The OpenSees Community

Hello everyone!
I am struggling to understand how "loadconst" works. For a static pushover analysis, it is clear to me that you need to set
loadConst -time 0.0
after the application of the gravity loads in order to keep the gravity loads constant during the application of the lateral displacements.

However, I have a different situation. I am trying to model a cyclic loading experiment in which load control and displacement control protocols are alternated. I apply a displacement, after that a force, after that another displacement and so on.
Do I need to apply "loadConst" in any instance of the code in this case? The force never remains constant during the test, but I am afraid I am missing something here, because the results are very far from the experiment. Any help?

This is not related to your question (I think), but I was just taking a look at your code and I noticed that you have both node 1 and 3 fully restrained if I am not mistaken. Your Zero Length Section element seems to be fully restrained, but only the shear deformations should be restrained. Am I right?

What is the best way of modelling a reinforced concrete plate in OpenSees?
I have tested many alternatives, but I am not satisfied with the results. The one that seems to be closer to the experimental results, consists of the following sequence of commands:

nDMaterial PlaneStressUserMaterial ....
nDMaterial PlateFromPlaneStress .... (this requires an out of plane shear modulus which I am not sure how to take)...
uniaxialMaterial Steel01 ....
nDMaterial PlateRebar ....
section LayeredShell .....

and then:
element ShellMITC (or almost equally successful: ShellNLDKGQ ...).

The response seems almost twice as stiff compared to the laboratory results. I have checked geometry and materials, so the problem should not be there.

In this model, I am not sure how to take the out of plane shear modulus, which seems to have some influence on the convergence and the accuracy of results. Any suggestion related to it would be greatly appreciated.

What is your experience with modelling RC plates in OpenSees? What do you think about the sequence of commands above?

Before running the dynamic analysis, I would suggest you to simply try an eigenvalue analysis before gravity loading and after gravity loading (if you are applying gravity loads). See if you find anything strange by comparing the results.

EricsonEncinaZ wrote:
> I can run 3 models at the same time.... I'm able to run 3 models at the same time, and the remaining
> core and RAM is used to keep using the computer.
> Please note that what I am doing is not parallel computing (OpenSeesMP) it
> is just running different instances of OpenSees.exe at the same time.

This is just what I am doing too. Anyway, I guess using OpenSeesMP is a better choice and it is worth trying.

I run multiple models at the same time because they finish in less time than it takes to run them one by one. I don't think I am getting my job done 4 times faster, but it is noticeably faster (never had time to actually measure though).

Thank you very much ismailqeshta for your reply and for the link. I read the discussion in researchgate.net, but it doesn't quite answer my question. I am not sure if I understand the terminology you are using. Let's say I have a 1 storey frame, a 3 storey frame, a 5 storey frame and a 7 storey frame. What I did was to run opensees.exe and upload the source tcl file for the 1 storey frame. The analysis started to run. In the meantime, I ran opensees.exe again and started a new analysis (3 storey frame) and repeated this for all four frames. What would be a better alternative? Is this damaging the computer in any way? Let's assume that the analysis duration is acceptable for me (so, I would change what I am doing only if this is harmful or could lead to erroneous results).

Hello! First of all, I apologize if this is not the appropriate forum for my question, but I thought there was no better option. Secondly, I know that I am not benefiting from the full potential of OpenSees and I probably should immidiately start using "parallel processing". Here is the deal: I ran 4 different analyses (simultaneously) in four instances of Opensees.exe. Two of the instances were writting output on the same folder, but on different output files. I noticed that with 3 analyses the CPU was working at 70%, so I thought it could handle one more analysis. With 4 instances of opensees.exe, the CPU was running at 100%. Everything seemed to work fine. The analyses were completed successfully in less time than it would have taken to run them one by one. This is sufficiently good for me (for now). However, the next day I started having problems with my computer. More specifically, I think I am having problems with the harddisk drive.
Any opinions on my situation? Could the damage to the harddisk be related to the way I am using Opensees?

Thank you very much for your reply. I added the equalDOF line and it works now.

Hello!
I was wondering if I could use the PinchingLimitStateMaterial in a non-conventional way. Can I assign a PinchingLimitStateMaterial to a spring that is located nowhere near the element in which I am monitoring the drifts? For example:

http://imgur.com/SW4cqtZ

Here is a piece of code I was trying to use. This doesn't work. The analysis fails even for gravity loading. Do you have any idea?

model BasicBuilder -ndm 2 -ndf 3

node 1 0.0 0.0
node 2 0.0 1000.0
node 3 0.0 2000.0
node 4 300.0 1000.0
node 5 300.0 1000.0
node 6 1300.0 1000.0

geomTransf Linear 1

#vertical elements

#element elasticBeamColumn $eleTag $iNode $jNode $A $E $Iz $transfTag <-mass $massDens> <-cMass>
element elasticBeamColumn 1 1 2 [expr 550.0*550.0] 35000.0 [expr 550.0*550.0*550.0*550.0/12] 1
element elasticBeamColumn 2 2 3 [expr 550.0*550.0] 35000.0 [expr 550.0*550.0*550.0*550.0/12] 1


# horizontal elements
element elasticBeamColumn 3 2 4 [expr 550.0*550.0] 35000.0 [expr 550.0*550.0*550.0*550.0/12] 1
element elasticBeamColumn 4 5 6 [expr 550.0*500.0] 35000.0 [expr 550.0*550.0*550.0*550.0/12] 1

#limitCurve RotationShearCurve $crvTag $eleTag $ndI $ndJ $rotAxis $Vn $Vr $Kdeg $rotLim
limitCurve RotationShearCurve 60 2 3 2 3 60000000.0 -1 -10.0 0.1

#uniaxialMaterial PinchingLimitStateMaterial $matTag $nodeT $nodeB $driftAxis $Kelas $crvTyp $crvTag $YpinchUPN $YpinchRPN $XpinchRPN $YpinchUNP $YpinchRNP $XpinchRNP $dmgStrsLimE $dmgDispMax $dmgE1 $dmgE2 $dmgE3 $dmgE4 $dmgELim $dmgR1 $dmgR2 $dmgR3 $dmgR4 $dmgRLim $dmgRCyc $dmgS1 $dmgS2 $dmgS3 $dmgS4 $dmgSLim $dmgSCyc
uniaxialMaterial PinchingLimitStateMaterial 511 3 2 1 500000000000000000000.0 2 60 0.3 0.2 0.0 0.3 0.2 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3

element zeroLength 88 4 5 -mat 511 -dir 3

fix 1 1 1 1

... the rest is just the analysis and recorders

Thanks a lot for sharing this.

I tried some options, but not all. What options would you suggest for this situation, for a time history analysis?

selimgunay wrote:
> You can use elastic-perfectly plastic material with a large enough initial
> stiffness


That's what I have been trying to do. However, I can't find a large enough initial stiffness for which both a) no numerical errors occur and b) the stiffness is large enough. Any ideas?

Hi everyone!
I am trying to model a rotational spring that has a moment rotation curve as follows:
first point: rotation=0, moment=0
second point: rotation=0, moment=M; (this is the tricky part)
third point: rotation=x, moment = M.
other points: (...not important).

Does anyone know how to do this? Ideally I would like to use the Pinching4 uniaxial material, but if you have the answer for some other material, please let me know. I have tried defining the curve using the exact values (as described above) but it doesn't work (no convergence). I tried with small value of rotation for the second point and it still doesn't work. Any idea?