SimCenter Forum

Dear all,

I am working on a large bridge model in OpenSees, and to conduct time-history analysis I use OpenSeesSP but with only one core (I cannot use more than one core due to some issues in OpenSees source which has not been addressed). However, since in OpenSeesSP, I can switch to the Mumps system the analysis is much faster than those performed with regular OpenSees.

Now I would like to know if there is any way to substitute the regular OpenSees in the QuoFEM directory with OpenSeesSP and do the forward propagation analysis with the QuoFEM user interface. I know the OpenSees executable file in the QuoFEM directory can be easily replaced by OpenSeesSP executable, but I do not know how can I change the commands to call OpenSeesSP and define the MPI executable and number of processors. (i.e. OpenSeesSP needs to be executed with the following command line: mpiexec -np numProcs? OpenSeesSP model.tcl )

I would be thankful if you can provide me with some details about how to apply this procedure if it is feasible.


Best regards,

Hello all,

I am getting this error when I use QuoFEM for my Opensees model subjected to a ground motion input to do forward propagation analysis.

"Error(s) encountered reading results file workdir.32\results.out for Evaluation 32:-- Expected 1 function value(s) but found 0."

I checked the workdir folders and the analysis converges to the end for all the samples, but for some reason, QueFEM is unable to generate the results.out correctly and the file is empty. I generated the postproc.tcl myself and it only needs two input files from Opensees which are correctly generated and available in workdir folder. If I run the postproc manually it gives me the correct results.out file with no issue, so I am confused why I am getting this error. I checked the opt.out file to and I saw the following error when the postproc file is running by QuoFEM:

missing operand at _@_
in expression "(0.498816-_@_)"
    (parsing expression "(0.498816-)")
    invoked from within
"expr ($TopDisp($i)-$BaseDisp($i))"
    ("for" body line 2)
    invoked from within
"for {set i 1} {$i <= [expr ($m-2)]} {incr i} {
    set RelativeDisp [expr ($TopDisp($i)-$BaseDisp($i))]
   set RelativeDisp [expr abs($RelativeDisp)]
   i..."
    (file "TowerEQPostProc.tcl" line 24)
    invoked from within
"source TowerEQPostProc.tcl"
    (file "SimCenterInput.tcl" line 7)

Since both of the required input files for this postproc are available in the directory and both have 5000 data points, I am not sure why there is an error in reading them. Surprisingly, if I use other ground motions sometimes I get the same error and sometimes it works perfectly and gives me the results. To me, it seems there is an issue by QuoFEM when it calls the postproc. In other words, it seems the postproc is called before the Opensees analysis is over and as a result, the data is not available. However, this is only my guess and I might be wrong. Would you please help me to find the source of this problem and fix it?

Thank you so much,

 

Dear all,

I am working on a long-span 3D bridge model and since the model is large, I prefer to use OpenSees SP as an interpreter to run the analysis with quoFEM. Is there any way that I can do it using my local machine or Design-Safe supercomputers?


Best,