column base stiffness

[sandhya]

Hi, my analysis runs fine when the columns of the bridge are kept fixed. But, when I am having convergence issues when I try to introduce linear springs along transverse and long directions (partial fixity)
is this the right way to introduce base springs?
node 7 $l_p6 -$height_p6 0; #master node
node 1000 $l_p6 -$height_p6 0;
fix 1000 1 1 1 1 1 1
uniaxialMaterial Elastic 1111 [expr 500*$kip/$in]
element zeroLength 1112 7 1000 -mat 1111 -dir 3

I get a warning that says "failed to converge but going on"

[fmk]

yes. the failed to converge going on is coming from the script or choice of convergence test flag. it is sometimes ok to fail in one step and go to the next step, just look carefully at the results for that step and the next few to make sure you don't deviate too far from the solution path.

if you want to track down the error further, i suggest a very stiff elastic material to simulate fixed condition in all 3 directions to start. reduce these in the indiviual direction till you find out what the problem is.

[sandhya]

I tried giving a very high stiffness value for the restrained directions. I see that it converges without any hitch for some ground motions and fails for a few others:
I have also tried increasing the number of elements (displacement based) in the columns- again, convergence for some and failure for some other earthquake motions.
#--------GROUND MOTION ANALYSIS STARTED--------#
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 98.4304 (max: 1e-008)
Norm deltaX: 0.122172, Norm deltaR: 93089
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 369.597 (max: 1e-008)
Norm deltaX: 0.42552, Norm deltaR: 145865
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 1688.96 (max: 1e-008)
Norm deltaX: 0.991045, Norm deltaR: 192799
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 1725.97 (max: 1e-008)
Norm deltaX: 1.07352, Norm deltaR: 155083
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 7170.58 (max: 1e-008)
Norm deltaX: 2.48504, Norm deltaR: 304393
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 32236 (max: 1e-008)
Norm deltaX: 4.3365, Norm deltaR: 559782
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 174232 (max: 1e-008)
Norm deltaX: 12.6396, Norm deltaR: 993063
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 61413.1 (max: 1e-008)
Norm deltaX: 5.07976, Norm deltaR: 557131
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 38130.6 (max: 1e-008)
Norm deltaX: 3.38694, Norm deltaR: 511173
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 42798.4 (max: 1e-008)
Norm deltaX: 7.10244, Norm deltaR: 585174
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 75955.1 (max: 1e-008)
Norm deltaX: 8.71828, Norm deltaR: 634266
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 49313.6 (max: 1e-008)
Norm deltaX: 5.07063, Norm deltaR: 599509
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 62716.5 (max: 1e-008)
Norm deltaX: 3.7195, Norm deltaR: 477698
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 1.68051e+149 (max: 1e-008)
Norm deltaX: 1.5095e+073, Norm deltaR: 2.28409e+077
WARNING: CTestEnergyIncr::test() - failed to converge but goin on - current Ener
gyIncr: 1.#QNAN (max: 1e-008)
Norm deltaX: 1.#QNAN, Norm deltaR: 1.#QNAN
WARNING SuperLU::solve(void)- Error 7 returned in factorization dgstrf()
WARNING ModifiedNewton::solveCurrentStep() -the LinearSysOfEqn failed in solve()

DirectIntegrationAnalysis::analyze() - the Algorithm failed at time 2.64

[fmk]

you obviously need to modify the script .. that error is not going away on the next step, but problems are getiing worse .. the very large unbalance can be an indiation that you went too large on the large stiffness.

anyways, you might want to remove the just go on part of the script and try things if the analysis fails, e.g.

http://opensees.berkeley.edu/wiki/index ... e_Analysis