Problem with zerolength element recorders

[mugekuleli]

Hi Dr. Terzic,

I am having a problem with getting the hysteresis of a Hyperbolic element. I have defined the material and recorders like this:

-------------------------------------------------------------------------------------------
node 1 0 10.7
node 21 87.2 10.7

node 27 0 10.7
node 28 87.2 10.7

set Kmax 24000000; # The initial tangent stiffness [N/m/m]
set Kur 24000000; # Unloading/reloading stiffness[N/m/m]
set Rf 0.8; # Failure ratio [dimensionless]
set Fult -550000; # Ultimate (maximum) passive resistance [N/m]
set gap -0.0254; # Initial gap [m]

uniaxialMaterial HyperbolicGapMaterial 1 $Kmax $Kur $Rf $Fult $gap

element zeroLength 26 27 1 -mat 1 -dir 1
element zeroLength 27 21 28 -mat 1 -dir 1

# Bondary Conditions

fix 1 0 1 0
fix 21 0 1 0
fix 27 1 1 1
fix 28 1 1 1

#recorders

recorder Element -file Output/LeftAbutmForceHGM1.out -time -ele 27 force;
recorder Element -file Output/LeftAbutmDefHGM1.out -time -ele 27 deformation;

---------------------------------------------------------------------------

I conducted time history analysis. However, Force output is all zero.

So, i couldn't achieve the hysteresis curve. I suppose I don't have any mistake in defining the material and element, because I solved the eigenvalues and they are correct.
Which recorders should I use to obtain hysteresis curve? Thank you.

[vesna]

Your recorders are good. There may be a bug related to Hyperbolic material.

Why don't you test a hyperbolic material on a model with zeroLength element only.

[mugekuleli]

Dr. Terzic, I appreciate your quick response. I was able to record the force, after I posted my problem here. But this time the forces are smaller than they should be, where the displacements are correct.
For instance, I recorded a peak force about 500 kN, but it should be about 7500 kN.
So, I went ahead and checked my units, but I couldn't find any problem. Also given I found the correct eigenvalues and displacements, there shouldn't be a problem in units.
I recorded the forces utilizing -xlm to see if I'm recording the correct force component and they are correct too. There are force recordings in axial force columns of the output, and all others are zero as they should be.
I didn't understand what I'm doing wrong or what the problem might be.

[mugekuleli]

Could this be an issue of the units in recorder?

Because the ultimate passive resistance force [Fult] unit in the input of hyperbolicgapmaterial is in terms of [force/m] which means [per meter of abutment width] as indicated in the following page:
http://opensees.berkeley.edu/wiki/index ... p_Material

Is it possible that the recorded force output unit for this material is in terms of [force per meter of abutment width]?

[vesna]

As I previously sad, test Hyperbolic material independently from your model to see if it is working as supposed. The problem may be there.

[mugekuleli]

Dr. Vesna,

Thank you for your reply.

I ran a pushover analysis on the Hyperbolic material independently. It turned out with the following error:

WARNING bandgenlinlapacksolver::solve() -LAPACK routine returned 1
DisplacementControl::newStep<void> - failed in solver
StaticAnalysis::analyze() - the Integrator A failed at iteration: 0 with domain at load factor 0.1
analyze failed, returned: -2 error flag

Then I went ahead and ran an eigenvalue analysis to check the model with some mass applied on node 1 with the following command: "eigen -fullGenLapack 1". There were no error messages and the eigenvalue turned out 0.

I seems like the stiffness cannot be initiated somehow. What would you suggest on this issue?

Also, if you also agree that there is some kind of bug with the use of this material in version 2.4.0, is there anyway that I can obtain some earlier version for immediate use of this material until the bug is fixed?

Thank you very much.

[vesna]

Would you post your model?

[mugekuleli]

# HyperbolicGapMaterial Pushover
# units kg - m - N

wipe; # clear opensees model
model basic -ndm 2 -ndf 3; # 2 dimensions, 3 dof per node

# Nodal Coordinates
# node $nodeTag (ndm $coords) <-mass (ndf $MassValues)>

#Left Deck

node 1 0 11.6

# Nodes For Zerolength Element Definition

node 27 0 11.6

# Transformation
# The OpenSees Geometric Transformation Command defines how the element coordinates correlate to the global model coordinates.
# In a 2D problem, element orientation does not need to be considered, and can be the same for all elements.
# The linear transformation will be used in this demonstration:
#geomTransf Linear $transfTag <-jntOffset $dXi $dYi $dXj $dYj>

geomTransf Linear 1

#Elements

# Abutment

set Kmax 24000000; # The initial tangent stiffness [N/m/m]
set Kur 24000000; # Unloading/reloading stiffness[N/m/m]
set Rf 0.8; # Failure ratio [dimensionless]
set Fult -550000; # Ultimate (maximum) passive resistance [N/m]
set gap -0.0254; # Initial gap [m]

uniaxialMaterial HyperbolicGapMaterial 1 $Kmax $Kur $Rf $Fult $gap

element zeroLength 26 1 27 -mat 1 -dir 1 -doRayleigh 1

# Bondary Conditions
# fix $nodeTag (ndf $ConstrValues)
# DX DY RZ

fix 1 0 1 1
fix 27 1 1 1

# Define RECORDERS ------------------

recorder Element -file Output/HGMForce26.out -time -ele 26 force;
recorder Element -file Output/HGMDef26.out -time -ele 26 deformation;
recorder Node -file Output/HGMDisp1.out -time -node 1 -dof 1 disp;

# Pushover

pattern Plain 2 Linear {
load 1 1000 0.0 0.0
}

# displacement parameters
set IDctrlNode 1; # node where disp is read for disp control
set IDctrlDOF 1; # degree of freedom read for disp control (1 = x displacement)
set Dmax 0.15; # maximum displacement of pushover: drift
set Dincr [expr 0.005]; # displacement increment

# analysis commands
constraints Plain; # how it handles boundary conditions
numberer RCM; # renumber dof's to minimize band-width (optimization)
system BandGeneral; # how to store and solve the system of equations in the analysis (large model: try UmfPack)
test NormUnbalance 1.0e-6 400; # tolerance, max iterations
algorithm Newton; # use Newton's solution algorithm: updates tangent stiffness at every iteration
integrator DisplacementControl $IDctrlNode $IDctrlDOF $Dincr; # use displacement-controlled analysis
analysis Static; # define type of analysis: static for pushover
set Nsteps [expr int($Dmax/$Dincr)];# number of pushover analysis steps
set ok [analyze $Nsteps]; # this will return zero if no convergence problems were encountered
puts "Pushover complete"; # display this message in the command window

wipe all;

[mugekuleli]

# HyperbolicGapMaterial Pushover
# units kg - m - N

wipe; # clear opensees model
model basic -ndm 2 -ndf 3; # 2 dimensions, 3 dof per node

# Nodal Coordinates
# node $nodeTag (ndm $coords) <-mass (ndf $MassValues)>

#Left Deck

node 1 0 11.6

# Nodes For Zerolength Element Definition

node 27 0 11.6

# Mass

mass 1 4000.6 0 0

# Transformation
# The OpenSees Geometric Transformation Command defines how the element coordinates correlate to the global model coordinates.
# In a 2D problem, element orientation does not need to be considered, and can be the same for all elements.
# The linear transformation will be used in this demonstration:
#geomTransf Linear $transfTag <-jntOffset $dXi $dYi $dXj $dYj>

geomTransf Linear 1

#Elements

# Abutment

set Kmax 24000000; # The initial tangent stiffness [N/m/m]
set Kur 24000000; # Unloading/reloading stiffness[N/m/m]
set Rf 0.8; # Failure ratio [dimensionless]
set Fult -550000; # Ultimate (maximum) passive resistance [N/m]
set gap -0.0254; # Initial gap [m]

uniaxialMaterial HyperbolicGapMaterial 1 $Kmax $Kur $Rf $Fult $gap

element zeroLength 26 1 27 -mat 1 -dir 1 -doRayleigh 1

# Bondary Conditions
# fix $nodeTag (ndf $ConstrValues)
# DX DY RZ

fix 1 0 1 1
fix 27 1 1 1

eigen -fullGenLapack 1

[vesna]

Here is the modified code that works. You need to set up the number of analysis steps to deform it as much as you need. Check if the output is what you expect. Be aware that it is a compression only material and that you need to load it in compression to see the behavior.

# HyperbolicGapMaterial Pushover
# units kg - m - N

wipe; # clear opensees model
model basic -ndm 2 -ndf 3; # 2 dimensions, 3 dof per node

# Nodal Coordinates
# node $nodeTag (ndm $coords) <-mass (ndf $MassValues)>

#Left Deck

node 1 0 11.6

# Nodes For Zerolength Element Definition

node 27 0 11.6

# Transformation
# The OpenSees Geometric Transformation Command defines how the element coordinates correlate to the global model coordinates.
# In a 2D problem, element orientation does not need to be considered, and can be the same for all elements.
# The linear transformation will be used in this demonstration:
#geomTransf Linear $transfTag <-jntOffset $dXi $dYi $dXj $dYj>

geomTransf Linear 1

#Elements

# Abutment

set Kmax 24000000; # The initial tangent stiffness [N/m/m]
set Kur 24000000; # Unloading/reloading stiffness[N/m/m]
set Rf 0.8; # Failure ratio [dimensionless]
set Fult -550000; # Ultimate (maximum) passive resistance [N/m]
set gap -0.0254; # Initial gap [m]

uniaxialMaterial HyperbolicGapMaterial 1 $Kmax $Kur $Rf $Fult $gap

element zeroLength 26 1 27 -mat 1 -dir 1 -doRayleigh 1

# Bondary Conditions
# fix $nodeTag (ndf $ConstrValues)
# DX DY RZ

fix 1 0 1 1
fix 27 1 1 1

# Define RECORDERS ------------------

recorder Element -file Output/HGMForce26.out -time -ele 26 force;
recorder Element -file Output/HGMDef26.out -time -ele 26 deformation;
recorder Node -file Output/HGMDisp1.out -time -node 1 -dof 1 disp;

pattern Plain 1 Linear {
sp 1 1 -1.0
}
constraints Penalty 1.0e14 1.0e14
integrator LoadControl 0.001
numberer Plain
system BandGeneral
#system UmfPack
test NormDispIncr 1.0e-7 100 0
algorithm KrylovNewton
analysis Static

analyze 1000

puts "Analyis is done!"

[vijaykulhar]

for recording data we use command like this
recorder Node -file $dataDir/DFree.out -time -node 4 -dof 1 2 3 disp; # displacements of free node
this records data in .out file.
Now my question is that can we change this .out file to .xlsx (i.e. excel) directly ??

[mugekuleli]

Hi Dr. Terzic,

Sorry for my late answer. I ran the code that you modified, but the element force output (element 26 in the code, which is the spring) is all zero (except the first column, which is the load applied).

P.S.: I was able to find the eigenvalues by utilizing an equivalent EPP material. I guess the eigenvalue problem is caused by the gap in HGM. (It is not possible to define the gap equal to zero in HGM).

[vesna]

Check if the gap is closed. Until you close the gap you will not be able to see any forces. You may need to increase number of analysis steps for the gap to be closed.