Weird answers from time history analysis!!

[Ameri]

Hi there,

I have already performed a lot of nonlinear dynamic analysis (time history, progressive collapse, etc.) in OpenSees and have never had any problems! However, this time it seems something silly is going on but I cannot figure out what is it ...?!
I have a 3D 6-stoty RC structure - which I have already used in progressive collapse analysis, without any problems! -, and now want to conduct a time history analysis on this building. The problem is that whatever record I use with whatever scale factor, I get EXACTLY THE SAME STORY DRIFTS, which are really small (something like 2.46087e-15). Even the decimal numbers are exactly the same! What I have done so far is that:

- I changed the model mass from -20*actual M to +20*actual M (though the actual M gives me a logical fundamental period, i.e. 0.68 Secs),
- I used different *.txt earthquake files with different scale factors (from 1 to 2000!) + different Path TimeSeries,
- different types of recorders, used for different nodes (nodal displacement and drift),

but it continues to give me exactly the same displacements/drifts which are too small. I would appreciate if somebody could make any comments on what might be going on! My time history script is as follows. Note that I don't receive any errors until the analysis successfully finishes! Thanks ..

# ----------------------------------------------------------- #
# ----------- Perform Time History Analysis ----------- #
# ----------------------------------------------------------- #

recorder Drift -file Drifts/Drifts.out -iNode 11 22 33 44 55 -jNode 22 33 44 55 66 -dof 1 -perpDirn 3;

# ----------------------------------------------------------- #
# Apply Ground Motion
# ----------------------------------------------------------- #

set accel "Series -dt 0.01 -filePath L1_ATC01_X.txt -factor 1"
pattern UniformExcitation 2 1 -accel $accel

# ----------------------------------------------------------- #
# Perform Analysis
# ----------------------------------------------------------- #

set dt 0.01

constraints Lagrange;
numberer RCM;
system UmfPack -lvalueFact 80;
test EnergyIncr 1.E-6 6;
algorithm Newton;
integrator Newmark 0.5 0.25;
analysis Transient;
analyze [expr int(30.00/$dt)] $dt

[Jeena]

Hi,

I might be wrong, but just couple of questions ...

1. Is the ground motion data stored in 'g' units ?? If it is in 'g' units, where u multiplying with 386.4 ??

2. You are applying ground motion only in one direction of the structure (in direction 1, which is your global X axis). When recording the drift, perpendicular direction is specified as '3' (which refers to global Z direction, out of the plane direction). Since you are applying ground motion only in X direction, then isn't the drift should be recorded with '2' as the perpendicular direction ?????

Thanks,
Jeena

[Ameri]

Hi Jeena,

Thanks for your comments.

1) I'm using MKS and the records have been already multiplied by 9.81. Besides, I even made the earthquake records 2000 times bigger using scale factors, but still get the same answer.
2) Nope! If you check the drift recorder in the command manual, it's written that "$perpDirn1 $perpDirn2 ... are the set of perpendicular global directions (1=X, 2=Y, 3=Z). Theses inputs are needed to calculate the length between the nodes whose drift is calculated." So it should be "3" not "2". By the way, I had already tried this and it didn't work. Also, the displacement recorders also show a very tiny displacement at nodes, which is completely unusual.