Natural period of a structure with dampers

Kave · Post by **Kave** » Tue Feb 06, 2007 3:14 am

Dear all
I have used some dampers in a steel moment frame using the viscous material and truss elements, but the natural period of structure calculated by OPENSEES did not change. My question is that what is wrong and how could I solve this problem?

silvia · Post by **silvia** » Tue Feb 06, 2007 7:55 am

how did you define the dampers and the damping.
have you tried just increasing the value by 100 to check if the influence is there?

Kave · Post by **Kave** » Wed Feb 07, 2007 8:48 am

Dear Silvia
I have done it, but no changes occur. In the OPENSEES analyzes the natural period does not correlate to dampers constants! Is it possible that in evaluating natural period of structure it does not consider the viscous damper elements?
I have put the dampers definition below:
# dampers
set c0 2e+6
#uniaxialMaterial Viscous $matTag $C $alpha
uniaxialMaterial Viscous 2 [expr 2*$c0] 1
uniaxialMaterial Viscous 3 [expr 2*$c0] 1
uniaxialMaterial Viscous 4 [expr 2*$c0] 1
#element truss $eleTag $iNode $jNode $A $matTag
element truss 100 2 5 1 2
element truss 200 3 6 1 3
element truss 300 4 7 1 4

silvia · Post by **silvia** » Wed Feb 07, 2007 1:22 pm

have you amplified them a lot to see if they do affect the period when large values are used?

why not just use the Rayleigh damping command on these elements only?

Kave · Post by **Kave** » Thu Feb 08, 2007 2:09 am

Yes I have tried a range of 1E+5 to 1E+7 which reduce the maximum drift of structure to less than half, and also I have modeled one of the samples of a passive control book in which the period of structure before and after using dampers is given, and in there I have get same results.
About the Rayleigh command you mean that I define the dampers elements then use the command to define their damping constant and then define other element which are not dampers with their damping? Is Rayleigh damping constant exactly works same as viscous damping constant? It is an applicable solution and thank you for your idea. Therefore the program do not modify the natural period considering the dampers defined by viscous dampers?

silvia · Post by **silvia** » Thu Feb 08, 2007 6:18 am

it seems that the program does notice your dampers.
The only problem, i believe, is the eigensolver, which may not be seeing them. i'll ask frank.
to apply damping to a region use the region command, see the manual:

region $regTag <-ele ($ele1 $ele2 ...)> <-eleRange $startEle $endEle> <-ele all> <-node ($node1 $node2 ...)> <-nodeRange $startNode $endNode> <-node all> <-rayleigh $alphaM $betaK $betaKinit $betaKcomm>

but it seems that what you are doing works.

endryus · Post by **endryus** » Fri Feb 09, 2007 3:22 am

Obviously the eigensolver uses only the mass and stiffness matrices. The complex modal analysis (taking into account the damping matrix as well) is not implemented.

andrea

silvia · Post by **silvia** » Fri Feb 09, 2007 2:50 pm

exactly. so you can use the rayleigh damping to find the resulting period, but otherwise you're okey with what you've got!

Kave · Post by **Kave** » Sun Feb 25, 2007 5:02 pm

I have used Rayleigh damping in several ways such as Rayleigh command, region command and using in Newmark integrator command. Then I have changed the value of them by 100, but the natural period of structure calculated by Eigen command did not change. It seems that the Eigen solver even do not consider the Rayleigh damping. Is that right? So it is no way to calculate natural period of structure considering damping?

silvia · Post by **silvia** » Mon Feb 26, 2007 8:15 am

you can do it the way we used to do it in my days: a free-vibration pull-back test.

Kave · Post by **Kave** » Mon Feb 26, 2007 11:11 am

Nice idea. Thank you.