OpenSees weirdness with axial load

[jwaugh]

Ok, when I run the following script it executes just fine.

model basic -ndm 3 -ndf 6

# Create nodes
node 3 0.0 0.0 377.77
node 5 72.0 0.0 377.77
node 7 144.0 0.0 377.77
node 9 216.0 0.0 377.77
node 11 288.0 0.0 377.77
node 13 360.0 0.0 377.77
node 15 432.0 0.0 377.77

node 4 0.0 0.0 -377.77
node 6 72.0 0.0 -377.77
node 8 144.0 0.0 -377.77
node 10 216.0 0.0 -377.77
node 12 288.0 0.0 -377.77
node 14 360.0 0.0 -377.77
node 16 432.0 0.0 -377.77

# Apply Boundary Conditions
fix 3 1 1 1 1 1 1
fix 4 1 1 1 1 1 1

# Assume floors only constrain lateral translations
#rigidLink bar 5 6
#rigidLink bar 7 8
#rigidLink bar 9 10
#rigidLink bar 11 12
#rigidLink bar 13 14
#rigidLink bar 15 16

#equalDOF 5 6 3
#equalDOF 7 8 3
#equalDOF 9 10 3
#equalDOF 11 12 3
#equalDOF 13 14 3
equalDOF 15 16 3

# Section definition
#section Elastic 99 29000 30.6 259 3100 11153.8 4.72
#section Elastic 19 29000 30.6 259 1550 11153.8 4.72
#
#uniaxialMaterial Elastic 500 3000
uniaxialMaterial Steel01 500 20 29000 .005
section Fiber 99 {

patch quad 500 36 36 -9 -9 9 -9 9 9 -9 9
}

section Fiber 199 {

patch quad 500 18 18 -7 -7 7 -7 7 7 -7 7
}

set GJ 1.0e20
uniaxialMaterial Elastic 50 $GJ
section Aggregator 19 50 T -section 99
section Aggregator 119 50 T -section 199

geomTransf PDelta 1 1 0 1
geomTransf PDelta 2 1 0 -1

# Define element

element nonlinearBeamColumn 3 3 5 5 19 1;# -iter 1500 1.0e-12
element nonlinearBeamColumn 4 4 6 5 119 2;# -iter 1500 1.0e-12
element nonlinearBeamColumn 5 5 7 5 19 1;# -iter 1500 1.0e-12
element nonlinearBeamColumn 6 6 8 5 119 2;# -iter 1500 1.0e-12
element nonlinearBeamColumn 7 7 9 5 19 1;# -iter 1500 1.0e-12;# Start Disp
element nonlinearBeamColumn 8 8 10 5 119 2;# -iter 1500 1.0e-12
element nonlinearBeamColumn 9 9 11 5 19 1;# -iter 1500 1.0e-12
element nonlinearBeamColumn 10 10 12 5 119 2;# -iter 1500 1.0e-12
element nonlinearBeamColumn 11 11 13 5 19 1;# -iter 1500 1.0e-12
element nonlinearBeamColumn 12 12 14 5 119 2;# -iter 1500 1.0e-12
element nonlinearBeamColumn 13 13 15 5 19 1;# -iter 1500 1.0e-12
element nonlinearBeamColumn 14 14 16 5 119 2;# -iter 1500 1.0e-12

# Record Floor level forces and accelerations
recorder Node -file FloorT.out -time -node 5 7 9 11 13 15 -dof 3 disp
recorder Node -file FloorT2.out -time -node 6 8 10 12 14 16 -dof 3 disp

# Record Base Level Forces
recorder Element -file Col1.out -time -ele 3 globalForce
recorder Element -file Col2.out -time -ele 4 globalForce

# Apply Axial Loads to the wall
# Define Selfwieght of the slabs
pattern Plain 1 "Constant" {
# load 15 -1.00 0.0 0.0 0.0 0.0 0.0
# load 16 -1.00 0.0 0.0 0.0 0.0 0.0
}

# ---------------------------------
# Gravity analysis
# ---------------------------------
# Define analysis parameters
integrator LoadControl 0 1 0 0
system SparseGeneral -piv
test NormDispIncr 1.0e-5 500 1
numberer Plain
constraints Transformation
algorithm KrylovNewton
analysis Static

#analyze 1

# Set Maximum Displacement
set maxU 20.00
set numIncr 2000
set dU [expr $maxU/$numIncr]

pattern Plain 2 "Linear" {
load 16 0.0 0.0 1.0 0.0 0.0 0.0
# sp 15 3 [expr 1.0*$dU]
# sp 16 3 [expr 1.0*$dU]
}

integrator DisplacementControl 15 3 $dU
#integrator LoadControl 1
analyze $numIncr
#analyze 1

But if I uncomment "load pattern 1" and the line "analyze 1", then the program does something illegal with memory and Windows puts a stop to it. Any idea why this is occurring?

[silvia]

I just think that you have to handle the serious discountinuty in the load very very very very very carefully, that's why we recommend applying a load, even an axial load, in a linear manner and then holding it constant!
see the examples manual.

[jwaugh]

I will give that a try, but that is weird because I have done it that way many times before and never had a problem with it.

[silvia]

which means there is a problem with this new model.
why would you apply a load like that? instead of ramping it up?

I always try to stay far far far away from discountinuities.

[fmk]

jon,

the problem is arising in the solver .. change the solver to something else, UmfPack works fine .. don't forget to remove comment on analyze 1 command after gravity analysis!