vesna »

Thanks for the explanation.

You do not have to specify integrator. Define acceleration and the opensees will integrate to get velocities and displacements using trapezoidal method. However, this is not the best way to get displacements.

It would be better if you calculate displacements using MatLab accounting for base line correction (opensees does not do it).
elisa silvestri
Posts: 5
Joined: Sat Jan 29, 2011 11:39 am
Location: padova

Re: timeseries integrator

Post by elisa silvestri »

Thank you very much!!

I'll try to do it with matlab, I hope I'll be able to do it..

Anyway, I tried to do it with the time series integrator in this way:

(one of the files "load pattern":

# LoadPattern_3.tcl

# LoadPattern "SismaLong01g": patternTag
pattern MultipleSupport 3 {
# GroundMotion gmTag Type TimeSeries
groundMotion 1 Plain -accel $AccelerogrammaLong01g

# ImposedMotion nodeTag dofTag gmTag
imposedMotion 1 1 1
imposedMotion 3 1 1
imposedMotion 29 1 1
imposedMotion 31 1 1
imposedMotion 57 1 1
imposedMotion 59 1 1
imposedMotion 105 1 1
imposedMotion 106 1 1
imposedMotion 108 1 1
imposedMotion 107 1 1
imposedMotion 109 1 1
imposedMotion 110 1 1
imposedMotion 111 1 1
imposedMotion 112 1 1
imposedMotion 521 1 1
imposedMotion 522 1 1
imposedMotion 523 1 1
imposedMotion 524 1 1
imposedMotion 525 1 1
imposedMotion 526 1 1
imposedMotion 497 1 1
imposedMotion 498 1 1
imposedMotion 499 1 1
imposedMotion 500 1 1
imposedMotion 501 1 1
imposedMotion 502 1 1
imposedMotion 503 1 1
imposedMotion 504 1 1
imposedMotion 505 1 1
imposedMotion 506 1 1
imposedMotion 507 1 1
imposedMotion 508 1 1
}

and the warning in the results is:

WARNING: GroundMotion::getDisp(double time) - integration is required to get the ground displacements from the ground velocities
WARNING: GroundMotion::getDisp(double time) - integration is required to get the ground displacements from the ground velocities
WARNING: GroundMotion::getDisp(double time) - integration is required to get the ground displacements from the ground velocities
WARNING BandGenLinLapackSolver::solve() -LAPACK routine returned 2275
WARNING NewtonLineSearch::solveCurrentStep() -the LinearSysOfEqn failed in solve()
DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 0.01
OpenSees > analyze failed, returned: -3 error flag

Thank you for your courtesy and disponibility.
vesna
Posts: 3033
Joined: Tue May 23, 2006 11:23 am
Location: UC Berkeley

Re: timeseries integrator

Post by vesna »

The problem is not with the time series. You have some error somewhere in your model. After you build your model try to do eigenvalue analysis. This will help you find the error in your model.
Post Reply