Damping in New Material

[Gholamreza]

Hello everyone,
I created new material that resistance force is proportional to relative acceleration (something like ViscousDamper material that force is proportional to relative velocity). everything looks good. The only weird thing is when I do transient analysis,(sine pulse loading....), the structural response is reducing a little bit while the damping is Zero. I compared the result of SDOF in MATLAB and opensees and I see there is no reduction in structural response in MATLAB but there is a smooth reduction in every cycle of response in opensees. I am guessing I made mistake in some part of my C++ code but I couldn't understand where !! I appreciate any idea.

[Jhno]

Did you impose numerical damping?

[Gholamreza]

Hi Jhno,
I did not use Rayleigh damping for sure. I am not sure what you mean by numerical damping ?

[Gholamreza]

I am using Newmark method with gamma = 0.5, so there is no numerical damping, right ?
How can I be sure there is no numerical damping In my C++ code ? Is there any numerical damping in Viscous material source code ?

[Jhno]

When you select the newmark method, you specify gamma and beta. If you use gamma>0.5, you induce numerical damping equivalent to gamma-0.5.
Beta should then be selected as 1/4*(gamma+1/2).
If you used gamma = 0.5 and Beta as 1/4, I don't think you did induce numerical damping.

[Gholamreza]

That’s right. I used beta = 0.5 and gamma = 0.25 and like you said there is no numerical damping.
Any idea why I have smooth reduction in my response while I don’t have any damping?

[Jhno]

No idea. One simple way thought to check if it comes from your material is to use another already built in material, like a elastic one.
If you get no damping using this alternative material, you know it comes from your code.
If you do get damping using this other material, you know it does not necessary comes from your code.
Hope it helps,
GL

[Gholamreza]

Yes, the issue is coming from the new material for sure.

[Jhno]

Assuming that you observe the reduction during the free oscillating phase of the loading, since you used the equivalent of the viscous damper but applied to acceleration, is it not normal that you observe this reduction in movement?

[selimgunay]

I also think that it comes from the damping forces. If you want to check there is nothing inelastic you can do an eigenvalue analysis before the dynamic analysis and after several time steps during the time history. If they are all the same, it means there is no inelasticity coming from your material.

[Gholamreza]

Thanks!