Possible bug in localForce recorder

[brendonbradley]

Hello,

When I run the simple program below I get the problem that the output in the file "Element1_udl.out" has a first line with zeros
When executed. The output is:

"
0.1 0 0 0 0 0 0
0.2 7500 -1942.07 -6456.1 -7500 1942.07 -12964.6
0.3 15000 -3884.15 -12912.2 -15000 3884.15 -25929.3
0.4 22500 -5826.22 -19368.3 -22500 5826.22 -38893.9
0.5 30000 -7768.29 -25824.4 -30000 7768.29 -51858.5
0.6 37500 -9710.37 -32280.5 -37500 9710.37 -64823.2
0.7 45000 -11652.4 -38736.6 -45000 11652.4 -77787.8
0.8 52500 -13594.5 -45192.7 -52500 13594.5 -90752.4
0.9 60000 -15536.6 -51648.8 -60000 15536.6 -103717
1 67500 -17478.7 -58104.9 -67500 17478.7 -116682
"

However, if I add another recorder for globalForce above this, i.e.
recorder Element -file Element1g_udl.out -time -ele 1 globalForce

Then the Element1_udl.out file now doesnt have the first line of zeros, but now for some reason Element3_udl.out does (which didnt in the original file).

Any advice????
Thanks,
brendon


##################################
original code

#This file provides the input for a one storey one bay frame as given in
#Module 1
#
#Written: Brendon Bradley Dec 2011
#
################################################
#first clear all previous information stored in OpenSees domain
wipe

#define the model domain in 2D with 3 DOF per node
model BasicBuilder -ndm 2 -ndf 3

#define the nodes of the structure
# nodeTag xcoord ycoord
node 1 0.0 0.0
node 2 0.0 10.0
node 3 15.0 0.0
node 4 15.0 10.0

#set the boundary conditions for the bases of the columns
# nodeTag fixX fixY fixZ
fix 1 1 1 1
fix 3 1 1 1

#set the masses of the nodes based on the column and beam masses
# nodeTag mX mY mZ
mass 2 1.15e5 1.15e5 0.0
mass 4 1.15e5 1.15e5 0.0

#now define the geometric transformation
# TransType TransTag
geomTransf Linear 1

#now define each of the elements, their properties and geometric transformation
# $eleType $eleTag $iNode $jNode $A $E $Iz $transfTag
element elasticBeamColumn 1 1 2 0.8 200.0e9 0.06 1
element elasticBeamColumn 2 2 4 1.0 200.0e9 0.08 1
element elasticBeamColumn 3 3 4 0.8 200.0e9 0.06 1


#end of the model
##############################################
#now set the recorders for node 2 and element 1

recorder Node -file Node2_udl.out -time -node 2 -dof 1 2 disp
recorder Node -file Node4_udl.out -time -node 4 -dof 1 2 disp

recorder Element -file Element1_udl.out -time -ele 1 localForce
recorder Element -file Element2_udl.out -time -ele 2 localForce
recorder Element -file Element3_udl.out -time -ele 3 localForce

#end of setting the recorders

#write to the screen that the model and recorders have been created
puts "Model and recorders created"
##############################################
#now define the gravity loads

# $nodeTag fX fY fZ
pattern Plain 1 Linear {
# $eleTag -type -beamUniform $Wz
eleLoad -ele 2 -type -beamUniform -10000.0
}

#now define all of the analysis features. For a gravity analysis there are no special
#features which are needed

#use the Transformation method for enforcing constraints
constraints Transformation

#renumber nodes using the Reverse Cuthill McKee algorithm
numberer RCM

#Store the system of equations in a banded form and solve using Gauss elimination
system BandGeneral

#use a linear algorithm to solve Ku=p
algorithm Linear

#apply the gravity loads in a load-controlled fashion in increments of 0.1 time the total
integrator LoadControl 0.1

#perform a static analysis
analysis Static

#execute the analysis using 10 increments (as load control in 0.1 steps)
analyze 10
puts "Analysis executed"

[vesna]

Thanks for reporting the bug. We fixed the problem and it will be build in the next version of OpenSees.

In the meanwhile, if you want to record local forces use any other solution algorithm other then Linear.