Sign Convention

[aliirmak]

Hello Openseesers,

I was trying to understand the sign convention of Opensees. For that i implemented a small script. In that script, there is a concrete block having one unit length and one unit area. I apply a load control on the specimen axially. Assuming that i applied the control load upwards and load control with positive increments, I get compression. Well, according to the setup I was expecting to get tensional behavior of the specimen.

I checked fiber element dimensions and did like in the manual.

Again to sum up, It is a very simple script and tried to understand sign convention. I assumed I was applying tension but recorder gives me such results that element is in compression.

Could you please help me about that problem?

Code: Select all

#-----MODELLING------

# SETUP
wipe;					
model BasicBuilder -ndm 2 -ndf 3	

# Node Assignment
node 1 0 0 
node 2 0 1 

#constraints
fix 1 1 1 1  ;

#definition for material properties
set fcc -6.6
set epsicc -0.0085
set fcu -5.7
set epsicu -0.0277
set fu 0.8
set fts 0.6
set Est 1000

# Material parameters for concrete
uniaxialMaterial Concrete02 1 $fcc $epsicc $fcu $epsicu $fu $fts $Est

# Fiber section for RC section
section fiberSec 1 {
	patch quadr 1 5 5 -0.5 -0.5 -0.5 0.5 0.5 0.5 0.5 -0.5 ;
}

# perform geometric tranformation
geomTransf Linear 1 

# element connectivity:
element dispBeamColumn  1 1 2 10 1 1;


# loading pattern is defined
pattern Plain 1 Linear {
	load 2 0.0 1.0 0.0
}

recorder Node -file node2dis.out -time -node 2 -dof 2 disp
recorder Element -file element.out 

# analysis object algorthims
integrator LoadControl 0.1
test EnergyIncr 1.0e-6 10
algorithm Newton
numberer Plain
constraints Plain
system SparseGeneral
analysis Static

# displacement control is used
analyze 66

[silvia]

first of all, your element recorder command is not complete.
second of all, make sure you are using the most recent version of opensees.
thirdly, and this applies to your model, the fiber patch vertices need to be specified in counter-clockwise order, as specified in the manual.

[aliirmak]

Well, i am sorry for the misleading input file. the given fiber patch was one of the trials while i was trying to reveal the sign convention.
About element recorder you might be right. Still it is a copy-paste error. Sorry for that.

Below I gave the code I was experiencing the problem. I am using Opensees 1.7.5, the last version so far. Again when the direction of the control load and increment value is negativ i am getting tension forces. I was expecting compression...

Code: Select all

#-----MODELLING------

# SETUP
wipe;               
model BasicBuilder -ndm 2 -ndf 3   

# Node Assignment
node 1 0 0
node 2 0 1

#constraints
fix 1 1 1 1  ;

#definition for material properties
set fcc -6.6
set epsicc -0.0085
set fcu -5.7
set epsicu -0.0277
set fu 0.8
set fts 0.6
set Est 1000

# Material parameters for concrete
uniaxialMaterial Concrete02 1 $fcc $epsicc $fcu $epsicu $fu $fts $Est

# Fiber section for RC section
section fiberSec 1 {
	patch quadr 1 5 5 -0.5 -0.5 0.5 -0.5 0.5 0.5 -0.5 0.5 ;
}

# perform geometric tranformation
geomTransf Linear 1

# element connectivity:
element dispBeamColumn  1 1 2 10 1 1;


# loading pattern is defined
pattern Plain 1 Linear {
   load 2 0.0 -1.0 0.0
}

recorder Node -file node2dis.out -time -node 2 -dof 2 disp
recorder Element -file element.out -time -ele 1 section 1 fiber 0.10 0.10 1 stressStrain

# analysis object algorthims
integrator LoadControl 0.1
test EnergyIncr 1.0e-6 10
algorithm Newton
numberer Plain
constraints Plain
system SparseGeneral
analysis Static

# displacement control is used
analyze 66

[silvia]

the output in your file does have negative values -- compression, as expected.

[aliirmak]

These are my first increments using the following setup
load 2 0.0 -1.0 0.0
integrator LoadControl 0.1

0.1 -6.43939e-005
0.2 -0.000129525
0.3 -0.000195169
0.4 -0.000261336
0.5 -0.000328038
0.6 -0.000395289
0.7 -0.000463103
0.8 -0.000531493
0.9 -0.000600476


Here is another setup
load 2 0.0 1.0 0.0
integrator LoadControl -0.1
That case is also referring to compression.

-0.1 -6.43939e-005
-0.2 -0.000129525
-0.3 -0.000195169
-0.4 -0.000261336
-0.5 -0.000328038
-0.6 -0.000395289
-0.7 -0.000463103
-0.8 -0.000531493
-0.9 -0.000600476


For this case i was expecting loads having the same sign. Displacements are ok.

So could you tell me then what is the relationship between the load pattern and load control?

[silvia]

the load increment is what controls the direction of loading. the load pattern is a pattern -- specifying the proportion of loads between the nodes. opensees applies a factor of that load.

[aliirmak]

OK but the problem I am having with the above two sample runs is that when I change the direction of both the load increment and the load pattern, the nodal force from the recorder also changes its direction (as expected); however, the nodal displacement does not change its direction.

I cannot understand this sign convention.

By the way, in both cases the element fiber stress-strain recorder gives me compression, which is probably because I changed the direction of both the load increment and the load pattern.

Thank you for your reply..

Ali

[silvia]

what controls the analysis is the load increment, not the load pattern, so don't touch that.
did you fix the definition of your fiber section as i suggested at the beginning?

[aliirmak]

Thanks for the answer.

Yes, the new input file is an updated version of old one and has the correct fiber definition.

The posted results(first ten increments for both two cases) correspond to the that updated input file.

[silvia]

are you working on a NEES project?

[aliirmak]

I am sorry,
i am not

[silvia]

i think you just need to sit down with pencil and paper and go through each case.