KrylovNewton algorithm

[mohsenkh]

Dear

I modeled soil-structure interaction. The structure is RC moment frame (2-D). Beams and columns were modeled by forceBeamColumn element. The soil was modeled quad element. Analysis is nonlinear time history. The gravity and transient analysis were applied as following:

-----------------------------------------------------------
# Static Analysis
# -----------------------------------------------------------
system ProfileSPD
test NormDispIncr 1.e-3 500 0
algorithm ModifiedNewton
constraints Transformation
integrator LoadControl 1 1 1 1
numberer RCM
analysis Static
analyze 1

puts "End Gravity"

# -----------------------------------------------------------
# Switch the Material to Plastic
# -----------------------------------------------------------
for {set i 1} {$i <= $Nlayer} {incr i} {
updateMaterialStage -material $i -stage 1
}
system ProfileSPD
test NormDispIncr .08 50 0
algorithm ModifiedNewton
constraints Transformation
integrator LoadControl 1 1 1 1
numberer RCM
analysis Static
analyze 1

puts "End Update"

.
.
.
#-----------------------------------------------------------
# Transient Analysis
#-----------------------------------------------------------
loadConst -time 0.0

.
.
.
wipeAnalysis

set tFinal [expr $NumTransiant*$deltaT]
set Tol1 1.0e-4

test NormDispIncr $Tol1 1000
algorithm KrylovNewton
integrator Newmark 0.5 0.25
system ProfileSPD
constraints Transformation
numberer RCM
analysis VariableTransient
set ok [analyze $NumTransiant $deltaT [expr $deltaT/100]$deltaT 1000]

% ----------------------------------------

Can I use the KrylovNewton algorithm as I display this algorithm as top? Is this right? Can anyone introduce example that explained the KrylovNewton algorithm?

Thanks.

[vesna]

The way you define it is correct.

Here you can find a paper where KrylovNewton algorithm is explained:
http://scitation.aip.org/getabs/servlet ... s&gifs=yes