Problem with series material in joint2D element

[sirwan]

Hi
Problem regarding Eigen value analysis!
I hove modeled a 3 story-3 bay RCS frame composed of nonlinear elements with fiber sections (steel02, concrete02) for beams and columns and joint2D elements with in series connected uniaxial and hysteretic materials for their central node, the mass and loads are applied on nodes, and the P-delta effects are considered through the leaning column technique.
The problem is When I put Eigen value analysis prior to applying gravity loads the periods (first mode 1.4 sec) are different from that when I put the eigen value analysis after applying gravity loads (firs mode 1 sec!!!!) (no lateral loads are applied), but actually there should be not such difference ( as it is so in all Examples presented by berkely ...). finally regarding the model described above what probable errors in the model may stand for such difference? ( The material? Type of Elements? loads?....)
Thanks for your help and giving time in Advance!

[sirwan]

Hi
Doing more analysis I found out that if I put the Eigen vaue analysis before gravity analysis, it can not grasp the stiffness of Joint2D elements! I mean that what ever I change the parameters of uniaxialMaterials of joint2D element the results of Eigen value analysis would be the same as I have joint2D elements with very small stiffness, unless I do Eigen value analysis after doing gravity load analysis that makes the parameters of joint2D element be included effectively.
However I have constructed joint2D elements completely according to Arash.Al dissertation I don't know if the problem is with joint2D elements or the Eigen value analysis!
I mention that we could define mass for joint2D elements both at central node or top or bottom node, and I have tried both, but the results are about the same.

[sirwan]

Well! Another step forward!
I find out that the problem was with series material used in joint2D elements, because when you delete the series material (Hysteretic plus uniaxialMaterial) and model a joint2D element with only uniaxialMaterial or a hysteretic material for the central node of joint2D element, all the things goes right. finally it sounds that the hysteretic material is not compatible with uniaxialmaterial(steel02) and actually it gives the series material a zero stiffness. now please some help! How I can solve problem with this series material! I put down the codes I used to define it:

#steel02 material parametrs
#story 1
set PSh_matTag 21
set PSh_Mps [expr 3031021936*$N*$mm]; #joint moment strength
set PSh_Kes [expr 1.638390236e12 *$N*$mm]; # joint stiffness
set PSh_b 0.02
set PSh_R0 18.5
set PSh_cR1 0.96
set PSh_cR2 0.15
uniaxialMaterial Steel02 $PSh_matTag $PSh_Mps $PSh_Kes $PSh_b $PSh_R0 $PSh_cR1 $PSh_cR2 0 1 0 1;

#Hystretic Material Parametrs
set H_matTag 25;
set s1p [expr 4564406925*$N*$mm];
set e1p 0.00108;
set s2p [expr 5071563250*$N*$mm]; #joint bearing strength
set e2p 0.05
set s3p [expr 329651613*$N*$mm];
set e3p 0.1
set s1n -$s1p
set e1n -0.00108
set s2n -$s2p
set e2n -0.05
set s3n -$s3p
set e3n -0.1
set pinchX 0.2
set pinchY 0.5
set damage1 0.015
set damage2 0.0
uniaxialMaterial Hysteretic $H_matTag $s1p $e1p $s2p $e2p $s3p $e3p $s1n $e1n $s2n $e2n $s3n $e3n $pinchX $pinchY $damage1 $damage2
set SeriesMat1 105;
uniaxialMaterial Series $SeriesMat1 $H_matTag $PSh_matTag

[kamranmosleh]

Hi,
I have used beamcolumnjoint element but faced the same gordian knot as you,
since your post is for about a year ago, I'm wondering if you have find a solution?
with regards

[baikunthasilwal]

Hi Eveyone,

I have issue with pushover analysis as mentioned below.

When I model a simple one story frame with joint2d element for the beam column joint considering five uniaxial elastic materials, and elastic beam and column elements, I got the reasonable time period performing eigen analysis after having run the gravity analysis. However, I got the following message while running pushover analysis:

-------------------------------------------------------------------------------------
T1 = 0.6803021840219462 s
T2 = 0.07118745616480307 s
Running Pushover...
WARNING: CTestNormUnbalance::test() - failed to converge
after: 4000 iterations
NewtnRaphson::solveCurrentStep() -the ConvergenceTest object failed in test()
StaticAnalysis::analyze() - the Algorithm failed at iteration: 0 with domain at
load factor 0.0181057
OpenSees > analyze failed, returned: -3 error flag
Pushover complete
--------------------------------------------------------------------------------------

Please advice us for improving the analysis for pushover analysis.

[fmk]

what constraint handler are you using? need to be Transformation or Penalty as those MP_Constraints are changing at every time step.

[baikunthasilwal]

Dear Dr. Frank,

Thank you for your quick response.

I use the following script for running pushover analysis.
-------------------------------------------------------------------------------------
# Pushover Analysis #
if {$analysisType == "pushover"} {
puts "Running Pushover..."
set lat2 [expr 0.125*$Floor2Weight/4.0]; # force on each frame node in Floor 2
pattern Plain 200 Linear {
load 1101 $lat2 0.0 0.0;
load 2101 $lat2 0.0 0.0;
load 3101 $lat2 0.0 0.0;
load 3104 $lat2 0.0 0.0;
}
# display deformed shape:
set ViewScale 5;
DisplayModel2D DeformedShape $ViewScale ;

# displacement parameters
set IDctrlNode 1102;
set IDctrlDOF 1;
set Dmax [expr 0.15*$HBuilding];
set Dincr [expr 0.01];

# analysis commands
constraints Penalty 1e12 1e12;
numberer RCM;
system UmfPack;
test NormUnbalance 1.0e-3 4000;
algorithm Newton;
integrator DisplacementControl $IDctrlNode $IDctrlDOF $Dincr;
analysis Static;
set Nsteps [expr int($Dmax/$Dincr)];
set ok [analyze $Nsteps];
puts "Pushover complete";
}
-------------------------------------------------------------------------------------

However, I got some reasonable pushover result while removing joint2d element with modified solution algorithm.

Please advice us for improving the analysis for pushover analysis using joint2d element.

[fmk]

the problem is probably use of Penalty method and a convergence test that uses Norm of the Unbalance (the Penalty method imposes large forces to get the 'stiff' elements it adds to impose the required displacements .. use NormDispIncr instead.

[baikunthasilwal]

Dear Dr. Frank,

I would like to thank you for your valuable insights that works.