asdeasoft.net

Hi everyone,
I'm trying to use the corotTruss element and I wanted to know if the procedure I follow is correct because in the analysis I have an error. After building the model, I generated FIBER sections and assigned them to the elements. Then I created the element truss_corot_elements and assigned it to the elements of my structure. Is the procedure correct? Why launching a pushover analysis I get this error:


WARNING BandGenLinLapackSolver::solve() -factorization failed, matrix singular U(i,i) = 0, i= 2
WARNING NewtonRaphson::solveCurrentStep() -the LinearSysOfEqn failed in solve()
StaticAnalysis::analyze() - the Algorithm failed at iteration: 0 with domain at load factor 1
OpenSees > analyze failed, returned: -3 error flag
Initial displacement at node 43 - dof 3
TOTAL DURATION: 1.0
MPCORecorder warning: iterative guess of num fibers: reached maximum number of iteration, giving up...

thanks all

These errors are saying that your model is singular, in particular the 3rd DOF as 0 in the diagonal of the stiffness matrix that cannot be factorized.

WARNING BandGenLinLapackSolver::solve() -factorization failed, matrix singular U(i,i) = 0, i= 2
WARNING NewtonRaphson::solveCurrentStep() -the LinearSysOfEqn failed in solve()
StaticAnalysis::analyze() - the Algorithm failed at iteration: 0 with domain at load factor 1
OpenSees > analyze failed, returned: -3 error flag

The last line is a warning coming from the MPCO Recorder

MPCORecorder warning: iterative guess of num fibers: reached maximum number of iteration, giving up...

This happens when you use a fiber cross section with a truss or corotTruss. In this case STKO generates trussSection or corotTrussSection elements, that are the same as truss or corotTruss, but instead of using an area and an uniaxial, they use a fiber cross section. Unfortunately those elements have a small bug in the report of the results, so that the MPCO recorder cannot understand how many fibers are there.

You can simply solve this problem by using a uniaxial material + Area, by the way, the truss does not need bending.

However the first issue of singularity of the stiffness matrix is not related to the last errror.

OK, but with corottruss elements is it possible to grasp the geometric and material nonlinear behavior with OpenSees without using the STKO postprocessor? Thank you

If you refer to the possibility of using trussSectiona and corotTrussSection instead of truss and corotTruss, you can do it in both STKO and manually in OpenSEES.

The problem is that if you use a section instead of a uniaxial material + area, you won't be able to see the output on fibers because of a bug in the way OpenSEES gives back information about fibers in a trussSection.

We identified the problem in OpenSEES, so we are going to fix that issue and release a modified version of OpenSEES. However since you are from an italian institution you cannot download our version of OpenSEES, so if you want we can tell you how do change the source code and you should re.compile it by yourself. If you need it let us know.

I have errors which are saying that your model is singular, when I use truss elements. When I use other than truss elements like elastic elements I am not having this issue.

How can I use trusses in this geometry.

thank you very much.

Hi, the main issue in the model comes from the definition of the custom section. The geometry properties, such as Iz, Iy, alphaY etc, can not be zero, as you defined. Please review your section properties.