Simple Model - Simple Problem?

[bandwidth0zero]

Hi everyone.
I'm trying to impose displacement pattern at a cantilever's (3d - 6 dofs) free end :

pattern Plain 2 "Series -dt $dt -filePath $filename -factor $factor" {
sp 2 1 1.0
}

The integrator is LoadControl, and I use constraints "Penalty". I am getting the error message:
" ** On entry to DGETRF, parameter number 1089 had an illegal value ", then opensees.exe collapses.

I have no idea what I'm doing wrong...

[fmk]

post the simple model.

[bandwidth0zero]

wipe
model BasicBuilder -ndm 3 -ndf 6
#Input Data
set path Results_CANTI_SP
file mkdir $path
logFile log_cant_SP.txt
#Geometry
set L 3.0;
set bcol 0.4 ;
set hcol 0.4 ;
#
set nGP_col 6
set sectag_c 1
set GeomTransfType Corotational
# Materials
# Steel
set fy 500.0e3;
set Es 200.0e6;
set b 0.001
uniaxialMaterial Steel01 1 $fy $Es $b
#Concrete01
set fc 20.0;
uniaxialMaterial Concrete01 10 [expr -*$fc*1000] -0.002 [expr -1.0*$fc*1000] -0.0035
# Geometric Non-Linearity
geomTransf $GeomTransfType 1 0 0 -1
# Nodes
node 1 0.0 0.0 0.0
node 2 0.0 $L 0.0
#
# Section:
set pi 3.141593
set cover 0.025;
set dbl_1 0.018;
set dbw 0.010;
set d1 [expr $cover+$dbw+$dbl_1/2];
set c_nf_y 150;
set c_nf_z 150;
section fiberSec $sectag_c {
patch rect 10 $c_nf_y $c_nf_z [expr -$hcol/2.] [expr -$bcol/2.] [expr $hcol/2.] [expr $bcol/2.] ;
layer straight 1 2 [expr $pi*$dbl_1**2./4.] [expr -$hcol/2.+$d1] [expr $bcol/2.-$d1] [expr $hcol/2.-$d1] [expr $bcol/2.-$d1] ;
layer straight 1 2 [expr $pi*$dbl_1**2./4.] [expr -$hcol/2.+$d1] [expr -$bcol/2.+$d1] [expr $hcol/2.-$d1] [expr -$bcol/2.+$d1] ;
}
#
#Torsional Stiffness
set SecTagTorsion 100
set TorsStiff 1.0e9
uniaxialMaterial Elastic $SecTagTorsion $TorsStiff
set newsectag_c 99
section Aggregator $newsectag_c $SecTagTorsion T -section $sectag_c
# Elements
element forceBeamColumn 1 1 2 $nGP_col $newsectag_c 1 ;
# Boundary Conditions
fix 1 1 1 1 1 1 1
#
# Recorders
recorder Node -file $path/Displacement.out -time -node 2 -dof 1 2 3 4 5 6 disp;
recorder Node -file $path/Reaction.out -time -node 1 -dof 1 2 3 4 5 6 reaction;
#
# Axial Load
set N -500.0 ; #
pattern Plain 1 "Linear" {
load 2 0.0 $N 0.0 0.0 0.0 0.0
}
# Analysis parameters
set Tol 1.0e-4;
integrator LoadControl 0.1
system BandGeneral
test EnergyIncr $Tol 25
numberer RCM
constraints Plain
algorithm Newton
analysis Static
#Axial Load Analysis
analyze 10
loadConst -time 0.0 ;
#Displacement Pattern
set dt 1.0;
set filename spx.txt
set factor 0.1;
pattern Plain 2 "Series -dt $dt -filePath $filename -factor $factor" {
sp 2 1 1.0
}
#Solve
system BandGeneral
constraints Penalty 1.0e12 1.0e12
numberer Plain
set dl $dt ;
integrator LoadControl $dl 1 $dl $dl
test NormDispIncr 1.0e-6 10
algorithm Newton
analysis Static
analyze 2500

[fmk]

i cannot repeat the error message with a simple load case .. does it appear for you at first step or a few steps into analysis .. if a few steps post the first few lines of the time series file .. if a long way in you will need to email file if long.

[bandwidth0zero]

This error message appears with every pattern I tried, before the beginning of the analysis [dead loads have no issue]. The weird thing is that opensees.exe stops working.

[fmk]

the last pattern i replaced with:

pattern Plain 2 "Linear -factor $factor" {
sp 2 1 1.0
}

and it works for me .. if it fails for you tell me the version and operating system you are using.

[bandwidth0zero]

Thank you Dr. Mc Kenna for your reply.

The problem is caused by the fiber Section OR/AND Section Aggregator, with an elastic 3D section works fine. I tried all the possible ways to create a fiber Section [patch rect, patch quad] and all kinds of elements [forceBeam, nonlinearBeam etc]
I tried also the "BuildRCrectSection" routine from Example 7 with same results.

I use OpS 2.4.6 (rev 6062) , ActiveTcl 8.5.18.0 (64-bit) and Windows 7 Ultimate SP1 x64.
edit: OpS 2.5.0 crashes too.

[bandwidth0zero]

upd: OpenSees crashes right after the 1st step of the Lateral Loads analysis, whatever loads I apply. If I write "analyze 1" there is no problem.
Any ideas ?

[bandwidth0zero]

Sorry for the double post. I discovered the source of the problem: using num_fiby x num_fibz = 100 x 99 [or less] runs without problem. Using 100 x 100 fibers crashes...