PlainHandler warning

[teymoorian]

Hi everyone
I modeled a 3-story building. each floor has 16 nodes which are the intersections of 4x4 resisting frames. I used "RigidDiaphragm" command for nodes in same level. but opensees returns this warning: "WARNING PlainHandler::handle() - constraint matrix not identity, ignoring constraint for node 121
non-varyng assumed ". may anybody tell me what's the problem?

[fmk]

you need to use a different constraint handler .. try Transformation .. the Plain constraint handler cannot deal with a constraint matrix that is not equal to the identity matrix.

[teymoorian]

Thanks alot fmk, I changed the constraint handler to Transformation and System command to SparseSYM. The opensees doesn't reply any problem, but I don't trust the results. something is very strange. could you please tell me how it is possible that two nodes with same distance from "master node" in "RigidDiaphram" command can have different displacements?

[teymoorian]

I really hope that somebody gives me a clue. how it is possible that two nodes with same distance from "master node" in "RigidDiaphram" command can have different displacements?

[fmk]

use UmfPack as opposed to SparseSYM, it handles poorly conditioned matrix better .. the difference in displacements would be due to the rotation of the master node .. fix the master from rotating and they should be the same.

[teymoorian]

the point is that the structure is asymmetric and master node rotation same as roof rotation is very important. you have mentioned before that "the displacements are determined using the typical relations for a rigid diaphragm .. the matrix relating dof's is something like:

1 0 -dY
0 1 dX
0 0 1

where dX and dY are difference in X and Y coordinates and the matrix does not change over time. so for example displacement at constrained node in x dirn equals displ in x at retained dof added to the product of the retained nodes rotation and -dY, dY being the difference in y dirn coordinates."
I am recording two nodes displacements in X direction which have same distance in Y rotation from master node. according to procedure above both nodes should have same displacements in X dir, shouldn't they?

[fmk]

they should .. check your input if not getting the same results .. for example example 5.1.tcl

http://opensees.berkeley.edu/WebSVN/log ... 78&peg=178


Node: 5
Coordinates : -120 120 144
Disps: -9.39655 -10.1474 -0.961455 0.00577853 -0.00453376 3.1871e-06
Velocities : -4.3347 0.424389 0.0279332 -0.00119562 -0.00421446 -2.37622e-05
commitAccels: 65.2757 52.2248 5508.21 0.773226 -17.7734 0.000256999
unbalanced Load: 0 0 -129.6 0 0 0
ID : 0 1 2 12 13 14


Node: 6
Coordinates : 120 120 144
Disps: -9.39655 -10.1466 -0.765492 0.00577323 -0.0050397 3.1871e-06
Velocities : -4.3347 0.418686 0.0609142 -0.00119815 -0.00420065 -2.37622e-05
commitAccels: 65.2757 52.2865 6717.35 1.19867 18.6162 0.000256999
unbalanced Load: 0 0 -129.6 0 0 0
ID : 3 4 5 12 13 14


Node: 7
Coordinates : 120 -120 144
Disps: -9.39578 -10.1466 -0.995415 0.00477596 -0.00503292 3.1871e-06
Velocities : -4.3404 0.418686 -0.0244045 -0.000911622 -0.00420293 -2.37622e-05
commitAccels: 65.3374 52.2865 4780.2 0.189915 18.0232 0.000256999
unbalanced Load: 0 0 -129.6 0 0 0
ID : 6 7 8 12 13 14

[teymoorian]

yes, that's right fmk. it works correctly.
I face a new problem each step. I fixed "rigiddiaphragm command" master node's 6th dof but I couldn't record its reaction. recorder returns 0 value!!! is there a way to obtain the force imposed to this dof to be fixed?

[fmk]

no. other than usinga penalty approach or lagrange multiplier approach and looking at the forces.